Hot Mix Asphalt Materials Mixture Design and Construction

• C
HOT MIX ASPHALT

MATERIALS, MIXTURE
DESIGN,AND
CONSTRUCTION
Second Edition
1996
Freddy L. Roberts
Prithvi S. Kandhal
E. Ray Brown ,
National Center for Asphalt Technology
Auburn University, Alabama
Dah-Yinn Lee
Iowa State University
Ames, Iowa
Thomas W. Kennedy
University of Texas
Austin, Texas
National Asphalt Pavement Association

Research and Education Foundation
Lanham, Maryland
Copyright © 1991, National Asphalt .Pavement Association
Research and Education Foundation
This book or any part thereof shall not be reproduced without

permission of the publisher.
Library of Congress Catalog Number: 91-60083
Printed in the United States of America
First Edition,(1991)
First Printing-2/91
Second Printing-5/91
Third Printing-3/92
Fourth Printing-3/94
Second Edition (1996)
First Printing-11/96
Second Printing-I 0/97
Third Printing---4/00
Fourth Printing-3/03
National Asphalt Pavement Association Research

and Education Foundation
5100 Forbes Boulevard
Lanham, Maryland 20706-4407
L
PREFACE TO FIRST EDITION
This textbook has been prepared to assist the college professor in pro-
viding technical information on Hot Mix Asphalt (HMA) to engineering
students. In the past, professors have had to obtain information from many
sources to teach even a short course involving HMA. Many faculty who
have had n.o formal training in HMA have deleted the subject from their
_ Materials Courses completely or have reduced the amount of material cov-
ered so that it has little benefit to the student. This book should help alle-
viate these problems.
This book has been written to _cover most topics important to those
who plan to work in the HMA industry. It can be used for teaching under-
graduate as well as graduate courses. The primary purpose of the book is
to provide all information necessary for teaching Hot Mix Asphalt to
undergraduate students to eventually improve the quality of engineers in
the industry. A comprehensive undergraduate course should expose the
student to some of the material in Chapters 2, 3, 4, and 6. Obviously all
of this material cannot be presented to a student in one course. For exam-
ple, Chapters 4, 5, and 6 have been used as the basis for teaching three
graduate level courses.
This book is also useful to the practicing engineer who is to some
extent involved with Hot Mix Asphalt. This includes personnel that work
for the Federal Highway Administration, State Departments of
Transportation, Federal Aviation Administration, Corps of Engineers, Air
Force, Navy, County Engineers, City Engineers, Consulting Engineers,
and Contractors.
Much research is presently being performed on the subject of Hot Mix
Asphalt, under the Strategic Highway Research Program (SHRP). This
research effort alone is spending $50 million dollars over a period of five
years to study asphalt cement and asphalt mixtures. It is anticipated that
once this study nears completion, many additions and updates will need to
be made to this textbook. Hence it is expected that the second edition of
this book will be forthcoming within the next five years.
iii
PREFACE TO SECOND EDITION
The second edition of this book was published in 1996. The primary
purpose of the second edition was to include information on Stone Matrix
Asphalt, Superpave performance grade (PG) asphalt binders, Superpave
mix design, and asphalt modifiers. Other portions of the book were
reviewed and modified where necessary to bring each chapter up to date
with the latest technology.
The modifications were prepared by E. Ray Brown and Prithvi S.
Kandhal. The major modifications were reviewed by Jon Epps, Dale
Decker, and John Bukowski. Lois Flowers typed the revised manuscript
and made corrections as needed. Rajib Malli,!=k revised the index.
Much research is currently underway to validate the research from the
Strategic Highway Research Program. Other research is underway to eval-
uate laboratory and field wheel tracking devices. As this work is complet-
ed, it is anticipateq. that the textbook will again need to be revised. It is
expected that the third edition of the book will be initiated in approxi-
mately five years.
iv
ACKNOWLEDGMENTS
The authors thank all those individuals and organizations that provid-
,ed support for the preparation of this textbook. Those providing signifi-
. cant co.ntributions to the preparation of this book are identified below.
All funds for preparation and publication of this book were provided
by the National Asphalt Pavement Association (NAPA) Education
Foundation. General guidance was provided by the National Center for
Asphalt Technology (NCAT) Board of Directors and the NCAT
Applications Steering Committee.
All typing, editing, xeroxing, and drafting were coordinated by Joyce

Woodfin. She also typed some of the original manuscript and made cor-
rections as needed. Other individuals involved with typing the manuscript
included Mary Kathryn Gaylor, Mary Lin, and Helena Cross. Various por-
tions of the book had to be copied many times prior to sending to review-
ers. Figures and tables had to be inserted in the correct position in the
manuscript for easy reference. Leigh Anne Bumgardner, Missy Bridges and
Shannon Jones• assisted in copying, sending manuscripts out for review,
and inserting tables and figures
All figures were prepared under the direction of Wally Ridgeway, who
also designed the book cover and dust jac~et. Others involved with prepar-
ing figures included Robert Stroud, Julia Freeman, Mike Cole, and Hugh
Smith.
Many people reviewed either portions of the textbook or the entire

book for technical content. Those involved with reviewing individual
chapters are as follows: Chapter I-John Gray and Ronald Kenyon;
Chapter 2-Woodrow Halstead, R.D. Pavlovich, Claine Petersen, and Don
Beck; Chapter 3-Truman Jones, Frank Nichols, Jr., and Charles Marek;
l
Chapter 4-Paul Teng, John Carroll, Maqbool Khatri, and Dick Smith;
Chapter 5-Harold Von Quintus; Chapter 6-Mike Geller, Stuart Murray,
Jay Winford, Jay Gabrielson, and Ted Rapallo; Chapter 7-Leonard Wood
_____a_n_,d Cnarlesrlughes; Cnapter 8=Fred-Fir1n-and Marvin Traylor; and
Chapter 9-Donald Jordison and Mohammed Shahin.
Those individuals reviewing several chapters included Richard Davis,

Fred Fehsenfeld, Gail Jensen, Charles Potts, and G.E. Ramey.
V
A special thanks to Ronald Terrel, Jon Epps, and Campbell Crawford
who reviewed the entire book. Their comments were very helpful for tech-
nical content, organization, and readability of the text.
The final manuscript was edited for grammar and accuracy by Wallace
Whatley, Ann Morton, Kim Toney, and Kristy Jones.
Thanks to Brown Printing Company for working closely with the

NCAT Staff in typesetting and printing. Employees directly involved
included Charles McDevitt; Betsy Beale and Eddie Hudson, typesetting;
and Barbie Storey, design and artwork.
,~
Thanks are also due to Judy, Uma, and Shirley for their support dur-
ing the preparation of this book
vi
TABLE OF CONTENTS
CHAPTER 1. INTRODUCTION ........................ 1
ORIGIN OF BOOK ................................ 1

~
IMPORTANCE OF SUBJECT ............. , ......... .3

'
DESCRIPTION OF CONTENTS ..................... .4
REFERENCES ....................................6
CHAPTER 2. ASPHALT REFINING, USES, AND

PROPERTIES ............................7
HISTORICAL BACKGROUND AND REFINING ......... 7

Historical Background ............................7
Asphalt Refining ................................. 8
Crude Petroleum Variation ...................... 9
Distillation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10
Solvent Deasphalting ......................... 12
Solvent Extraction (ROSE) ..................... 12
Continuous Air-Blowing ....................... 13
Several Manufacturing Options .................. 14
ASPHALT TYPES AND USES ....................... 15

Asphalt Cements ................................ 16
Emulsified Asphalts .............................. 17
Cutback Asphalts ............................... 19
ASPHALT CEMENT PHYSICAL TESTS .............. .23

Consistency Tests ...............................23
Absolute Viscosity at 140°F .....................23
Kinematic Viscosity at 275°F ...................25
-----------IJe·rretration ........ ;-;-.......................26 '
Softening Point ..............................27
Ductility Test ...............................28
Durability Tests ...............................•29
Thin Film Oven Test .........................29
vii
Rolling Thin Film Oven Test ... : ................30
Purity Tests ....................................31 w
Safety Tests .................................... 31

Other Tests ....................................32
Specific Gravity .............................32
Spot Test ..................................32
ASPHALT CEMENT GRADING SYSTEM ............. .33

Historical Background ........................... 33 ~
Viscosity Grading System ........................ .37
AR Viscosity Grading System ..................... .41
RHEOLOGICAL PROPERTIES AND PAVEMENT

PERFORMANCE ..............................42
Rheological Properties ............................42_
Age Hardening ............................. .42
Penetration ................................ .46
Ductility ..................................47
Viscosity ..................................48
Temperature Susceptibility .................... .49
Shear Susceptibility ........................... 54
Stiffness ................................... 54
Pavement Performance Related to Asphalt Cement \.'..
Rheological Properties .........................64
Ravelling .................................. 64
Cracking: .................................. 64
Rutting ...................................70
Stripping ..................................70
SUPERPAVE ASPHALT BINDER TESTS AND

SPECIFICATION ..............................72
Superpave Physical Tests for Asphalt Binders ........... 73
Rolling Thin Film Oven .......................73
Pressure Aging Vessel ........................ .75
Rotational Viscometer ......................... 76
Dynamic Shear Rheometer ........ .- ............ 78
_ _ _ _ _ _ _ _Bending-Beam-Rhee-met-er ...................... 86,--------
Direct Tension Tester ......................... 92
Superpave Asphalt Binder Specification ............... 93
Permanent Deformation ....................... 97
viii
Fatigue Cracking ......................... , .. 97
Low Temperature Cracking ..................... 98
Other Common Specification Criteria .............98
Selection of Superpave Binder Grades ............. 99
ASPHALT CHEMISTRY ........................... 102

Origin of Asphalt .............................. 103
Elemental Composition and Molecular Structure ....... 103
Functional or Polar Groups ....................... 104
Conceptual Compositional Model .................. 105
Asphaltenes ............................... 105
Resins ................................... 106
Oils ..................................... 106
Asphalt as a Colloidal System ..................... 106
Analytical Procedures ........................... 107
Chemical Precipitation Method ................. 107
Selective Adsorption - Desorption Method ........ 109
Molecular Size Distribution .................... 111
SHRP Research in Asphalt Chemistry ............... 113
REFERENCES ........ , ......................... .114
CHAPTER 3. AGGREGATES ......................... 121
INTRODUCTION ............................... 121
AGGREGATE PRODUCTION ...................... 123

Quarry Operations ............................. 123
AGGREGATE SAMPLING ......................... 128
MINERALOGY & CHEMICAL PROPERTIES ......... 129

Aggregate Mineralogy ........................... 129
Common Minerals Found in Aggregates .......... 129
Effect of Mineralogy on Performance ............. 130
Petrographic Examination_ ... : ... , , ............ 134
CliemicarProperties of Aggregates .................. 134
PHYSICAL PROPERTIES OF AGGREGATES .......... 136

Toughness and Abrasion Resistance ................. 138
ix
Durability and Soundness ........................ 139
Particle Shape anal Surface Texture .................. 140
Cleanliness and Deleterious Materials ............... 143
Sand Equivalent Test ......................... 144
Clay Lumps and Friable Particles ................ 144 .
Plasticity Index ............................. 145
Specific Gravity ................................ 145
Specific Gravity and Absorption of
. Coarse Aggregate ........................ 147
Specific Gravity and Absorption of
Fine Aggregate .......................... 148
Size and Gradation .......... ,.,. .................. 153
Desired Gradation Curves ..................... 153
Maximum Aggregate Size ..................... 155
Typical Specification Bands and Tolerances ........ 157
Sieve Analysis ................, .............. 157
Aggregate Blending to Meet Specifications ......... 160
Surface Area .................................. 169
REFERENCES ...................................171
CHAPTER 4. HOT MIX ASPHALT MIXTURE DESIGN

METHODOLOGY ..................... .173
INTRODUCTION ...............................173
HISTORY OF MIX DESIGN ....................... 173

The Pat Test .................................. 174
Large Stone Mixes .............................. 174
Dense Gradation ............... ~ ............... 175
Hubbard-Field Method ................... : ...... 175
Hveem Mix Design Method ...................... 176
Marshall Mix Design Method ..................... 177
Superpave Mix Design Method .................... 181
-------oBJEeTIVES-ANB-E-E-EM:EN-T:-S-0F-MIX-BESIGN-.-.-.-.-.-l-R.----
Objectives .................................... 181
Gradation and Maximum Density Curves ............ 185
Asphalt Film Thickness .......................... 187
Laboratory Compaction ......................... 189
X
Weight Volume Relationships ..................... 192
Definitions ................................ 192
Calculations ............................... 199
MARSHALL MIXTURE DESIGN METHOD ..........204

Marshall Criteria ...............................212
Voids in the Mineral Aggregate .................214
Voids in Total Mix (VTM) ....................215
Density ..................................216
Stability ..................................217
Flow ....................................218
Percent Voids Filled with Asphalt ................218
HVEEM MIXTURE DESIGN METHOD .............218
SUPERPAVE MIX DESIGN METHOD .............. ,.234

Background ..................................234
Materials .................................... 235
Selection of Aggregate Blend ......................239
Superpave Gyratory Compactor .................... 244
Compaction of Samples .........................245
Selection of Optimum Asphalt Content .............. 246
OTHER CONSIDERATIONS .......................249

Moisture Susceptibility ..........................249
Hot Recycling .................................250
Frictional Resistance of Surface· Mixes ............... 252
EXAMPLES OF MIXTURE DESIGN ................. 254

Marshall Mix Design ............................255
Hveem Mix Design .............................262
Superpave Mix Design Example Problem .............263
Binder Evaluation ..............................263
Aggregate Evaluation ............................263
Evaluation of Aggregate Structure .................. 265
Trial Blends ..................................266
- - - - - - - - -Determme Optimum Asplial.t Content ..............269
REFERENCES ...................................273
xi
CHAPTER 5. CHARACTERIZATION OF ASPHALT
MIXTURES ........................... .275
INTRODUCTION ............................... 275
MATERIAL CHARACTERIZATION TECHNIQUES ....276
CRITERIA FOR ASPHALT MIXTURE TESTS ......... 277

Material Variability and Project Size ................. 277
Variability ................................. .278
Project Size ................................278
Ability to Estimate Fundamenta:1:-Properties ...........280
Ease of Testing ............................... .280
Reproducibility of Test Results .......... ·...........280
CURRENT TESTS ...............................281

Physical Pi'.operties .............................283
Air Voids ................................. 283
Voids Filled with Asphalt Cement ............... 283
Voids in Mineral Aggregate .................... 283
Aging ...................................283
Mechanical Tests ...............................284
Marshall Stability ...........................284
Hveem Stabilometer ......................... 285
Cohesiometer ..............................286
Modulus of Elasticity .........................286
Resilient Modulus (Repeated Load Indirect Tensile)
(MiJ .................................286
Stiffness Modulus by Indirect Means ............. 289
Dynamic Complex Modulus ...................289
FlexuralTest ...............................292
Indirect Tension ............................298
Creep and Permanent Deformation .............. 301
GyratoryTesting Machine (GTM) ............... 304
Laboratory Wheel Tracking Tests ................ 305
-t---------"M~o1'-"-·st,,.,,u'""re~Su~s=ceRtibilicy: Tests ................... 306
REFERENCES ................................... 310
xii
CHAPTER 6. EQUIPMENT AND CONSTRUCTION ... 313
INTRODUCTION ............................... 313
HOT MIX ASPHALT FACILITIES .................. .313

Typical Layout ................................ 313
Drum Mix Facility .......................... 313
· Batch Facility ............................. .316
Drum Mixers ................................. 318
Aggregate Feed ............................. 318
Drum.· ................................... 318
Burners ..................................321
Drying and Heating Process ................... 323
Asphalt Cement Injection .....................325
Fines Feed.er System ......................... 326
Batch HMA Facilities ........................... 327
Aggregate Dryer ............................ 327
Burners ..................................328
Screens and Hot Bins ........................ 328
Weigh Hopper and Asphalt Bucket .............. 330
Pugmill Mixer ............................. 330
Air Quality Control System ....................... 331
Primary Dry Collectors ....................... 332
Wet Collectors/Scrubbers ..................... 333
Fabric Filter (Baghouse) ...................... 335
TRANSPORTATION AND LAYDOWN OPERATIONS . .337

Transportation Operations ........................ 338
End Dump Vehicles ........................ .338
Bottom Dump Vehicles ....................... 339
Live Bottom Vehicles ....................... .339
Truck Loading Procedures ..................... 339
Protection During Haul ..................... .341
Factors Affecting Truck Cycle .................. 341
Asphalt Pavers ................................. 344
Principles of the Asphalt Paver ................. .344
.....__ _ _ _ _ _ _ _ _ _Auro:marid'\1.ver-S-cree-d-eon:trols-:-:: ............. 351
Types of Grade Reference Systems ............... 351
Effect of Specifications on Paver Operations ........ 353
xiii
· COMPACTION ................................... 354
Mixture Factors Affecting Compaction .............. .354
Compaction Equipment ......................... 356
Steel-Wheeled Rollers ........................359
Pneumatic-Tired Rollers ..................... .360
Vibratory Steel-Wheeled Rollers ................ 363
Field Conditions Which Affect Compaction .......... .365
Rolling Patterns ............................... 366
Types of Density Specifications Used for
Construction of HMA ....................... 369
Density Specified as a Percent of
Laboratory Density ... _.................... 371
Density Specified as a Percent of
Theoretical Maximum Density .............. 375
Density Specified as a Percent of Control Strip ..... .376
Measurement of Density ......................... 378
Core Method ............................. .378
Nuclear Gauge Method ....................... 378
MIXTURE SEGREGATION ........................ 379
CONTRACT AND MATERIAL SPECIFICATIONS ...... 380

General .....................................380
Bidding Methods .............................. 381
Low Bid ..................................381
Best Bid .................................. 381
Negotiated Bid ............................ .381
Specification Development ....................... 382
Method Specifo:ation ........................382
End-Result Specification ......................382
Sampling of Aggregate and Calibration of Feeders ·...... 385
Aggregate Sampling .........................386
Aggregate Feeder Calibration .................. .387 ·
Asphalt Cement Sampling .. . ---:-:-:-: ................. 390
Sampling HMA .............................. .390
STATISTICAI:;-Ce>NeEPTS ........................ 3:9·1c1--- - - -
QUALITY CONTROL/QUALITY ASSURANCE

(QC/QA) ............. :, ......................399
Aggregate Gradation ......................... 399
xiv
Asphalt Content ............ '............... .400
Field Management of Volumetric Properties .......... .402
Construction .................•............403
Testing and Quality Control Charts .................404
Testing Frequency ...........................404
Technician Qualifications .....................404
Quality Control Charts ...................... .404
REFERENCES .................................. .408
CHAPTER 7. SPECIAL MIXTURES, RECYCLING, AND

ADDITIVES ........................... .411
OPEN-GRADED HMAMIXES .................... .411

Open-graded'Base Mixes ........................ .411
Open-graded Friction Course (OGFC) ............. .413
Stone Matrix Asphalt (SMA) ..................... .415
Introduction .............................. .415
Description .............................. .416
Materials .................................417
Mix Design .............................. .419
Construction ..............................422
Performance .............................. .423
DENSE-GRADED LARGE STONE MIXES ........... .423,

Development of Modifications to
Marshall Procedures ........................ .425 ,)
Analysis of Comparative Data .................... .426
Typical Mix Designs and Field Trials ................428
ASPHALT PATCHING MIXTURES ................. .430

Desirable Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . .431
Challenges of Mix Design ....................... .433
New Concepts ............................... .433
Improved Formulation .............. . r. • • • • • • • .' • .434
Production, Inspection, and Storage ................ .437
HMA MIXTURES FOR HYDRAULIC STRUCTURES .. .439
xv
HOT MIX ASPHALT RECYCLING ................. .439
Introduction ..................................439
Candidates for Recyding .........................440
Design Considerations for Recycled Mixtures ..........442
Production of Recycled Asphalt Mixtures- ............ .444
ADDITIVES AND MODIFIERS IN HMA ............ .448

Ideal Pavement Binder ......................... .449
Classification .................................451
Fillers .....................................451
Extenders ................................... .454
Polymers .....................................455
Rubber .................. '-:- . ...............456
Plastic ...................................460
Combinations ............................. .461
Fibers .....................................461
Oxidants ....................................463
Antioxidants ..................................463
Hydrocarbons ................................ .464
Recycling Agents .......................... .464
Hardening Agents .......................... .465
Antistripping Agents ........................... .466 . ,,
Liquid Antistripping Additives ................ .466
Lime Additives ............................ .467
Waste Materials ................................468
Roofing Shingles ........................... .469
Scrap Rubber from Tires ..................... .469
Waste Glass .............................. .469
Miscellaneous ....... '. ..................., ..... .470
Silicone ..................................470
Deicing Calcium Chloride Granules .............471
REFERENCES ...................................472
CHAPTER 8. PERFORMANCE/DISTRESS OF HMA ... .477
----S'ERVICEABI-Hfl/-PE-RF'ORMANGE-Gf>NGEP-'IS-.-.-;-.-.-A-7--7------
Pavement Roughness ........................... .479
Road Roughness Measurement Methods ............ .481
Smoothness Requirements for Construction ...........486
xvi
DISTRESSES IN HMA PAVEMENTS ... ,. ........... .489
Cracking ....................................494
Fatigue Cracking ...... _- ................... .494
Low Temperature (Thermal) Cracking ........... .495
Longitudinal Cracking ...................... .496
Block Cracking .............................498
Reflection Cracking ........................ .499
Slippage Cracking .......................... 501
Distortion ................................... .502
Rutting .................................. 502
Shoving and Corrugation ..................... 506
Disintegration ................................. 507
Ravelling ................................ .507
Wear Loss ................................. 509
Stripping ................................ .510
Potholes .................................. 517
Loss of Frictional Resistance ..................... .517
REFERENCES ............... " .................. .522
CHAPTER 9. MAINTENANCE, REHABILITATION, AND

RECONSTRUCTION OF HMA .......... .527
INTRODUCTION ............................... 527
MAINTENANCE ............................... .531

Preventive Maintenance .......................... 531
Fog Seals .................................. 532
Rejuvenators ............................... 535
Chip Seals or Surface Treatments ................ 535
Slurry Seals ................................ 540
J,r Crack Sealing .. ~ ........................... .543
Corrective Maintenance ......................... 544
Patching . . .- . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .544
Thin HMA Overlays ....................... .548
- - - - - - - ~ 0 N .................. ~~., ........ 550

Hot Surface Recycling .................... .' ..... .551 .,
Open-Graded Friction Courses ................... .552
Cold Milling ................................. .552
xvii
Structural Recycling ........................... .554
Structural Overlays ............................ .555
Crack Relief Layers and Systems ................... .556
Thick Overlays ............................ .559
Treating the Existing Pavement ................ .559
Treatment of Existing Cracks and Joints .......... .560
Special Overlay Design Considerations ........... 561
Techniques for Cracking Concrete Slabs to Reduce
Reflection Cracking ..................... .563
Equipment for Cracking, Breaking, and Rubblizing
Concrete Pavement ...................... .564
Crack Pattern and Piece Siz~ .................. .568
Seating of Cracked Slabs ..................... .570
Other Considerations in Cracking and Seating ..... .571
Saw and Seal ................ , ............ .572
RECONSTRUCTION ............................ .574

Cold Mix Recycling with Overlay ................ .' .574
Hot Mix Recycling with Overlay .................. .575
Drainage ................................... .575
REFERENCES ............................ '. ..... .576
INDEX .............................................579
xviii
CHAPTER 1. INTRODUCTION
ORIGIN OF THE BOOK
This textbook is a direct result of the concern of U.S. Hot Mix Asphalt
(HMA) contractors and the National Asphalt Pavement Association
(NAPA) about the lack of material on asphalt technology available to engi-
neering students in American universities and colleges and to technicians
in technical schools. In fact, two of the principal educational activities of
the National Center for Asphalt Technology (NCAT) have been directed
toward (1) providing the opportunity for professors currently on U.S. Civil
Engineering faculties to learn more about asphalt technology by attending
an annual summer course offered at NCAT, and (2) preparing a book suit-
able for use in college courses in asphalt technology.
Technology as used in this book means the systematic treatment of a
subject. The primary subject of this book is HMA. The topics covered
include manufacture of the asphalt and aggregates, design of HMA mix-
tures, characterizing HMA in terms of engineering properties, construc-
tion, performance, and maintenance and rehabilitation of aging HMA sur-
faces. This book has been prepared for use by engineers; however, it has
much to offer technicians in advancing their understanding of the topics
covered in the book.
This book is the first comprehensive treatment of the subject in many
years. Based on a cursory review of the NCAT library, it appears w'be the
first textbook in over 20 years. The last textbook that was published on
--------.1.B-.LJMA in rbe U.S. was in 1967 by Wallace and Martin (1) titled Asphalt
Pavement Engineering. Just prior to that, Bituminous Materials in Road
Construction was prepared by the staff of the Road Research Laboratory in
the United Kingdom in 1962 (2). In fact, the 1960s was the last decade in
which substantial printing of reference books dealing with asphalt materials
2 Hot Mix Asphalt Materials, Mixture Design and Construction
occurred. A sample of the books and book series printed in the 60s
includes the following:
1. 6th Edition of H. Abraham's five volume set on Asphalt and Allied

Substances (3) in 1960:
Volume I: Historical Review and Natural Raw Materials
Volume II: Industrial Raw Materials
Volume III: Manufactured Products
Volume IV: Methods ofTesting-Part I-Raw Bituminous
Materials
Volume V: Methods of Testing-Part 2-Fabricated Products
2. R.N. Traxler's Asphalt.ilts Composition, Properties and Uses (§J in ·

1961.
3. E.J. Barth's Asphalt Science and Technology (2) in 1962.
4. A.J. Hoiberg edited a three volume set on Bituminous Materials:
Asphalts, Tars, and Pitches (..Q) in 1965:
Volume I: General Aspects
Volume II: Asphalts
Volume Ill: Tars and Pitches
While all four of these books and book sets were excellent as reference doc-
uments for the experienced asphalt technologist, they were not written for
use in the classroom as textbooks.
These comments should not lead the reader to believe that there is a
shortage of technical information on asphalt materials. In , the
Transportation Research Information Service (TRIS) maintained by the
Transportation Research Board (TRB) of the National Academy of
Sciences, there are tens of thousands of referen~es, papers, and research
reports on asphalt technology. Indeed, every year there are many papers
presented and published by the TRB, and the Association of Asphalt
Paving Technologists publishes a proceedings of their annual meeting,
which typically contains 20 to 30 of the most recent papers on asphalt
technology. In addition, the International Society for Asphalt Pavements
(ISAP) sponsors an international meeting and publishes a proceedings
every 5 years to focus on some aspect of the design and performance of
---asprratr_--pavemems:--Ad-ditiona±ly,-dre-Natrorra:1-Asphah-Paveme•-n-r-------
Association and the Asphalt Institute regularly publish and revise a whole
host of publications dealing with all aspects of asphalt pavement design,
construction, testing, and issues of particular concern to their individual
constituencies. However, with all this information available, little has been
Chapter !-Introduction 3
prepared specifically for the classroom. This book is targeted specifically for
that audience. Therefore, in preparing this textbook, the authors have
extracted the best current technology from the literature and put it into a
form suitable for classroom use.
IMPORTANCE OF THE SUBJECT
In the United States in 1988 there was a total of approximately 4 ·mil-

lion miles of roads of which 2.3 million were surfaced with asphalt or con-
crete. The balance of the roads were surfaced with either gravel, stone, or
soil (1.3 million miles), or they were non-surfaced (0.4 million miles). Of
the hard surfaced roads, approximately 96 percent (almost 2.2 million
miles) had asphalt surfaces. More than 95 percent of the 2 trillion annual
vehicle miles of travel occurred on these hard surfaced roads (Z).
In 1988, expenditures for highways were ov:er $68 billion at all levels
of government. Since 1981, these expenditures have grown at an annual
rate of over $3.5 billion. The $68 billion spent in 1988 included $32 bil-
lion for capital outlays and $20 billion on maintenance activities.
During 1988, the amount of HMA produced and placed in the U.S.
was estimated to be 500 million tons (.i) valued at some $10.5 billion. The
HMA industry directly employs some 300,000 people and indirectly
accounts for an additional 600,000 jobs (B.). When combined with the
state and federal employees associated with the construction and mainte-
nance of asphalt surfaced roads, the industry today has a significant impact
on the economic vitality of the nation.
Additionally, many engineers and technicians from all sectors of the
asphalt business who entered the work force during the early days of the
construction of the interstate system are reaching the end of their careers.
These people must be replaced by engineers and technicians who are
knowledgeable of asphalt materials and their proper use to successfully
rebuild the highway sector of the transportation infrastructure.
This textbook has been prepared to help those individuals teaching
asphalt materials to ensure that future engineers and technicians do a bet-
ter job. Students should recognize that the contents of this book cannot be
taught in any one course. They are encouraged to make this book a part of
-------Ltheir...p.erman.enue.chnkaUihtai:y_an_d to _use it as they experience new chal-
lenges and face new obstacles that must be overcome in the practice of their
chosen profession.
DESCRIPTION OF THE CONTENTS
Chapter 2 deals with the subject of asphalt cement as an individual

component of a HMA paving mixture. Included in that discussion are the
manufacture and uses of asphalt materials, types of asphalt materials used
in pavement engineering, asphalt chemistry, asphalt grading systems, and
tests for evaluating the properties of asphalt cement and relationships
between those properties and field performance.
The subject of Chapter 3 is aggregate processing and properties that
affect the quality of a HMA paving mixture. Included in the discussion are
the different types of rocks that are suitable for use as aggregates, types of
crushers typically used to process quarried rQck, sizing and blending oper-
ations, mineral and chemical properties of aggregates and their relationship -
to acceptability of the aggregate as a paving material, physical property
tests and requirements for use in HMA, aggregate blending to meet speci-
fications, and characteristics of gradation as it affects HMA properties and
construction.
Chapter 4 describes the three most commonly used methods for mix-
ture design ofHMA materials, the Hveem, Marshall, and Superpave meth-
ods. Included in that discussion are the background on development of the
methods and a description of how that relates to considerations of specifi-
cation or criteria modification; elements of mixture design and develop-
ment of criteria for Marshall, Hveem, ap.d Superpave methods; a detailed
description of all three methods including a step by step fl.ow chart of activ-
ities, tests, and analyses of results; a description of factors which affect mix-
ture design; and a comprehensive example problem showing how all cal-
culations and analyses are performed.
Once an HMA mixture has been designed using the empirical proce-
dures described in Chapter 4, there is a need to determine the engineering
properties of these mixtures for use in evaluating performance, in the
mechanistic-empirical thickness design methods, and for forensic.analysis
of pavement failures.
Chapter 5 describes the common tests used for determining engineer-
ing properties, the analysis techniques for using those properties, and the
general area of moisture susceptibility and tests being used.to evaluate the
______,Ffopensitys,f HMA mixtures to stri .
The topic of Chapter 6 is the facilities used to produce HMA p us t e
construction process and the associated equipment. Included is a discus-
sion of general HMA facility layouts; each of the elements included in both
batch and drum mix facilities; vehicles used to transport the HMA to the
paving site and economic considerations of trucking; principles of opera-
Chapter !-Introduction 5
tion of the paver; troubleshooting mat problems; compaction equipment

and procedures as well as the effect of various factors on compaction; types
of compaction specifications and the advantages and disadvantages of each;
statistical concepts that form the basis for statistical quality control/ quali-
ty assurance programs; arid sampling locations at the HMA facility and
sampling errors.
Chapter 7 deals with special mixtures and processes not included in
the previous 6 chapters. Included are descriptions of the purpose and
design of open-graded HMA mixtures including friction courses, Stone
Matrix Asphalt (SMA) mixtures, special large aggregate mixtures being
used experimentally for roadways with very heavily loaded axles, HMA
mixtures for hydraulic applications, maintenance/ cold mix design and
uses, recycling and its implications on mixture design and construction,
and various additives (or modifiers) being used to prevent stripping and to
provide property enhancement for the asphalt cement.
The subject of Chapter 8 is distress and performance of HMA road-
ways. The concept of serviceability and performance is discussed along
with techniques for measuring smoothness and the types of smoothness
specifications being implemented by many of the State Departments of
Transportation (DOTs) t9day, the common distresses experienced by
HMA pavements. due to traffic loads and environment, and finally, the
probable causes of each distress are enumerated as well as typical tech-
niques for alleviating the problem.
As HMA pavements get well into their service lives, there is a need to
perform periodic maintenance and rehabilitation activities even on the
facilities which provide the best service. Chapter 9 contains detailed
descriptions of the proper procedures for performing maintenance on typ-
ical flexible pavement distresses such as rutting, cracking, ravelling, and
stripping. In addition, other maintenance treatments including fog seals,
slurry seals, and various other surface treatments are discussed .. Milling is
discussed as a method of removing material for both maintenance and
recycling purposes. Various techniques used to minimize reflection crack-
ing are also discussed including interlayers; large stone crack relief layers;
crack or break and seat; saw and seal; and rubblizing concrete pavements
to prevent reflection cracking. Overlays are also discussed as techniques for
both maintenance and strengthening the pavement structure.
REFERENCES
1. Wallace, H.A., and J.R. Martin. Asphalt Pavement Engineering.

McGraw-Hill Book Co., New York, NY, 1967.
2. Bituminous Materials in Road Construction. Road Research
Laboratory, Department of Scientific and Industrial Research,
London: Her Majesty's Stationery Office, 1963.
3. Abraham, H. Asphalt and Allied Substances. 6th Edition, D. Van
Norstrand Co., Inc., Princeton, NJ, 1960:
Volume I: Historical Review and Natural Raw Materials
Volume II: Industrial Raw Materials
Volume III: Manufactured Products
Volume IV: Testing Raw Bituminous Materials
Volume V: Testing Fabricated Products.
4. Traxler, R.N. Aphalt-Its Composition, Properties and Uses. Reinhold
Publishing Corporation, New York, NY, 1961.
5. Barth, E.J. Asphalt Science and Technology. Gordon and Breach
Science Publishers, New York, NY, 19,62.
6. Hoiberg, A.J., Editor. Bituminous Materials: Asphalts, Tars, and
Pitches. John Wiley and Sons, Inc., New York, NY, 1965, Reprinted
Edition 1979 by R.E. Kriger Publishing Co., Inc., Huntington, NY:
Volume I: General Aspects
Volume II: Asphalts
Volume III: Coal Tars and Pitches.
7. Selected Highway Statistics and Charts 1988, Federal Highway
Administration, U.S. Department of Transportation, 1988.
8. Hotmixfax 1988. National Asphalt Pavement Association, 1988.
CHAPTER 2. ASPHALT REFINING,
USES, AND PROPERTIES
HISTORICAL BACKGROUND AND REFINING
Historical Background
Asphalt cement and tar are considered bituminous materials. Quite

often, these two terms are used interchangeably due to misconceptions
resulting from their similarity in appearance and in some parallel applica-
tions. However, asphalt cement and tar are two distinctly different materi-
, als with different origins and different chemical and physical characteris-
tics. Asphalt cement is a dark brown to black cementitious material that is
either naturally occurring or is produced by petroleum distillation. Tar, on
the other hand, is primarily manufactured from the destructive distillation
of bituminous coal and has a very distinct odor. Asphalt cement is used
principally in the United States in paving applications. Tar is hardly ever
used in paving because of (a) some undesirable physical characteristics such
as very high temperature susceptibility, and (b) significant health hazards
such as severe eye and skin irritation when _exposed to its fumes.
Asphalt cement is man's oldest engineering material. Its adhesive and
waterproofing properties were known at the dawn of civilization
W . It was used by a thriving shipbuilding industry in Sumeria about 6000
B.C. An ancient civilization in the Indus Valley (northwestern India) used
asphalt cement in the construction of large public baths or tanks about
3000 B.C. They were built by placing a one-inch layer of asphalt cement
between two courses of baked brick. As early as 2600 B.C., Egyptians used
_ _ _ _ _ _....
n=atu.ive..as.phalts£0Lw.aterp.ro.ofing,-m.ummification,.and building structures.
The use of naturally occurring asphalts as a mortar for building and paving
blocks, caulking for ships, and numerous waterproofing applications con-
tinued in later years in various parts of the world.
Commercial types of asphalt can broadly be classified into two cate-
gories (1):
1. Natural Asphalts: These were laid down in geologic strata and occur
both as relatively soft asphalt material and also as a hard, friable,
black material in veins of rock formations, or impregnated in vari-
ous limestones, sandstone formations, and the like. The relatively
soft asphalt material, almost like heavy petroleum, is typified in the
Trinidad Lake deposit on the Island of Trinidad, in Bermudez Lake,
Venezuela, and in the extensive "tar sands" throughout western
Canada.
Hundreds of thousands of tons of asphalt have been removed
from the Trinidad Lake without its showing any appreciable sign of
loss (Z). As the material is removed, pressure deep in the earth forces
more of the heavy residue to the surface, where nature continues its
refining process. Trinidad asphalt contains about 50 to 57 percent
asphalt. The remainder is colloidal day of volcanic origin, plus some
inert organic matter. By the last two decades of the nineteenth cen-
tury, considerable amounts of asphalt from the Trinidad Lake
deposit wdre used for road building in the United States and else-
where in the world (1). The first sheet asphalt pavement using
Trinidad Lake asphalt was laid in 1876 on Pennsylvania Avenue in
Washington, D.C. Trinidad asphalt has a very low penetration
(3-10) and therefore was fluxed back for use in paving in the United
States. Normally, it is refined so .that the volcanic ash is between 3
and 5 percent. -
The hard variety of natural asphalts such as Gilsonite includes
asphaltites which are solid asphalts without impurities (silts, days,
etc.). Gilsonite is a black, brittle and hard asphalt material mostly
mined in Utah.
2. Petroleum Asphalts: These are colloidally dispersed hydrocarbons in
crude petroleum and are obtained by refining petroleum crudes.
Following the discovery of the refining process in the early 1900s
and the popularity of automobiles, large quantities of petroleum
asphalt were processed by the oil companies. Gradually these refined
asphalts became plentiful and of good quality. This forced the native
(natural) asphalts into a position of relative unimportance (1).
---AsphaltRefining--------------------'----
The material for this section on asphalt refining has largely been
obtained from papers by Corbett (1.) and Hodgson (j).
Almost all paving asphalt cement used today is obtained by processing
crude oils. Many refineries in the Unit~d States are located near water sup-
Chapter 2-Asphalt Refining, Uses, and Properties 9
plied transport or are supplied by pipelines from marine terminals.

Therefore, the world supply of crude oil is significant to the supply of
asphalt. A substantial amount of crude oil used in the United States is
imported from foreign countries at the present tim<;. The main countries
or regions from which the heavy asphalt-bearing crucle oils are imported
are Mexico, Venezuela, Canada, and the Middle East. There is also consid-
erable production of asphalt-bearing heavy crude oils in the United States,
primarily in the following areas: Gulf Coast, Mid-Continent, Rocky
Mountain, West Coast, and the North Slope of Alaska, as shown in Figure
2-1.
Figure 2-1 Location of Heavy Oil Fields in the United States

(after Hodgson, S)
Crude Petroleum Variation. Crude petroleums vary in composition

from source to source. They yield different amounts of residual asphalt
cement and other distillable fractions. Figure 2-2 shows typical volume
percentages of these fractions in three crudes. The residual asphalt (bitu-
men) contents are shown by the shaded areas.
Crude oils may be arbitrarily classified according to their API
(American Petroleum Institute) gravity, which can be used to estimate the
asphalt yield. The API gravity is an arbitrary expression of the density or
weight of a umt volume of material expressed -at 60°F and is obtained as
follows:
API Gravi-ty (de~ _ _1_4_1._5_ _ - 131.5

Specific Gravity
Hot Mix Asphalt Materials, Mixture Design and Construction
BOSCAN ARABIAN NIGERIA

VENEZUELA HEAVY LIGHT
API Degrees 10.I 28.2 38.1
SP. Gravity .999 .886 .834
% Sulfur 6.4 2.8 0.2
Gasoline
Kerosene 6
Lt. Gas Oil 7 33
Hv. Gas Oils 26
20
16
Bitumen
Residuum
30
I
Figure 2-2 Make-up of crude petroleum (after Corbett, 1)
The API gravity for water is 10. Asphalts have an API gravity of
approximately 5-10, whereas the API gravity of gasolines is about 55.
Low API gravity crudes (API less than about 25) yield relatively low
percentages of distillable overhead fractions and high percentages of
asphalt cement. On the other hand, the high API gravity crudes (API more
than about 25) yield relatively high percentages of overhead fractions and
low percentages of asphalt cement. The low gravity crudes are generally
known in the industry as heavy crudes, or as sour crudes if they contain
high sulfur content. The high gravity crudes are referred to as light crudes,
or sweet crudes if they contain low sulfur content. Therefore, a refiner
must choose the crude type depending on the types and amounts of end
pro uct esire to e pro uce , w 1c may e ctate by the marRe--t_ _ _ _ __
demand. Each crude. is identified by name or source along with its API
gravity. Table 2-1 lists some crude types and typical sources.
Distillation. The first step in the processing of all crude petroleum is

Chapter 2-Asphalt Refining, Uses, and Propertie 11
straight reduction by distillation. Figure 2-3 shows the distillation princi-

ple which is used to separate various crude fractions which have different
boiling ranges. Because asphalt cement is primarily made up of the highest
boiling fractions, it becomes the residuum (also called "resid" in the indus-
try) from the vacuum tower. The crude oil is heated in a large furnace to
about 650°F and partially vaporized. Then it is introduced into
Table 2-1. Arbitrary Crude Types and Typical Sources

(after Corbett,~
Type API Source

A 34 Arabian Light, Saudi Arabia
32 Kuwait
B 28 Hawkins, USA
26 Tiajuana, Venezuela
C 19 Galan, Colombia
18 Cyrus, Iran
D 16 Lloydminster, Canada
15 Obeja, Venezuela
E 12 Panuco, Mexico
10 Boscan, Venezuela
a distillation tower where the lightest components vaporize, rise to the top,
cool, condense, and are drawn off for further processing. At various heights
in the tower different fractions reach their boiling point and then, as the
temperature is reduced, condense on trays inside the tower. The interme-
diate or mid-tower components are drawn off at those levels and are treat-
ed to make jet fuel, kerosene, and diesel. The residuum from this distilla-
tion is usually fed into a vacuum distillation unit where heavier gas oils are
produced. The reduced pressure (typically 55 mm Hg, 1.0 psi) in the vac-
uum tower assists in flashing these oils at lower temperatures to prevent
thermal cracking of the asphalt cement. In cracking, asphalt molecules are
thermally broken apart. Often steam is added to the bottom of the tower
-----~·o_fur.ther-reduce..the.-P-r:ess.ur:e-fr0m-5-0-mm-Hg-to about 12 mm Hg, 0.24
psi. The bottom fraction from this unit is called vacuum processed, steam
refined asphalt cement. The grade of asphalt cement, whether it is mea-
sured by penetration or viscosity, is controlled by the amount of heavy gas
oil removed. All of this is done by a continuous flow operation.
Figure 2-3 shows a typical range of distillation temperatures to obtain

different overhead fractions. The figure also shows a cut point 800-1050°F
(427-565°C) at which the asphalt residuum is fractionated from the over-
head fractions above it in the vacuum tower. The cut point is the atmos-
pheric equivalent vapor temperature to effect this fractionation.
60-325 F
Gasoline Treating Gasoline
300-400
Naphtha Reforming
325-500 Jet Fuel
Kerosine
Heating Oil
450-600
LI Gas Oil ~__racking
575-700
Hv Gas Oil
650-750
Crude •-.-J Voe Gas Oil
Heat Fract 700~850 Fuel Oil
Gas 011 or Blending
0.:::::J 800-1050 F cut Paint
Heat
Residuum
Paving Bitumen AC-2.5 to AC-4O
or
Solvent Process Precipitated Bitumen
or
Air Blowing Paving or Special
or
Alternate Process Coke Fuel
Figure 2-3. Crude Initially Separated by Distillation (after Corbett, 1)
Solvent Deasphalting. A vacuum residuum is subjected to solvent

deasphalting (SDA) to extract additional amounts of high boiling fractions
(deasphalted oil) for applications such as lube manufacture. Figure 2-4
shows a typical flow process. Extracting solvents generally used are propane
and butane. Vacuum residuum is contacted at 100 to 150°F (38 to 66°C)
with the solvent in a countercurrent fashion in an extracting (EXT) tower.
A high softening point, hard asphalt cement (precipitated asphalt)· is
obtained in this process. This hard asphalt can be used as a blending com-
ponent for producing paving grade asphalt cements. An excessive amount
of precipitated asphalt component is often detrimental because it increases
----athe-t~mperatuie-Suscep.tibilicy_oLth.e_bleJ:i.d~_d as12=h=al=t_;c=e=m::.::e:=:n:.;:;t•=-------------
/ Solvent Ex.traction (ROSE). A newer process called the residuum oil

supercritical extraction (ROSE) provides a greater flexibility in resid frac-
tion characteristics. The ROSE process is shown in Figure 2-5. Aresid feed
is mixed with a low-boiling hydrocarbon solvent (for example, normal pen-
Chapter 2-Asphalc Refining, Uses, and Properties 13
~ - - - ~ - - Solvent
Stripping
Vacuum Towers STA
Residuum
Extracting ,....__ _ _ _ _ Precipitated Bitumen

Tower 140-230 F Softening Point
..__---+------+-- Deasphalted Oil
Figure 2-4 Solvent Deasphalting (SDA) (after Corbett, 1)
tane) under supercritical conditions. This mixture is then fed into a sepa-
rator (SEP) at a predetermined controlled temperature (T1) and pressure,
to separate an asphaltene concentrate. Asphaltenes (first fraction) are then
recovered by stripping (STR) off the solvent used. The extract portion
from the preceding operation is then fed into a second separator where at
a controlled higher temperature (T2) a decrease in solubility results,
SEP Solvent
T, STA STA T
3
STA
Vacuum
·Residuum
Asphaltenes Resins Oils

(200-350. F SP)
SEP = Separator
STA = Stripping
SP = Softening Point
Figure 2.:_5. Supercritical extraction (ROSE) (after Corbett, 1)
causing the precipitation of resins (second fraction). Oils (third fraction)

are then similarly separated and recovered. Asphaltene or resin fraction
------£=ro=m=-.,,;th1=-e=-ROS-E process is used as a blenciing-component for asphalt
cements to meet the specification requirements.
Continuous Air-Blowing. Continuous air-blowing is normally used

when viscosity of vacuum resids must be increased as in producing roofing
7 I
asphalts, or when the temperature susceptibility of the paving asphalt

cement needs to be improved. A schematic for an air-blowing unit is
shown in Figure 2-6. The process involves the continuous pumping of a
vacuum resid (flux) through a heater where its temperature may be raised
to approximately 450 to 500°F (232 to 260°C) into an air-blowing tower.
Air is injected at the bottom of the resid and flows upward through the
resid. Since the reaction is exothermic (and therefore produces heat), water
is usually sprayed onto the surface to cool the asphalt. Under normal cir-
cumstances it is potentially hazardous to add water to hot asphalt cement
or vice versa because of the tremendous foaming problem. It can be done
only in specially designed and controlled processes like this.
Environmental considerations require the gases from the air-blowing
operation to be, sent to a scrubber where fumes are cooled and the oil is
skimmed off. Non-condensable vapors, which may be quite odorous, are
sent on to an incinerator. Roofing asphalts are manufactured with the
process of batch air-blowing. Continuous air-blowing, which involves a
SCRUBBER
INCINERATOR
TO OIL
SKIMMER
___.!!PHAL T FLUX--1-.RS"
250·300"F
AIR
BLOWER
BLOWN ASPHALT
Figure 2-6. Unit for Air-Blowing of Asphalt (after Hodgson, .i)
lesser degree of chemical conversion (oxidation), is normally used for

paving asphalt cement manufacture.
Several Manufacturing Options. Depending on the crude type and

available process capabilities, several manufacturing options are available to
produce asphalt cements. Although straight reduction to grade is the most
direct and most common practice, Figure 2-7 shows other options that
may be used. Blending of two or more crude feeds is quite common.

However, if these crudes are refined separately with one yielding a low vis-
cosity residuum and the other yielding a high viscosity residuum, the resids
could be blended in suitable proportions to meet the specification require-
ments. Also, a low viscosity resid could be blended with a precipitated
asphalt or air-blown asphalt to increase the viscosity. A high viscosity resid
could be blended back with a gas oil or a similar fraction to reduce the vis-
cosity. However, the selection of a process by the refinery is generally
dependent on the type of available crude(s) and the asphalt cement speci-
fication requirements. Hard and soft asphalt cements can also be blended
in an asphalt terminal to meet a specified grade.
CRUDEA-C-E
(OR MIX)
VAC Reduced to Grade

RESID
Resid Blending AC-2.5

to
AC-40
Blend with Ppt. Bitumen
Air Blow to Grade
Figure 2-7. Asphalt Cement Prepared by Several Methods (after

Corbett, 1J
ASPHALT TYPES AND USES
Asphalts most commonly U:sed in flexible pavement construction can

-e-divi:d·etl-into-rhree-ryp . -- - ---- - -
1. Asphalt cements;
2. Emulsified asphalts; and
3. Cutback asphalts.
Asphalt Cements
As discussed earlier, asphalt cement is obtained by the distillation

process from crude petroleum using different refining techniques. At ambi-
ent temperatures asphalt cement is a black, sticky, semisolid, and a highly
viscous material. It is a strong and durable cement with excellent adhesive
and waterproofing characteristics. It is also highly resistant to the action of
most acids, alkalies and salts. The largest use of asphalt cement is in the
production of Hot Mix Asphalt (HMA), which is primarily used in the
construction of flexible pavements throughout the world. The asphalt
cement can readily be liquefied by applying heat for mixing with mineral
aggregates to produce HMA. Being very sticky,_ it adheres to the aggregate
particles and binds them to form HMA. After cooling to ambient temper-
ature, HMA is a very strong paving material which can sustain heavy traf-
fic loads such as those experienced on interstate highways and heavy duty
airfields. Of the 2.3 million miles of paved roads in the United States, 2.2
million miles, or 96 percent, are surfaced with asphalt of some type.
In the literature, HMA has numerous synonyms. It has been referred
to as asphalt concrete, asphaltic concrete, asphalt cement concrete, asphalt
mix(ture), asphalt paving mix(ture), bituminous concrete, bituminous
mix(ture), bituminous paving mix:(ture), etc. In this textbook an attempt
has been made to use the term HMA as much as possible.
Several standard grades of asphalt cement; based on consistency, are
available commercially. Three methods based on penetration, viscosity or
performance are used to classify asphalt cements into different grades. The
penetration grading of asphalt cement is specified in ASTM D946 and is
primarily controlled by the penetration test. Five standard penetration
graded asphalt cements, 40-50, 60-70, 85-100, 120-150; and 200-300,
are used for paving grade asphalt. The penetration test and the penetration
grading system will be discussed in detail later in this chapter. It will suf-
fice to mention here that the penetration test involves measurement of
penetration of a standard needle into a sample of asphalt cement under
standard temperature, time, and load. Obviously, the higher the penetra-
tion, the softer the asphalt cement. Therefore, 40-50 is the hardest grade, '
and 200-300 is the softest grade.
+ - - - - - - ___ The second method of classifying asphalt cements is by viscosity grad-
ing, as specified in ASTM D3381. This grading is based on either the vi-s;-::--------
cosity of the original asphalt cement or on the viscosity of the asphalt
cement after aging in the rolling thin film oven (RTFO) test. Both vis-
cosities are measured at 140°F (60°C) and reported in poises. The viscosi-
ty grades based on original asphalt cements include AC-2.5, AC-5,
Chapter 2-Asphalt Re~ning, Uses, and Propertie 17
AC-10, AC-20, AC-30, and AC-40. The numerical values indicate vis-
cosity at 140°F (60°C) in hundreds of poises. The viscosity grades based on
the asphalt residue (AR) from the RTFO test include AR-1000, AR-2000,
AR-4000, AR-8000, and AR-16000. The numerical values indicate vis-
" cosity at 140°F (60°C) in poises. The viscosity tests and the viscosity grad-
ing systems will be discussed in detail later in this chapter.
The ,third method of classifying asphalt binders is performance based
and w:as developed in the Strategic Highway Research Program (SHRP). It
is discussed later under the section titled "Superpave binder tests and
specification."
Emulsified Asphalts
Emulsified asphalt (also simply called emulsion) is a mixture of asphalt
cement, water, and emulsifying agent. Because the asphalt cement will
ASPHALT CEMENT
IN FORM OF
PERFECT
SPHERES
WA TER---tl:!'.'"
CONTAINING
EMULSIFYINGt:=:...=.-=:=====-==-=-=~-t:.=..;:==:.:..==::L..J
AGENT
Figure 2-8. Emulsified Asphalt
not dissolve in water, asphalt cement and water exist in separate phases as
shown in Figure 2-8. Hot asphalt cement and water containing the emul-
sifying agent (soap is one example) are passed under pressure through a col-
loid mill to produce extremely small (less than 5-10 microns) globules or
------drop-I-ets-of-aspha:lt-cement-whieh-a:re-st1-spefl:cl:ed in water (Figure 2-9). The
emulsifying agent imparts an electric charge to the surface of the droplets
which causes them to repel one another, and thus the globules do not coa-
lesce. Emulsified asphalts are also categorized as liquid asphalts because,
unlike asphalt cements, they are liquid at ambient temperatures.
Emulsions are made to reduce the asphalt viscosity for lower application
temperatures.
Two most commonly used types of emulsified asphalts are specified in
ASTM D977 and ASTM D2397:
I. Anionic-electro-negatively charged asphalt droplets; and
2. Cationic-electro-positively charged asphalt droplets'.
If the emulsifying agent is anionic, the asphalt droplets bear a negative
charge. If the emulsifying agent is cationic, the droplets bear a positive
charge.
Most mineral aggregates bear a positive or a negative or mixed charge
on the surface. Most siliceous aggregates, such as sandstone, quartz and
siliceous gravel, are negatively charged and th;refore are generally compati-
ble with the positively charged cationic emulsified asphalts. On the other
hand, some aggregates such as limestone bear a positive surface charge and
are therefore generally compatible with the negatively charged anionic emul-
sified asphalts. This happens because opposite charges attract one another.
HOT
ASPHALT
TANK COLLOID MILL
:==-=====OPUMP
WATER
+
- -
STORAGE STORAGE STORAGE
Figure 2-9. Manufacture of Emulsified Asphalt
When an emulsified asphalt is mixed with an aggregate, it "sets" or

"breaks" because the asphalt droplets react with the surface of the aggregate
-----.
and-coalesce, squeezing out the water between them. The evaporation of
water is the primary method which finally causes the anionic emulsified
asphalt to "break'' or "set" and produce a continuous film of asphalt on the
aggregate or pavement. Cationic emulsified asphalts break primarily by
electro-chemical processes.
Both anionic and cationic emulsified asphalts are further graded

according to their "setting" rate. The anionic emulsified asphalts include
rapid setting (RS), medium setting (MS), and slow setting (SS) as specified
.in ASTM D977 and shown in Table 2-2. The setting rate is controlled by
the type and amount of the emulsifying agent. The anionic grades are:
RS-I, HFRS-2, RS-2, MS-I, HFMS-2, MS-2, MS-2h, SS-I, and
SS-Ih. The h designation means a harder base asphalt cement is used in
the emulsion. The HF designation refers to a high. float residue, which is
an indication of chemical gelling of the emulsion residue.
The cationic emulsified asphalts which are specified in ASTM D2397
and shown in Table 2-3, also include rapid-setting (CRS), medium-
setting (CMS), and slow-setting (CSS) grades. The cationic grades are
CRS-I, CRS-2, CMS-2, CMS-2h, CSS-I, and CSS-I h.
Selection and uses of emulsified asphalts are given in ASTM D3628.
Generally, they are used as follows:
I. Rapid-Setting Grades: Surface treatments and penetration macadams;
2. Medium-Setting Grades: Open-graded cold asphalt-aggregate
mixtures; and
3. Slow-Setting Grades: Tack coat, fog seal, dense-graded cold
asphalt-aggregate mixtures, and slurry seals.
Cutback Asphalts
.Cutback asphalts are liquid asphalts which are manufactured by

adding (cutting back) petroleum solvents (also called cutter stock or dilu-
ent) to asphalt cements. They are made to reduce the asphalt viscosity for
lower application temperatures. Application to aggregate or pavement
causes the solvent to escape by evaporation, thus leaving the asphalt
cement residue on the surface. Based on the relative rate of evaporation,
cutback asphalts are divided into three types:
I. Rapid-Curing (RC)-produced by adding a light diluent of high
volatility (generally gasoline or naphtha) to asphalt cement. These
are used primarily for tack coat and surface treatments.
2. Medium-Curing (MC)-produced by adding a medium diluent of
intermediate volatility (generally kerosene) to asphalt cement. These
are generally used for prime coat, stockpile patching mixtures, and
---------rc-,o--,-a➔
a-_-m,..,1xi,,.,,-,,n_g_o_p_e-r~ations.
3. Slow-Curing (SC)-produced by adding oils oflow volatility (gen-

erally diesel or other gas oils) to asphalt cement. They are also called
road oils. They are generally used for prime coat, stockpile patching
mixtures, and as dust palliatives.
Table 2-2. SeL I
Requirements fur AnionkEmulsi~ A.phalu ~
Tests Rapid-Setting Medium-Setting

I Slow-Setting
I
RS-1 RS-2 HFRS-2 MS-1 MS-2 MS-2h HFMS-1 HFMS-2 HFMS-2h HFMS-2s SS-1
I SS-lh
Tests on emulsionJ
Viscosity, Saybolt 20-100 - - 20-100 100+ 100+ 20-100 100+ 100+ 50+ 20-100 20-100
Furol at 77°F
iB:::
Viscosity, Saybolt - 75-400 75-400 - - - 20-100 100+ 100+ 50+ 20-100 20-100 ,l"
Furol at 122°F
Minimum Residu. 55 63 63 44 65 65 55 65 65 65 57 57
i
by distillation, o/o
Tests on residue
from distillation:
t
lf -
~
t)
Penetration at 77' F, 100-200 100-200 100-200 100-200 100-200 40-90 100-200 100-200 40-90 200+ 100-200 40-90
100 g, 5s ~-
l
Aooua,, I40"F, r - - 1200 - - - 1200 1200 1200 1200 - -
f
j.
::,
Table 2-3. Selected Requirements for Cationic Emulsified Asphalts
Tests Rapid-Setting Medium Setting Slow-Setting
CRS-1 CRS-2 CMS-2 CMS-2h CSS-1 CSS-lh
Tests on emulsions:
Viscosity, Saybolt
Purol at 77°P - - - - 20-100 20-100
Viscosity, Saybolt
Purol at 122°P 20-100 100-400 50-450 50-450 - -
Distillation:
Oil distillate, % 3 3 12 12 - -
Minimum residue,% 60 65 65 65 57 57
Tests on residue
from distillation test:
Penetration at 77°P, 100-250 100-250 100-250 40-90 100-250 40-90

100 g, 5s
Cutback asphalts are commercially available iri different grades as

shown in Figure 2-10. The thinnest and most fluid grade is designated by
the suffix number 30, which is available in MC type only. Other suffix
numbers, 70, 250, 800, and 3000, denoting increasingly higher viscosities,
are available in all types. These suffix numbers represent the minimum
kinematic viscosity in centistokes at I 40°F for the particular grade.
Specifications for SC, MC, and RC type cutback asphalts are given in
ASTM D2026, ASTM D2027, and ASTM D2028, respectively. Some
specification requirements for RC and MC type cutback asphalts are given
in Tables 2-4 and 2-5, respectively.
Emulsified asphalts are increasingly being used in lieu of cutback
asphalts for the following reasons:
1. Environmental regulations. Emulsions are relatively pollution free.

-------------"-nlike cufb-ack:-a:sph-a±trth:-ere--are relatively small amounts of
volatiles to evaporate into the atmosphere other than water.
2. Loss of high energy products. When cutback asphalts cure, the dilu-
ents which are high energy, high price products are wasted into the
atmosphere.
Table 2-4. Selected Requirements for Rapid-Curing (RC) Cutback

Asphalts
Tests RC-70 RC-250 RC-800 RC-3000
Kinematic viscosity
at 140°F(60°C),cSt 70-140 250-500 800-1600 3000-6000
Flash Point (Tag Open-cup)

°F(0 C), minimum - 80(27) 80(27) 80(27)
Residue from distillation

to 680°F(360°C),
percent by volume,
minimum 55 65 75 80
Tests on residue from

distillation:
Viscosity at 140°F(60°C), P 600-2400 600-2400 600-2400 600-2400
Ductility at 77°F(25°C),
cm, minimum 100 100 100 100
Table 2-5. Selected Requirements for Medium-Curing (MC) Cutback

Asphalts
Tests MC-30 MC-70 MC-250 MC-800 MC-3000
Kinematic viscosity
at 140°F(60°C),cSt 30-60 70-140 250-500 800-1600 3000-6000
Flash Point (Tag Open-cup)

°F(°C), minimum 100(38) 100(38) 150(66) 150(66) 150(66)
Residue from distillation

to 680°F(360°C), percent
by volume, minimum 50 55 67 75 80
11
Tests on residue
from distillation:
Viscosity at 140°F(60°C), P 300-1200 300-1200 300-1200 300-1200 300-1200
i Ductility at 77°F(25°C),
cm,minimum 100 100 100 100 100
Ii
II
3. Safety. Emulsions are safe to use. There is little danger of fire as

compared to the cutback asphalts, some of which have very low
flash points.
4. Lower application temperature. Emulsions can be applied at rela-
tively low temperatures compared to cutback asphalts, thus saving
"fuel costs. Emulsions can also be applied effectively to a damp pave-
ment, whereas dry conditions are required for cutback asphalts.
GRADE 30 70 250 800 3000

NAPTHA FOR RC { {
KEROSENE FOR MC SOLVENT
LIGHT OIL FOR SC
ASPHALT CEMENT
APPROX. FUROL VIS., 140" F, SEC., 15-30 35-70 125-250 400-800 1500-3000
APPROX. KINEMATIC VIS. 140"F, CS 30-60 70-140 250-500 800-1600 3000-6000
Figure 2-10. Composition of Cutback Asphalts
ASPHALT CEMENT PHYSICAL TESTS

Physical testing of asphalt cements can be categorized as follows:
1. Consistency tests;
2. Durability tests;
3. Purity tests;
4. Safety tests; and
5. Other tests.
Consistency Tests
Consistency describes the degree of fluidity of asphalt cement at any
particular temperature. Since asphalt cement is a thermoplastic material, its
consistency varies with the temperature. Therefore, it is necessary to mea-
sure the consistency of different asphalt cements at the same temperature
and shear loading conditions if comparisons are to be made. The following
consistency tests are performed on asphalt cements:
--------Ahsolute-V-iseasity-at-1-40~F--(60"C). Viscosity can simply be defined

as resistance to flow of a fluid. Viscosity grading of asphalt cements is based
on viscosity measurement at 140°F (60°C). This temperature was selected
because it approximates the maximum Hl\1A pavement surface tempera-
ture during summer in the United States.
A capillary tube viscometer is used to perform the viscosity test at

140°F (60°C). Two viscometers, the Cannon-Manning vacuum viscometer
(Figure 2-11) and the Asphalt Institute vacuum viscometer (Figure 2-12),
are commonly used. ASTM D2171 test method describes the test proce-
dures. The viscometer is mounted in a thermostatically controlled; con-
stant temperature water or oil bath which is maintained at 140°F (60°C).
The viscometer tube is charged with asphalt cement through the large side
until the level of asphalt cement reaches the filling line. After the filled vis-
cometer tube is kept in the bath for a prescribed period of time to obtain
TIMING
MARKS
FILLING
LINE
Figure 2-11. Cannon-Manning Vacuum Viscometer

the equilibrium temperature of 140°F (60°C), a partial vacuum is applied
to the small side of the viscometer tube to cause the asphalt cement to flow.
Application of partial vacuum is necessary because the asphalt cement is
too viscous to flow at this temperature. A vacuum control device and a vacuum
pump are needed as part of the testing equipment. After the asphalt
Iii Figure 2-12. The Asphalt Institute Vacuum Viscometer
I,i1
cement starts to flow, the time (in seconds) required for it to flow between
two timing marks is measured. The measured time (in seconds) is multi-
plied by the calibration factor for the viscometer tube to obtain the value
for viscosity in poises, which is the standard unit for measuring viscosity.
Manufacturers calibrate their viscometer tubes with standard oils and fur-
nish the calibration factor with each tube.
When testing a material such as asphalt cement of an unknown vis-
cosity; the shear stress is the same as that when calibrating the viscometer
tube. The only difference is the measured rate of shear, which
is inversely proportional to the time required to pass a liquid through the
tube. The following relationship can be used to determine the viscosity:
or
V=T(½)
2 2 T
1
where,
VI viscosity of standard material;
Tl time for standard material to pass through the tube;
vz viscosity of unknown material;
Tz time for unknown material to pass through the same
tube; and
V1
calibration factor for tube.
TI
Kinematic Viscosity at 275°F (135°C). ASTM D2170 is used to mea-
sure the kinematic viscosity at 275°F (135°C) using a Zeitfuchs
Cross-Arm viscometer (Figure 2-13). At this temperature the asphalt
cement is sufficiently fluid to flow through the capillary tube under gravi-
tational forces alone, and there is no need to apply any partial vacuum.
This temperature was selected because it approximates the mixing and lay-
down temperatures used in the construction of HMA pavements.
Asphalt cement is poured into the large opening of the viscometer
--------u-1'.l,til-it-r.eaches...the filling line Th.,._viscometer is then placed in a clear oil
bath medium to obtain the equilibrium temperature. A slight vacuum is
applied to the small opening or a slight pressure is applied to the large
opening to induce the flow of the asphalt cement over the siphon section
just above the filling line. Then the gravitational forces cause the asphalt
cement to flow downward in the vertical section of the capillary tube. A

stopwatch is used to measure the time (in seconds) required for the asphalt
cement to flow between the two timing marks. The kinematic viscosity in
centistokes is obtained by multiplying this time by the calibration factor
FILLING
LINE
ASPHALT
TIMING
MARKS
Figure 2-13. ~itfuchs Cross-Arm Viscometer
supplied with the viscometer tube. The unit of centistokes is used in kine-
matic viscosity at 275°F (135°C) because gravitational forces induce flow
and the density of the material affects the rate of flow. Absolute viscosity
(such as obtained at I 40°F or 60°C) can be calculated from kinematic vis-
cosity by multiplying it by the density of asphalt cement.
Example 2-1: An asphalt cement has a kinematic viscosity of 800 centis- ·
tokes. Its specific gravity is 1.03. What is its absolute viscosity in poises?
1 stoke = 100 centistokes
Absolute viscosity in poises = kinematic viscosity in stokes x

specific gravity
= 8 X 1.03
= 8.24
Penetration. The penetration test is an empirical test used to measure

the consistency of asphalt cement. Usually penetration is measured at 77°F
(25°C) which also approximates average service temperature of the HMA
pavements. At this temperature no simple test method is currently avail-
able to measure the consistency of asphalt cement in terms of viscosity,

although viscosity is the preferred form of measurement. The standard
penetration test is shown in Figure 2-14. A container of asphalt cement is
brought to the standard test temperature (usually 77°F or 25°C) in a ther-
mostatically-controlled water bath. The sample is placed under a needle of
prescribed dimensions. The needle is loaded with a 100 g weight and is
allowed ·to penetrate the asphalt cement sample for 5 seconds. The depth
of penetration is measured in units of 0.1 mm (dmm) and is reported as
penetration units. For example, if the needle penetrates 8 mm, the pene-
tration of asphalt cement is 80.
BEFORE AFTER 5 SECONDS

Figure 2-14. Penetration Test
The penetration test is run in accordance with ASTM D5. The test
can also be run at other temperatures such as 32, 39.2, and l15°F (0, 4,
and 46°C). However, the needle load or penetration time or both are then
varied. At low temperatures, such as 39.2°F (4°C), the asphalt cement is
very stiff and therefore penetration of the needle is significantly lower if it
is loaded with a 100 g weight and is allowed to penetrate for 5 seconds. At
these temperatures, a 200 g weight and 60 seconds penetration time have
been used to obtain penetration values.
-------Sa-ft-ening-Point.-£0ftmi-ng-p0-i-nt-is measured by ring and ball (R &

B) method in accordance with ASTM D36. It can simply be defined as the
temperature at which an asphalt cement cannot support the weight of a
steel ball and starts flowing. Its purpose is to determine the temperature at
which a phase change occurs in the asphalt cement. Although its use in
specifying paving asphalt cement is quite common in Europe, it is mostly

used in the United States for high viscosity roofing asphalts.
As shown in Figure 2-15, the test consists oftaking a brass ring filled
with asphalt cement and suspending it in a beaker filled with water or ethy-
lene glycol. A steel ball of specified dimensions and weight is placed in the
center of the sample. The bath is heated at a controlled rate of 5°C/minute.
When the asphalt cement softens, the ball and asphalt cement sink toward
the bottom of the beaker. The temperature is recorded at the instant when
the softened asphalt cement sinks the prescribed distance and touches the
bottom plate. The test is conducted with duplicate specimens. If the differ-
ence between the two temperatures exceeds 2°F, the test must be repeated.
Thermometer
J-s,eelBall
~Sample
Figure 2-15. Softening Point Test
Ductility Test. Many asphalt paving technologists consider ductility

as an important property of asphalt cement. Its significance will be dis-
cussed later in this chapter. The ductility test, run in accordance with
ASTM Dl13, measures the distance in centimeters that a standard bri-
quette of asphalt cement will stretch before breaking (Figure 2-16). The
cross section of the briquette at its smaller dimension is one square cen-
timeter. The test sample is brought to temperature in a water bath which
is maintained at the standard temperature of 77°F (25°C). The two ends
---~of the sample are separated at the rate ot5 cm/mmute until rupture. Tn:t:--=-e---~--
water must be at the same specific gravity as the asphalt to prevent floating
or sinking of the stretched sample. Salt is added to increase the specific
gravity, and alcohol is added to decrease the specific gravity of the water.
Some engineers believe in running the ductility test at lower temperatures
such as 60 and 39.2°F (16 and 4°C) rather than 77°F (25°C). The pulling
rate is usually lowered to 1 cm/minute when the test is run at 39.2°F (4°C)
to allow for more stretching prior to break.
Durability Tests
Asphalt cements undergo substantial short-term aging (hardening)
when they are mixed with hot mineral aggregates in a HMA mixing facil-
ity. Their long-term aging continues during the life of HMA pavements
which ·are subjected to environmental and other factors discussed later in
this chapter. The following two tests are generally used to approximate the
short-term hardening conditions which occur in normal HMA mixing
facilities:
Centimeter Scale
Asphalt Sample
Figure 2-16. Ductility Test
Thin Film Oven Test. The thin film oven (TFO) test is conducted by
placing a 50 g sample of asphalt cement in a cylindrical flat-bottom pan
(5.5 inches or 140 mm inside diameter and 3/8 inch or 9.5 mm deep). The
asphalt cement layer in the pan is about 1/8 inch (3 mm) deep. The pan
containing the asphalt cement sample is transferred to a shelf in a ventilat-
ed oven maintained at 325°F or 163°C (Figure 2-17). The shelf rotates at
5 to 6 revolutions per minute. The sample is kept in the oven for 5 hours,
------arrd-rhen--transfer-red.-t-0-a-s-1:1:i1;al,le-e0fl-1Ca:ine-F-feF measuring penetration or
viscosity of the aged asphalt cement. The test method is described in
ASTM D 1754. The aged asphalt cement is usually required to meet spec-
ified minimum percentage retained penetration or maximum viscosity. A
loss or gain in weight of the test sample is also measured and reported.
325•F
5.J.• diameter x j• deep

sfiallow pans
Figure 2-17. Thin Film Oven Test
Rolling Thin Film Oven Test. A variation of the thin film oven test
is in use by agencies in the western United States, although it has the same
purpose. Figure 2-18 shows the rolling thin film oven (RTFO) used in
ASTM D2872. A specified amount of asphalt cement is poured into a bot-
tle, which is placed in a rack in the oven maintained at 325°F (163°C). The
FAN
MOTOR
1725 RPM
SENSOR
Figure 2-18. Rolling Thin Film Oven Test

rack rotates at a prescribed rate around a horizontal axis. The rotating bot-
tle continuously exposes fresh asphalt. The orifice of the sample bottle
passes in front of an air jet during each rotation. The vapors accumulated
in the sample bottle are purged by the heated air from the jet.
The RTFO achieves about the same degree of hardening (aging) as the
TFO but in less time (only 75 minutes). Also, it can accommodate a larg-
er number of samples .than the TFO. Some western states use the viscosity
of the aged residue (AR) of the asphalt cement from this test to grade their
paving asphalt cements.
Purity Tests
Refined asphalt cements consist of almost pure bitumen which, by def-
inition, is entirely soluble in carbon disulfide. Only very little amounts of
impurities are generally present in refined asphalt cements. To determine
the purity of asphalt cement, a solubility test (ASTM D2042) is conduct-
ed. An asphalt cement sample of known weight is dissolved in trichloroeth-
ylene (rather than carbon disulfide which is flammable and presents a safe-
ty hazard), and is then filtered through a glass fiber pad. The insoluble
material retained on the pad is washed, dried and weighed. The insoluble
material constitutes the impurity in the asphalt cement sample.
Specifications for paving asphalt cements generally require a minimum of
99.0 percent solubility in trichloroethylene. It is also desirable that the
.asphalt cement is free of moisture because it can cause the asphalt cement
to foam when it is heated above 212°F (I00°C).
Safety Tests
If asphalt cement is heated to a high enough temperature, it gives off
enough vapor to flash (ignite) in the presence of a spark or open flame.
Flash point indicates the temperature to which asphalt cement may be safe-
ly heated without the danger of instantaneous fl.ash in the presence of an
open flame. This temperature is below that of the fire point, which is the
temperature where a material combusts. Although the flash point of paving
asphalt cements is well above the temperatures normally used in HMA pro-
duction, it is necessary to measure and control it for safety considerations.
The Cleveland Open Cup method (ASTM D92) is generally used for
---determinmg-tl1:e~flash-p0i-rtt--e,f'-pav-i-ng-aspha:l-t-cements- (Figure-2-19). A
brass cup partially filled with asphalt cement is heated at a prescribed rate.
A small flame is passed over the surface of the cup periodically. The tem-
perature at which sufficient vapors are released to produce an instanta-
neous flash is reported as the flash point.
Thermometer
Figure 2-19. Cleveland Open Cup Flash Point Test
Other Tests
Specifi~ Gravity. Specific gravity is defined as the ratio of the mass of

the material at a given temperature to the mass of an equal volume of water
at the same temperature. Specific gravity of the asphalt cement changes
when the asphalt cement expands on heating. Therefore, specific gravity
determinations are useful in making temperature-volume corrections or
determining the weight per unit volume of asphalt cement heated to its
application temperature. ·
The pycnometer method is used to determine the specific gravity of
asphalt cements (ASTM D70). Since the specific gravity varies with the
temperature, test results are expressed in terms of specific gravity (Sp. Gr.)
at a given temperature for both the asphalt cement and the water used in
the test. For example, Sp. Gr. 1.02.at 60°/ 60°F (15.6°/15.6°C) means that
the asphalt cement specific gravity is 1.02 when both the asphalt cement
and the water are at 60°F (15.6°C). Specific gravity at 60°/60°F
----1•r,.-6°1T5:6--o-q-is normallyusea for making monetary corrfpensationro..------
transactions concerning asphalt cements.
Spot Test. The purpose of the spot test is to determine whether an

asphalt cement has been damaged during processing due w overheating,
resulting in cracking. In cracking, the molecules are thermally broken

apart. According to some engineers, the asphalt cement thus produced is
inferior and is more susceptible to weathering or aging. Since cracking is
not likely to occur in the current refining practices, the spot test is usually
not a requirement in the specifications. Also, cracked asphalt cements tend
to have poor ductilities and high age hardening rates.
This' test, developed by Oliensis, a chemist, is a crude form of paper
ch_romatography in that it involves a visual evaluation of a spot of asphalt
cement dissolved in a standard solvent (such as naphtha). A drop of the
asphalt/solvent mixture is placed on a special filter paper and the results
observed visually. If the spot formed is a uniformly brown stain, the test is
reported as negative and the asphalt cement is acceptable. However, i£
within the spot, there is a darker area at the center, the test is reported as
positive and will disqualify the asphalt. The significance of the spot test is
still in dispute. The spot test cannot be applied to an asphalt cement which
has been extracted (recovered) from HMA.
ASPHALT CEMENT GRADING SYSTEM
Historical Background
Halstead and Welborn have given a very good account of the history
of the development of asphalt specifications in Reference Q.
From 1878 to 1900 most of the asphalt cement used in the construc-
tion of sheet asphalt, penetration macadam, and mixed macadam (now
called asphalt concrete) was obtained from the Trinidad Lake. The Barber
Asphalt Paving Company was the primary supplier ofTrinidad Asphalt and
builder of asphalt pavements since the beginning of 1878. Bermudez Lake
asphalt found in Venezuela came into use during the 1890s and was a
major source by 1901. Like the Trinidad Lake asphalt, the Bermudez
asphalt had to be mined, and water and extraneous inorganic and organic
matter had to be removed. Both Lake asphalts were found to have remark-
able uniformity.
The first specification for asphalt cement in the United States was
------basecl-0n-t:he-app€a-i:a-11-G€----0-£.th€----Gr-114e-Tri.nidad.. asphalt and.on analytical
tests to determine amounts of bitumen (soluble in carbon disulfide) and
insoluble organic and inorganic matter. Such specifications were devised
merely to identify the source of asphalt cement at the exclusion of other
source materials. As the HMA industry grew and petroleum asphalts came
into use in the United States about 1900, it became evident that the ele-
ment of uncertainty in material requirements, especially the consistency,
m'ust be removed. The native asphalts were solid or semi-solid materials,
and it was necessary to add fluxing materials to reduce their consistency for
use as binders in HMA mixtures. Standardized methods were needed to
determine the consistency of the paving binders obtained from different
sources.
In 1888, H.C. Bowen invented the Bowen Penetration Machine, the
forerunner of the penetrometer, to determine consistency of the asphalt
cement. Previous to Bowen's invention, the method of testing the proper
degree of softening of the asphalt cement was by chewing. Even after the
invention of the Penetration Machine, the chewing method, crude as it
may now appear, served as a valuable check.An asphalt man generally prid-
ed himself on the fact that by chewing he could achieve approximately the
same results as those obtained by the machine (Q).
After several modifications of the initial Penetration Machine, by 1910
the penetrometer became the principal means of measuring and control-
ling consistency of semi....:solid asphalts at 77°F (25°C). The Bureau of
Public Roads (now the Federal Highway Administration), since 1901, and
the American Society for Testing and Materials (ASTM), since 1903, have
been instrumental in making the penetration test a standard for control-
ling the consistency of paving asphalt cements.
In 1918, the Bureau of Public Roads introduced the penetration grad-
ing system by developing various penetration grades suited to different cli-
matic conditions and applications. The American Association of State
Highway Officials (MSHO) published the standard specifications for
penetration graded asphalt cements in 1931.
The next major change in asphalt cement grading specifications was
initiated in the early 1960s by the FHWA, ASTM, MSHTO, industry,
and a number of state highway departments that wan~ed the asphalt
cements to be graded by viscosity at 140°F (60°C). The primary objectives
were: (a) to replace the empirical penetration test with a rational scientific
viscosity test, and (b) to measure the consistency at 140°F or 60°C (rather
than 77°F or 25°C), which approximates the asphalt pavement maximum
surface temperature on a hot summer day in most of the United States.
Various viscosity grades were developed to suit the different climatic con-
mons an app 1cat10ns. ---·
During the early 1960s when the viscosity grading system was being
developed, the California Department of Highways was developing a par-
allel aged residue (AR) viscosity grading with the cooperation of the Pacific
Coast User Producer Group. AR viscosity grading was based on the vis~
Chapt;r 2-Asphalt Relining, Uses, and Properties 35
cosity of the aged residue resulting from the rolling thin film oven (RTFO)
test rather than the viscosity of the original asphalt cement as recommend-
ed by the FHWA. The California Department of Highways had experi-
enced mix setting problems (tender mix) with some asphalt cements whose
original viscosity did not increase during the plant mixing phase as much
as that of other asphalt cements. Therefore, they wanted to grade the
asphalt after aging in the RTFO so that all asphalts after mixing would
behave about the same during construction.
Figure 2-20 shows the grading systems used in the U.S. during 1985
(.Z). Several states were using more than one grading system. A detailed dis-
cussion of the three grading systems follows. Performance based asphalt
binder grading system developed in the Strategic Highway Research
Program (SHRP) will be presented later.
C.=:J AC Grading
- Penetration Grading
AR Grading
More Than One Method
Figure 2-20. Grading Methods for Asphalt Cements Used during

1985 (after Terrel et al, 2)
Table 2-6 gives the requirements for penetration graded asphalt

------eements-a-s-fo1:1-aEl-i£-A-£+-M--D~46.-Fiv-e-penetration grades are specified:
40-50, 60-70, 85-100, 120-150, and 200-300. As mentioned earlier,
one penetration unit equals one tenth of a millimeter (0.1 mm). The lower
the penetration, the "harder" the asphalt. The_ "softest" grade is 200-300,
which is used in cold climate areas such as Canada. Two grades, 60-70 and.
85-100, are most commonly used for paving in the U.S. Other require-
ments, such as ductility at 77°F (25°C) and retained penetration after thin
film oven (TFO) test, are also listed in Table 2-6.
Table 2-6. ASTM Requirements for Penetration Graded Asphalt

Cements
Penetration Grade
40-50 60-70 85-100 120-150 20·0-300
Min Max Min Max Min Max Min Max Min Max
Penetration at 77°F (25°C),
100 g, 5 s 40 50 60 70 85 100 120 150 200 300
Flash point, °F
(Cleveland open cup) 450 450_ - 450 425 350
Ductillg at 77°F
(25° , 5 cm/min, cm 100 100 100 100 100
Solubility in
trichloroethylene, % 99.0 99.0 99.0 99.0 99.0
Retained penetration after
thin-film oven test, % 55+ 52+ 47+ 42+ 37+
Ductili% at 77°F (25°C),
5 cm min, after thin-film 50 75 100 100* -
oven test, cm
*If ductility at 77°F (25°C) is less thanl00 cm, material will be accepted if ductility at 60°F (15.5°C)
is 100 cm minimum at the pull rate of 5 cm/min.
The penetration grading system has the following advantages:
1. Grading is based on the consistency of the asphalt cement at 77°F

(25°C), which is close to the average pavement service temperature.
Testing at 77°F (25°C) may provide a better correlation with low
temperature properties than the viscosity test whie;h is measured at
140°F (60°C).
2. Testing time is relatively short.
3. Penetration testing is adaptable to field conditions. It can be used by
the HMA contractor to detect contamination of the asphalt cement
(usually from other petroleum distillates or asphalt cutbacks in the
transport units). '
4. Equipment costs are relatively low.
5. Precision limits for the penetration test are well established.
6. Temperature susceptibility (change in asphalt cement consistency
·----with-co-rresp·o·u-ding-change-in-rem:pera:mre7-of-the-a:sphal-t-cement·-----
can be determined by measuring penetration at temperatures other
than 77°F (25°C). Temperature susceptibility is discussed in detail
later.
However, the penetration grading system has the following disadvan-

tages:
1. Penetration is an empirical test and does not measure the consisten-

cy of asphalt cement in fundamental units such as viscosity.
2. Shear rate is high during the test.
3. Shear rate is variable because it depends on the consistency of the
asphalt cement.
4. Similitude at 77°F (25°C) can be deceptive to performance at high-
er and lower service temperatures (Figure 2-21).
5. No viscosity is available to establish mixing and compaction tem-
peratures.
Viscosity Grading System

This system, which is based on the .viscosity of original (as supplied)
asphalt cement, is most widely used in the United States. Table 2-7 gives
the requirements for asphalt cement viscosity graded at 140°F (60°C) as
found in ASTM D3381. ASTM Tables 1 and 2 pertain to the following
viscosity graded asphalt cements: AC-2;5, AC-5, AC-10, AC-20,
A - Low Temp. Susceptibility

Asphalt
B - Medium Temp. Susceptibility
Asphalt
C - High Temp. Susceptibility
Asphalt
Cl)
Cl)
w
z
LL
A
LL
77
TEMPERATURE (°F)
Figure 2-21. Temperature Susceptibility of Penetration Graded
Asphalt Cements
38 Hot Mix Asphalt Materials, Mixture Design and Consuuction
Table 2-7 ASTM Requirements for Viscosity Graded Asphalt Cements

(D 3381-83)
TABLE 1 Requirements for Asphalt Cement, Viscosity Graded at 140°F {60°C)
NOTF.-Grading based on original asphalt.
Viscosity Grade
Test
AC-2.5 AC-5 AC-10, AC-20 AC-40
Viscosity. 140°F {60°C), P 250: 50 500: 100 1000 :t:200 2000 :t 400 4000 ± 800
Viscosity, 275°F (135°C), min, cSt 80 110 150 210 300
Pene1rat1on, 77°F (25°C), 100 g, 5 s. min 200 120 70 40 20
Flash pomt. Cleveland open cup, min, °F ( C)
0
325 (163) 350 (177) 425 {219) 450(232) 450 (232)
Solubility in trichloroethylene, min,% 99.0 99.0 99.0 99.0 99.0
Tests on residue !rem thin-film oven test:
Viscosity, 140°F (60°C). max, P 1250 2500 5000 10 000 2o ooo
Ductility, 77°F (25°C), 5 cm/min, min. cm 1QQA 100 50 20 10
"'II ductiUty is less than 100, material win be accepted if ductility at 60°F (15,5°C) is 100 minimum at a pull rate ol 5 cm/min,
TABLE 2 'Requirements for Asphalt Cement Viscosity Graded at 140°F (60°C)

NOTE-Grading based on original asphalt.
Viscosity Grade
Test
AC-2.5 AC-5 AC-10._ AC-20 AC-30 AC-40
Viscosity, 140°F (60°C), P 250 :!: so 500:: 100 1000:: 200 2000:: 400 3000:: 600 4000:: 8Ci0
Viscosity, 275°F (135°C). min, cSt 125 175 250 300 350 400
Penetration. TT°F (25°C), 100 g. 5 s, min 220 140 80 60 50 40
Flash point, Cleveland open cup. min, °F (°C) 325 (163) 350{177) 425(219} 450(232) 450(232) 450 (232)
Solubility in triehloroethylene. min, "o 99.0 99.0 99.0 99.0 99.0 99.0
Tests on rP.s1due from th1n-f1lm oven test
Viscosity, 140°F (60°C), max, P 1250 2500 5000 10 000 15 000 20 000
Duct1hty, 77°F (25°C), 5 cm/min. min. cm 100 .. 100 75 50 40 25
,., If ductility 1s less than 100. mat~rial will be accepted if ductility at 60°F (15.5°C) 1s 100 m1n1mum a1 a pull rate of 5 cm/min.
T~BLE 3 Requirements for Asphalt Cement Viscosity Graded at 140°F (60°C)

NOTE-Grading b:ised on residue lrom rollin~ lh1n-lilm ovcm tc~r.
V1!;ce:;11y Cr,,cJc
Tests en Residue from Rolling Thin-Film Oven Test:"
AR,1000 AR-2000 AR--4000 AR,8000 AR-16000
Viscosity. t40°F (60°C),· P 1000 :<:: 250 2000:.':500 4000 ~ 1000 8000: 2000 16000: 4000
Viitcos1ty, 275°F (135°C), min. cSt 1•10 200 275 400 550
~~:1~!~:~\~~~ ~2s50
1 65 40 25 20 20
=e~t~~~ 1~;j ~).m~fn 40 45 50 S2
D1.1cti!ity, 77°F (25°C), 5 cm/min, min, cm ~00 8 100 8 75 75 75
Testsononginal asphalt:
Flash point. Cleveland open cup, min. °F (0 q 400 (205) 425(219) 440 (22i} 450(232) 460(238)
Solub1l1ty in trichloroethy!ene, min, 't. 99.0 99.0 99.0 99.0 99 0
"Thin-film even test may be used but the rolling thin-film oven test shall be the referee method.
8 lf ductility is less than 100, matenal wm be accepted II ductility at 60°F (15.5°C) is 100 minimum at a ;,1.111 rate of 5 cm/min.
AC-30, and AC-40. ASTM Table 3 pertains to the AR (aged residue) vis-
cosity grading system, which is discussed later. AASHTO Specification
M226 on viscosity graded asphalt cements is similar to ASTM D3381 with
minor differences.
The poise is the standard unit of measurement for viscosity. The lower
the number of poises, the less viscous the asphalt cement. AC-2.5 (asphalt
cement with a viscosity of 250 poises at 140°F or 60°C) is "softer" than
AC-40 (asphalt cement with a vis~osity of 4000 poises at the same tem-
perature). AC-20 grade is most commonly used for paving in the United
----5tates;-AG=3fl-grade-wa-s---reeently-added-to-the-specifi:cations-and-is-tlsed-by------
some southeastern states. Other test requirements such as penetration at
77°F (25°C), viscosity at 275°F (135°C), viscosity at 140°F (60°C) and
ductility at 77°F (25°C) of the residue from thin film oven test (TFOT),
and flash point are also given in the specifications. The penetration at 77°F
(25°Cj controls the consistency of the asphalt cement near the average ser-
vice temperature, and the viscosity at 275°F (135°C) controls its consis-
tency near the mixing and compacting temperatures. Together these spec-
ified requirements also control the temperature susceptibility of the asphalt
cement as shown in Figure 2-22. The primary difference between ASTM
Tables 1 and 2 (within Table 2-7) is that the asphalt cements in Table 2 are
less temperature susceptible because the specified minimum penetration
values at 77°F (25°C) and the minimum viscosity values at 275°F (135°C)
are higher than those in Table 1.
LOW TEMP. AVERAGE HOT MIXING-

CRACKING SERVICE SUMMER COMPACTION
TEMP.
!
U)
CJ)
I
I
I
I
w
z :,-60 PEN
LI..
LI..
i=
~
>-
I-
U)
0
u
CJ) 1600 T
> I
I
I
I
I
I
I
I
-10 77 140 275

TEMPERATURE ' F -
Figure 2-22. Temperature Susceptibility of Viscosity Graded Asphalt

Cements
The viscosity grading system has the following advantages:

1. Viscosity is a fundamental property, rather than an empirical test
and therefore is independent of the test system and the sample size.
2. It is suitable to a wide range of environments (pavement tempera-
tures 77 to 140°F or 25 to 60°C).
3. It is based on viscosity at 140°F (60°C) which is near the maximum
pavement surface temperature generally experienced in the United
States. This temperature is critical for pavement performance dur-
ing hot summer days.
4. There is a reduced overlap with other grading systems (Figure

2-23).
5. A wide range of test instruments is. available.
6. Test standards are available with established precision limits.
7. Temperature susceptibility of the asphalt cement can be determined
since the consistency is measured at three temperatures.
The disadvantages of the viscosity graded system are listed below:
I. Grading at 140°F (60°C) is deceptive to performance at average or

low service temperatures (Figure 2-22).
2. It is not adequate to safeguard against low temperature cracking
although within Table 2-7 the specifications in Table 2 are superior
to those in Table- I.
3. The test system is slightly more expensive than the penetrometer.
V)
II.I
V)
5 Pen. AC-40
a.
0
: 100
40
50
T
C, AC-20
ea
II.I
:::)
a 50
Pen.
60
T
in 70 Pen. AC-10
~
l-
o
LL.
l-
85
100
Pen.
T
AC-5
a::
G:
~::::
120
l~O
T
PenJ AC-2.5
u 10
b
(l)
r
200
300 T
I-
ii> 5
0
u
Figure 2-23. Comparison of Penetration Grades and Viscosity Grades

of Asphalt Cement (based on RTFOT Residue for
AR-Grades and Penetration Grades; TFOT Residue for
AC-Grades)
4. Testing time is longer.

5. TFOT residue viscosity can vary considerably within the same
grade. For example, AC-20 asphalt cements from two different
sources can have TFOT residue viscosities of 3,500 and 10,000
poises. These asphalts are likely to behave differently during and
after construction.
AR Viscosity Grading System
The AR viscosity grading system, which is based on the viscosity of the

aged residue (AR) from the rolling thin film oven (RTFO) test, is used by
some states in the western United States. Table 3 (lower table) within Table
2-7 gives the requirements for the following five AR grades: AR-1000,
AR-2000, AR--4000, _AR-8000, and AR-16000. AR-1000 represents a
viscosity of 1000 poises at 140°F (60°C) of the aged residue. Specifications
also require minimum penetration at 77°F (25°C) and minimum viscosity
at 275°F (135°C), again for the aged residue. Penetration of the original
asphalt cement (although not specified) is determined to check the speci-
fied minimum requirement for percentage of retained penetration after
RTFO test. It should be noted that no consistency values at any tempera-
ture are specified for original asphalt cements.
' The AR viscosity grading system has the following advantages in addi-
tion to those listed previously for the viscosity grading system:
1. It represents asphalt properties after the HMA is manufactured in a

typical HMA batch facility.
2. Reasonable uniform behavior of asphalt cement is expected from
different sources but within the same AR grade during laydown and
. compaction.
The following are the disadvantages of the AR grading system:
1. It is highly regional.
2. It requires more testing equipment (RTFO).
3. Testing time is longer because aged residue from the RTFO test
---------cm=u=s=tc-1b""'e,_o-=-1b~t='ainea first.
4. Detecting contamination is a problem because there are no consis-
tency requirements on original asphalt cement.
5. Aged residue may not relate to hardening that occurs in the drum
mix process.
RHEOLOGICAL PROPERTIES AND

PAVEMENT PERFORMANCE
Rheological Properties
It has been well established that the rheological properties of the

asphalt cement binder affect the pavement performance. Since the rheo-
logical properties of asphalt cements change during HMA production and
continue to change subsequently in service, the phenomenon of aging (or
age hardening) is discussed first.
Age Hardening. The first significant harclening of the asphalt cement

takes place in the pugmill or drum mixer where heated aggregate is mixed
with hot asphalt cement. During the short mixing time, the asphalt
cement, which is in very thin films, is exposed to air at temperatures which
range from 275 to 325°F (135 to 163°C). Substantial rheological changes
such as a decrease in penetration and an increase in viscosity of the asphalt
cement take place during this short mixing period from both air oxidation
and loss of more volatile components. Age hardening of the asphalt cement
continues, although at a much slower rate, while the HMA is processed
through a surge or storage silo, transported to the paving site, laid, and
compacted. After the HMA pavement has cooled and been opened to traf-
fic, the age hardening process continues at a significantly slower rate for the
first 2-3 years until the pavement approaches its limiting density under
traffic. Thereafter, the rate of age hardening is further reduced and longer
time periods are needed to discern the changes in the rheological proper-
ties of the asphalt cement. Age hardening in service takes place at an accel-
erated rate if the HMA pavement has a higher air void content than orig-
inally designed, which provides for· easy entry of air, water and light.
Thicker asphalt cement films around the aggregate particles harden at a
slower rate compared to thin films. Figure 2-24 shows a typical plot of
changing viscosity at 140°F vers4s time for six asphalt cements used in a
durability study (.B).
The following six factors (not necessarily in order of importance) have
been reported to contribute to the age hardening of asphalt cement during
-----1mi-xciRg-an<lt0,r-in-s&W-ice-~, J . 0 1 - - · - - - - - - - - - - - - - - - - - - - - -
l. Oxidation;
2. Volatilization;
3. Polymerization;
4. Thixotropy;
5. Syneresis; and
6. Separation.
These factors have been defined as follows (2):
1. Oxidation is the reaction of oxygen with asphalt cement, the rate

depending on the character of the asphalt cement and the tempera-
ture.
2
-- - -
en
w
en
0 104 . ···----· --···--
--- --- -- ----·-
--
.. --- ------
a..
lJ..
0
0
¢
...
9
8
7
6
5
----- __.
.,,,,,---.--------
,,,:::.---- - -
.-•-- --·
-----------
I- 4 . I
<(
>-
I-
3
ci5 I
0
u
en
'/
/
ASPHALT
ASPHALT
I ----
2 ------ ---
> ASPHALT 3-----
ASPHALT 4-·-·-·-·-
ASPHAL T 5 -··· - · · · -
ASPHALT 6-·-·-·-·-·-
IXING
. . 60 80
TIME - MONTHS
Figure 2-24. Viscosity at 140°FVersus Time in Months (after Kandhal
et al, B)
2. Volatilization is the evaporation of the lighter constituents from asphalt

cement and is primarily a function of temperature. It is usually not a
significant factor contributing to long-term aging in the pavement.
3. Polymerization is a combining oflike molecules to form larger mol-
ecules, causing a progressive hardening. There is no scientific evi-
dence that this is a significant factor during the low temperature
aging of asphalt in pavements in spite of such speculation in the lit-
erature.
4. Thixotropy is a progressive hardening due to the formation of a
structure within the asphalt cement over a period of time, which can
be destroyed to a degree by reheating and working the material.
Thixotropic hardening (also called steric hardening) is generally
associated with pavements which liave little or no traffic, and its
magnitude is a function of asphalt composition.
5. Syneresis is an exudation reaction in which the thin oily liquids are
exuded to the surface of the asphalt cement film. With the elimina-
tion of these oily constituents, the asphalt cement becomes harder.
6. Separation is the removal of the oily constituents, resins, or
asphaltenes from the asphalt cement as caused by selective absorp-
tion of some porous aggregates.
Traxler (11) has suggested nine additional factors, including several

effects of light, water, chemical reaction with aggregate, microbiological
deterioration, and adsorption of heavy asphalt components on the surface
of the aggregates.
The rate at which th~se, or possibly other reactions not specifically
identified, occur appears to be extremely complicated and has been argued
at length in the literature. For the engineer, these arguments may be super-
fluous because each reaction seems to lead to an undesirable change or ,
embrittlement of the asphalt cement, which in turn has been associated
with HMA of poor durability properties UflJ.
Many long term pavement performance studies involve periodic core
sampling and testing to determine the changing asphalt cement properties
such as penetration at 77°F (25°C) and viscosity at 140°F (60°C), and
pavement properties such as, percent air voids in the pavement. Changes in
such properties have been known to affect pavement performance with
ti1r1e;--and-have-been-fot1:nd-to-foHow-a-hyper-bol-i-G-ffJ:0del-(-J-2,±2,:M)-.~----
According to this theory, the changes in these physical properties follow a
hyperbolic function with time and approach a definite limit as time
increases. The following equation expresses the age hardening of asphalt in
the field:
T
~Y=-- (Eq.1)
a+bT
or
T
- =a+ bT (Eq. 2)
~y
where,
l:!.Y = change in test property (such as penetration and viscosity)
with time T or the difference between the zero-life value
and the value at any significant time;
T = time;
a =constant, the intercept of the Eq. 2 line on the ordinate;
b =slope of the Eq. 2 line; and
lib = the ultimate change (limiting value of change) of the prop-
erty at infinite time.
The extent of age hardening can be quantified in terms of penetration

(percent retained penetration) or viscosity (aging index) as follows:
% Retained Penetration = Penetration of aged asphalt x 100

Penetration of original asphalt
Aging Index = Viscosity of aged asphalt

Viscosity of original asphalt
Both percent retained penetration and aging index have been used to
evaluate relative aging of asphalt cements of different grades and/ or from
different sources.
The thin film oven (TFO) test and rolling thin film oven (RTFO) test
generally simulate the asphalt cement hardening which occurs in the pug-
mill of a HMA batch facility. Because the asphalt cements from different
sources age differently in service after pugmill mixing, the tests on TFO or
RTFO residue are not completely reliable to predict the long-term asphalt
aging or durability. There is an urgent need to develop an accelerated lab-
_ _ _ _ _ _o_r_a_to_ry_,__a-ogi_gg procedure for asphalt cements that can closely simulate the
hardening that takes place in the pavement under certain climatic condi-
tions (15). Field observations indicate that the asphalt cement does not age
as rapidly in a drum HMA facility as it does in a batch facility because of
the presence of water vapor in the drum, which reduces oxidation.
., 46 Hot Mix Asphalt Materials, Mixture Design and Construction
Penetration. The penetration test measures the consistency of paving

asphalt cement which is expressed as the distance in tenths of a millimeter
that a standard needle vertically penetrates a sample of the asphalt cement
under known conditions of loading, time, and temperature. Test proce-
dures for measuring penetration at 77°F (25°C) and lower temperatures
(usually 60°F or I6°C and 39.2°F or 4°C) are given in ASTM D5. It is an
empirical test. In general, the penetration test involves higher stresses and
shorter loading time than do most viscosity tests. Therefore, attention
must be given to stress levels or shear rates, or both, to attain direct com-
parison of penetration and viscosity results. Penetration at 77°F (25°C) has
been widely used in asphalt cement specifications because no simple
method of viscosity determination at 77°F (25°C), or lower temperatures,
is currently available. Even viscosity ,graded asphalt cement specifications
require minimum penetration values at 77°F (25°C). As discussed earlier,
penetration is also used to determine the temperature susceptibility of
asphalt cements in terms of penetration index (PI) or pen-vis number
(PVN).
Penetration at 77°F (25°C) generally gives the consistency of asphalt
cement near average yearly service temperature, thus it has some influence
on the overall performance of HMA pavement. A vast storehouse of data
on penetration of aged asphalt cements in the pavements under service is
available in the literature (J.Q). It has been observed that aging of the HMA
pavement results in progressively lower penetration with time as shown in
Figure 2-25 (ll). Hubbard and Gollomb (1!l) concluded the following
from their study showing the effect of asphalt cement hardening on the de-
velopment of cracks in HMA pavements:
I. When the penetration of the asphalt cement (77°F or 25°C) falls

below 20, serious pavement cracking may occur.
2. Some cracking may occur when the penetration is between 20 and
30.
3. High resistance to cracking may occur when a mixture is well
designed and properly compacted and the penetration of the asphalt
cement is well above 30. ,
4. To ensure long life, one should use as soft an asphalt cement as pos-
sible without reducing stability below the minimum required to
prevent displacement under traffic.
Most studies (16, 12) generally support the above conclusions.

Obviously, other factors such as asphalt cement film thickness, asphalt
cement temperature susceptibility, asphalt cement ductility, air void con-
Chapter 2-Asphalt Refining, Uses, and Properties · 47
tent, age of the pavement, traffic, and climatic conditions also influence
this general relationship between penetration and cracking.
70
60 -BEAVER CO.
-------LYCOMING CO.
LL -·-WASHINGTON CO.
I'- -···-LEBANON CO.
I'- 50 X CRACKING FIRST OBSERVED
,, I-
<(
z ' ' ' ·...
0 40 ',~
I-
<(
et::
I- ~
w 30
z
w
Q..
·--·---
20.
·--~"::: ::.::.::.it~.::=::.-:..- -
10
0 20 40 60 80 100 120 140
LIFE OF PAVEMENTS, MONTHS
Figure 2-25. Penetration Versus Time in Months (after Kandhal, 12)
Ductility. The ductility of a paving asphalt cement is measured by the

distance to which it will elongate before breaking when two ends of a bri-
quette specimen are pulled apart at a specified speed and temperature.
ASTM D113 gives the test procedure to measure ductility at 77°F and
lower temperatures.
The significance of the ductility test as a means of asphalt cement qual-
ity control has been debated because of its empirical nature and poor
reproducibility. Exactly which property is being measured is difficult to
decide. Welborn et al (12) and Kandhal (Il) reported good correlation
between ductility and shear susceptibility at 45 and 60°F (7 and 16°C) for
------:var-iG-us..as-p-hal-t-cemen.tS-,-J;eg.a.i::d.l.ess-o.fsoui::ce...Barth (1) stated that the duc-
tility is a valuable indicator of the physiochemical state of the asphalt col-
loid system.
Several studies have related the ductility of asphalt cement to pave-
ment performance. Doyle (2Q) measured ductility at 55°F (13°C) and
observed extensive pavement cracking when the ductility dropped below 5

cm. Halstead (21) demonstrated that the pavements containing asphalt
cement with penetration in the range normally considered satisfactory (30
to 50), but with low ductilities are likely to show poorer service than pave-
ments containing asphalt cements of the same penetration, but with high
ductilities. Reporting on the Zaca-Wigmore Project, Hveem et al (22)
found increasing evidence that the ductility test on asphalt cements recov-
ered from pavement during its service life is an important method for judg-
ing pavement performance. The ductility results for unsatisfactory asphalts
on the Zaca-Wigmore Project were very low. Serafin (23) reviewed data on
cores taken after 7 years service on the Michigan Bituminous Experimental
Road and reported that the section containing the lowest ductility had the
most pitting and cracking, whereas there did not appear to be any signifi-
cant differences in penetration. Vallerga and Halstead (21) studied 53
highway pavements throughout the United States and reported that severe
ravelling occurred in cold climates when the ductility at 60°F (16°C)
dropped to 3 cm or less. Kandhal (1Z) studied 10 test pavements in
Pennsylvania and concluded that aging of the pavement results in progres-
sively lower penetration and higher viscosity values. However, the accom-
panying decrease in low-temperature ductility was determined to be an
important factor. After the penetration of asphalt cement drops below 30
because of hardening, the pavements containing asphalt cement with low
ductilities showed poorer service than pavemef,].ts containing asphalt
cements of equal penetration but with high ductilities.
A subsequent study by Kandhal and Koehler U5) of six experimental
test sections in Pennsylvania indicated that lower ductility values were asso-
ciated with higher incidence of load-associated longitudinal cracking.
The preceding discussion indicates that the ductility test can discern
the differences between different asphalt cements more effectively if it is
run on highly stiff asphalt (either by lowering the test temperature below
77°F (25°C) and/or further aging of asphalt after TFO or RTFO testing).
Present ASTM or AASHTO specifications require this test to be run at
77°F (25°C) on TFO residue. There is a need to develop a more funda-
mental test (such as shear susceptibility) for evaluating the quality of
asphalt cement.
Viscosity. Viscosity at any given temperature and shear rate is essen-

tially the ratio of shear stress to shear strain rate. At high temperatures such
as 275°F (135°C), asphalt cements behave as simple Newtonian liquids;
that is, the ratio of shear stress to shear strain rate is constant. At low tem-
peratures, the ratio of shear stress to shear strain rate is not a constant, and
the asphalt cements behave like non-Newtonian liquids. Unlike the empir-
ical tests of penetration and ductility, viscosity is a fundamental consisten-
cy measurement in absolute units that is generally not affected by changes
in test configurations or geometry of the sample (.25). Standard ASTM test
methods which use capillary viscometers are available to determine asphalt
cement viscosity at 140°F or 60°C (ASTM D2171) and 275°F or 135°C
(ASTM D2170) on a routine basis. Large vacuum capillary viscometers
have been used to test asphalt cement viscosity at 77°F (2£). However, the
method needs to be refined further. The cone and plate viscometer is used
in ASTM D3205 to measure viscosity in the range from 103 to 10 10 poises
and is therefore suitable for use at temperatures where viscosity is in the
range indicated. The shear rate may vary between approximately 103 to 10 2
sec1, and the method is suitable for either Newtonian or non-Newtonian
asphalt cements. This method is not used routinely at the present time.
Asphalt cement viscosity at 140°F (60°C) has some influence on the
performance of HMA pavements during hot summer days when the pave-
ment surface temperatures are near 140°F (60°C). A low viscosity at 140°F
(60°C) can induce flushing and/or rutting if other factors are identical.
This is especially true for mixes that have a relatively high asphalt content.
Therefore, proper viscosity grade selection appropriate to prevailing cli-
matic conditions must be made. For example, use of AC-10 in the south-
ern part of the United States will most likely result in an increase in flush-
ing and/or rutting ofHMA pavements when compared to AC-20. On the
other hand, an unusually high viscosity (or stiffness) at low service tem-
peratures (such as -20°F or -29°C) in the northern portion of the United
States) will result in non-load associated transverse shrinkage cracking.
It has been observed that aging of the HMA pavement results in pro-
gressively higher viscosity with time, as shown in Figure 2-24. However,
vis~osity of the aged asphalt cement alone does not necessarily control .the
pavement performance. Aging indexes based on viscosity at 77°F (25°C)
were found to be more meaningful to indicate comparative aging and pave-
ment performance of test pavements in Pennsylvania. Ravelling was
------,obs&FV-e-d--GH1-t.h~™-pa-¥emts-w.h.en...this-agingJndex exceeded 12 (11).
Temperature Susceptibility. Asphalt cement is a thermoplastic mate-

rial. Its consistency changes with temperature. Temperature susceptibility
is the rate at which the consistency of an asphalt cement changes with a
change in temperature and is a very important property of asphalt cement.

Asphalt cements highly susceptible to temperature change are not desirable
because (a) their viscosity at 275°F (135°C) can be very low, resulting in
occasional tender mix problems during compaction, and (b) their viscosi-
ty (stiffness) at the lowest service temperatures are usually very high, result-
ing in low temperature shrinkage cracking. Three different approaches for
determining temperature susceptibility of asphalt cement are currently
used.
I. Penetration Index (PI):

(_
Pfeiffer and van Doormaal (22) expressed the temperature susceptibil-

ity quantitatively by a term designated as "penetration index" (PI).
The PI is determined from an asphalt cement's softening point (ring
and ball test), its penetration at 77°F (25°C), and an assumption that the
penetration of an asphalt cement at its softening point is 800. This is
referred to as PI (pen/R&B). Several investigators have reported that the
penetration of a substantial number of asphalt cements at their softening
point can vary widely from 800, especially for waxy asphalts and asphalt
cements having high softening points and high PI values. Therefore, it is
considered prudent to measure penetration at some other temperature in
addition to 77°F (25°C) rather than relying on this assumption.
Penetration can be measured at two temperatures to obtain this index.
Logarithm of penetration is plotted against the test temperature in degrees
C which normally gives a straight line. The slope A of this line is calculat-
ed as follows:
A = log Pen at T1 - log Pen at T2

Tl - T2
The following empirical expression is then used to calculate the pene-

tration index termed PI (pen/pen):
PI= 20 - 500A
1 + 50A.
Chapter 2-Asphalt refining, Uses, and Properties 51
Example 2-2: The penetration of an asphalt cement at 77°F (25°C) is

120. Its penetration at 40°F (4.4°C) is 10. What is the penetration index
(PI) ?
A = log Pen at T1 - log Pen at T2 log 120 - log 10

~~--~- = 0.052
T1 - Tz 25 - 4.4
PI= 20 - SOCIA = 20 - 500 (0.052) = -1.67

1 + 50A 1 + 50 (0.052)
Note that the temperatures used in the equation are in degrees centigrade.
The lower the PI value of an asphalt cement, the higher its tempera-
ture susceptibility. Most paving asphalt cements have a penetration index
between + 1 and -1. Asphalt cements with a PI below -2 are highly tem-
perature susceptible, usually exhibit brittleness at low temperatures, and
are very prone to transverse cracking in cold climates.
2. Pen-Vis Number (PVN):
McLeod proposed PVN to determine the temperature susceptibility of

asphalt cements (211). This number is based on penetration at 77°F (25°C)
and viscosity at either 275°F or 140°F (135°C or 60°C) which are usually
specification requirements for paving asphalt cements, thus the data is
readily available to calculate the PVN. The folk-wing formula is used to
calculate PVN:
L-X
PVN = - - (-1.5)
L - M
where,
X = the logarithm of viscosity in centistokes measured at

---------:2-;7-¥F-(-:l-3-5£G-)-;~----
L = the logarithm of viscosity at 275°F (135°C) for a PVN of
0.0; and
M= the logarithm of viscosity at 275°F (135°C) for a PVN of
'-1.5.
52 Hot Mix Asphalt Materials, Mixrure Design and Construction
The viscosity values of L and M can be read from the graph (Figure
2-26) developed by McLeod by plotting penetration at 77°F (25°C)
against viscosity at 275°F (135°C) for "typical" asphalts. This graph con-
tains two lines with the assigned PVN values of 0.0 and -1.5, which are
used. However, the following equations (based on least squares line) can be
used to calculate more accurate values of L and M:
The equation for the line representing a PVN of 0.0 is
log V = 4.25800 - 0.79670 log P
The equation for the line representing ~PVN of - 1.5 is
log V = 3.46289 - 0.61094 log P
....
~ 5000 1-...-,PV---'~---+--4-<<1--E-."·~,.........,Haa:::+:--+=:.;;;;:::1-+-+-+-+·;;.;;i··1---+--+-+-++t-++1 C
~ 3000 ~"o.-"-,;;;::--+--F.i
.. f--M--+A---+:.'"4·::r--llt++<::::.._.+--1-+-fwi<···l----+--+-+-t-H+H
Cl) 2000 i-0. ~ , , I .. ···· ···· ....
0 1
I- ,- 'o. ,""'~~'-F-<81-+.=l--+t;;r--+.-····+--t---····+·--HH-+:;;;a:::1---+-+-+--t-H-++I
rau I000---a ,,, r--... ,,r-. = ..
----: •.o - . .
.... ..... . ..
s:;.3ca:,' ~ ' ~ ~ ~::: :::::

500 -'I.~'
Z
v~ . . , :·.. : r-...:: ~', ....... .....
.... .
LL
0
IO
r---
(\J
~ I 00 1--4n,50 --· ..
>-
I-
Cl)
0 --,so12ou - ·· :\ ,~ ';1:,, _
u
Cl)
3 0 ,..300/40r • .. ···· ···· ::: ', .. ,....-~--+-+-+++11
20 --aoo11000· ·· ···· ···· ··· · ~,.,. "'o
> I 1 •• •··· •••• ••• I~♦.;.~...
10 : I i:-: - ; -frH,-Hf-- 1-1:-:-:-:11----1--1 -1~--
1o · 20 5o I,:>o 200 500 1ooo 2000 5000

PENETRATION AT 77°F
Figure 2-26. A Chart for the Determination of Approximate Values for
PVN for Asphalt Cements (after McLeod, 21!)
Where Vis the viscosity in centistokes at 275°F and Pis the penetration at
770F.
Example 2-3: An asphalt cement has a penetration of 100 at 77°F

(25°C) and a viscosity of 400 centistokes at 275°F (135°C). What is its
PVN? Use McLeod's chart (Figure 2-26).
L-X
PVN = - - (-1.S)
L - M
X = log 400 = 2.602

L= log 450 = 2.653 (from McLeod's chart)
M=log 180 = 2.255 (from McLeod's chart)
PVN = 2.653 - 2.602 (-1.S) = _0 _19

2.653 - 2.255
The lower the PVN value of an asphalt cement, the higher its temper-
ature susceptibility. Most paving asphalt cements have a PVN between
+0.5 to -2.0.
Although PVN can readily be determined from the specification tests,
some asphalt technologists believe that the temperature susceptibility
obtained in the higher temperature range (77-275"F) cannot be extrapo-
lated to be applicable at temperatures below 77°F. One notable difference
between PI and PVN is that the PI changes on aging (during mixing and
subsequently in service), whereas the PVN remains substantially the same
(22).
3. Viscosiry-Temperature Susceptibiliry (VTS):

For determining VTS, a double logarithm of viscosity in centistokes is
plotted against the logarithm of the absolute temperature in degrees K
(empirical Walther's equation). Such plots generally result in straight lines,
with the slope of the line being equal to VTS as follows:
VIS = log log viscosity at T2 - log log viscosity at T1

log T1 - log T2
Larger VTS numbers indicate higher temperature susceptibility.

Generally, the slopes of lines at temperatures lower than 140°F (60°C) tend
to deviate from the slopes established between 140°F (60°C) and higher
temperatures: This is because shear dependent viscosities are encountered
below 140°F (60°C). The numerical differences between the VTS values of
different asphalt cements are not large. The VTS value ranged from 3.36
to 3.98 for 52 asphalts sampled in the United States (3.Q). Very little use of
VTS is found in the literature.
Shear Susceptibility. At high temperatures most asphalt cements

exhibit purely viscous flow (Newtonian) in which rate of shear strain is
proportional to stress. In other words, viscosity is independent of shear
rate. However, at low temperatures and/ or after aging, most asphalt
cements exhibit non-Newtonian or viscoelastic flow. That is, their viscos-
ity is dependent on the shear rate, the viscosity increases as the rate of shear
increases. The rate of change of viscosity with-rate of shear is referred to as
shear susceptibility which is considered to be an intrinsic property of the
asphalt cement. The shear susceptibility (or shear index) is the tangent of
the angle of log shear rate (X - axis) versus log viscosity (Y - axis) plot
determined by conducting the viscosity test at different shear rates. Shear
susceptibility values of asphalt cements tend to increase at different rates on
aging, depending on their chemical composition (11).
, Shear susceptibility or viscosity of the aging asphalt alone does not
necessarily indicate the pavement performance. The rate of gain in shear
susceptibility relative to increases in viscosity at 77°F seems to be one of the
major factors affecting pavement performance. Relatively lower gain in
shear susceptibility with the corresponding increase in viscosity has been
reported to be associated with better pavement performance (11).
Stiffness. Stiffness (or stiffness modulus) is the relationship between

stress and strain as a function of time of loading and temperature; this rela-
tionship between stress, strain and time is also referred to as the rheologi-
cal behavior of asphalt cements or mixtures. In many applications of
HMA, its stiffness characteristics must be known not only to assess the
behavior of the mix itsel£ but also to evaluate the performance of an engi-
neering structure of which ,the mix is a part, such as a highway or airfield
pavement (lJJJ. Ideally, for a highway pavement surface course, increased
asphalt cement stiffuess is desirable at high service temperatures (near
140°F or 60°C) to avoid rutting, and decreased asphalt cement stiffness is
---desi-rable-a:t-rhe-lew-s~i:-v-iGe-t©mper..atur:e..to-r:esistlow_temp_er_at.uJ.e shrink~ge_____ _
J
cracking. At high temperatures, increased resistance to rutting can be
obtained by proper selection of the aggregate component and mix design
because the asphalt cement stiffness has a secondary importance. However,
asphalt cement stiffness is primarily responsible for cracking at low service
temperatures.
Figure 2-27 shows a simplified diagram illustrating the time-of-load-

ing dependence of the stiffness of HMA for a particular temperature. At
very short loading times, it ,is noted that stiffness is essentially time inde-
pendent; in this case, the stiffness approaches the elastic modulus. For an
intermediate range on the time scale, the stiffness decreases with an
increase in time of loading. At very long loading times, the stiffness may
still decrease, but at a uniform rate, and the behavior may be considered to
CD
15
0
1/)
Elastic Behavior
Delayed or Retarded
Elastic Behavior
DI
E
-
~
0
CTI
Cl)
..
1/)
1/)
..
Q) (I) At Short Loading Times• S = E
-ci5-= (2) At Intermediate Loading Times•

(3) At Long Loading Times•
S = E(t)
s = )..
t
Where• A = Viscous Traction = 311
Ca Measure of Viscous Deformation)
Time, t Clog scale)

Figure 2-27. Idealized Time-of-Loading Dependence of the Stiffness
(Stress/Strain) Characteristics of an Asphalt Material
Subjected to an Axial Tensile Stress (after Finn, JO)
be purely viscous. The stiffness under these circumstances gives a measure

-------of-mcfl-ow-characteri-sric:s-of-the-mixtn-re:-ln--Figure 2-27 ,··this viscous
deformation has been characterized by viscous traction to indicate behavior
under axial stress; as noted in the figure, this viscous traction is approxi-
mately three times the viscosity, which is a measure of flow resulting from
shear stresses U!)).
The response to a three-dimensional system of stress which occurs in

a HMA pavement layer subjected to loading is rather complex. When the
response also depends on the time or rate of loading and the temperature,
the characterization is even more difficult (1J}).
In 1954, Van der Poel recommended to concentrate on a single stress
and its resulting strain because for many purposes this is reasonably ade-
quate (12). He suggested a single parameter termed stiffness (S) as follows:
S (t, T) = cr/£
where,
S = stiffness, in psi or kg/ c~;
· cr = axial stress;
£ = axial strain;
t = time of loading; and
T = temperature.
At very short times of loading and/ or low temperatures, the behavior

of HMA is almost elastic in the classical sense, and the stiffness, S, is anal-
ogous to an e\astic modulus, E (Figure 2-27). At longer times of loading
and higher temperatures, the stiffness is simply a relation between the
applied stress and the resulting strain. Conversely, if the stiffness corre-
sponding to a particular time and temperature and either the stress or the
strain are known, an estimate can be made of the strain or the stress devel-
oped in the asphalt mixture (1.J}).
Asphalt cement stiffness can be determined by the following two

approaches:
C
1. Indirect Methods: These methods provide an estimate of stiffness
without direct laboratory measurements. They use routine asphalt
cement test data for estimating stiffness by use of nomographs.
2. Direct Methods: Asphalt cement stiffness is obtained by direct test-
ing, by creep, relaxation, or constant-rate-of-strain testing in either
tension or compression. Dynamic or flexural testing methods can
----als0-be...use_______________ ~---~·-··
1
',.
A discussion of these two approaches follows.
1. Indirect Methods: Van der Poel conducted extensive creep and dynamic
tests using well compacted (approximately 3 to 5 percent air voids)
dense-graded HMA mixtures. He indicated that the stiffness of the mix is

dependent on the stiffness of the asphalt cement which it contains and the
volume concentration, Cv, of the aggregate (Figure 2-28).
Cu = volume of compacted aggregate 100 - VMA

volume of aggregate + asphalt 100 - air voids
10•-----------------------
.E
u
0, 0
..>::
Cl) I0 4 1-----:i.-----,,,.'--~-~------I------I
....
.2
X
~
-0
II)
II)
1031---..,C....--IIC--..,C....------1------1----....,..-1
Cl)
-
.E
U)
10----.i...i...i.i.......i..--...i..i.~.i......i...i..i..i.....i.--i....1..1.......i......i........
10 10 2 10 3 10 4 10~ 10•
2
Stiffness of Asphalt, kg/cm
Figure 2-28. Relation between ·stiffness of Asphalt Cement and

Stiffness of Mixtures (after Van der Poel, J2)
Based on the test results, he developed a nomograph (Figure 2-29)

which can be used to determine the stiffness of asphalt cement at any tem-
perature and at any rate of loading if the softening point (R&B) and the
·erretra:rion-ind-ex-Pl-Epen·/-R:&B7-of-the-asphalt··cement are known. An
"accuracy'' on the order of a factor of 2 was reported and has been con-
firmed by some researchers. Once the asphalt cement stiffness is obtained
from the nomograph, the stiffness of the HMA can be determined from
Figure 2-28 'using the volume concentration of the aggregate.
.. •
r~ .
i~ •
fw;I~
I
.IA ;.
~
;
r~ 1 .
!: ::;~:l
!!i2 1 ~
a
!f; l C
0 , . . 11 ......L
]~! ..
iI .
r•o:!ti
~ i~
a:0.1-.::
..
~,,id
1 .
': ,, o;, '! ,:, , ':'

!ii~i~
r·11t
1l!™
ii.
r~··
m~~
!i.1i::
_J
OlflllCll,l.'nl,l,JN3,,I
Figure 2-29. Nomograph for Determining the Stiffness of Bitumens

(after Van der Poel, J2J
Heukelom and Klomp W studied Van der Poel's method in detail

and modified the relationship between asphalt cement stiffecss (Sb,,) and
mix stiffness (Sm,.) as shown in Figure 2-30. Later Heukelom (..31) sug-
gested to measure penetration at three temperatures, plot the data on bitu-
men test data chart (BTDC), and determine the "corrected" PI, which he
called PI (pen/pen), and the "corrected" ring-and-ball softening point.
The "corrected" softening point is obtained at the intersection of the
extended penetration straight line and the 12,000 poise ordinate on the
BTDC. These "corrected" softening points are then used in the Van der
Poel nomograph. These corrections were deemed necessary because erro-
10 3
8
6
Cv = 0.90
4
o.aa'\..
.-0~
o.~§, '\
2
if;-_ 0 0,t:S!
r---......0.8~
~ (\
2
10 -0.74 -...... ...
- .....- ........ ' "
"'-
8 o,.,zg ... ' "
-
~
---- - --- - ~ .........

'' '
............ '''' '
---- ----....____ ............
6 u.,u ......._ ................. .....
' '''\."'\'\. ~
'
-
::0
en
X
E
4
_0;65 - -......... .....................
-----... ~~ ~"""
~ ~~
\\
"
::::--.......-........:-
---------
en 2 Cv=0.60~ ~
10
- ~ ~ ..._,~
__ ,
8 -- -......
-......
,~~
~ ,,
.......
~
........_,
6
4 '
2
I
----------1-Ge-... ----1----+110 I0 2 3
I Qe-------110~ 10 9
sbih kg/cm2
Figure 2-30. a Function of sbit and Cv (after Heu.kelom
smix / sbit as
and Klomp, JJ)
60
neous results are possible when the R&B softening point is used for waxy
asphalt cements.
McLeod (2.i.) further modified the Heukelom method for determining
the stiffness of asphalt cement. The major difference is the use of pen-vis
number (PVN) by McLeod in lieu of penetration index PI (pen/pen) used
by Heukelom. Figures 2-31 and 2...:32 are used to determine asphalt stiff-
ness (28). McLeod's method requires measurement of asphalt cement pen-
etration at 77°F (25°C) and viscosity at 275°F (135°C). Stiffness moduli
135 400
w
a: 130
::,
~ 125 ii!:
a:
w 120 w
~ 115
a:
::,
w
I-
110 ~ ~
105 "'
W
ffia..
~ ~
w
1:1:! 1- 100 <t 1-
::,~
~w ~ ~
~ / z
ffi~ 90 ,;r / ~
~
~~ 85 ~ / >-
z<(
I-~
W I-
80 ~
;s-i? .,.6u / /
~ w 75 ~~ ~
~~ 10 I~ "s / 20 ~·
z w
~
LP~ .,. . / ~
w~
!;!;! I-
,... !1:- ~
o,3"
~
..,,..3' 15 0
I- a: 60 ii.~- . .,. , ;!;
~ 55
~ @ 50
e ""~
Js,(i) ,,",
,, ,? 10 f3
8 3
<(
ii!:~ 45 R:-(,ij O >
w .J ~ 6
~~ 40 ,!/, 5
I-
fa1-
ffiw 35 -9"' ,? 4
z
lt 30 _p '.3
3
0
0
25 /.!ff GlYEt:I ~
a:
~ <t PENETRATION AT 25"C=90 1-
~ 20 PEN-VIS NUMBER=-1 ·0 2 w
zw
~ I5Il:::IEli a..
------'W----t-------J=R.9M-GRAP't1--------
~ IO TEMPERATURE DIFFERENCE=20"C
~ 5 A~
O BASE TEMPERATURE•25•20=45"C
Figure 2-31. Relationship Between Penetration, PVN, and Base
Temperature for Asphalt Cements (after McLeod, 2!l)
values obtained by McLeod's method have been reported generally within

a factor of 2 of the diametral creep stiffness moduli obtained by direct test-
ing (22).
Although indirect methods are quick and easy to use, they have the fol-
lowing limitations (35):
a. The scale of nomographs requires that extreme care be used in inter-
polating the data and results .. ~ dull pencil can lead to appreciable
errors.
b. Determining stiffness of asphalt cement and mix by a series of trans-
formations, starting from routine test data on the asphalt cement,
appears to be somewhat "unscientific" to a number of engineers.
c. There is no estimate of possible errors in determining stiffness by
indirect methods. However, no precision statements are available at
the present time even for direct methods.
Despite the foregoing limitations, the indirect methods can be used for
initial estimates of stiffness modulus if direct testing is not possible.
2. Direct Methods: At high temperatures, viscosity can be used as a measure
of asphalt cement stiffness. At 140°F (60°C) and higher, capillary vis-
cometers are suitable because of the Newtonian behavior of asphalt cement
in these temperature ranges. However, at low temperatures, a viscometer
must be able to quantify viscosity as a function of both temperature and
shear stress or rate. This complicated factor has resulted in the develop-
ment of a large number of devices for measurement of shear-dependent
viscosities (25).
In 1974, Schweyer (16) presented a pictorial overview of many exper-
imental viscometers that have been used to study the viscoelastic behavior
of asphalts. These included (a) rotational types, which utilize coaxial cylin-
ders or a cone and plate; (b) rheometers, a specialized capillary type in
which a piston is used to drive the asphalt through a capillary tube; and (c)
miscellaneous types such as the sliding-plate microviscometer. Schweyer
and others have done considerable work on the use of the capillary
rheometer (32) and the development of several generations of the con-
stant-stress rheometer Ufl), which has been used to examine both asphalt
cement stiffness and viscosity at low temperatures.
The methods most widely used to determine the asphalt cement vis-
osity-adow tempera:tnres-are-proba:b1y--clte-cone and plate viscometer or
the sliding-plate microviscometer. Both were adopted as ASTM standard
test methods and are capable of examining viscosities over a wide range of
shear rates. ASTM has recently adopted a capillary tube method (D4957)
for measuring the apparent viscosity of non-Newtonian bitumens.
The Shell sliding-plate rheometer (12) conveniently measures

low-temperature asphalt cement stiffness and has been used to determine
the accuracy of different asphalt cement stiffness prediction techniques
(1.Q). Recently, the sliding-plate rheometer was modified to extend its
upper stiffness limit from lxl0 8 to l.5xl0 9 N/m2 (41) Basically, the instru-
ment (Figure 2-33) operates as follows: a specimen of asphalt cement vary-
ing in thickness from 2 to 10 mm is formed between two aluminum plates
10 1001,00010,000
VISCOSITY, CP
I \
I \
I \
,BASE TEMl:ER.4TURE, \
•c ABOVE BASE T ~ T/ 'C BELOW BASE TEMPE{fATURE
,'9.?, ~, ,, ~~}.P,.~?..~~.'f,.,~.. ~,?., :A.~~,,,., . ,~
DIFFERENCE IN 'C BETWEEN ANY SPECIFIED TEMPERATURE
/ AND BASE TEMPERATURE \
I \
I \
I \
I \
I \
I
I \
I \
\
I \
I 2 4 2
\
/ I kg/cm =14-2psi=9-81 xi 0 N/m \
I \
I \
I
FREQUENCY,CY/SEC
\
'ioo _ 10 .. _ 1' I SEC I MIN
- - - - - , · 864 --2- - - - - - . . - 2 - - - - - - - - - - - - - - - - - - - - - - - - - - - , - - - - - -
24A 24A 24~ 2 4 • 24U 248
0.001 Q.01 O,I I 1,000 10,000 100,000 1,000,000
LOADtlG TME, SEC
Figure 2-32. Nomograph for Determining Modulus of Stiffness of

Asphalt Cement (after McLeod, J.i)
(20x30 mm size). One of the plates is clamped vertically to the frame. At

time, t=o, the support beneath the asphalt cement specimen and the other
undamped plate is quickly removed, and the asphalt cement supports the
total load of the undamped plate and other accessories such as load adap-
tor and steel needle. The measurement of specimen shearing is done by
means of a displacement transducer.
Rheometrics mechanical spectrometer (RMS), an instrument orig-
inally designed for polymer research, has been used to make accurate
low-temperature dynamic viscoelastic measurements on asphalt cement
(12). The dynamic shear modulus was measured in forced torsion at sever-
ORIGINAL RHEOMETER
APPLIED 0
LOAD
MODIFIED RHEOMETER
MODIFIED
~~-1'-BACK-PLATE
r---,_...._, CLAMP
MODIFIED
BACK-PLATE
Figure 2-33. Original and Modified Rheometers (after Gaw, 11)

al frequencies and, utilizing the extreme low-temperature capability of the

RMS, at temperatures down to -94°C. However, the RMS is very expen-
sive and the stiffness moduli are obtained at very short loading times.
Torsional rheometer (from Carrie Med) is also available to measure the
dynamic viscoelastic properties of asphalt cement over a wide range of tem-
peratures.
Pavement Performance Related to Asphalt Cement

Rheological Properties
There are some specific types of HMA pavement distresses (affecting

the pavement performance) which are related to the rheological properties
of the asphalt cement. These distress types are a1scussed below.
Ravelling. Ravelling of a HMA pavement surface is usually caused by

one or a combination of the following factors: (a) deficient asphalt content
(lean mix), (b) insufficient amount of fine aggregate matrix to hold the
coarse aggregate particles together, (c) lack of compaction (high air void
content), and (d) excessively aged (oxidized/brittle) asphalt cement binder.
Only the last factor is discussed here. It was mentioned earlier that age
hardening of asphalt cement in pavement results in progressively lower
penetration or higher viscosity. This causes a progressive increase in the
brittleness (and lack of binding characteristics) of asphalt cement and thus
induces ravelling. High air void contents in the HMA pavements, when
constructed, accelerate the age hardening, and premature ravelling results
(12). The relationship between air void content and extent of ravelling is
shown in Figure 2-34. All other things being equal, asphalt cements from
some sources age and harden faster than the others and thus induce ravel-
ling earlier. Increased asphalt film thickness can significantly reduce the
rate of aging and offset the effects of high air voids.
Cracking. Over the years, engineers have been able to categorize crack-
ing under two broad groups: load associated and nonload associated,
although most pavement cracks can be described according to their geom-
etry such as longitudinal, transverse, polygon (also alligator and map) and
block, or by the mechanism that causes the cracking such as slippage,
--~hrinkage..a.nclr.eflec.tio.n_(.ii)_._ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _~
I. Load-associated cracking
The principal class of load-associated cracking has been described as
fatigue cracking (alligator cracking), the phenomenon of fracture under
I
I
, I
,,
I
w
0:: I
w
,, •
I
> CORR. COEFF. =- 0.92 I
w , 1-1 , I
I,
,, ,
Cl) I
, I
I
, ,,
,
2-1,
. ,
I
I
, I
, I
(!) w , I
, I
z I- I
,,
I
~
:J w I
I
,,
I I
..J
w a
0
2-2 ,
I
> , ,,
I I
~,, ,
::I
<
0:: I
I
I
1-2/
I ••
u.. / 3
0 ,,
I
z
I- I-
:r:
, I
w S2 ,'7
I- ...J
(/)
,
I
X I
W, , I
,,
I
I
I 6 ,
r
I
I
.,, •'
,, ,,
I
I
,
w
z , ,I
0
z ,,
6 8 10 12 14 16 18
AIR VOID CONTENT, PERCENT
Figure 2-34. Air Void Content Versus Extent of Ravelling
(after Kandhal, 43)
-------1.--,~pea-t~d-odluctuating-StresS-ha.ving a roaxioium_v.:alu.eJes.Hhan the t~1J..sile

strength of the material. Some of the factors which influence the develop-
ment of fatigue cracks are: (1) composition of structural section, (2)
asphalt cement consistency, (3) asphalt content, (4) air voids and aggregate
characteristics in asphalt mix, (5) in-place properties of untreated aggre-
66 Hor Mix Asphalt Materials, Mixture Design and Construction
gates or cement-treated materials, and in-place properties' of foundation

soils, (6) temperature and (7) traffic (J1). However, only asphalt cement
properties will be discussed here.
Numerous researchers have conducted fatigue tests on HMA mixes.
However, laboratory test results have been influenced by the mode of testing
(constant stress or constant strain) and the failure criteria. In general, constant
stress tests will respond with an increasing fatigue life to any asphalt
cement or mix property which increases the stiffness of HMA. For exam-
ple, lower asphalt cement penetrations or higher viscosities of asphalt
cement generally result in a higher stiffness modulus and a greater number
of cycles to failure, all other things being equal. For constant strain tests,
the effect of stiffness modulus is reversed; th_at is, higher penetrations or
lower viscosities of asphalt cement result in inZreased fatigue life. Based on
the interpretation of numerous test results, the following hypotheses have
been made relative w pavement thickness, stiffness modulus, and fatigue
performance (11.):
a. For thin asphalt pavement, mixes of low stiffness (low viscosity

asphalt cements) should be used and the fatigue life should be eval-
uated in the constant strain mode of testing. Obviously, the stiffness
should not be too low to cause rutting in the surface courses.
b. For thick pavements (125mm and greater), mixes of high stiffness
(high viscosity asphalt cements) should be used and the fatigue life
should be evaluated in the constant stress mode of testing.
This means that asphalt cements which become hard through aging
will demonstrate poor fatigue life (develop fatigue cracking) in thin HMA
surfacings.
2. Nonload associated cracking

r
Of many types of nonload associated cracking, low-temperature
cracking is of considerable interest and concern to highway engineers (15).
It manifests itself through transverse shrinkage cracking in the HMA pave-
ment layer. Field observations and measurements have indicated the crack-
_ _____,·ng-st-a:rt-s-at-d1e-sur-fa-ee-a:a-cl-p:r0gresses-cl-e-wn-wi-1:h-1:ime-.--'-T-his-happ~11-------
because the low ambient temperatures chill the road surface first. HMA
pavements subjected to high cooling rates and low temperatures develop
tensile stresses due to shrinkage. If these stre~ses exceed the fracture
strength of the HMA pavement layer, transverse cracking develops. HMA
l
mixes which have high stiffness modulus at low temperatures are very
prone to cracking. Mix stiffness at low temperatures is primarily dependent
on the stiffness of the asphalt cement. Therefore, high asphalt cement stiff-
ness at low temperatures is the predominant cause of this type of cracking.
Other factors such as thickness of the HMA layer and subgrade soil type
influence the frequency or extent of these cracks.
Laboratory research supported by field observations indicates that
asphalt cement consistency, largely reported in terms of penetration at
77°F (25°C), and temperature susceptibility (PI or PVN) are the major
asphalt cement properties influencing low temperature cracking.
Increasing the asphalt cement penetration is suggested as the most direct
designer-controlled variable for minimizing cracking. However, care must
be exercised to avoid asphalt cements that are highly temperature suscepti-
ble (that is, asphalt cements that could become extremely stiff at low tem-
peratures regardless of the initial penetration). Table 2 in MSHTO
Specifications M226 for Viscosity Graded Asphalt Cements generally pro-
vides an asphalt cement with lower stiffness at lower temperatures than
that provided in Table 1. However, Table 2 is not adequate for most states
in the northern portions of the United States and Canada. More stringent
specifications to control the temperature susceptibility of asphalt cements
are needed to minimize or eliminate low temperature cracking.
Many researchers have recommended limiting stiffness values of
asphalt cement (40, 46, 1ZJ or HMA (111) for a particular temperature
range based on field experiments. The Pennsylvania Department of
Transportation modified Table 2 of AASHTO Specifications for Viscosity
Graded Asphalt Cements in 1977, based on a limiting stiffness modulus of
275 kg/cm2 (3900 psi) at the minimum pavement temperature of -10°F
(-23°C) and 20,000 sec loading time (1Z). The limiting stiffness modulus
was established based, on the performance of six test pavements. Figure
2-35 shows two test pavements in adjacent lanes. One developed extensive
transverse cracks. during the first winter and the other was free of cracks.
Using the limiting stiffness of 275 kg/cm2 and McLeod's nomograph
- - - - - - - < , - • · - k 'ethod..~~ roioiro11m_allo:wable...EY1Sls_w.e.r.e_determined for va,rious pen-
etration values. Minimum kinematic viscosities were then determined

from the corresponding penetration and PVN values. By specifying the
minimum kinematic viscosity thus determined for each penetration value,
it was ensured that the PVN was not lower than the permissible value.
Some values are given below:
Specified Minimum
MinimumPVN Viscosity at 135°C
Penetration at 25°C Allowable centistokes
60 -0.80 390
65 -0.95 330
70 -1.10 290
75 -1.25 250
It can be seen that higher temperature susceptible asphalt cements can
be used if their penetration values get higher in accordance with those list-
ed above. At higher penetration values, the temperature viscosity lines are
shifted relatively lower at 77°F (25°C), thus the asphalt cement does not
exceed the limiting stiffness modulus at -10°F (-23°C) in spite of steeper
slopes (higher temperature susceptibility) of these lines.
Figure 2-35. Test Pavements in Adjacent Lanes-Foreground Lane

Shows Transverse Cracks, No Cracks in the Background
Lane (after Kandhal, 12) ·
------------'---------'----------------·- ·----
The St. Anne Test Road in Canada has been the most comprehensive
full scale project reported in the literature (1.2). Many researchers have
based the critical asphalt cement stiffness at low temperatures on data from
this project (1Q). The St. Anne test pavements cracked at an asphalt binder
r.
L
stiffness of lxl0 9 N/m2 (145,000 psi) at 0.5-h loading time (that is, equiv-
alent to approximately 6xl0 8 N/m2 at 2.0-h loading time; 5xl0 8 N/m2 at
10,000-s loading time). Readshaw (5JJJ based the British Columbia
asphalt cement specifications on a critical asphalt cement stiffness of 2xl08
N/m 2 at 2.0 h loading time, a value arrived at largely from the St. Anne
Test Road data. Fromm and Phang (1§J suggested l.4xl0 8 N/m2 (that is,
20,000 psi) at 10,000-s loading time as the critical asphalt cement stiff-
ness.
Gaw (if!) has recommended that the low-temperature asphalt cement
specifications be based on a specific nomographic cracking temperature
selected from Figure 2-36, at a performance level appropriate to the antic-
ipated minimum ambient temperature or pavement surface temperature
for the region. The asphalt cement specification limits are represented by a
single line drawn on a logarithm plot of asphalt cement penetration at
25°C (77°F) and 5°C (41 °F) similar to Figure 2-36. Candidate asphalt
cements for which the penetration at 25 and 5°C appear on, or to the right
u
0
300
10
(\J
.,_ 200 I
<1: I
I
z 150
0
.,_
<1: 100
.,_
0:::
80 I
I
w I
I
z -25 I
w 60 I
I
a.. I
I
I
I- I
I
_J 40 I
<1: I
I
I I
a..
CJ)
I
I
I
I
<1: I
I
-0- - 3 4 6 8. 10 15~2~0___ 30 40
2
ASPHALT PENETRATION AT 5°C
· Figure 2-36. Nomograph for Predicting Cracking Temperatures in °C
from asphalt Penetration (after Gaw, 1!l)
of, the specification line are considered acceptable. Both unaged and aged
asphalt cement penetrations yield similar nomographic cracking tempera-
tures, and either can be used to establish the specification limits. This
approach requires the penetration measurement at one additional temper-
ature (that is, at 5°C).
Rutting. Rutting is caused by the progressive movement of materials

under repeated loads either in the asphalt pavement layers or the underly-
ing base. This can occur either through consolidation or through plastic
flow.
Consolidation is the further compaction of HMA pavement by traffic
after construction. When compaction is poor {JO to 12 percent or more air
voids for a mix designed at 3 to 5 percent), the channelized traffic provides
a repeated kneading action in the wheel track areas and completes the con-
solidation to the designed air voids level (usually 3 to 5 percent). A sub-
stantial amount of rutting can occur if very thick asphalt layers are consol-
idated by the traffic.
Rutting also results from lateral plastic flow (permanent deformation)
of the HMA from the wheel tracks. Use of excessive asphalt cement is the
most common cause for this. phenomenon. Too much asphalt cement in
the mix causes the loss of internal friction between aggregate particles and
results in the loads being carried by the asphalt cement rather than the
aggregate structure. Plastic flow can be minimized by using large size aggre-
gate, angular and rough textured coarse and fine aggregates, and providing
adequate compaction at the time of construction.
The consistency (penetration or viscosity) of asphalt cement plays a
relatively small role in the rut resistance of HMA if well-graded, angular
and rough textured aggregates are used. Some increased resistance to rut--
ting can be obtained by using stiffer (high viscosity or low penetration)
asphalt cements. Some states in the southeastern United States have start-
ed to use higher viscosity AC-30 grade in place of AC---20 to improve the
resistance of the mix to rutting. However, stiffer asphalt cements are more
prone to cracking during winter in cold regions especially if they are used
in the surface courses. Certain mineral fillers (or baghouse fines) also
increase the apparent viscosity of asphalt cement at 140°F or 60°C (51)
----,nd---rhns-i:nake-th:e-mix-more-resistan-t-to-rtttting;--However,use-efexcessive-------·-- ----
1
fines should be avoided. .
Stripping. Stripping (more appropriately called moisture-induced dam--

age) can be defined as the weakening or eventual loss of the adhesive bond
L
usually in the presence of moisture between the aggregate surface and the
asphalt cement in a HMA pavement or mixture. Strength of the mixture is
derived from the cohesional resistance of binder and grain interlock and fric-
tional resistance of the aggregate. The cohesional resistance is only fully avail-
able if a good bond exists between the binder and the aggregate (52). If the
bond is poor, failure occurs at the binder-aggregate interface and may result
in premature failure of the mixture and the HMA pavement.
There may be as many as five different mechanisms by which stripping
of asphalt cement from an aggregate surface may occur. Those five mecha-
nisms include detachment, displacement, spontaneous emulsification, pore
pressure, and hydraulic scouring (.53). The stripping phenomenon is dis-
cussed in detail later in the book.
Stripping is a complex problem depending on many variables, includ-
ing the type and use of mix, asphalt cement characteristics, aggregate char-
acteristics, environment, traffic, construction practice, and the use of anti-
strip additives; however, the presence of moisture is the common factor to
all stripping (23, 51, 5..i).
The chemistry of both the asphalt and aggregate at the as-
phalt-aggregate interface plays an important and primary role in the strip-
ping phenomenon. The effect of asphalt cement characteristics especially
the physical properties, although secondary in role, is discussed here. The
most frequently referenced relationship between the characteristics of the
asphalt cement in a paving mixture and the tendency of the mix to strip
relates stripping resistance to the viscosity of the asphalt binder in service.
Binders of high viscosity have generally been observed to resist displace-
ment by water much better than those of low viscosity. Low viscosity, how-
ever, is desirable during mixing operations because a low viscosity fluid has
more wetting power than one of high viscosity. Moreover, high viscosity
asphalt cements cannot be used in many instances because of other con-
siderations such as low-temperature cracking in cold regions and potential
reduction in fatigue life of the surface courses. There is a need to under-
stand the fundamentals of aggregate-asphalt adhesion so that the problem
can be minimized by other means rather than increasing the asphalt
cement viscosity, which is not effective in all cases and which may result in
otl:rerp-e-rformance-probtem. ------· - -
SUPERPAVE ASPHALT BINDER TESTS AND

SPECIFICATION
The current asphalt cement physical tests and specifications have

many limitations as noted below.
1. Tests such as penetration and ductility are empirical and are not
directly related to HMA pavement performance. Any relationship
between the empirical tests and performance has to be gained by
experience as discussed previously.
2. Tests are conducted at one standard temperature such as penetration
at 77°F (25°C) and viscosity at 140°F {60°C) in spite of different
climatic conditions at project sites or geographical regions.
3. The entire range of pavement temperatures typically prevalent at a
project site is not considered. For example, there is no test method
or specification requirements for asphalt cement stiffness at low
temperatures to control thermal cracking.
4. Current test methods and specifications consider short term aging
of asphalt cements during HMA production and construction by
conducting thin film oven (TFO) or rolling thin film oven (RTFO)
tests. However, long term aging of asphalt cements in service is not
considered. As discussed previously, asphalt cements continue to
stiffen during service thereby affecting the HMA performance in
terms of fatigue cracking and low temperature cracking.
5. Asphalt cements can have different temperature and performance
characteristics within the same grading system whether it is a pene-
tration grading or a viscosity grading system.
6. Current tests and specifications have been developed for asphalt
cements based on experience. They are not suitable or applicable to
modified asphalt binders which are gaining. in popularity. See
asphalt additives and modifiers in Chapter 7.
Recognizing the preceding limitations of the current system, a 5-year

Strategic Highway Research Program (SHRP) was launched in 1987,
which included a $50-million research effor.t to develop performance-
based tests and specifications for asphalt binders and HMA mixtures. The
so-called Superpave (Superior Performing Asphalt Pavements) binder tests
and specification (iQ, 5ZJ have 'the following salient features.
1. Tests and specifications are intended for asphalt "binders" which

include both modified and unmodified asphalt cements.
_I_
2. The specified criteria remains constant, however, the temperature at

which the criteria must be met changes in consideration of the
binder grade selected for the prevalent climatic conditions.
3. The physical properties measured by Superpave binder tests are
directly related to field performance by engineering principles.
4. The Superpave binder specification requires the asphalt binder to be
tested after simulating its three critical stages: (a) the first stage is
represented by original asphalt binder which has to be transported,
stored, and handled prior to mixing with aggregate, (b) the second
stage is represented by the aged asphalt binder after HMA produc-
tion and construction (short-term aging), and (c) the third stage is
represented by the asphalt binder which undergoes further aging
during a long period of time in service.
5. The entire range of pavement temperatures experienced at the pro-
ject site is considered.
6. Tests and specifications are designed to eliminate or control three
specific types of HMA pavement distresses: rutting, fatigue crack-
ing, and thermal cracking. Rutting typically occurs at high tem-
peratures, fatigue cracking at intermediate temperatures, and ther-
mal cracking at low temperatures.
7. The Superpave asphalt binder test procedures and specification were
developed in SI units which will be used in this section without
English units.
Superpave asphalt binder technology is still in an emerging phase at

this time (1996). Some tests and criteria are likely to be revised in the
future as this technology is implemented.
Superpave Physical Tests for Asphalt Binders
Table 2-8 gives the list of testing equipment to conduct various

Superpave physical tests, the related purpose for testing, and the related
performance parameter being partly influenced by the asphalt binder. It
must be recognized that the HMA mixture properties and the pavement
structure significantly affect the HMA performance in addition to asphalt
binder properties.
Rolling Thin Film Oven (RTFO). The rolling thin film oven
(RTFO) test, AASHTO T240 and ASTM D2872, was described earlier in '
this chapter. Like the thin film oven (TFO) test, the RTFO simulates the
asphalt binder aging during the manufacture and construction of HMA
I
Table 2-8. Superpave Asphalt Binder Testing Equipment and Purpose
8 8-: ....,
e:... ::r, '"O
• 'Tj~ ~ el "'
~
>-g::io...,::ro
>;rlrl(bo-8
8"' 0 0 : 3 ... , 0
Equipment Purpose Performance Parameter I .8. ~ ~ e:. 5.. g
.... ;:I (b " '
.... o' ;:I A- .....
I>,) .... ~ .... 8 ....,
Rolling Thin, Film Oven (RTFO) Simulate binder aging (hardening) during Resistance to aging (durability ) during s,. 8
(b (b
::r~
... • (b ::r ::r
I
HMA production and construction construction .... P..8--::;i ~ (b
! s,. ~ rl (b rl :;q
I>,) ::rP-~....,
;:i::t"o,-., 'Tl
.... . ;:I A-
Pressure Agjg Vessel (PAV) Simulate binder aging (hardening) during Resistance to aging (durability) during ::n
< g.. g:
;:I
s,. e: • 0
o,-.•p,.o>-t~
HMA service life service life
5~~ e; g:ie; ::r:
~
i:: o::r ~ "'
.... p,.'"O
..., 0-
. Rootfoo,l lnmct<t (RV) Measure binder properties at high con- Handling and pumping " ' ~ o e:.,.P-h
o
ri I>,) "' rl i:: (")
...a (fQ ;:I 0- ::i. ~
2:::
x·
i);-
struction temperatures s. s· g s·~ A- "Cl
::r
ri (£Q 8 Ft" ...,
o- !'¾-
A-p.,o>-t@:~ 2:::
'::I") ;:I ~ A- .... I>,)
Dynamic Sh~ar Rheometer (DSR) Measure binder properties at high and Resistance to permanent deformation 0 A- Vi:: ;:I i:: ~
::! 0::: ~- (b (£Q ~
i
intermediate service temperatures (rutting) and fatigue cracJcing I ~'--'P-8 ?,....._
.--o 1--t..._.,,e,. 2:::
0
...., rl (b
"Ti r-t rn }:::: S
:=;· ~-
Bending Beajn Rheometer (BBR) Measure binder properties at low service Resistance to thermal cracking 0 ~;:I0::ip,.n .... 0 ~
0
rl"'rlCtQ,-.,0 t;)
temperatures I ~ 0 e:... I>,) ~ g
r-t a o 0a.~ i::s· ~-
!,l
'-< ~
~
V
Direct Tension Tester (DTT) Measure binder properties at low service Resistance to thermal cra~king '--I I>,) ? ;:;.; e:... A.
Vl ;:I,....._ i:: ~ ()
temperatures I 9'-<~~
,..,1:i::fl-~
;:I .... ;:I '"O
0
~
:::C
gj;3">~&;~ 0
a.
~@ if9 ~ 0
::,
The RTFO test provides an aged asphalt binder for further testing by
the dynamic shear rheometer (DSR), and allows the determination of the
mass of volatiles lost from the binder during the test. The amount of
volatiles lost indicates the amount of aging that may occur during HMA
production and construction. Some asphalt binders gain weight, rather
than lose weight, during the RTFO aging because of the oxidative prod-
ucts formed during the test.
Pressure Aging Vessel (PAV). Pressure aging has been used in asphalt
research and for aging rubber products for many years. A method involv-
ing the pressure aging was developed by Dr. D. Y. Lee at Iowa State
University for long term aging of asphalt cements, and is currently a stan-
dard method for aging rubber products (ASTM D454 and ASTM D572
in Reference 5.fi). Pressure aging of asphalt binders has the following advan-
tages: (a) limited loss of volatiles, (b) the oxidative process is accelerated
without resorting to high temperatures, (c) an adequate amount of asphalt
binder can be aged at one time for further testing, and (d) the test is prac-
tical for routine laboratory testing.
The pressure aging vessel (PAV) was developed in SHRP to simulate
the asphalt binder aging that occurs during 5-10 years of in-service HMA
pavements (i2). Since the asphalt binder in the HMA pavement has
undergone short-term aging during production and construction, the PAV
is used to age RTFO residue. Figure 2-37 shows a schematic of the PAV
and related accessories.
air
pressure
temperature
probe
/ asphalt
~
pressure vessel sample rack sample pan
Figure 2-37. Pressure Aging Vessel (2Q)

76 Hot Mix Asphalt Materials, Mixture Design and Constructio.n
The RTFO residue is exposed to high air pressure and temperature for
20 hours in the PAV to simulate the effect oflong term aging. A cylinder
of dean, dry compressed air with a pressure regulator, release valve, and a
slow-release bleed valve is used to supply and regulate air pressure. The
PAV composed of stainless steel must be able to operate under the pressure
(2070 kPa) and temperature conditions (90°, 100°, or 110° C) of the test.
The vessel accommodates a sample rack on which ten sample pans can be
placed for aging. The PAV is placed in an oven for 20 hours to maintain
the desired temperature inside the vessel. PAV equipment which incorpo-
rates the pressure aging chamber and the oven built as one unit, is also
available.
RTFO-aged asphalt binder is heated until fluid, stirred and pla~ed in
the PAV pan which can accommodate a 50-gram sample. The sample rack
containing sample pans is placed in an unpressurized PAV preheated to the
desired test temperature depending on the design climate. The lid of the
PAV is secured quickly to minimize heat loss. When the PAV reaches with-
in 2°C of the desired temperature, a pressure of 2070 kPa is applied using
the valve on the air cylinder. After 20 hours, the air pressure is released
slowly (over a period of 8-10 minutes) using the bleed valve. A quick
pressure release may cause foaming in the binder samples. The sample
pans are then placed for 30 minutes in an oven maintained at 163°C to
remove entrapped air from the samples. The samples are then stored for
further testing (such as dynamic shear rheometer, bending beam rheome-
ter, and direct tension tester).
Rotational Viscometer (RV). The rotational viscometer (RV) has been

adopted in Superpave for determining the viscosity of asphalt binder at
high construction temperatures (above 100°C) to ensure that the binder is
sufficiently fluid for pumping and mixing. Most asphalt binders behave as
Newtonian fluids (stress response not dependent on shear rate) and have a
totally viscous response at such high temperatures. Therefore, a viscosity
measurement is sufficient to represent workability of the binder@). The
RV is more suited for testing many modified asphalt binders (such as those
containing crumb rubber modifier) compared to capillary viscometer
because the latter can get dogged up partially inhibiting smooth flow. The
Superpave binder specification limits the viscosity to 3 Pa.sat 135°C.
This method of measming viscosity--is-given---in---det-ail-i-n-A8'.:f-M----
Method D4402 or AASHTO TP48, "Viscosity Determination of Unfilled
Asphalts Using the Brookfield Thermosel Apparatus." Figure 2-38 shows
the principle of rotational or Brookfield viscometer. Rotational viscosity is
determined by measuring the torque required to maintain a constant rota-
tional speed (20 RPM) of a cylindrical spindle while submerged in an

asphalt binder at a constant temperature. The measured torque is directly
related to the viscosity of the binder sample, the latter is automatically
determined and displayed by the viscometer.
torque
spindle
Figure 2-38. Principle of Rotational Viscometer (56)
The test equipment to measure rotational viscosity consists of two

items as shown in Figure 2-39: Brookfield viscometer and Thermosel™
system. The Brookfield viscometer is comprised of a motor, spindle, con-
trol keys, and digital readout. Many sizes of the spindle are available for
the Brookfield viscometer to suit the viscosity range of the binder being
tested. Most asphalt binders can be tested with only two spindles: Nos.,
21 and 27, the latter is used most frequently. The Thermosel™ system
consists of a stainless steel sample chamber, a thermo container equipped
with electric heating elements, and a temperature controller to set the test
temperatures. ,
Brookfield
digital viscometer
readout
control
keys spindle extension
ermo;;
container
Figure 2-39. Schematic of Brookfield Viscometer (56)

Approximately 30 g of asphalt binder is heated in an oven so that it is

sufficiently fluid to pour. The binder is weighed into the sample chamber.
The amount of binder used varies with the size of the spindle. Less binder
will be required with a larger spindle. Typically, no more than 11 grams
are used. The sample chamber containing the binder sample is then placed
in the thermo container and is ready for testing when the desired temper-
ature (135°C for specification) stabilizes. The spindle is lowered into .the
chamber containing hot binder sample, and is coupled with the viscome-
ter. After the sample temperature stabilizes again, the viscometer motor is
turned-on at 20 RPM. The viscometer reading is displayed digitally along
with spindle number, RPM, and temperature. The digital display of the
rotational viscosity is in units of centipoise (cP) whereas the Superpave
binder specification uses Pa.s. The following factor is used to convert from
cP to Pa.s. ·
1000 cP = 1 Pa.s
Therefore, the Brookfield viscosity output in cP is multiplied by 0.001
to obtain viscosity in Pa.s.
Most agencies use equiviscous temperatures for mixing and com-
paction during HMA mix design (see Chapter 4). This can be accom-
plished by measuring rotational or Brookfield viscosity at one more tem-
perature other than 135°C; and plotting the temperat~re-viscosity curve.
Dynamic Shear Rheometer (DSR). The dynamic shear rheometer

(DSR) is used to characterize the viscous and elastic behavior of asphalt
binders at high and intermediate service temperatures. The DSR which is
also generically known as dynamic rheometer or oscillatory shear rheome-
ter, has been used in the plastics industry for many years. The DSR mea-
sures the complex shear modulus G* (G star) and phase angle 8 (delta) of
asphalt binders at the desired temperature and frequency of loading.
Complex modulus G* can be considered as the total resistance of the
binder to deformation when repeatedly sheared. As shown in Figure 2-40,
complex modulus G* consists of two components: (a) storage modulus G'
(G prime) or the elastic (recoverable) part, and (b) loss modulus G" (G
double prime) or the viscous (non-recoverable) part (26).
Both temperature and frequency ofloading significantly affect the val-
ues of G* and 8 for asphalt binders. Asphalt binders behave like elastic
solids as shown by the horizontal arrow in Figure 2-41 at verr-_,l=ow_,_,_t=e=m,.._-_ _ _ __
- + - - - - · peramres. On the ~th~r hand, asphalt binders behave like viscous fluids at
high temperatures (well above pavement temperatures) as shown by the
vertical arrow in Figure 2-41. However, at usual pavement service tem-
peratures most asphalt binders behave like viscoelastic materials (repre-
Ch_apter 2-Asphalt Refining, Uses, and Properties 79
sented by arrows 1 and 2 in Figure 2-41), that is, they simultaneously act
like elastic solids and viscous liquids. When viscoelastic materials are
loaded, part of the deformation is elastic (recoverable) and part is viscous
(non-recoverable). The DSR gives a complete picture of the behavior of
asphalt binders at pavement service temperatures by measuring both G*
ando.
Viscous
Part
Viscous
Part
G"
G"
G'
Elastic Part
Elastic Part
Asphalt A Asphalt B
Figure 2-40. Components of Complex Modulus G* (i6)
Viscous Behavior
both viscous and

elastic behayior
Elastic Behavior
_ _ _ ____,F.....igur.e...2d:l.-Yisco_us__andElastkBehavior of Asphalt Binders (i6)
Figure 2-40 shows two asphalt binders with the same complex modu-
lus G* (indicated by length of the diagonal) but with different phase
angles ◊. However, Asphalt B has a larger elastic component compared to
Asphalt A. When both asphalt binders are loaded, Asphalt B will display
more elastic (recoverable) deformation and less viscous or non-recoverable
(permanent) deformation compared to Asphalt A. This example demon-
strates that complex modulus (G*) alone is not sufficient to characterize
asphalt binders, phase angle (3) is also needed.
Applied Stress
Osei!::~~))
Position of
Oscillating Plate B
Plate Fixed Plate
C
1 cycle
Figure 2-42. Basics of Dynamic Shear Rheometer (.2Q)
The elastic component or storage modulus is related to the amount of

energy stored in the sample during each testing cycle. The viscous com-
ponent or loss modulus is related to the energy lost during each testing
cycle through permanent flow or d~formation (5JJ.). It is important to note
that a real viscoelastic material does have a significant amount of delayed
elastic response which is recovered completely when the sample is no
longer loaded. Both the storage and loss moduli contain a portion of this
delayed elastic response. In this sense, the storage and loss moduli cannot
be considered purely elastic or viscous respectively (§.D.
The principle of operation of DSR is simple. The test procedure is
given in AASHTO TP5. As shown in Figure 2-42, the asphalt binder
sample is sandwiched between a fixed plate and an oscillating plate. When
torque is applied to oscillating plate, it starts from point A and moves to
point B. From point B, the plate moves back and goes to point C passing
the point A. From point C it returns back to point A. This movement
comprises one cycle of oscillation. If two oscillation cycles occur in one
----· second, the frequency of0seillat:i0n-wot1:lcl-be-two-cydes-perseco11d-ortwo
hertz (Hz). The frequency of oscillation can also be expressed as the cir-
cumferential distance (radians) traversed by the oscillating plate in one sec-
ond. All Superpave DSR tests are conducted at a frequency of 10 radians
per second which is equivalent to about 1.59 Hz.
Chapter 2-Asphalt Refining, Uses, and Properties '81
Two types of oscillatory shear rheometers are available: constant stress

and constant strain. ,Constant stress rheometers use a fixed torque to oscil-
late the top plate or spindle and, therefore, the strain will vary. Constant
strain rheometers work by moving the plate a fixed distance from point A
to point B and measuring the necessary torque to cause this movement.
Superpave binder tests are conducted in the constant stress mode. "
Elastic: 8 = 0 deg Viscous: 8 = 90 deg
'tmax
I
v•mo
Applied
:,_v•
I,
Shear
Stress I
Ii time lag = 8
Resulting ~
Shear
Strain
Figure 2-43. Stress-Strain Output of a Constant Stress Rheometer (i6).
The thickness of asphalt binder sample sandwiched between the fixed

plate and the oscillating spindle depends on the test temperature. High
test temperatures (46°C or greater) require a small gap of 1000 microns
(1 mm). Lower test temperatures in the 4° to 40°C range require a larger
gap of 2000 microns (2 mm). Also, two spindle diameters are used: large
spindle (25 mm) for high test temperatures and small spindle (8 mm) for
low test temperatures (i6).
After the asphalt binder sample is sandwiched between the fixed plate
and the spindle the desired test temperature has to be achieved within
±0 .1 °C using a water or air bath. When the spindle is oscillated back and
forth with constant stress, the resulting strain is monitored. The relation-
ship between the applied stress and the resulting strain (Figure 2-43) is
used to compute G*. G* is the ratio of maximum shear stress (tma.J to
maximum shear strain (rmaJ or 'tmax + 'Ymax· The time lag between the
applied stress and the resultmg stram 1s the phase angle 8~ · For a perfectly
elastic material; there is an instant responseand, therefore, the time lag
between the applied stress and the resulting strain or phase angle 8 is zero.
When a viscous liquid such as hot asphalt binder is tested, the time lag is
large and the phase angle 8 approaches 90 degrees. However, asphalt
binders display a stress-strain response between these two extremes at

pavement service temperatures as shown in Figure 2-44. The resulting
phase angle is between Oto 90 degrees (2Q).
The following formulas are used by the DSR software to calculate 'tmax
and "!max·
2T
nr
'Ymax
' er
h
where,
T maximum applied torque,

r radius of binder specimen/plate (either 12.5 or 4 mm),
0 deflection (rotation) angle,
h specimen height (either 1 or 2 mm).
Viscoelastic: O < 8 < 90 °

'tmax
Applied
Shear __________.......,_ _ __
Stress
time
'tmax
G*= 'Ymax
8 = time lag
Resulting
Shear
Strain--------·--------
time
j
Figure 2-44. Stress-Strain Response of a Viscoelastic Material
In constant stress rheometers used for Superpave binder testing, the con-
stant stress used in testing should be such that the resulting shear strain is
approximately 10-12 percent when relatively soft materials are tested at

high temperatures (for example, unaged binders and RTFO aged binders),
and about one percent when relatively hard materials (for example, PAV
residues tested at moderate temperatures) are tested (2Q). The DSR auto-
matically sets the constant stress levels for the desired percentage of strain.
The DSR also computes and reports G* and o on completion of the test.
The DSR is used in the Superpave specification to measure the prop-
erties of the asphalt binder at high and intermediate pavement service tem-
peratures. The actual test temperatures are determined based on the antic-
ipated in-service temperatures in which the asphalt binder will function.
A weather database of about 7500 North American weather stations over a
20-year period is used to determine the anticipated pavement service tem-
perature at the project. The high temperature (maximum design temper-
ature) is determined by averaging the maximum HMA pavement temper-
ature over a seven-day period throughout the summer. Average values for
successive seven-day periods are then determined and the largest
seven-day average maximum pavement·temperature is chosen as the max-
imum design temperature (@. The high temperature DSR testing is
accorn:plished at this temperature. The intermediate testing temperature is
established as an approximate average of the seven-day average maximum
and the minimum design temperatures. The minimum design tempera-
ture is chosen as the minimum pavement temperature expected to occur at
the design location (.62.). This expected minimum is obtained from the
low temperatures recorded in the database over a period of 20 years. The
Superpave computer software is used to determine these HMA pavement
design temperatures from the air temperatures in the weather database.
Physical property measureipents such as G* and o for a given asphalt
binder are made three times with the DSR. The asphalt binder is tested in
the DSR in its original (unaged), oven aged (RTFO residue), and PAV
aged conditions. Original and RTFO aged asphalt binder samples are test-
ed at the maximum design temperature to determine the binder's ability to
resist rutting. PAV aged samples are tested at the intermediate design tem-
perature to determine the asphalt binder's ability to resist fatigue cracking.
A discussion of parameters used to quantify the asphalt binder's contribu-
tion in resisting rutting (permanent deformation) and fatigue cracking fol-
lows.
Rutting Parameter:
For rutting resistance, a high complex modulus G* value and low

phase angle oare both desirable (60). The higher the G* value, the stiffer
and thus the more resistant to rutting the asphalt binder will be. The lower
the o value, the more elastic the asphalt binder.
The increased elasticity makes the asphalt binder more resistant to per-
manent deformation. The high temperature measurements for rutting are
made on the original and RTFO aged samples. The RTFO aged samples
represent the. asphalt binder in the early stages of its life immediately after
placement and before long-term oxidative aging begins. Since long term
oxidative aging increases the stiffness of an asphalt binder, it continues to
increase the binder's resistance to rutting. Testing of the asphalt binder
prior to long term oxidative hardening, therefore, is critical with respect to
rutting. The specification for DSR testing of the original asphalt binder
was added as a safeguard against those asphalt-binders whose RTFO aged
residues do not indicate the actual aging during HMA production and
placement in the field.
The SHRP researchers considered rutting as a stress controlled, cyclic
loading phenomenon in determining the rutting parameter chosen for
specification purposes. With each traffic loading cycle, work is being done
to deform .the HMA pavement surface. A part of this work is recovered by
elastic rebound of the surface while some is dissipated in the form of per-
manent deformation and heat. In order to minimize permanent deforma-
tion (rutting), the amount of work dissipated during each loading cycle
must be minimized. Mathematically, the work dissipated per loading cycle
at a constant stress can be expressed as follows (§fl}:
W =1txcr 2 1 ]
C O [G* /sino
where
We work dissipated per load cycle,

, cr0 stress applied during the load cycle,
o
G* complex modulus,
= phase angle. j
1
This equation indicates that the work dissipated per loading cycle is ·,
mversely proportional to G*/smo. For this purpose, the Gorysino_p_ar_am--e--- - - - - - - -
ter was chosen as a Superpave asphalt binder specification. It can be seen
from the preceding equation that the work dissipated per loading cycle can
be decreased by either increasing the value of G* and/or decreasing the
value of o. This appears logical because increasing the G* of an asphalt
binder makes it stiffer and thus more resistant to rutting. Also, decreasing
the O of an asphalt binder makes the asphalt binder more elastic, again
increasing its resistance of rutting.
Fatigue Parameter:
To determine the parameter for resistance to fatigue cracking for spec-

ification purposes, fatigue cracking was considered a strain controlled phe-
nomenon. Fatigue cracking is typically considered a stress controlled phe-
nomenon in thick HMA pavement layers and a strain controlled phenom-
enon in thin HMA pavement layers. Since fatigue cracking is known to be
more prevalent in thin pavements, the SHRP researchers assumed that it
should be considered mainly a strain controlled phenomenon @.
Mathematically; the work dissipated per loading cycle at a constant strain
can be expressed as follows:
where E is the strain and the other variables are as previously described.
This equation indicates that as G* and/or o are increased, more work will
be dissipated per traffic loading cycle. As G* decreases, the asphalt binder
becomes less stiff and thus able to deform without building up large stress-
es. Additionally, asphalt binders with low o values will be more elastic and
thus can regain their original condition without dissipating work. Work
can be dissipated in several different ways; heat dissipation, plastic flow,
cracking, and crack propagation can all dissipate work. However, all of •
these contribute to pavement distresses. The lower the amount of energy
dissipated per cycle, the lower the likelihood of fatigue cracking or any
other damage phenomena to occur. The G*sino parameter was, therefore,
chosen for Superpave specification purposes to limit the total amount of
energy dissipated thereby minimizing fatigue cracking @. PAV aged
samples are tested in the DSR for determining the asphalt binder's ability
to resist fatigue cracking. As the asphalt binder ages during its service life,
it becomes stiffer thus causing it to become more susceptible to fatigue
cracking. PAV aged samples are, therefore, used to test the asphalt binder
to simulate the most critical in-service condition.
nere are tlrree-s-ep-a:rate--specification--requirements for the DSR test
parameters. When the unaged asphalt binder is tested, the G*/sino value
must be a minimum of 1 kPa. When the RTFO residue is tested for the
same asphalt binder, the G*/sino value must be a minimum of 2.2 kPa.
These two minimum limits were established for rutting purposes as previ-
ously discussed. The third DSR specification requirement is that the G*sin
8 value of the PAV aged asphalt binder must be no more than 5000 kPa.
This maximum specification limit was established to address fatigue crack-
ing in HMA pavements.
Each of these specification limits was established by consensus of the
Federal Highway Administration Asphalt Binder Expert Task Group
(ETG). This group consists of persons from industry, specifying agencies,
and academia. The group fir;st established the minimum limit of 1 kPa for
original asphalt binders. This was done by referencing the conventional
viscosity data for paving asphalt cements at 60°C. This pavement service
temperature was thought to be representative of a moderate climate where
the use of AC-10 viscosity graded asphalt cement (viscosity of 1000 Poises)
would be common. When unaged AC-10 asphalt cements are tested in
the DSR at 10 rad/second, their G*/sin8 values are approximately 1 kPa.
Therefore, _the minimum limit of 1 kPa for the unaged asphalt binder was
considered reasonable from a rutting standpoint because AC-10 asphalt
cements were providing reasonable service in moderate climates.
The minimum limit of 2.2 kPa for the G*/sin8 value of RTFO aged
asphalt binders was established based on the 1 kPa value of the unaged
asphalt binder. RTFO test data on asphalt cements indicated that, on the
average, the aging index (viscosity after RTFO/viscosity before RTFO) for
asphalt cements ranged from 2 to 2.5. In other words, asphalt binders
become from 2 to 2. 5 times stiffer when aged in the RTFO. Therefore, the
average of the range, 2.2 was used to establish the 2.2 kPa minimum limit
for RTFO aged asphalt binders.
The original specificatioti. for fatigue cracking called for a maximum
limit of 3000 kPa. However, when the test results of the 42 asphalt binders
used to set the limits were investigated, half would not meet the 3000 kPa
maximum criterion. The ETG, therefore, deemed the 3000 kPa maximum
a
limit to be too restrictive. It was decided to use higher value of 5000 kPa
_J
until field validation data could be gathered and analyzed. Since about 15
percent of the 42 asphalt binders would not meet this limit, the value of
5000 kPa appeared reasonable at the time.
Bending Beam Rheometer (BBR). The Bending Beam rheometer

----. (BBR) tests asphaltbinders.ado.:w_pavement service temperatures to deter-
mine the binder's propensity to thermal cracking. Thermal cracking of li
HMA pavements results when the temperature drops rapidly at cold tern- 1
peratures. As the pavement contracts, stresses begin to build up within the

HMA pavement layers. If the contraction occurs very rapidly the stresses
can build and eventually exceed the stress relaxation ability of the HMA
Chapter 2-Asphalt Refining, Uses, and Propertie~ 87
pavement. When this occurs, the pavement develops cracking as ; way to

relieve stress. Thermal cracking can result from one thermal cycle when a
critically low temperature is encountered, or from" thermal cycling where
the temperature cycles up and down, but always remains above the critical
low temperature. Cracking caused by a single thermal cycle is related to
the asphalt binder stiffness at the temperature at which cracking occurs.
This temperature is referred to as the limiting stiffness temperature. The
limiting stiffness temperature is, therefore, defined as the pavement service
temperature at which a certain stiffness value is reached after a specified
loading time.
A device to determine the properties and response of asphalt binders
at low temperatures was initially developed in the 1980's (§3). This equip-
ment was modified and updated as part of the SHRP binder research. The
resulting machine was named the Bending Beam Rheometer (BBR). The
testing procedure is given in MSHTO TP 1.
The BBR uses a transient creep load, applied in the bending mode, to
"load an asphalt beam specimen held at a constant low temperature. The
temperature is maintained by a fluid bath filled with a mixture of ethylene
glycol, methanol, and water. Both the specimen geometry and the loading
technique closely resemble ASTM D 790, Test Methods for Flexural
Properties of Unreinforced and Reinforced Plastics and Electrical
Insulating Materials (§1). The BBR consists of a loading mechanism, tem-
perature control bath, and data acquisition system as shown in the
schematic in Figure 2-45. The dimensions of the beams were selected in
conformity with ASTM D 790 and so that the elementary Bernoulli-Euler
/ Deflection
1-----1■1/ Transducer
r
,,,,.-- Air Bearing
Control and
Data Acquisition Load Cell
/ Fluid
r--'--'--~/ Bath Loading
Frame
Supports
Figure 2--45. Schematic of Bending Beam Rheometer (2Q,)

theory of bending prismatic beams would apply (52). The BBR uses a
blunt-nosed shaft to apply a midpoint load to the asphalt beam resting on
two end supports. The load cell mounted on the loading shaft is enclosed
in an air bearing to eliminate frictional resistance. A transducer affixed to
the shaft monitors deflections of the beam. Loads are applied by pneu-
matic pressure. The data acquisition system records the load and deflec-
tion test results and calculates two parameters: (a) creep stiffness, S(t)
which is a measure of how the asphalt binder resists the constant (creep)
loading, and (b) m-value which is a measure of the rate at which the creep
stiffness changes with loading time.
Asphalt binder beams measuring 125 mm in length, 6.25 mm in
width, and 12.5 mm in height are formed bya,pouring heated binder into
aluminum or silicon.e molds. The beam specimens are demolded after
cooling the mold assembly in a freezer for 5-10 minutes. The beam is then
kept in the test bath for 60±5 minutes before testing. After pre-loading
procedures (iZ), a 100-gram load is applied to the beam for a total of 240
seconds. The deflection of the beam is recorded during this loading peri-
od as shown in Figure 2-46. Load and deflection versus time plots are also
generated on the computer screen. After 240 seconds, the BBR software
performs all calculations to determine creep stiffness and m-value as fol-
lows.
Classic beam analysis theory is used to calculate the creep stiffness of

the asphalt binder beam at 60 seconds loading time (5.i):
S(t) = pp
4 bh 3 8(t)
where,
S(~ =' creep stiffness at time, t = 60 seconds
P = applied constant load, 100 g (980 mN)
L distance between beam supports, 102 mm
b beam width, 12.5 mm
h beam thickness, 6.25 mm
8(t) = deflection at time, t = 60 seconds
Although creep stiffness-is-desireclar._the___minimum HMA eaverri.ent

design temperature after two hours loading time based on past research,
the testing time will be too long. It was determined from time tempera-
ture superposition principle discussed later that by raising the test temper-
ature l 0°C, an equal creep stiffness is obtained after a 60-second loading
time which is more realistic for routine testing.
Deflection
simulates stiffness
/ after 2 hours at
/ 10 Clower temp
60
Log Time, s
Figure 2-46. Deflection versus Time for BBR Ci@
The second parameter obtained by the BBR is the m-value. The

m-value indicates the rate of change of the stiffness, S(t), with loading time
as shown in Figure 2-47. In other words, them-value is the slope of the
log creep stiffness versus log time curve at any time, t. The Superpave
binder specification requires m-value be greater than or equal to 0.300
when measured at 60 seconds (26).
Log Creep
Stiffness, S
/ slope = m-vaiu,
60
Log Loading Time, s
Figure 2-47. m-value from Bending Beam Rheometer (i@

A discussion of the BBR test development and specification require-

ment follows.
In developing the BBR test method to determine the rheological para-
meters that affect thermal cracking, the SHRP researchers performed a
series of tests to determine the critical factors involved such as time and
temperature dependency, physical hardening and conditioning time, and
the effect ofload levels (52).
To develop performance based specifications, the laboratory test
intended for thermal cracking specification purposes should simulate field
conditions. However, most researchers (Q, 15) have correlated thermal
cracking of in-service HMA pavements with asphalt binder stiffness values
estimated at loading times of 3,600-20,000 se£_onds (1-5.5 hours). These
loading times are rather excessive for a laboratory test procedure used for
specification purposes. Therefore, the time-temperature superposition
principle was used to shorten the laboratory test time. Using this princi-
ple, results are obtained for a shorter loading time using one or more tem-
peratures and the response at the longer loading time estimated.
Preliminary tests showed that a loading time of 240 seconds was reason-
able. A series of tests performed at different temperatures indicated that at
the lowest pavement temperatures, the time-temperature superposition
shift functions for many asphalt binders are similar regardless of loading
time. This indication led to the establishment of a single offset tempera-
ture for estimating the response_at longer loading times. It was determined
that an offset of l 0°C above the lowest pavement temperature could be
used to equate the BBR stiffness at 60 seconds loading time to the asphalt
binder stiffness at 7,200 seconds loading time in the field at the lowest
pavement temperature (iJ!). A limiting stiffness based on 2 hours (7,200
secqnds) was selected as one of the criteria based upon the literature (@.
The BBR loads the beams for 240 seconds and reports the stiffness val-
ues at loading times of 8, 15, 30, 60, 120, and 240 seconds. These values
were chosen because they are fairly equally spaced on a logarithmic time
scale (52). These data points, along with the following equation, are used
to determine the shape of the master stiffness (creep compliance) curve for
the asphalt binder.
_ S(t) =A+ B log(t) + C [log(t)] 2
where,
S(t) = the asphalt binder stiffness,
t = time,
A, B, and C = constants.
Chapter 2-Asphalt Refining,_?ses, and Properties 91
Since the time dependency of asplialt binders varies, the shape of the
master stiffness curve as well as the stiffness at 2. hours loading time are
important. Therefo~e, the slope of the master stiffness curve is also used
for specification purposes (§2). The stiffness, S(t), is a measure of the ther-
mal stresses developed in the HMA pavement as a result of thermal con-
traction. The slope of the stiffness curve, m, is a measure of the rate of
stress relaxation by asphalt binder flow as discussed previously (52). The
effect of these two specification parameters, S(t) and m-value, on thermal
cracking is analogous to the effect of G* and 3 on rutting and fatigue
cracking. As S(t) increases, the thermal stresses developed in the pavement
due to thermal shrinking also increase, and thermal cracking becomes
more likely. On the other hand, as the m-value decreases, the rate of stress
relaxation decreases. In other words, as the slope of the asphalt binder stiff-
ness curve flattens, the ability of the HMA pavement to relieve thermal
stresses by flow decreases. This again would increase the propensity of
thermal cracking in the pavement. Accordingly, the Superpave binder
specification requires a maximum limit of S(t) at 60 seconds and a mini-
mum .m-value. The numerical limits for these two parameters were cho-
sen as S(60)=300 MPa maximum, and m=0.300 minimum. These limits
were established based on data from previous studies as well as the data
obtained by SHRP (52). A recent report (§5) based on six asphalt cement
binders used in Pennsylvania in 1976 (15) has offered some insight into
the field validation of these two parameters. The six test sections, which
used AC-20 asphalt cements from different sources, were monitored for
thermal cracking for a period of 7 years. Samples of the original asphalt
binders were retained and tested in 1995 using the Superpave binder
equipment. The results indicated that the S and m-values chosen in
Superpave binder specifications were generally reasonable. The low tem-
perature behavior of all but one of the AC-20 asphalt cements could be
explained by these values (§5).
Physical hardening is a phenomenon which was observed by the
SHRP researchers. The phenomenon is believed to be caused by delayed
volume shrinkage and results in significant time-dependent hardening
(@. How this phenomenon affects the HMA pavement performance is
not known. Therefore, no specification limit could be established for
physical hardening (@. However, the results of the physical hardening
---------tc-h.a.t-occm-o:v.er-2--tio.urs_ofiso..thermaLs.to.rage_can_be reported according
to Superpave binder specification. To minimize the effect of physical hard-
ening on the BBR test data, a standard conditioning time of 60±5 minutes
was selected as mentioned before.
For the BBR test results to be valid, the measurements must be made
in the linear viscoelastic region. A small experiment which involved test-

ing replicate beams with loads ranging from 20-150 grams was conducted
by SHRP researchers to ensure that this was the case. It was concluded that
in this load range, stress and strain. were independent. Based on the test
results, the researchers established a standard load of 100 grams 02).
Direct Tension Tester (DTT). There is a fairly good relationship

between the stiffness of asphalt binders and the amount of stretching (duc-
tility) they undergo before breaking at low temperatures. Typically, softer
asphalt binders are more ductile and stiffer asphalt binders are more brittle
(those that break without much stretching). Creep stiffness of the asphalt
binders as measured by the BBR is not adequate to completely character-
ize the low temperature behavior of the asphalt-binders in terms of thermal
cracking. There are asphalt binders, particularly some modified asphalt
binders, which may have high creep stiffness (more than 300 MPa) but do
not crack because they can stretch further before breaking. Therefore, the
DTT was developed to address these stiff, ductile asphalt binders. The
DTT is only used for testing asphalt binders which have a BBR creep stiff-
ne.ss between 300 and 600 MPa at low pavement service temperatures. If
the creep stiffness is less than 300 MPa, the DTT need not be used. The
test procedure is given in AASHTO TP3.
t
C)
Load.
-~-- change in length (AL)
failure strairilcr) = =============~~=======--------
effective gauge length (Le)
Figure 2-48. Direct Tension Test Ci@

Chapter 2~Asphalt Refining, Uses, and Properties 93
The DTT measures the low temperature ultimate tensile strain of the
asphalt binders, typically in the 0° to -36°C range. Like the BBR test, the
asphalt binder is subjected to RTFO and PAV aging before the DTT test.
A small dog-bone shaped specimen is loaded in tension at a constant rate
(I mm/min.) as shown in Figure 2-48 until it breaks. The failure strain in
the specimen is the change in length (LlL) divided by the effective gauge
length (Le). The failure strain in the DTT corresponds to the stress where
the load on the specimen reaches its maximum value, and not necessarily
the load when the specimen breaks. The asphalt binder must exhibit a fail-
ure strain of at least 1.0 percent to meet the Superpave binder specification.
At this time (1996), the DTT equipment is being modified for a bet-
ter control of the test temperature and a more accurate measurement of the
failure strain.
Superpave Asphalt Binder Specification

The Superpave asphalt binder specification (AASHTO MPI-93) is
shown in Table 2-9. The physical properties included in the specification
are measured by the test equipments described earlier. As mentioned
before, the physical properties remain constant for all performance grades
(PG), but the temperature at which these properties must be achieved
varies from grade to grade depending on the climate in which the asphalt
binder is expected to perform (Figure 2-49). For example, a PG 64-28
grade is intended for use in an environment where an average seven-day
maximum pavement temperature of 64°C and a minimum pavement
design temperature of -28°C, are likely to be experienced.
Performance Grade PG52 PGSS
-10 -16 -22 -28 -34 -40 -46 -16 -22 -28
Average 7-day Maximum Pavement
Desi nTem ,C <52 <58
Minimum Pavement
Desi n Tern erature, C >-10 >-16 >-22 >-28 >-34 >-40 >-46 >-16 >-22 >-28
Ori inal Binder
Flash Point Tern , T48: Minimum, C 230
Viscosity, ASTM D 4402:
Maximum, 3 Pa-s (3000 cP) 135
Test Tern , C
Dynamic Shear, TP5:
G*/sino, Minimum, 1.00 kPa 52 58
Test Tern @ 10 rad/sec, C
Spec Requirement Test Temperature

Remains Constant Changes
Figure 2-49. Superpave Binder Specification Example (i6)
I
Table 2-9. ~erformance Graded Asphalt Binder Specification ':e
Perfommnce Grade PG46 PG 52 PG 58 PG64

;
; -34 -40 -46 -10 -16 -22 -28 -34 -40 -46 -16 -22 -28 -34 -40 -10 -16 -22 -28 -34 -40
Average 7-<lay Maximum Pavement
Design Temperaturp, °C a <46 <52
- <58 <64
Minimum Pavement
Design Temperaturi''• °C a >-34 >-40 >-46 >-10 >-16 >-22 >-28 >-34 >-iO >-46 >-16 >-22 >-28 >-34 >-40 >-10 >-16 >-22 >-28 >-34 >-40
I Original Binder
Flash Point Temp, T48: Minimum °C 230

Viscosi[)', ASTM D fl402: D
Maximum, 3 Pa•s (3000 cP), ::c:
g
Test Temp, °C [ 135
Dynamic Shear, TPp: "
G•/sin 6, Minin]um, 1.00 kPa
f
Test Temperature @ 10 rad/s, °C, 46 52 58 64 l
I
Mass Loss, Maxim4m, o/o
Rolling Thin Film Oven (T 2qOJ o, TI1in Film On,n CT 179) Residue
1.00
i
Dynamic Shear, TPr5:
G•/sin 6, Min"i'?>um, 2.20 kPa
Test Temp @ 1 , rad/sec, °C 46 52 5J 64 [
I
f
Pressure Aging Vessel Residue (PP!)
PAV Al!inl! Temnf'.hture, 0 cd 90 90 100 100
Dynamic Shear, Ttj5:
G•sin 6, Max"n\1m, 5000 kPa t)
Test Temp @ 1 rad/sec °C 10 7 4 25 22 19 16 13 10 7 2'i 22 19 16 u 31 28 2~ 22 19 16 &l
<§'
Phvsical Hardenin~ e Renort
Creep Stiffness, TP:1: f g_
S, Maximum, 300 MPa (')
m-value, lvlinin'ium, 0.300 0
Test Temp, @ <$o sec, °C -24 -30 -36 0 -6 -12 -18 -24 -30 -36 -6 -12 -18 -24 -30 0 -6 -12 -18 -2-l -30
Direct Tension, TP3: 1
Failure Strain, Minimum, 1.0%
Test Temo @ 11.o mm/min °C -24 -30 -36 0 -6 -12 -18 -24 -30 -36 -6 -12 -18 -24 -30 0 -6 -12 -18 -24 -30
I.
::,
-
Table 2-9. Performance Graded Asphalt Binder Specification (continued) 9
I ~
Performance Gjde PG70 PG76 PG82 a
N
-10 -16 -22 -28 -34 -40 -10 -16 -22 -28 -34 -10 -16 -22 -28 -34
Average 7-day taxirnum Pavement ~
...,,
Design Tempe ture, °C a <70 <76 <82
I
Minimum Pavelnent Design
i
Temperature, 0 ~ a >-10 >-16 >-22 >-28 >-34 >-40 >-10 >-16 >-22 >-28 >-34 >-10 >-16 >-22 >-28 >-34 ~
I:!'>
I Original Binder J·
Flash Point Terµp, T48: Minimum °C 230 s
Viscosity, ASTM D 4402: b .fl
Maximum, BPa•s (3000 cP), 135 g_
Test Temp,l°C
Dynamic Sµea~, TP5: c ...,,a°
G•/sin o, ~nimum, 1.00 kPa 70 76 82
Test Tern ,rature @ 10 rad/s, °C I·
I Rolling TI1in Film Oven (T 2-i0J or Thin Film Oven <T 179) Residue
Mass Loss, Maipmum, % 1.00
Dynamic Sheat, TP5:
o,
G•/sin Minimum, 2.20 kPa 70 76 82
Test Temp:@ 10 !"Jd/sec, °C
Pres.~ure Aging \'essel Residue (!'Pl J
PAV Aging Temperature, 0 cd 100(110) 100( I 10) 100(110)
Dynamic Shear, TP5:
G•sin o, Maximum, 5000 kPa 34 31 28 25 22 19 37 34 31 28 22 40 37 34 31 28
Test Temp @ 10 rad/sec, °C
Physic-JI Hardening e Report
Creep Stiffness, TPl: f
S, Maximum, 300 MPa
m-value. Minimum, 0.300 0 -6 -12 -18 -24 -30 0 -6 -12 -18 -24 0 -6 -12 -18 -24
Test Temp. @ 60 sec, °C
Direct Tension, TP3: r \D
V\
Failure Strain, Minimum, 1.0% 0 -6 -12 -18 -24 -30 0 -6 -12 -18 -24 0 -6 -12 -18 -24
Test Temp @ i.0 mm/min, °C
I
\D
Table 2-9. ~erformance Graded Asphalt Binder Specification (continued} 0\
l
Notes: '
a. Pavement teriperaturcs can be estimated from air temperatures using an algorithm contained in the
Superpave™ sof;tware program or may be provided by the specifying agency, or by following the procedures
as outlined in PPX.
b. This require~ent may be waived at the discretion of the specifying agency if the supplier warrants that
the asphalt binder can be adequately pumped and mixed at temperatures that meet all applicable safety
~uality LI of unmodffied a,pl,ak cement production, measurement of the vOCOSty of. the origm,I
asphalt cementlmay be substituted for dynamic shear measurements of G•/sin 6 at test temperatures where
~
~
the asphalt is a Newtonian fluid. Any suitable standard means of viscosity measurement may be used,
including capill ry or rotational viscometer (AASHTO T 201 or T 202). i
d. The PAV agipg temperature is based on simulated climatic conditions and is one of three temperatures
9Q°C, 100°C or /110°C. The PAV aging temperature is 100°C for PG 64- and above, except in desert climates,
where it is 110°C.
I
~
~
e. Physical Harening - TP 1 is performed on a set of asphalt beams according to Section 13.1 of TP 1 , tJ
except the con~tioning time is extended to 24 hrs ± 10 minutes at 10°C above the minimum performance ~-
temperature. re 24-hour stiffness and m-value are reported for information pu~s only.
~
p..
f. If the creep ~ness is below 300 MPa, the direct tension test is not required. If the creep stiffness Q
between 300 _3rd 600 MPa the direct tension failure strain requirement can be used in lieu of the creep
stiffness requ~ment. The m-value requirement must be satisfied in both cases.
AASHTO Perfo.rmance Graded Binder Specification (MPl)
Ig.
::,
Since the Superpave asphalt binder specification is meant to be per-

formance based, it addresses three .primary performance parameters of
HMA pavements: permanent deformation (rutting), fatigue cracking, and
low temperature (thermal) cracking. In addition, the specification address-
es concerns related to safety, pumping and handling, and excessive aging of
.... the asphalt binders. A discussion of these parameters addressed in the
Superpave binder specification follows.
Permanent Deformation (Rutting). The Superpave binder specifica-

tion uses a rutting factor, G*/sin8, which is a measure of asphalt binder's
stiffness or rut resistance at high pavement service temperatures. The sig-
nificance of this factor was addressed when discussing the DSR. As shown
in Figure 2-50, the rutting factor must be at least 1.00 kPa for the origi-
nal asphalt binder and a minimum of 2.20 kPa for the RTFO aged asphalt
binder when tested by DSR at the maximum pavement design tempera-
ture.
Perfonnance Grade
Average 7-day Maximum Pavement

Design Temperature, C
Minimum Pavement Design
Temperature, C
Flash Point Temp, T48: minumum C

Viscosity, ASTM D 4402: 0
Maximum, 3 Pa-s (3000 cP),
TestTemu,C
Dynamic Shear,TPS:c
G*/sin 5, Minimwn, 1.00 kPa ---::
Test Temperature @ 10 rad/s, C Spec Requirements to
Rolling Thin Film Oven (T240)
Control Rutting
Mass Loss, Maximum % ./"
Dynamic Shear, TPS: /
G*/sin 6, Minimum, 2.20 kPa
Test Temp @ 10 rad/sec, C
Figure 2-50. Superpave Specification Rutting Factor Requirements

(W
Fatigue Cracking. The specification uses a fatigue factor, G*. sin8,

-----~hich-represents-a-sph-a:l-t-binclels-resist-a:nee-t0 fatigue cracking. As shown
in Figure 2-51, the specification has a maximum limit of 5000 kPa for G*
sin8 for the asphalt binder subjected to RTFO and PAV aging, and tested
at intermediate pavement service temperature. For example, the interme-
diate pavement service temperature for a PG 64-28 grade is 22°C.
PAV A in Tern ,C
Dynamic Shear, TP5:
G*sin 6, Maximum, 5000 kPa
Test Tern @ 10 rad/sec, C
Physical Hardening e Specification requirement
Creep Stiffness, TPI / to control fatigue cracking
S, Maximum, 300 MPa
m-value, Minimum, 0.300
Test Tern , @60 sec, C.
Direct Tension, TP3:
Failure Strain, Minimum, 1.0%
Test Temp @ 1.0 mm/min, C
Figure 2-51. Superpave Specification Fatigue Cracking Factor

Requirements (jQ}
Low Temperature Cracking. Figure 2-52 gives the specification

requirements for resistance to low temperature cracking. When the BBR
is used to test the RTFO as well as PAV aged asphalt binq.er at the lowest
pavement design temperature, the creep stiffness (S) must not exceed 300
MPa and them-value must be at least 0.300. However, if the creep stiff-
ness exceeds 300 MPa but is less than 600 MPa, the asphalt binder is
acceptable if the failure strain is more than 1.0 percent when tested by the
DTT. The m-value requirement must be satisfied in both cases. The
DTT is not required at all if the creep stiffness is less than 300 MPa.
PAV Aging Temp, C

Dynamic Shear, TP5:
G*sin 6, Maximum, 5000 kPa
Test Temp @ IO rad/sec, C
Physical Hardeninge Specification requirements
Creep Stiffness,TPI: _-?, to control low temperature
S, Maximum, 300 MPa
m-value, Minimum, 0.300 cracking
Test Temp, @60 sec, C
Direct Tension, TP3:
Failure Strain, Minimum, LO%
Test Temp @ 1.0 mm/min, C
Figure 2-52. Superpave Specification Low Temperature Requirements
Other Common Specification Criteria. The Superpave binder speci-

fication requires a minimum flash point of 230°C for all PG grades to
address safety concerns. AASHTO T-48, "Standard Method of Test for
Flash and Fire Points by Cleveland Open Cup", is used to determine the
flash point of unaged asphalt binders.
The specification also contains a maximum viscosity requirement of 3
Pa.sat 135°C for original binders to ensure pumping and handling of the
asphalt binders (especially modified asphalts) at the HMA facility.
A maximum mass loss of 1.00 percent applies to RTFO residues of all
PG grade binders. This requirement ensures that excessive aging
(volatilization) of the asphalt binder does not occur during HMA produc-
tion and construction.
Selection of Superpave Binder Grades. The Superpave performance

based binder grades are selected based on the climate prevailing at the pro-
ject site. Table 2-10 gives the binder grades available in the specification.
PG 76 and 82 are intended only to accommodate slow transient or stand-
ing loads (such as those near intersections), or excessive truck traffic, as dis-
cussed later.
The selection of an appropriate PG grade for a specific project is facil-
itated· by the Superpave software which provides the following three alter-
nate methods:
l. By Geographic Area: The designer makes an input of project loca-

tion (such as state and county).
Table 2-10. Superpave Performance Based Asphalt Binder Grades
High Temperature Grades (0 C) Low Temperature Grades (0 C)

PG46 -34,-40,-46
PG 52 -10,-16,-22,-28,-34,-40,-46
PG 58 -16,-22,-28,-34,-40
PG 64 -10,-16,-22,-28,-34,-40
PG70 -10,-16,-22,-28,-34,-40
PG76 -10,-16,-22,-28,-34
PG 82 -10,-16,-22,-28,-34
2:-BrPavement-Temperatttre:-T-he-designer-must-need-to know design

pavement temperatures.
3. By Air Temperature: The designer determines air temperatures
which are converted by the Superpave software to design pavement
temperatures.
The Superpave software has a database of weather information for

about 7,500 reporting weather stations in the U.S. and Canada. The
hottest seven-day period was identified for each year at these weather sta-
tions to calculate the average maximum air temperature for high tempera-
ture design condition. For all the years of operation (weather stations with
less than 20 years of operations were not used), the mean and standard
deviation of the seven-day average maximum air temperature have also
been computed. Similarly, the one-day minimum air temperature for each
year and the means and standard deviation of the average one-day mini-
mum air temperatures were also determined.
Since the design temperatures for selecting PG grade are based on
pavement temperatures rather than air temperatures, the Superpave soft-
ware converts the air temperatures to pavement 'temperatures using theo-
retical analyses of actual conditions performed with models (62). At the
present time (1996), the Superpave criteria considers the minimum pave-
ment temperature equal to the minimum air temperature. This is likely to
change in the future based on field data. It is believed that the minimum
pavement temperature is generally higher (warmer) than the minimum air
temperature.
The Superpave software also allows the designer to use different levels
of reliability concerning the design pavement temperatures. Reliability has
been defined as the percent probability in a single year that the actual tem-
perature (one-day low or seven-day average high) will not exceed the cor-
responding design temperatures. Consider the example of Topeka, Kansas
as shown in Figure 2-53 (jg). It has a mean seven-day maximum air tem-
perature of 36°C and a standard deviation of 2°C. According to the fre-
36
40
Air Temperature, Topeka, KS
-31 -23
-40 -30 -20 -10 0 10· 20 30 40 so 60

Figure 2-53. Distribution of High and Low Design Air Temperatures
for Topeka, KS (2Q)
quency distribution, there is a 50% probability in an average year that the

seven-day maximum air temperature will exceed 36°C. However, the
, chance that the temperature will exceed 40°c is only two percent. ln other
words, a design high air temperature of 40°C provides a 98 percent relia- '
bility. Similarly, the average 1-day coldest temperature is -23°C with a
standard deviation of 4°C which means a very cold winter could have a
minimum air temperature of -31 °C at 98% probability.
The high and low air temperatures for Topeka, Kansas are converted
to pavepient temperatures by the Superpave software as shown in Figure
2-54. The high pavement temperature is calculated 20 mm below the
pavement surface and the low temperature is calculated at the surface for
the wearing course in this example. Figure 2-55 shows how PG grades are
selected at 50 and 98% minimum reliability levels. For at least 50% min-
56
60
Pavement Temperature, Topeka, KS
~31 -23
-40 -30 -20 -10 0 10 20 30 40 50 60 70
Figure 2-54. Distribution of High and Low Design Pavement

Temperatures for Topeka, KS (if?)
PG 64-34 (98% minimum reliability)
PG 58-28 (50 % minimum reliability)
-40 -30 -20 -10 0 10 20 30 40 50 60 70
Figure 2-55. Binder Grades for Topeka, KS (2Q)

imum reliability, the high temperature grade must be PG 58. Since there
is no PG 56 grade it has to be "rounded up" to PG 58, which actually
results in a probability of about 85 percent rather than 50 percent. On the
other hand, the low temperature grade for at least 50% reliability must be
-28°C which gives almost 90% reliability rather than 50%. For 98 per-
cent reliability, a PG 64 grade is needed for high temperature and a PG-34
grade for low temperature resulting in a PG 64-34 grade.
The preceding binder selection process based on climate only assumes
that the H.MA pavements will be subjected to fast moving traffic. The
DSR tests to determine the rutting factor G*/sino are conducted with a
loading rate of 10 radians per second which corresponds to a traffic speed
of about 90 kilometers per hour. However, .significantly slower loading
rates are experienced by HMA pavements at some locations such as inter-
sections and toll booths. There are cases such as bus stops where loads are
also stationary. In such cases of slower loading rates an asphalt binder must
exhibit a higher stiffness to minimize rutting. To meet these situations, the
high temperature grade should be increased by one grade for slow transient
loads and by two grades for stationary loads. For example, if a PG 64-22
grade was selected based on climate for fast moving traffic, a PG 70-22
grade would be used for slow transient loads, and a PG 76:-22 grade would
be used for stationary loads. The selected low temperature grade is not
affected by the loading rate. Since the highest possible pavement temper-
ature in North America is about 70°C, it was necessary to add two high
temperature grades, PG 76 and PG 82, to accommodate slow traffic load-
ing rates (5§).
Most state highway agencies are planning at the present time (1996)
to adopt the Superpave asphalt binder system beginning in 1997.
ASPHALT CHEMISTRY
The importance of asphalt chemical composition, although not well

understood, cannot be disputed. The performance of asphalt as a binder in
HMA pavements is determined by its physical properties (discussed earli-
er in the chapter) which in turn are determined directly by chemical com-
position. An understanding of the chemical factors affecting physical prop-
- - - -ertles is inusfLindamental to an understanding of the factors that contro·i-------
asphalt performance. This section on asphalt chemistry is largely adapted
from the papers authored by Petersen (g.i) and Halstead@. An attempt
has been made to avoid detailed discussion of complex chemical phenom-
ena and to minimize the use of terminology unfamiliar to civil engineers.
Origin of Asphalt
Asphalt cement is refined from the crude petroleum which is primari- ·

ly formed by nature from plant life. The process of transformation from
plant life to crude oil occurs over millions of years under varied tempera-
ture and pressure conditions. Although all petroleums are basically hydro-
carbons (chemical combinations of carbon and hydrogen), the amount and
r.iature of hydrocarbons varies from crude to crude. Since the asphalt
cement is obtained by distillation from the crude, its chemical composition
and properties also vary from source to source.
Elemental Composition and Molecular Structure
Carbon and hydrogen are the principal elements present in asphalt

cement molecules. Sulfur is the next most abundant element. Nitrogen and
oxygen are usually present in very small amounts. Heavy metals such as
vanadium and nickel may also be present in trace amounts. Elemental
analyses of four representative petroleum asphalt cements are given in
Table 2-11.
Most asphalt molecules consisting of carbon and hydrogen contain
one or more of the following so-called heteroatoms: sulfur, nitrogen and
oxygen. The type of molecular structure is more important than the total
amount of each· element. Since the petroleum is transformed from diverse
living organic matter under different conditions, asphalt cements have
extremely diverse molecular structures depending on the crude source.
Table 2-11. Elemental Analyses of Representative Petroleum

Asphalts (.M)
B-2959 B-3036 B-3051 B-3602

Elements Mexican Blend Arkansas-Louisiana Boscan California
Carbon, percent 83.77 85.78 82.90 86.77

Hydrogen, percent 9.91 10.19 10.45 10.93
Nitrogen, percent 0.28 0.26 0.78 1.10
Sulfur, percent 5.25 3.41 5.43 0.99
Oxygen, percent 0.77 0.36 0.29 0.20
Vanadium, ppm 180 7 1380 4
Nickel,ppm 12 0.4 109 6
Asphalt cements contain a combination of the following three

arrangements by which the carbon atoms are linked with each other:
1. Straight or branched chains. Such asphalt cements are generally

called "aliphatic" or "paraffinic" types.
2. Simple or complex saturated rings. "Saturated" means the highest
possible hydrogen/ carbon ratio in the asphalt molecules. These
asphalt cements are usually referred to as "naphthenic" types.
3. One or more stable six-carbon condensed, unsaturated ring struc-
tures. These asphalt cements are called "aromatic" types. Benzene
and naphthalene have such structures.
Functional or Polar Groups
As mentioned earlier, heteroatoms (sulfur, nitrogen and oxygen) are

attached to carbon atoms in asphalt molecules in different configurations
and in form of different compounds. These configurations (molecules) are
polar because there is an imbalance of electrochemical forces within the
molecule which produces a dipole. Therefore, each polar molecule has elec-
tropositive and electronegatiye characteristics similar to a magnet which
has north and south poles. Since like charges repel each other and unlike
charges attract each other, molecular interactions are induced in asphalt
cements which strongly influence their physical or "engineering" proper-
ties and performance. These configurations of heteroatoms thus impart
functionality and polarity to asphalt molecules and are therefore called
functional or polar groups. Functionality (presence of function groups)
relates to how the asphalt molecules interact with each other or with sur-
faces and/ or molecules of other materials. For example, selective adsorp-
tion of functional or polar groups by aggregate surfaces is important to
understand the stripping phenomenon. Although present in small
amounts, these polar groups vary in characteristics from asphalt to asphalt,
and have a large effect on the performance of asphalt cement from differ-
ent sources.
Also, the non-polar components of asphalt cement, which act as sol-
vents-or disp-ersents-forthe--p-olarorfunctiona:l-groups,pla.y.r-major-role-i-n-- - - - - -
determining the effect that the polar groups will have on the physical and
aging properties of the asphalt cement.
The types of sulfur, nitrogen, and oxygen compounds which have been
identified in asphalt cements are given in Table 2-12.
Chapter 2-Asphalt Refining, Uses, an"d Properties 105
Table 2-12. Sulfur, Nitrogen, and Oxygen Compounds
Sulfur Compounds Polysulfides

Sulfides
Thiols
Thiophenes
Nitrogen Compounds Pyridinic

Pyrrole, indole, carbazole
Porphyrins
Oxygen Compounds Carboxylic acids/naphthenic acids

Phenols
Ketones
Esters
Ethers
Anhydrides
Conceptual Compositional Model
According to the most simple and generally accepted concept of

asphalt composition, asphalt cements are considered to be made up of
asphaltenes, resins, and oils as shown in the following chart:
Asphalt Cement
I
Asphaltenes Maltenes
(Petrolenes)
I
Resins Oils
Asphaltenes are insoluble (or precipitated) when the asphalt cement is

dissolved in a nonpolar solvent such as pentane, hexane, or heptane. The
comporren-r-which--is-disso1ved-i:s-ca±led7naltenes12 ( or 1'petrolenes"), and it
is comprised of resins and oils.
Asphaltenes. Asphaltenes are generally dark brown, friable solids. The

type of nonpolar solvent used to precipitate the asphaltenes affects the
determination of its total amount in the asphalt cement. Higher amounts

are precipitated by n-pentane than by n-heptane because the number of
carbon atoms in n-pentane is less than that in n-heptane. Asphaltenes are
the most complex components with the highest polarity. Therefore, they
have a very high tendency to interact and associate. They are mixtures of
many compounds that have a strong tendency to associate in conglomer-
ates. Asphaltenes play a major role as the viscosity-building ("bodying")
component of asphalt cements. Studies have shown that the amounts and
characteristics of asphaltenes vary considerably from asphalt to asphalt.
Low asphaltene content (less than about 10 percent) or weakly associating
asphaltenes have been linked with tenderness in HMA by some
researchers.
Resins. Resins are generally dark and semi-solid or soHd in character.

They are fluid when heated and become brittle when cold. They work as
agents that disperse (or "peptize") the asphaltenes throughout the oils to
provide a homogeneous liquid. They are soluble in n-pentane, but adsorb
on fuller's earth or alumina. On oxidation resins yidd.asphaltene type mol-
ecules. ' ·
Oils. Oils are usually colorless or white liquids. They are soluble in
most solvents. They have paraffinic and naphthenic structures with no oxy-
gen and nitrogen usually present. On oxidation they yield asphaltene and
resin molecules.
Asphalt as a Colloidal System
Asphalt cement is not a true solution, but is considered a colloidal or

micellar system. Nellensteyn (ZJ/J first recognized its colloidal nature by
stating that it has a dispersion of micelles in an oily medium. The relative
amounts and characteristics of asphaltenes, resins, and oils present in an
asphalt cement influence its physical properties and performance in a
HMA mixture. These influences make the asphalt act as a "sol," "sol-gel,"
or "gel."
"Sol:" "Sol" asphalt cements typify a system in which the resins keep the
asphaltenes highly "peptized" (or dispersed) in the oily phase. "Sol" asphalt
---cements largely exnioit Newtoman flow charactensncs. Accordmg to
Rostler (Z.l), nitrogen bases in asphalt cement are primarily responsible for
its "sol" characteristics.
"Sol-Gel·" It is intermediate between "sol" and "gel."

Chapter·z_:_Asphalt Refining, Uses, and Properties 107
"Gel·" "Gel" asphalt cements typify a system in which resins are not very
effective in "peptizing" asphaltenes. Excessive presence of paraffins in rela-
tion to nitrogen bases also tends to reduce solubility, leading to increased
"gel" characteristics, and suggesting increased separation of the dispersed
and dispersing phases (Zl). "Gel" asphalt cements exhibit largely
non-Newtonian behavior.
The degree to which an asphalt cement acts as a "sol" or as a "gel"
influences its rheological properties, aging characteristics, and its perfor-
mance in a H.MA mixture.
The compatibility and relationships of different components in the
microscopically homogeneous mixture controls the overall behavior of
asphalt cement rather than the quantitative amount of any single compo-
nent according to Petersen (@. Various components of asphalt cement
interact to form a balance or compatible system, which gives asphalt
cement its unique viscoelastic properties as a binder. It has long been rec-
ognized that this compatibility, or lack of compatibility, is related to
asphalt quality. Oliensis's spot test (Z2) was an attempt to determine the
"stability" of the asphalt components when dissolved in a solvent. Many
agencies adopted the Oliensis spot test as a specification test for homo-
geneity of asphalt cements and therefore as an indicator of asphalt cement's
durability in service. However, as was shown by Heithaus and Fink (Z3),
the spot test did not distinguish between "good" and "bad" asphalt
cements. This test did serve a good purpose when first introduced because
it identified some nondurable asphalt cements obtained from the cracking
of petroleum in certain refining processes.
Analytical Procedures
The two most frequently used fractionation methods are: (1) Rostler
and Sternberg's chemical precipitation method, and (2) Corbett's selective
adsorption-desorption (chromatographic) method. However, in all cases
the "generic" fractions are still complex mixtures and not defined chemical
species; thus the same generic fraction from different asphalt cements may
have a quantitatively different effect on the physical properties.
Chemical Precipitation Method. It is only an analytical method; it

-----d-oes-not-physiea:Hy--sepa-ra-te-the-eomponents··· in·· asphalt· cement. ·Figure
2-56 gives the flow chart for this method developed by Rostler and
Sternberg (Z1, ZS), who attempted to identify and quantify five components
in asphalt cement. Essentially, the asphaltenes are separated (precipitated)
by n-pentane and the solution of maltenes is sequentially treated with
I ASP~ALT I
n-Pentane
Precipitate Solution
NITROGEN
ASPHALTENES
BASES (85%)
1st ACIDAFFINS 1---~H,..=2-so-,.:...

4 ----I
(98%)
2nd AC IDAFFI NS 1----H...;;;2~S0,_.....;...

4 -----1
SO3
PARAFFINS
Figure 2-56. Chemical Precipitation Method (after Rostler and

Sternberg, Zi)
increasing strengths of sulfuric acid (H2S04) to measure the amounts of

four classes of molecular (or components) present in the maltenes. The fol-
lowing five components are quantified in this method:
Asphaltenes (A)
Asphaltenes are precipitated by n-pentane. They are considered as

"bodying" agents in asphalt cement.
Nitrogen Bases (N)
--------80-per-€@B-t-sHlfuFie-a€i-cl.-is-us@d-t-0-r-emo-v-e-0-r-pr-e-cip-i-tat-e-p0lar-ni-tr-0gen------
compounds (quite aromatic in nature) from the maltene solution .. This
fraction is called "nitrogetl bases" and is considered to be the peptizer for
the strongly associated asphaltenes in the solvating phase.
First Acidaffins (Al )

,.
98 percent sulfuric acid is then used to precipitate another aromatic
component. low in: nitrogen. Rostler called it "first acidaffins" which are
considered solvents for the peptized asphaltenes.
Second Acidaffins (A2)
Fuming sulfuric acid containing 30 percent SO3 is then used to pre-

cipitate aless r:eactive and less aromatic component of the maltenes called
"second acidaffihs." This component is also believed to be the solvent for
peptized asphaltenes.
Paraffins (P)
This is the final and oily constituent of the maltenes which is nonre-
active to the fuming acid. According to Rostler, paraffins are believed to act
as "gelling agents."
Many asphalt technologists with an engineering background have dif-
ficulty relating the results from this analytical procedure using sulfuric acid
of different strengths to the performance of asphalt cement. However,
Rostler has suggested the following relationships concerning the quality
and durability of asphalt cement:
Compatibili'ty R.atio = N
p
A compatibility ratio of more than 0.5 is considered good.
N+Al most reactive

Durabiliry Parameter =
P+A2 least reactive
A durability parameter of less than 0.40 is considered poor, and more

than 1.00 is considered good.
Selective Adsorption-Desorption Method
This method (ASTM D4124) developed by Corbett uses selective

adsorption-desorption procedures (Figure 2-57) and actually separates the
~ n-Heptane
Precipitate PETROLENES
(MALTENES)
ASPHALTENES Adsorption - Elution

Chromatography
on alumina
n-Heptane
SATURATES
(elute)
NAPHTHENE Benzene
AROMATICS (elute)
POLAR 1. Methanol-benzene
AROMATICS 2. Trichloroethylene
(elute)
Figure 2-57. Selective Adsorption-Desorption Method (after Corbett;

Z(i)
components virtually unaltered for further analysis (ZQ). The most polar
and least soluble asphalt cement components (asphaltenes) are first precip-
itated by n-heptane which is a nonpolar solvent. The heptane solution of the
remaining maltene fraction is then introduced into a chromatographic
column. Alumina in the column adsorbs the remaining three components.
Different solvents of increasing polarity are then used to sequentially
desorb (wash out) fractions of increasing polarity. Corbett's method sepa-
rates and quantifies the following four components:
_ _ _ _A_sphalten~e~s~(A~)'..__ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __
Asphaltenes precipitated by n-heptane are most polar and act as the

"bodying" fraction of asphalt cement as discussed earlier. They are solid or
semi-solid at ambient temperature.
Saturates (S)
Saturates are the first fraction to emerge from the column when elut-
ed (desorbed) with n-heptane. This fraction is not adsorbed by the alumi-
na in the column because it lacks polar chemical functional groups.
Saturates are equivalent to paraffins in the Rostler analysis. Often sulfur is
found incorporated in molecules of this fraction. Saturates are liquid at
ambient temperatures and hardly change with time. They have a negative
contribution to temperature susceptibility of asphalt cement.
Naphthene Aromatics (NA)
Naphthene aromatics emerge as the second fraction when eluted with

a more polar aromatic solvent such as benzene or toluene. Their molecular
structure is comprised of condensed nonaromatic and aromatic ring sys-
tems, and may contain the heteroatoms oxygen, nitrogen and sulfur. They
are liquid at ambient temperatures. They are considered to be the soften-
ing component in asphalt cement. They are also the aging fraction in
asphalt cement.
Polar Aromatics (PA)
They are the final fraction to emerge from the column when eluted
with a highly polar mixture of alcohol and benzene (or toluene). Alcohol
helps to debond this strongly adsorbed and most polar component of the
maltenes from the absorbent (alumina). Its molecular structure comprises
a higher percentage of condensed aromatic ring systems and functional
groups containing heteroatoms than the other fractions of maltenes. It is
solid or semi-solid at ambient temperatures. It has been related to ductili-
ty of asphalt cement and is also one of the aging fractions in asphalt
cement.
Molecular Size Distribution
High pressure liquid chromatography (HPLC), or high pressure gel

permeation chromatography (HP-GPC), has been used to determine the
molecular size distribution in asphalt cement. It is analogous to gradation
------1-ors:reve-arnrlysis)--ofmirreral-a:ggregares-.---- --
A chromatogram of the relative amounts oflarge, medium, and small
molecules in an asphalt cement is obtained by this procedure. Recent work
with this technique has been reported by Jennings at Montana State
University (ZZ). Essentially, a solution of asphalt cement (tetrahydrofuran
or THF has been used as solvent) is passed through gel permeation

columns. The system permits largest molecules to pass quickly through the
columns but successively retards the progress of the smaller molecules
which are slowed in their movement through the gel as they move in and
out of the pores of the gel while moving through the column.
A detector and recorder produces a chromatogram showing the rela-
tive amounts of molecules being eluted at different times.'Jennings has
classified the molecules eluted during the first third of the elusion period
as large molecular size (LMS), those eluted during the second third of the
period as medium molecular size (MMS), and those eluted during the final
third of the period as small molecular size (SMS).
Figure 2-58 illustrates two chromatograms obtained by this method
on two asphalt cements, A and B. These asphalt cements have significantly
different molecular size distributions, and therefore significantly different
performances would generally be expected.
Jennings (Z2) compared the relative amounts of LMS, MMS, and
SMS in asphalt cements recovered from poorly performing HMA pave-
ments with those from good HMA pavements. Results obtained from
Montana roads indicate greater relative amounts of LMS are associated
with poor performance. It is possible that a proper balance between the rel-
ative amounts of LMS, MMS, and SMS in an asphalt cement is desirable
for its durability and ability to perform as a binder. More recent work in
Legend
--Asphalt A
--Asphalt B
I
/
I
I
I
I
L
MMS SMS
0 ~ JO I~ 20
Elution Time ( lnutesl
Figure 2-58. Typical Chromatograms of Whole Asphalts (after

Jennings (Z2)
the Strategic Highway Research Program (SHRP) has shown that the large
molecular size (LMS) fraction is composed predominantly of a collection
(conglomerate) of smaller, highly associated polar molecules which thus
appear in the chromatogram as molecules of "apparent" large molecular
size. Thus, this technique has the potential of characterizing the strongly
associating molecular components that play a major role in determining
the rheological properties and aging characteristics of asphalt cement relat-
ed to pavement performance.
SHRP Researc~ in Asphalt Chemistry
Advances in analytical technology were used in SHRP to better under-

stand the asphalt chemistry.
Size exclusion chromatography (SEC) was used to isolate two fractions
(called SEC I and SEC II) from the asphalts. The SEC method was mod-
ified by using toluene as the solvent rather than the strong (polar) solvent
like tetrahydrofuran used by Jennings in earlier work (Z2). The weaker sol-
vent toluene did not disrupt the asphalt structure as much as tetrahydro-
furan. SEC I was believed to contain "associated molecules" and SEC II
was believed to contain "solvent molecules". Additional analysis demon-
strated that the majority of the polar materials were present in SEC I, while
the nonpolar materials were in the SEC II. Performance based asphalt
physical tests developed in the SHRP indicated that increased amounts of
SEC I generally caused brittleness, low temperature cracking, and fatigue
cracking (thin HMA pavements). On the other hand, increased amounts
of SEC II caused rutting and fatigue cracking (thick HMA pavements).
Ion Exchange Chromatography (IEC) was also used as a powerful ana-
lytical technique to understand molecular interactions in asphalts. In this
method, the asphalt is dissolved in a solvent and then passed through a
series of ion-exchange resins. Resin beds, basic and acidic in succession,
are used to attract the polar molecules according to their acidic or basic
character. The nonpolar molecules are not retained in the resin beds and
are collected as a separate fraction. The IEC thus permitted to separate
asphalt molecules based on the very characteristic (polarity) which con-
tributes to the performance of asphalt.
------.Accor.di..ng-to--5H.RE...r.es.ea.cchers,_asp.hal.t.is_b.elieYed to be a 3-dimen-
sional molecular matrix of highly polar molecules swollen by its more neu-
tral molecules. The 3-dimensional nature makes it somewhat elastic, but,
with sufficient stress (thermal or mechanical) the weak bonds rearrange
and the asphalt experiences permanent strain.
The IEC analytical technique was also used to isolate a fraction termed
"amphoterics". This fraction is believed to be the key building block of
asphalts. The term amphoterics is used to describe an asphalt molecule
which has both an acid and basic group in the molecule, but not at the
same site. Although amphoterics constitute only 10 to 15 percent of
asphalts, they are largely responsible for the viscoelastic properties of
asphalts based on the test data obtained in SHRP.
Asphalt chemistry is very complex. Researchers have debated in the
past and continue to debate on this very interesting subject. Reference Zfl.
contains SHRP research data on asphalt chemistry.
REFERENCES
1. Barth, E.J. Asphalt Science and Technology. Gordon and Breach

Science Publishers, New York, NY, 1962.
2. Wallace, H.A. and J .R. Martin. Asphalt Pavement Engineering.
McGraw-Hill Book Co., New York, NY, 1967.
3. Traxler, R.N. AJphalt: Its Composition, Properties and Uses. Reinhold
Publishing Corp., New York, NY, 1961.
4. Corbett, L.W Refining Processing of Asphalt Cement. TRB, Trans-
ponation Research Record 999, 1984.
5. Hodgson, R.S. Asphalt: What and From Where. Paper presented at
the Asphalt-State-of-the-Art Conference, Woodbridge, NJ, May
11, 1981.
6. Welborn, J.Y. and WJ. Halstead. Testing of Asphalts and Asphalt.
Mixtures. Proceedings AAPT, Vol. 43A, 1974.
7. Terrel, R.L., J.A. Epps, and C. Crawford. Making the Most of
TemperatureNiscosity Characteristics. National Asphalt Pavement
Association Information Series 102/88, 1988.
8. Kandhal, P.S., L.D. Sandvig, WC. Koehler and M.E. Wenger.
Asphalt Viscosity-Related Properties of In-Service Pavements in
Pennsylvania. ASTM, Special Technical Publication No. 532, 1973.
9. Vallerga, B.A., C.L. Monismith and K. Granthem. A Study of Some
Factors Influencing the Weathering of Paving Asphalts. Proceedings
A.APT,Vm.26~T9-~.--------------------i
10. Finn, Fred N. Factors Involved in the Design of Asphaltic Pave-
ment Surfaces. HRB, NCHRP Repon 39, 1967.
11. Traxler, R.N. Durability of Asphalt Cements. Proceedings AAPT,
Vol. 32, 1963.
12. Brown, A.B.; J.W. Sparks and 0. Larsen. Rate of Change of

Softening Point, Penetration, and Ductility of Asphalt in Bitu-
minous Pavement. Proceedings AA.PT, Vol. 26, 1957.
13. Lee, Dah-Yinn. Asphalt Durability Correlation in Iowa. HRB,
Highway Research Record 468, 1973.
14. Kandhal, P.S., and M.E. Wenger. Asphalt Properties in Relation to
Pavement Performance. TRB, Transportation Research Record
544, 1975.
15. Kandhal, P.S. and W.C. Koehler. Significant Studies on Asphalt
Durability: Pennsylvania Experience. TRB, Transportation Re-
search Record 999, 1984.
16. Welborn, J.Y. Relationship of Asphalt Cement Properties to
Pavement Durability. TRB, NCHRP Report 59, 1979.
17. Kandhal, P.S. Low-Temperature Ductility in Relation to Pave-
ment Performance. ASTM, Special Technical Publication 628,
1977.
18. Hubbard, P. and Gollomb, H. The Hardening of Asphalt with
Relation to Development of Cracks in Asphalt Pavements. Pro-
ceedings AAPT, Vol. 9, 1937.
19. Welborn, J.Y., E.R. Oglio andJ.A. Zenewitz. A Study ofViscosity
Graded Asph,alt Cements. Proceedings AAPT, Vol. 35, 1966.
20. Doyle, P.C. Cracking Characteristic of Asphalt Cement: Pro-
21. Halstead, W.J. The Relation of Asphalt Ductility to Pavement
Performance. Proceedings AAPT, Vol. 32, 1963.
22. Hveem, F.N., E. Zube and J. Skog. Proposed New Tests and
Specifications for Paving Grade Asphalts. Proceedings AAPT, Vol.
32, 1963.
23. Serafin, P.J. Discussion. Proceedings AAPT, Vol. 32, 1963.
24. Vallerga, B.A. and W.J. Halstead. Effects of Field Aging on Fun-
damental Properties of Paving Asphalts. HRB, Highway Research
Record 361, 1971.
25. Kandhal, P.S. et al. Low-Temperature Properties of Paving Asphalt
Cements. TRB, State-of-the-Art Report 7, 1988.
26. Kandhal, P.S. and M.E. Wenger. Measurement of Asphalt
Viscosity at 77°F with a Vacuum Capillary Viscometer.
Proceedings AAPT Vol. 44, 1975.
--:-Properties of7J..-sp-f1Cflri1:-m·tamm-;-J:-·Ph; Pfeiffer, Ed., Elsevier
Publishing Company, Inc., New York, 1950.
28. McLeod, N.W. Asphalt Cements: Pen-Vis Number and Its Ap-
plication to Moduli of Stiffness. ASTM Journal of Testing and
Evaluation Vol. 4, No. 4, 1976.
29. Kandhal, P.S. and WC. Koehler. Effect of Rheological Properties

of Asphalts on Pavement Cracking, ASTM, Special Technical
Publication 941, 1987.
30. Puzinauskas, VP. Evaluation of Properties of Asphalt Cements
with Emphasis on Consistencies at Low Temperatures. Proceed-
ings AAPT, Vol. 36, 1967.
31. Kandhal, P.S., L.D. Sandvig and M.E. Wenger. Shear
Susceptibility of Asphalts in Relation to Pavement Performance.
Proceedings AAPT, Vol. 42, 1973.
32. Van der Poel, C. A General System Describing the Viscoelastic
Properties of Bitumens and Its Relation to Routine Test Data.
Journal of Applied Chemistry, May 195'4.
33. Heukelom, W and A.J.G. Klomp. Road Design and Dynamic
Loading. Proceedings AAPT, Vol. 33, 1964.
34. Heukelom, W An Improved Method of Characterizing Asphaltic
Bitumens with the Aid of their Mechanical Properties. Proceedings
MPT, Vol. 42, 1973.
35. Haas, R.C.G. A Method for Designing Asphalt Pavements to
Minimize Low-temperature Shrinkage Cracking. The Asphalt
Institute, Research Report No. 73-1, January 1973.
36. Schweyer, H.E. A Pictorial Review of Asphalt Rheology. Pro-
ceedings AAPT, Vol. 43A, 1974.
37. Schweyer, H.E. and J.C. Busot. Experimental Studies on Viscosity
of Asphalt Cements at 77°F. HRB, Highway Research Record 361,
1971.
38. Schweyer, H.E., R.L. Baxley, and A.M. Burns. Low-Temperature
Rheology of Asphalt Cements-Rheological Background. ASTM,
Special Technical Publication 628, 1977.
39. Fenijn, J. and R.C. Kroosho£ The Slidit1g Plate Rheometer-A
Simple Instrument for Measuring the Visco-Elastic Behavior of
Bitumens and Related Substances in Absolute Units. Proceedings
Canadian Technical Asphalt Association, Vol. 15, 1970.
40. Gaw, WJ. Measurement and Prediction of Asphalt Stiffness and
Their Use in Developing Specifications to Control
----~Low=lemperarure-Fave.r:ne-nrrra:rrsve-rs-e-era.-cki:rrg:-AS.'.f-M;-Special-----~
Technical Publication 628, 1977.
41. Gaw, WJ. Measurement and Prediction of Asphalt Stiffness at
Low and Intermediate Pavement Service Temperatures. Pro-
42. Pink, H.S., R.E. Merz and D.S. Bosniack. Asphalt Rheology: Ex-
perimental Determination of Dynamic Moduli at Low Temper-
atures. Proceedings AA.PT, Vol. 49, 1980.
43. Kandhal, P.S. and W.C. Koehler. Pennsylvania's Experience in the
Compaction of Asphalt Pavements. ASTM, Special Technical
44. Finn, F.N., K. Nair and J.M. Hilliard. Minimizing Premature
Cracking in Asphaltic Concrete Pavement. TRB, NCHRP Report
195, 1978.
45. Kandhal, P.S. Low Temperature Shrinkage Cracking of Pavements
in Pennsylvania. Proceedings AA.PT, Vol. 47, 1978.
46. Fromm, H.J. and W.A. Phang. Temperature Susceptibility Control
in Asphalt Cement Specifications. HRB, Highway Research Re-
cord 350, 1971.
47. Kandhal, P.S. Evaluation of Low Temperature Pavement Cracking
on Elk County Research Project. TRB, Transportation Research
Record 777, 1980.
48. McLeod, N.W. The Case for Grading Asphalt Cements by Pen-
etration at 77°F. Proceedings Canadian Technical Asphalt As-
sociation, Vol. 20, 1975.
49. Burgess, R.A., 0. Kopvillem, and F.D. Young. Ste. Anne Test
Road-Relationship Between Predicted Fracture Temperatures
and Low Temperature Field Performance. Proceedings, AA.PT, Vol.
40, 1971.
50. Readshaw, E.E. Asphalt Specifications in British Columbia for
Low Temperature Performance. Proceedings AA.PT, Vol. 41, 1972.
51. Kandhal,·P.S. Evaluation of Baghouse Fines in Bituminous Paving
Mixtures. Proceedings AA.PT, Vol. 50, 1981.
52. Majiq.µ.deh, K. and F.N. Brovold. State-of-the-Art: Effect ofWater
on Bitumen-Aggregate Mixtures. HRB, Special Report 98, 1968.
53. Kandhal, P.S. Field and Laboratory Evaluation of Stripping in
Asphalt Pavements: State-of-the-Art-Report. Transportation
Research Record 1454, 1994.
54. Kandhal, P.S. Moisture Susceptibility of HMA Mixes:
-------~Identification_.of...Eroblem.-an.cl.Recomm.ended Solutions. NAPA
Quality Improvement Publication 119, December 1992.
55. Kandhal, P.S., C.W. Lubc5ld, and F.L. Roberts. Water Damage to
Asphalt Overlays: Case Histories. Proceedings AA.PT, Vol. 58,
1989.
56. McGennis, R.B., S. Shuler, and H.U. Bahia. Background of

Superpave Asphalt Binder Test Methods. FHWA, Report No.
FHWA-SA-94-069, July 1994.
57. Warren, RS., RB. McGennis, and H.U. Bahia. Superpave
Asphalt Binder Test Methods-An Illustrated Overview. FHWA,
Report No. FHWA-SA-94-068, July 1994.
58. American Society for Testing and Materials. 1994 Annual Book of
ASTM Standards, Vol. 04.04, Rubber, Natural and
Synthetic-General Test Methods, 1994.
59. Bahia, H.U. and D.A. Anderson. The Pressure Aging Vessel
(PAV): A Test to Simulate Rheological Changes Due to Field
Aging. ASTM Special Technical Publication 1241, 1994.
60. Bahia, H.U. and D.A. Anderson. The SHRP Binder Rheological
Parameters: Why Are They Required and How Do They Compare
to Conventional Properties. Transportation Research Board,
Preprint Paper No. 950793, January, 1995.
61. Christensen, D.W. and D.A. Anderson. Interpretation of
Dynamic Mechanical Test Data for Paving Grade Asphalt Cement.
Journal of the Association of Asphalt Paving Technologists, Vol.
61, 1992. ,
62. Anderson, D.A. and T.W. Kennedy. Development of SHRP
Binder Specifications. Journal of the Association of Asphalt
PavingTechnologists, Vol. 62, 1993.
63. Anderson, D.A., D.W. Christensen, R Dongre, M.G. Sharma,
and P. Jordhal. Asphalt Behavior at Low Service Temperatures.
FHWA Report No. FHWA-RD-88-078, March 1990.
64. American Society for Testing and Materials. 1994 Annual Book of
Standards, Vol. 08.01, Plastics, 1994.
65. Kandhal, P.S., R. Dongre, and M.S. Malone. Prediction of
Low-Temperature Cracking Using Superpave Binder
Specifications. Journal of the Association of Asphalt Paving
Technologists, Vol. 65, 1996.
66. Bahia, H.U. and D.A. Anderson. Glass Transition Behavior and
Physical Hardening of Asphalt Binders. Journal of the Association
of Asphalt Paving Technologists, Vol. 62, 1993.
?,Superpave f>erfor:man-ce-6raded:-Aspha:l-t-Binder-Speci-fieatiefl-a-fl:B------
Testing. Asphalt Institute Superpave Series No. 1 (SP-I), 1994.
68. Petersen, J.C. Chemical Composition of Asphalt as Related to
Asphalt Durability-State-of-the-Art. TRB, Transportation Re-
search Record No. 999, 1984.
Chapter 2__:_Asphalt Refining, Uses, and Propertie 119
69. Halstead, WJ. Relation of Asphalt Chemistry to Physical Prop-

erties and Specifications. Proceedings AAPT, Vol. 54, 1985.
70. Nellensteyn, EJ. Relation of the Micelle to the Medium in
Asphalt. Journal of the Institute of Petroleum Technologist, Vol.
14, 1928.
71. Rostler, ES. Fractional Composition: Analytical and Functional
Significance. (Chapter 6-Bituminous Materials-Asphalts, Tars,
and Pitches); Vol. II, Asphalts. Edited by Arnold J. Holberg.
Huntington, New York, Robert E. Krieger Publishing Company,
1979.
72. Oliensis, G.L. The Spot Test. Proceedings AAPT, Vol. 6, 1935.
73. Heithaus, J.J. and D.E Fink, An Examination of the Significance
of the Olierisis Spot Test. Proceedings AAPT, Vol. 28, 1959.
74. Rostler, ES. and H.W Sternberg. Compounding Rubber with
Petroleum Products. Industrial and Engineering Chemistry, Vol.
41, 1949. .
75. Rostler, ES. and R.M. White. Composition and Changes in Com-
position of Highway Asphalts, 85-100 Grade. Proceedings AAPT,
Vol. 31, 1962.
76. Corbett, L.W Composition of Asphalt Based on Generic Frac-
tionation Using Solvent Deasphalting, Elution-Adsorption Chro-
matography, and Densimetric Characterization. Analytical
Chemistry, Vol. 41, 1969.
77. Jennings, P.W High Pressure Liquid Chromatography as a
Method of Measuring Asphalt Composition. Report No.
FHWA-MT-7930 Montana State University, Bozeman,
Montana, 1980.
78. 'Branthaver, J.E, J.C. Petersen, R.E. Robertson, J.J. Duvall, S.S.
Kim, P.M. Harnsberger, T. Mill, E.K. Ensley, EA. Barbour, andJ.E
Schabron. Binder Characterization and Evaluation, Volume 2:
Chemistry.. Strategic Highway Research Program Report No.
SHRP-A-368, 1993.
CHAPTER3.AGGREGATES
INTRODUCTION
Aggregates used in highway construction are largely obtained from

local supplies of natural rock The natural rocks occur as either outcrops at
or near the surface or as gravel deposits usually along old stream beds.
Natural rocks are classified by geologists into three groups depending on
their origin-igneous, sedimentary, and metamorphic. Other types of
aggregate that are sometimes used for HMA are lightweight aggregate, pro-
duced by heating clay to very high temperatures, and slag, normally pro-
duced in the blast furnace during steel production. These two artificial
aggregates have provided good skid resistance when used in HMA.
Igneous rocks are primarily crystalline and are formed by the cooling
of molten rock magma as it moves toward or on the surface of the earth.
Igneous rocks are classified based on size of the crystal grains and on com-
position as either acidic or basic. The crystal grains are classified as coarse
if they are larger than 2 mm and fine when smaller than 0.2 mm.
Classification of igneous rocks based on composition is shown below:
Acidic Intermediate Basic

Silica,% > 66 55-66 < 55
Specific gravity < 2.75 > 2.75
Color Light Dark
Presence of free quartz Yes No
- - - - - - - -ecii-men-tat:y-r:ocks-ar:e-p.rimarily_fo.rmecLeither by the deposition of

insoluble residue from the disintegration of existing rocks or from deposi-
tion of the inorganic remains of marine animals. Classification is based on
the predominant mineral present as calcareous (limestones, chalks, etc.),
siliceous (chert, sandstone, etc.), or argillaceous (shale, etc.).
Metamorphic rocks are igneous or sedimentary rocks that have been

subjected to heat and/or pressure sufficient to change their mineral struc-
ture so as to be different from the original rock. Metamorphic rocks are
generally crystalline in nature with grain sizes from fine to coarse.
Gravel is formed from the breakdown of any natural rock. Gravel par-
ticles are found in existing or ancient waterways, and the particles are usu-
ally smooth and typically rounded or subrounded by wear as the material
is moved along the waterway by the action of water. Gravel is usually
required to be crushed prior to use in HMA.
Sands consist primarily of the most resistant final residue of the dete-
rioration of natural rocks. The predominant i.ineral in sand is normally
quartz. The grain size normally ranges from a No. 8 (2.36 mm) sieve down
to dust size particles (No. 200 or 75 µm sieve). Many sand deposits often
contain silt and/or clay particles in various quantities and may be required
to be washed prior to use in HMA. Many northern states have a great deal
of glacial sands and gravels which have less clay than most river gravels and
sands.
Slag is a byproduct of metallurgical processing and is typically pro-

duced from processing of steel, tin, and copper. Slags typically have the
properties of igneous rocks and their structure can vary from glassy to
honeycombed. Blast furnace slag produced during the processing of steel is
the most widely used of the slags for pavement construction applications.
This aggregate produces a high quality asphalt mix with good skid resis-
tance. The absorption of slag is often high; hence, the amount of asphalt
binder required when slag aggregates are used is usually higher than that
for naturally occurring aggregates.
For highway construction purposes it is not usually necessary to have

extensive geologic information about the rocks. The important characteris-
tics of the rocks relate to how well the materials serve in the various appli-
cations such as sub bases, bases, or in the various HMA courses used in pave-
ment construction. Therefore, for the most part, it is the physical properties
of the rocks which are important for highway construction. The chemical
properties, which are not well understood, are important to ensure goo··rt··-------
bond between the asphalt and aggregate. Since the natural rock materials
coming from different areas within a quarry or gravel pit can vary, it is
important to sample and test the material on a regular basis to ensure that
aggregate properties are consistent and meet the specified standards.
Chapter 3-Aggregates 123
AGGREGATE PRODUCTION
. An aggregate's physical characteristics, such as resistance to abrasion

and strength, are determined primarily by the characteristics of the parent
rock. However, the production process in the quarry can significantly
affect the quality of the aggregate by the elimination of weaker rock layers
and by the effect of crushing on the particle shape and gradation of the
aggregate.
Almost all quarries have a layer of overburden material that must be
removed before sound rock is encountered. The overburden typically con-
sists of soil and decomposed rock. The demarcation line between sound
rock and decomposed rock is often difficult to establish. In addition,
decomposition often occurs along cracks in the rock or at joints between
rock layers. Therefore it is important to eliminate any inferior rock by
stripping and wasting the soil overburden and decomposed rock and by
scalping the sound material off at the crushing operation and allowing the
unsound material to pass the screening unit (grizzley) and be wasted. To
provide an overall idea of a crushing operation, a schematic diagram and
brief description of a simplified crusher set-up are provided in the follow-
ing paragraphs.
Quarry Operations
The basis purpose of a quarry operation is to remove sound rock from

the face• of the quarry by blasting and then using a series of crushers, pul-
verizers, and,screening units to size the materials into a sufficient number
of components to produce the desired pavement construction materials. It
is also desirable to produce crushed aggregate that is essentially
cube-shaped and not flat or elongated. A simplified layout for a crushing
operation is shown in Figure 3-1. Notice that a jaw crusher is used as the
primary crusher, and a standard cone crusher is used as the secondary
crusher. These selected crushing units are chosen to:
1. Produce the desired product size and shape;

2. Reduce the largest size of rock normally received;
-------3-.-Hav-e-suffi.£-i.@-n-t-Gapac-i-1:y-t-e-haacll~d1e peak loads; -
4. Minimize choking and plugging;
5. Safely handle uncrushable material, like iron, by yielding instead of
breaking;
6. Require a minimum of personnel;
7. Satisfy crushing requirements without the need for additional

crushing stages and auxiliary equipment;
8. Provide lowest power demand per ton of finished product;
9. Resist abrasive wear;
1O. Operate economically with a minimum of maintenance; and
11. Offer long, dependable service life.
TRUCK Aggregate Grizzley Jaw

Feeder Crusher
Screening
Deck + lln.
II in. and ¾in.I
Minus ¾in.
Cone
Crusher
Figure 3-1. Simplified Crusher Set-up
The overburden is removed and wasted prior to blasting the quarry

face. Blasting is a dangerous operation and must be performed by a quali-
fied individual who is licensed to perform this type work. The rock,
removed from the quarry floor after blasting, is transported to the crusher
by truck:-:After-biasting~me-weatlrered-rod:s:--aromrd-crackscmd-at-other-----~
locations typically breaks into smaller pieces than the remaining sound
rock. These smaller pieces are removed with a grizzley prior to being
crushed. For the example shown in Figure 3-1, all material smaller than
one inch is removed and wasted. The material larger than one inch that
passes over the grizzley is fed into a jaw crusher (primary crusher) that
breaks the rock down to a size that can be handled by other crushers.
Suppose for this simplified crushing operation that four aggregate sizes will
be produced: 1 inch-3/4 inch (25.4 mm-19 mm), 3/4 inch-3/8 inch
(19 mm-9.5 mm), 3/8 inch-No. 4 (9.5 mm-4.75 mm), and minus No. 4
(4.75 mm). The next operation after the jaw crusher is to screen the aggre-
gate into various sizes. The first screen deck separates aggregate into three
sizes by using a 1-:--inch (25.4 mm) and a 3/4-inch (19 mm) screen. The
material larger than 1 inch (25.4 mm) is fed to the cone crusher for addi-
tional crushing. The material smaller than 1 inch (25.4 mm) but larger
than 3/4 inch (19 mm) is stockpiled. The material smaller than 3/4 inch
(19 mm) is fed to a second screen deck for further separation. The second
screen deck uses three screen sizes 3/4 inch (19 mm), 3/8 inch (9.5 mm),
and No. 4 (4.75 mm). The material larger than 3/4 inch (19 mm) is fed
back to the cone crusher for further crushing. The other three material sizes
are stockpiled as shown. This is a very simplified crushing operation that
identifies methods that are normally used to crush the aggregate and sepa-
rate it into sizes . Most operations are much more complicated than this
but, regardless of the complexity, the operation is almost entirely made up
of crushers and screening decks. Many operations also have the capability
to wash the aggregate as needed at certain points in the operation.
During the operation of a crushing facility, it is absolutely essential
that the properties of the final product be consistent. This requires that the
parent rock being fed into the crusher has consistent physical properties
such as specific gravity and absorption. When rock is removed from vari-
ous locations in the quarry, the physical properties of the aggregate may
vary substantially, thus producing problems in maintaining consistent air
voids in the HMA. Quality control during the crushing operation must
ensure that physical properties of the aggregate do not vary excessively.
Ideally, the awount of material being fed to the crushing operation should
be approximately constant. If the rate of material feed is increased or
decreased, the amount of overflow on particular screen sizes changes, espe-
cially when operating near full capacity for individual screens. For instance,
increasing the rate of flow of materials through the crusher normally results
in more overflow of finer aggregate over individual screen sizes. This over-
flow results in a stockpile with a finer aggregate gradation. The feed rate
must exceed some minimum amount for the finer aggregate to begin to
------<oJ-:v.ei:flo:w-the...scteen,_ _ _ _ _ _ _ _ _ __
The primary crusher provides the initial mechanical reduction of the
pit run material down to maximum sizes ranging from 8 inches to 1 inch.
Secondary and tertiary crushing refer to the additional stages of reduction
that are required to reduce the aggregate to some desired size.
Figure 3-2. Mechanical Reduction by Impact (1)
Mechanical reduction involves four different mechanisms: impact,

attrition, shearing and compression. Most er.ushers utilize a combination
of all four mechanisms. A brief description of each mechanism is provided
in the following paragraphs.
Impact refers to the sharp, instantaneous impingement of one object
against another. In crushing operations, impact may be by way of gravity
or dynamic impact (Figure 3-2). Most crushers use impact as one of the
mechanisms for size reduction (Figure 3-3).
CRUSHERS REDUCTION METHODS

.
Impact A- Shoartng ~
IMPACTORS
•
VERTICAL
>
IMPACT
PULVERIZERS
•
BRADFORD
BREAKERS •
HAMMERMILLS
• • •
RING-TYPE
GRANULATORS • • •
SINGLE-ROLL
CRUSHERS • • •
• •
-. •
GYRATORIES
J)(W ..CROSRERS"
--.- ----·- · · - -
CONE
• • •
Figure 3-3. Display of the Reduction Methods Employed by Different
Types of Crushers (1)
Figure 3-4. Mechanical Reduction by a Combination of Impact and

Attrition in a Hammermill (1)
Attrition refers to the reduction of materials by a rubbing action

between two hardened surfaces and is primarily a grinding method (Figure
3-4). Hammermills, gyratories, and cones use this type of reduction
method.
Shear crushing occurs as a result of a trimming or cleaving action such
as that produced by the single roll crusher shown in Figure 3-5. Shear
crushing is usually combined with other crushing mechanisms such as
impact and compression.
________.F..,,.ig:ur.e-3-=5, Mechanical Reduction by a Co~bination of Shear,

Impact, and Compression in a Single-roll Crusher (1)
Compression refers to the mechanical reduction by compressive forces

between two surfaces (Figure 3-6). Jaw crushers using compression are the
most effective system for reducing the size of larger particles of very hard
and abrasive rocks.
The types of crushers for a particular application must be carefully
selected to produce a material meeting the customer's specifications while
operating at minimum cost. Obviously the type of rock being processed
has a significant influence on the choice of crushing equipment to be used.
AGGREGATE SAMPLING
Prior to conducting any tests on aggregates, samples must be obtained
from some source using proper sampling techniques (ASTM D75). The
samples may be randomly selected or may be selected to be representative
(average of the material being sampled), depe~~ing on the purpose of sam-
ples. Usually for mix design, representative samples are taken, and for qual-
ity control, random samples are taken. When only one or two samples are
used to characterize a mixture property, representative samples are
required. However, when many samples are taken to evaluate mix quality,
random samples are normally taken. If proper samples are not taken, all
tests conducted on the aggregate are meaningless, and this may result in the
design or construction of a HMA mixture that has poor quality, resulting
in unsatisfactory performance. ,
Figure 3-6. Mechanical Reduction by Compression and Impact in a

Specially Designed Jaw Crusher that Crushes Without
Rubbing (1)
-a:rn:pt-enrf-aggYe-gare-are-rrorma±ly-ta:ken-fro-nrastnckpil-e;-b-eir,--lro
bins, or sometimes from a loaded truck. The person taking samples must
realize that the material segregates when stockpiled, loaded into trucks, or
loaded into bins. Hence, the best location to take a sample is from a con-
veyor belt; however, the full width of flow must be sampled since the aggre-
gate also segregates on the belt. In this case a random sample is taken at
one point in time from the belt. A representative sample is taken by com-
bining a number of random samples throughout some desired time peri-
od, such as one day for belt samples, or by taking samples from various
locations in a stockpile and combining these samples. Representative sam-
ples are almost always used to evaluate stockpiles. Again, samples must be
taken in a way to minimize the effect of stockpile segregation. The aggre-
gate at the bottom edge of a stockpile is usually coarser than the rest of the
stockpile. The preferred method of sampling a stockpile is to climb the side
of the pile to some point away from the top or bottom of the pile, remove
the surface aggregate and obtain a sample from underneath the surface. It
is sometimes necessary to place a form above the place being sampled to
prevent the aggregate from flowing into the area being sampled. Samples
should be taken at several locations throughout the stockpile and com-
bined for a representative sample.
After taking the required number and amount of samples and bring-
ing them back to the laboratory, each sample must be reduced to the
appropriate size for testing. Samples can be reduced with a sample splitter
or by quartering (ASTM C702). Both methods work successfully, but one
of these must be utilized if an accurate sample is to be obtained for tests.
The importance of sampling and splitting to test size cannot be overem-
phasized since all test results are dependent on the use of proper tech-
mques.
MINERALOGY AND CHEMICAL PROPERTIES

As discussed previously, it is primarily the physical properties of aggre-
gates (to a lesser extent, the chemical properties) that determine the suit-
ability for use in HMA. Basic physical/mechanical properties (such as den-
sity, porosity, and strength) and chemical/physicochemical properties (such
as wetting, adhesion, and stripping) are functions of the composition and
structure of the minerals in the aggregate. An understanding of the miner-
alogy and identification of minerals can provide information on the poten-
tial physical and chemical properties of an aggregate for a given use, and
can help avoid the use of an aggregate containing harmful mineral con-
stituents. The project specification requirements should be selected so that
aggregates having. undesirable mineral components are not accepted for
use.
Aggregate Mineralogy
Common Minerals Found in Aggregates. A majority of natural aggre-
gates are composed of a combination of minerals. Among the most impor-
tant minerals found in aggregates are silica minerals (quartz); feldspars

(orthoclase, plagioclase); ferromagnesian minerals (muscovite, vermi-
culite); carbonate minerals (calcite, dolomite); and clay minerals (illites,
kaolinites, montmorillonites). These minerals and associated rocks are
described in ASTM C294 (Table 3-1). Table 3-2 gives the average miner-
al compositions commonly found in aggregates U)
Effect of Mineralogy on Performance. Since minerals have definite

chemical composition and usually specific crystalline structure, the physi-
cal and chemical properties of aggregates can be expected to be associated
with mineralogy of the aggregates. Table 3-3 gives typical engineering
Table 3-1. Rocle and Mineral Constituen~_in Aggregates (J)
~ Igneous rocks Md1m11mhl~ rocks

Silica Granite Marble
Quartz Syenite Meta quartzite
Opal Diorite Slate
Chalcedony Gabbro Phyllite
Tridymite Peridotite Schist
Cristobalite Pegmatite Amphibolite
Silicates Volcanic glass Hornfels
Feldspars Obsidian Gneiss
Ferromagnesian Pumice Serpentinite
Hornblende Tuff
Augite Scoria
Clay Perlite
Illites Pitchstone
Kaolins Felsite
Chlorites Basalt
Montmorillonites
Mica S~!!im~ntl!o: rocks
Zeolite
Carbonate Conglomerate
Calcite Sandstone
Dolomite Quartzite
Sulfate Graywacke
Gypsum Subgraywacke
Anhydrite Arkose
Iron sulfide Claystone, siltstone,
Pyrite argillite, and shale
Marcasite Carbonates
Pyrrhotite Limestone
Iron oxide Dolomite
Magnetite Marl
Hematite Chalk
Goethite Chert
Ilmenite
Limonite
For brief descriptions, see "Standard Descriptive Nomenclature of Constituents of Natural Mineral
Aggregates" (ASTM C294).
~
I (")
::,'
-
j Rock Quartz Orthoclase Plagioclase Augite Hornblende Biotite Muscovite Epldote Rock Glass
I Igneous Rocks
C"'
n
\,I,)
i
v>
-
~
dranite 30 41 8 - - 3 3 1
rliorite
dabbro
8
0.5 -
7 30
44
3
28
27
9
4
2
0.1
-
5
1
-
-
fJ (JQ
tjiabase - - 44 46 - - - - 2 ~ ~
32 45 3 - 0.7 3 2 2 0.4 n (JQ
~yolite 8,,,
achyte
~ndesite
Basalt
3
0.6
-
42
-
-
1
48
36
2
14
35
6
3
-
0.5
-
-
-
-
-
8
3
-
0
13
21
~
~
I Metamorphic Rocks ::s
n
I .
guartztte
Feldspathic-
84 3 - - I 2 2 2 - ""
e.
~quartzite 46 27 1 - - 2 5 1 - n0
ornblende-
i:neiss 10 16 15 3 45 3 1 2 - .g
Granite-
~eiss
B otite-
37 32 3 - - 7 11 2 - ...
0
"'r:t.
hist 34 13 3 - I 38 3 2 - 0
::s
Mica-schist 37 16 I - - 13 26 2 -
slate 29 4 - - - 55 2 - g,
~arble 3 0.2 0.2 - - - Calcite=% - ""g:,
Ainphibolite 3 1 8 - 70 I 0.2 12 -
Sandstone 79 5 0.3
Sedimentary Rocks
- - 0.2 1 - -
8-
Feldspathic-
sandstone 35 26 2 - - 0.6 2 1 -
~
Calcareous-
sandstone 46 3 ·2 - - - Calcite= 42 - -
Chert 93 - - - - Calcite = 1 - -
Limestone 6 - - - Dolomite= 8 Calcite= 83 - -
Dolomite 5 - - - Dolomite = 82 Calcite = 11 - -
....
.......,
Table 3-3. Summary of Engineering Properties of Rocks (i)
Mechanical Chemical Surface Presence or

Type or Rock Strength Durability Stability Characteristic Undesirable
Impurities
Igneous:
Granite, syenite, Good Good Good Good Possible
diorite
Felsite Good Good Question- Fair Possible
able
Basalt, diabase, Good Good Good Good Seldom
gabbro
Periodotite Good Fair Question- Good Possible
able
Sedimentary:
Limestone,dolomite Good Fair Good Good Possible
Sandstone Fair Fair Good Good Seldom
Chert Good Poor Poor Fair Likely
Conglomerate, Fair Fair Good Good Seldom
breccia ,_
Shale Poor Poor Poor Good Possible
Metamorphic:
Gneiss, schist, Good Good Good Good Seldom
Quartzite Good Good Good Good Seldom
Marble Fair Good Good Good Possible
Serpentinite Fair Fair Good Fair to poor Possible
Amphibolite Good Good Good Good Seldom
Slate Good Good Good Poor Seldom
properties for common aggregate types, and Table 3-4 summarizes desir-
able properties of rocks for HMA aggregates. Most aggregates are com-
posed of several minerals, often with variable compositions. Even with
aggregates of uniform mineralogy, the properties may be altered by oxida-
tion, hydration, leaching, weathering and foreign coatings. Therefore,
mineralogy alone cannot provide a basis for predicting the behavior of an
aggregate in service. Petrographic examinations are useful, and past perfor-
mance of similar aggregates under similar environmental and loading con-
ditions can be helpful in evaluating aggregates.
In addition to the properties listed in Table 3-3 and Table 3-4, min-
eral composition of coarse aggregates also affects the skid resistance of
HMA. Quartz and feldspar are harder and more polish resistant minerals
which are normally found in igneous rocks such as basalt and gabbro. On
the other hand, calcite and dolomite, which occur in limestones, are exam-
ples of soft minerals. Limestones that have a high percentage of soft mate-
rials tend to polish more rapidly than most other aggregate types. The acid
insoluble residue test (ASTM D3042) has been used to measure .the
amount ofh:arcler materials present in carbonate aggregates. Many highway
agencies specify a minimum of 10 percent acid insolubles to assure accept-
able frictional properties. Other agencies use a polishing test (ASTM
D3319 or E660 and E303) or petrographic examination (ASTM C295) to
evaluate the polishing potential of an aggregate.
Another factor which affects an aggregate's usefulness in HMA, some-

what related to mineralogy, is the presence of surface coatings and other
deleterious substances (1'). These deleterious substances may include clay,
shale, silt, iron oxides, gypsum, water-soluble salts, and other soft particles
that affect proper bonding with asphalt. Deleterious materials may also
increase the moisture susceptibility of an asphalt mixture. Aggregates with
deleterious substances are undesirable for HMA and should not be used
unless the amount of foreign matter is reduced by washing or other means.
Specifications for aggregates normally list the deleterious materials (e.g.,
clay lumps, shale, soft particles) and limit their quantity (e.g., 3 to 6 per-
~ent). Other agencies specify a minimum sand equivalent requirement
(ASTM D2419) or a maximum plasticity index (ASTM D4318) which is
discussed later. AASHTO M283 suggests that coating on aggregates after
dry sieving should not exceed 0.5 percent when tested by AASHTO Tl I
(ASTM C117).
One of the most important effects of aggregate mineralogy on the per-
formance of HMA is its influence on adhesion and moisture damage.
Asphalt cement bonds better to aggregates with certain mineral types. For
instance, asphalt cement normally bonds better to carbonate aggregates
(limestone) than to siliceous aggregates (gravel).
Table 3-4. Desirable Properties of Rocks for HMA (i)
Hardness, Resistance to Surface Crushed

Rock Type Toughness Strlppingt Texture Shape
Igneous
Granite Fair Fair Fair Fair
Syenite Good Fair Fair Fair
Diorite Good Fair Fair Good
Basalt (trap rock) Good Good Good Good
Diabase (trap rock) Good Good Good Good
Gabbro (trap rock) Good Good Good Good
Sedimenta,y
Limestone, dolomite Poor Good Good Fair
Sandstone Fair Good Good Good
Chert Good Fair Poor Good
Shale Poor Poor Fair Fair
Metamo,phic
Gneiss Fair Fair Good Good
Schist Fair Fair Good Fair
Slate Good Fair Fair Fair
Quartzite Good ...EaiL________ ...-Good Good
Marble Poor Good Fair Fair
Se,pentine Good Fair Fair Fair
tAggregates that are hydrophilic (water-loving) tend to strip more readily since water more easily replaces an asphalt film.
Freshly crushed aggregates with many broken ionic bonds also tend to strip more readily.
Petrographic Examination. A petrographic examination of an aggre-

gate is a visual examination of individual aggregate particles using an opti-
cal microscope, sometimes supplemented by X-ray diffraction, differential
thermal analysis, electron microscope, or chemical analysis. The petro-
graphic examination can be used to measure the relative abundance of rock
and mineral types, the physical and chemical attributes (such as particle
shape, surface texture, hardness, pore characteristics, and chemical activi-
ty), and the presence of harmful contaminants. ASTM C295 describes the
detailed procedure for a petrographic examination of aggregates. This test
has been used for many years to evaluate aggregates for use in concrete but
has not been widely used for aggregates for HMA. The mineralogy of an
aggregate has a great~r effect on performance of concrete mixes than it does
for asphalt mixes. -1
Chemical Propertie~ of Aggregate

Chemical properties of an aggregate identify the chemical composition
and/or indicate the transformation that an aggregate can undergo due to
chemical action. Chemical compositions of aggregates based on chemical
analysis are usually given in terms of oxides, regardless of whether such
oxides are actually present in the sample. These data are not as informative
as mineral composition and sometimes may be confusing and misleading.
Some aggregates contain substances which (1) are soluble in water (e.g.,
gypsum); (2) are subject to oxidation, hydration, and carbonation; and (3)
can react with compounds in Portland cement, although reactivity of these
aggregates with asphalt cement has not been conclusively established.
Chemical properties of-aggregates have little effect on their suitability and
performance, except as they affect adhesion of the asphalt binder to the
aggregate and compatibility with antistripping additives that may be incor-
porated into the asphalt binder.
Asphalt cement must wet the aggregate surface, stick to the aggregate,
and resist stripping of the asphalt film in the presence of water. Therefore,
the surface chemistry ofthe aggregate particles plays an important role in
performance. The adhesion of asphalt cement to aggregate and the dis-
placement of the asphalt film by water (stripping) is a complex phenome-
non involving the physicochemical interaction between many parameters.
Several theories have been suggested on the mechanisms of adhesion and
stripping (1). It appears thus far that none can fully explain the phenome-
__ Il:on__ ~nd_ tha~~?_i:_e~~~e mechanism may be operating'-a_t_th_e_sa_m_e_______
time. However, it has been firmly established that mineralogy and chemi-
cal composition of the aggregate are of primary importance in the stripping
phenomenon.
There has been abundant evidence to indicate that some aggregates
appear to have a greater affinity for water than for asphalt cement, and
asphalt films on these aggregate particles may become detached or stripped

after exposure to water (Q). These aggregates are called hydrophilic
{water-loving), and they tend to be acidic in nature. On the other hand,
aggregates having an affinity for asphalt cement are called hydrophobic
(water-hating), and they tend to be basic in nature. It is commonly accept-
ed that the nature of electric charges on the aggregate surface when in con-
tact with water significantly affects the adhesion between the aggregate and
the asphalt cement and its resistance to moisture damage. Most siliceous
aggregates such as sandstone, quartz and siliceous gravel become negative-
ly charged in the presence of water, while limestone and other calcareous
materials bear a positive charge in the presence of water. Many aggregates
contain both types of charges because they are composed of minerals such
as silica with negative charges, and also calcium, magnesium, aluminum,
or iron with positive charges. Typical aggregates that bear mixed charges
include trap rock, basalts, porphyries, and the siliceous limesi:ones.
Dolomite and quartz are examples of extreme cases of electropositive and
electronegative aggregates, respectively. A classification of aggregates lying
between these extremes is shown in Figure 3-7 (2).
Since stripping is one of the major distresses affecting Hlv[A. perfor-
mance, the effect of aggregates on the moisture sensitivity of Hlv[A. must
be evaluated. A number of practical tests are available to predict the com-
patibility of aggregate and asphalt cement. These tests are discussed in
Chapter 5.
There has been evidence to suggest that asphalt-aggregate interactions
may cause asphalt molecular structuring at the aggregate surface which
affects the bonding and setting properties of Hlv[A. (B.). However, the pre-
cise mechanisms and the role of chemical/mineralogical composition of the
aggregate on the properties of asphalt cements and asphalt mixes are not
yet well understood.
CONTENT OF SILICA SiO 2, %
0 10 20 30 40 !50 60 70 80 90 100
• I I I
I I( SILICIOUS LIMESTONE yftjj;
BASALTS PORPH IES '
I
►! :11( ►: I
I II(
SILICA~!
~ I I I
I I I I
f"ll◄1----- POSJTIVE --'----1►
,1f-oCMIXEf► l 111' NEG-f' TIVE -;.---1>-!~!
; ; : ! j
111' SANDSTONE >-1
J◄ LIMESTONE )lls j DIORITES : 1
(CONTENT OF ~01 • ~PHITEl° ◄GRANITES ►
I 'If ' _. : I
'I
100 90 80 70 60 !50 40 30 20 O 0
CONTENT OF ALKALINE OR ALKALINE EARTH OXIDE, %
Figure 3-7. Classification of Aggregates by Surface Charges (2')
PHYSICAL PROPERTIES OF AGGREGATES

Aggregates for HMA are usually classified by size as coarse aggregates,
fine aggregates, or mineral fillers. ASTM defines coarse aggregate as parti-
cles retained on a No. 4 (4.75 mm) sieve, fine aggregate as that passing a
No. 4 sieve (4.75 mm), and mineral filler as material with at least 70 per-
cent passing the No. 200 (75 µm) sieve. Other agencies may use the No.
8 (2.36 mm) sieve (e.g., the Asphalt Institute) or the No. 10 (2.00 mm)
sieve as the dividing line between coarse and fine aggregates.
Aggregates for HMA are generally required to be hard, tough, strong,
durable (sound), properly graded; to consist of cubical particles with low
porosity; and to have clean, rough, hydrophobic surfaces. Table 3-5 sum-
marizes the various properties that an aggregate must possess in order to
fulfill its function and the relative importance of aggregate in HMA (2)
Specifications for coarse aggregates, fine aggregates, and mineral fillers
are given in ASTM D692, D1073 and D242, respectively. The suitability
of aggregates for use-in HMA is determined by evaluating the following
material characteristics:
1. Size and gradation;
2. Cleanliness/deleterious materials;
3. Toughness/hardness;
4. Durability/ soundness;
5. Surface texture;
6. Particle shape;
7. Absorption; and
8. Affinity for asphalt.
Affinity for asphalt is related to surface chemistry and was discussed
earlier. Other physical properties and the tests commonly used for evalua-
tion are discussed in the following sections. In addition, the various types
of specific gravity of aggregates are discussed since these are used to calcu-
late voids in mix design. The most important characteristics of aggregates
for HMA, their significance, commonly used tests to evaluate these prop-
erties, and recommended test requirements are listed in Table 3-6. At the
present time (1996), The National Center for Asphalt Technology
(NCAT) is conducting a major research project (NCHRP 4-19) titled
''Aggregate Tests Related to Asphalt Concrete Performance in Pavements"
which is scheduled to be completed in March 1997. The objective of the
---rc--cec-cse~arcli 1s to recommend a set of aggregate tests related to the performance
of HMA used in pavement construction. This research will include the
evaluation of existing aggregate tests to assess their ability to predict pave-
ment performance and, where this predictive ability or clear relationship to
performance is lacking, the development of new tests.
Table 3-5. Summary of Aggregate Properties to Meet Function of

System (2)
Function Aegn,gate Property Relative lmportance1 or
Property in Hot Mix Asphalt
Have adequate internal strength and 1. Mass stability I

stability to distnbute surface pressures 2. Particle strength I
and to prevent extensive surface 3. Particle stiffness I
deflections 4. Particle surface texture I
5. Particle shape I
6. Grading I
7. Maximum particle size I
Resist deteriorating effects of weather l. Resistance to attack by u

and chemicals chemicals
2. Solubuity u
3. Slaking I
4. Resistance to wetting.drying u
S. Resistance to freezing•thawing u
6. Pore structure I
Resist deteriorating effects of applied

loads 1. Resistance to degradation I
Resist the effects of internal forces, such 1. Volume change-thermal N

as expansion, contraction and warping 2. Volume change~wening and diying N
3. Pore structure N
4. Thermal conduaivicy N
Be compatible with binder used in system 1. Chemical compounds reactivity I
2. Organic material reactivity N
3. Coatings I
4. Volume stability N
S. Base exchange I
6. Surface charges I
7. Pore structure N
Retain acceptable standards of

performance at the surface by:
a. Maintaining adequate skid resistance 1. Particle shape I
2. Particle surface texture I
4. Particle strength I
S. Wear resistance I
6. Pa1ticle shape or abraded I
fragments
7. Pore structure I
b. Having acceptable surface roughness 1. Maximum particle size I

characteristics 2. Grading I
c. Minimizing glare and light reflectivity 1. Reflectivity I

2. Glare I
d. Preventing the occurrence of loose 1. Resistance to degradation I

material 2. Specific gravity I
e. Minimizing tire wear l. Particle shape I

2. Particle surface texture I
f. Minimizing rolling resistance 1. Maximum particle size I

2. Particle shape I
g. Minimizing noise level 1. Maximum particle size I

2. Particle shape 1
h. Preventing electrostatic buildup 1. Electrical conductivity I

Retaining properties during the 1. Maximum particle size 1
construction process that suppon all 2. Resistance to degradation 1
-other-fimctions-of-the-syste J:-lntegriry-d1.rring-1reartng---· - I·· I
'lillm
I - Important
N - Not Important
U - Importance Unknown
Table 3-6. Characteristics and Tests of Aggregates for HMA
Characteristics Slgnillcance Test Requirement
Hardness/toughness Resistance to abrasion and ASTMC131 Maximum percent weight

degradation; quality of ASTM C535 Joss; ASTM D692; D1073;
particles AASHTOM283
Soundness Resistance to freeze and ASTMC88 Maximum percent wejght loss;

thaw; wet and dry AASHTOT103 ASTM D692; Dl073;
AASHTOM283
Particle shape and surface Friction; skid-resistance; ASTM D3398 Minimum percent of crushed
texture compaction; mix stability ASTMD479! particles; ASTM D692;
ASTMC1252 AASHTOM283
Resistance to polishing Skid and wear resistance ASTM D33!9 Maximum polish value;
ASTM E303 Minimum acid insoiubles
ASTM E660
ASTM D3042
Durability Resistance to effect of ASTM D3744 ,_ Maximum durability index

weather and aging
Resistance to stripping Resistance to effect of ASTM D1664 Minimum percent of coated

water NCHRP 246; 274 area; Percent retained strength
AASHTOT283 or modulus; Number freeze-
ASTM D1075 thaw cycles to' cause crack;
AASHTO Tl82 AASHTO M283
Specific gravity and water Mix design calculations; ASTM C!27

absorption porosity ASTMC!28 -
Asphalt absorption Mix design calculations; ASTM D2041
durability ASTM D4469 -
Gradation and size Mix stiffness, stability, ASTM Cll7 Minimum and maximum
workability, fatigue ASTM C!36 & C702 percentages passing standard
resistance, fractural AASHTOT30 sieves; ASTM D3515
strength, durability,
permeability, voids,
compactibility
Mineral composition Wettability and stripping ASTM C294

potential; porosity ASTM C295 -
Cleanliness and Resistance to weathering ASTM C!l7 & D422 Maximum clay lumps and soft
deleterious substances and effects o( water ASTM C!23 particles of one percent;
ASTM Cl42 Minimum sand equivalent of
ASTM D2419 25-45%;
ASTM D4318 P.I. not exceeding 4-6; Percent
fines (ASTM D242; D1073);
AASHTO M283
Toughness and Abrasion Resistance

Aggregates, through internal friction, must transmit the wheel loads to
the underlying layers and also be resistant to abrasion and polishing due to
traffic. Aggregates are subject to crushing and abrasive wear during manu-
facturing, placing, and compaction of HMA. They must be hard and
tough to resistcmslrirrg~degradati:on;--an-d-disintegration-when-sto-ckpiled-,- - - - -
fed through a HMA facility, placed with paver, compacted with rollers, and
travelled over with trucks.
The Los Angeles (L.A.) abrasion test is most often used to obtain an
indication of the desired toughness and abrasion characteristics.
Methods that are available include ASTM C131 for coarse aggregate
smaller than 1.5 inch (38 mm) and ASTM C535 for testing aggregates
larger than 0.75 inch (19 mm) and up to 3 inch (76 mm) maximum size.
ASTM Cl31 is performed by mixing 5,000 g of aggregate blended to meet
one of several gradings, along with a charge of six to twelve steel balls, each
approximately 2 inches (51 mm) in diameter and approximately 400 g in
weight. This mix is tumbled inside a steel drum for 500 revolutions at a
speed of 30-33 rpm. This is the method that is normally used for HMA
aggregates. ASTM C535 is performed with 10,000 g of blended aggregate,
twelve steel balls, and 1,000 revolutions. After tumbling, the fines that pass
a No. 12 (1.68 mm) sieve are weighed, and the percentage loss by weight
of original sample is calculated as the Los Angeles abrasion. Initially, none
of the material would pass a No. 12 sieve. Hence if 40 percent of the mate-
rial passed a No. 12 sieve after the test, the L.A. abrasion would be 40.
Typical test values range from 10 percent for extremely hard igneous
rocks to 60 percent for soft limestones and sandstones. Obviously, a num-
ber of aggregates will exceed this typical range. The maximum wear for
coarse aggregate used in high-type HMA is typically limited by specifica-
tions to 40 percent by some agencies to as much as 60 percent by other
agencies. ASTM D 1073 does not specify a maximum abrasion loss for fine
aggregates for any purpose.
The Los Angeles abrasion test is primarily a measure of the resistance
of coarse aggregate to degradation by abrasion and impact; however, field
observations do not show a good relationship between L.A. abrasion loss
and performance. This test does not appear to be fully satisfactory for use
with slags, cinders, and other lightweight aggregates. Experience has shown
that many aggregates such as slag and soft limestones provide excellent per-
formance even though the Los Angeles abrasion is high. One item that
usually must be dealt with when producing HMA with aggregate having a
high L.A. abrasion is the generation of dust during handling of aggregates
and production of HMA. A high dust content will likely produce envi-
ronmental problems as well as mixture control problems.
Durability and Soundness
Aggregates must be resistant to breakdown or disintegration under

------<S1,eti-on-0f-wetti-ng-ancl-cl:ry-i-l'l-g-and-for-freezing··-and ·thawing (weathering).
The soundness test (ASTM C88) is an empirical screening test that is
intended to provide an indication of durability due to weathering and is
useful for evaluating new sources of aggregate for which no service records
are available. This test involves submerging the aggregate in a solution of
sodium or magnesium sulfate. During this test salt crystals grow in the aggre-
gate pores and cause particles to disintegrate in some aggregates. Presumably
this crystal formation simulates that of ice crystals and hence to some extent
simulates the effect of freezing and thawing. An oven-dried sample of the
aggregate is separated into specified sizes and immersed in a saturated solu-
tion of sodium or magnesium sulfate. The aggregate remains in the solution
at a constant temperature for approximately 18 hours. The sample is then
removed from the solution, dried to constant weight at 230 ± 9°F (110±5°C)
and cooled. This cycle is typically repeated five times, after which the sample
is washed to remove the salt and is dried. The loss in weight for each size frac-
tion is determined by sieving, and the weighted average percent loss for the
entire sample is computed (see ASTM C88 for sample calculation).
A number of tests to evaluate the effects of freezing and thawing have
been proposed. Typically the procedures are similar to that of ASTM C88,
except that different sized samples are immersed in water or 0.5 percent
water-alcohol solution and then alternately subjected to freezing and
thawing for the required number of cycles (16 to 50). MSHTO T103 is
one such procedure.
When specifications for aggregates contain a requirement for sound-
ness by the sodium or magnesium sulfate test, the specifications generally
state both the number of cycles and the maximum allowable percentage of
loss. For example, ASTM D692 specifies a maximum of 12 percent loss
after 5 cycles when using sodium sulfate and 18 percent loss when using
magnesium sulfate since the action of the magnesium sulfate is more
severe. The sulfate soundness test has been an accepted method of testing
for many years. Despite this acceptance, it has been widely criticized, and
numerous reports have appeared which describe its inability to accurately
predict field performance for specific aggregates. Many engineers believe
that freezing and thawing cycles do not create problems with aggregates in
HMA. The aggregate is initially dried during production and hence should
have little or no moisture in the pores immediately after production.
(Aggregates with high absorption values probably do have some free mois-
ture in the pores.) During HMA production, the aggregates are coated
with a film of asphalt binder which should prevent the aggregate from
absorbing a significant amount of moisture during the life of the mixture.
Since there is little or no moisture in the aggregate, freezing and thawing
should not be a significant problem.
------------------------------
- - - - - - - - -----------
Particle Shape and Surface Texture

Aggregate particles suitable for use in HMA should be cubical rather
than fl.at, thin or elongated. In compacted mixtures, angular-shaped parti-
des exhibit greater interlock and internal friction, and hence result in
greater mechanical stability than do rounded particles. On the other hand,
mixtures containing rounded particles, such as most natural gravels and
sands, have better workability and require less compactive effort to obtain
the required density. This ease of compaction is not necessarily an advan-
tage, however, since mixtures that are easy to compact during construction
may continue to densify under traffic, ultimately leading to rutting due to
low voids and plastic flow.
The surface texture, like particle shape, also influences the workability
and strength of HMA. A rough, sandpaper-like surface texture, such as
found on most cru,shed stones, tends to increase strength and requires
additional asphalt cement to overcome the loss of workability, as compared
to a smooth surface found in many river gravels and sands. Voids in a com-
pacted mass of rough-textured aggregate are also usually higher, providing
additional space for asphalt cement. Smooth-textured aggregates may be
easier to coat with an asphalt film, but the asphalt cement usually forms
stronger mechanical bonds with the rough-textured aggregates.
The combined effects of particle shape and surface texture of an aggre-
gate can be determined using the particle index test of ASTM D3398. The
equipment required for this test is simple, consisting basically of a cylin-
drical steel mold 6 inches (152 mm) in diameter (diameter ranges from 2
inches or 51 mm for particles passing No. 50 sieve to 8 inches or 203 mm
for particles larger than 1 inch or 25 mm) by 7 inches (178 mm) high, and
a steel rod 5/8 inch (16 mm) in diameter by 24 inches or 610 mm long
(the tamping end is rounded to a hemispherical tip). Tamping rods weigh
between 34 g for the finest size fraction to 2,204 g for the largest size frac-
tion. The procedure requires that a clean, washed, oven-dried, single-size
aggregate fraction be used. The mold should be filled in three equal layers
with each layer receiving ten tamps with the tamping rod. Each tamp con-
sists of a drop with the tamping rod from 2 inches above the surface of the
layer being compacted. The weight of the contents of the mold in each case
is determined and the corresponding percentage of voids is calculated using
the bulk specific gravity of each aggregate fraction. This procedure is
repeated using the same material but applying 50 blows on each of the
. three layers. The particle index is then calculated using the following equa-
tion:
Ia= l.25V10 - 0.25V50 - 32.0
Ia = particle index value;

V 10 = percent voids in the aggregate compacted with 10 blows
per layer; and
V 50 = percent voids in the aggregate compacted with 50 blows

per layer.
The particle index of an aggregate containing several size fractions is

weighted on the basis of the percentage of the fractions in the original
grading of the aggregate. An aggregate composed of rounded particles with
smooth surface textures may have a low particle index of 6 or 7, while
aggregates consisting of highly angular crushed particles with rough surface
textures can have particle indices of 15 to 20 or more. Figure 3-8 shows
the effect of particle index on Marshall stability (JO). This test is very time
consuming and is therefore not widely used.
A HMA can generally be made more stable and resistant to rutting by
specifying that a significant portion (e.g., '90 percent) of the coarse aggre-
gate have at least one fractured face. For high-type HMA, the FHWA (1J)
recently recommended that specifications require at least 60 percent of the
plus No. 4 (4.75 mm) material have at least two mechanically induced,
fractured faces and that natural sand be limited to 15 to 25 percent.
Natural sand is limited because it is generally rounded and hence is less
desirable than crushed fine aggregates. Some agency specifications require
that at least 75 percent or 85 percent of the coarse particles have two
crushed faces and that at least 90 percent have one crushed face. The par-
ticle shape and texture of fine aggregates has more influence on the
strength of dense-graded HMA than that of coarse aggregate. A high per-
centage of fractured faces is important for dense-graded mixtures to pro- ,
vide adequate stability but is even more important for open-graded fric-
tion courses.
Currently there is no ASTM or MSHTO standard test procedure for
measuring the percentage of fractured faces for an aggregate. One must
split out portions of material retained on the No. 8 and larger sieves and
count the particles that have none, one, and two or more fractured faces.
This test cannot practically be conducted on aggregates smaller than a No.
8 (2.36 mm) sieve because the particle sizes are too small to be visually
evaluated. The percentage by weight of each size having one and two or
more fractured facfS is computed. Weighted averages are then computed
by multiplying these percentages by the percentages of the respective aggre-
gate sizes retained on the sieve size being evaluated. The results are totalled,
providing a weighted value for perce..o.co£aggregate-with-o.ne-and-t:w:0-0J:-"- - - -
more fractured faces.
The particle shape and surface texture of the fine aggregate can be
quantified by the use of ASTM C1252 or MSHTO TP33 (Uncompacted
Void Content of Fine aggregate). This test is very promising in predicting
the rutting potential of HMA based on the work done at the National
Center for Asphalt Technology (12, 1.3.). It has been included in the
Superpave mix design system. In this method, a 100-cm3 cylinder is filled
with fine aggregate of prescribed gradation by allowing the sample to fl.ow
through the orifice of a funnel into the calibrated cylinder. Excess mater-
ial is struck off and the cylinder with aggregate is weighed. Uncompacted
void content of the sample is then computed using this weight and the
bulk dry specific gravity of the aggregate. Three variations of the method
are used. Method A uses a graded sample of specified gradation. Method
Buses three individual size fractions: Nos. 8 to 16 (2.36 to 1.18 mm),
Nos. 16 to 30 (1.18 to 0.6 mm), and Nos. 30 to 50 (0.6 to 0.3 nim); and
the mean void content is determined. Method C uses the gradation as
received. Method A is used in the Superpave mix design system.
Flat and elongated particles impede compaction and thus may prevent
the development of satisfactory strength in HMA. AASHTO M283
requires that the portion of aggregate retained on the 3/8-inch (9.5 mm)
sieve shall not contain more than 15 percent of particles by weight that
have a ratio oflength to width or width to thickness that exceeds 5: 1. The
fl.at and elongated particles in coarse aggregates can be determined follow-
ing ASTM D4791, using a proportional caliper device. The shape of the
fine aggregate is also very important, and research is being performed to
develop tests that accurately quantify the particle shape and thus determine
the suitability of the fine aggregate blend for use in HMA.
..,
~
--:-400
- PARTICLE INDEX-15
,I I I I I I I
.,..0 PARTICLE INDEX-11.5 I
sf" I I I I I
,--- 300
PAR 111..LI,. ll'IUc..-.-t:J
I j
-< I I I
,->-
,-t-
~- 2900 I
-~
-Iii I
I
COMP. TEMP.-280"F
. I/
. ,
-..J I/ / J
-..J-
< 1o'o V VMA-15% lj j
,,,.
-~
-<
:::E ,_
- r.f... II-'
AIR VOIDS-4%
I I I I
I -86 88 90 92 94 95 98 10011021 I I
I I %LAB. COMP. SPEC. GRAV. I I I I I I
Figure 3-8. Effect of Particle Index on Marshall Stability (11!)
Cleanliness and Deleterious Materials
Cleanliness refers to absence of certain foreign or deleterious materials

that make aggregates undesirable for HMA. Washing the dirty aggregates
can usually reduce the amount of undesirable foreign matter to an accept-

able level. Typical objectionable materials include vegetation, shale, soft
particles, day lumps, day coating on aggregate particles, and sometimes
excess dust from the crushing operation. In addition to the petrographic
examination described earlier, the following tests can be used to identify
and measure the quantity of deleterious materials.
Sand Equivalent Test. The Sand Equivalent Test (ASTM D2419) is

used to determine the relative proportions of plastic fines and dust in fine
aggregates. In this test, 85 ml of aggregate passing a No. 4 (4.76 mm) sieve
is agitated in a water-filled transparent cylinder (1.25 inches or 32 mm
inside diameter, 17 inches or 432 mm high,· and graduated from the bot-
tom up to 15 inches or 381 mm by tenths) whkh is filled with a mixture
of water and a flocculating agent (a combination of calcium chloride, glyc-
erine and formaldehyde). After agitation and 20 minutes of settling, the
sand separates from the flocculated day, and the heights of day and sand
in the cylinder are measured. The sand equivalent is the ratio of the height
of sand to the height of day times 100. Cleaner aggregate ·will have a high-
er sand equivalent value. Specifications for aggregates in HMA often spec-
ify a minimum sand equivalent in the range of 25 to 35.
Clay Lumps and Friable Particles. Clay and friable particles (ASTM
C142) may be present in the form oflumps. The harm caused by these
lumps depends on their survival as lumps in HMA. These lumps in the fin-
ished HMA can break down from freezing and thawing or wetting and
drying and cause stripping, ravelling, or otherwise affect the durability of
the HMA. These particles can normally be removed during the aggregate
crushing operation at the grizzley or in. a washing operation. Some of the
particles break down when processed through the HMA facility. Natural
sands and gravels are the most likely source of these lumps. If the lumps
remain intact during placement of the HMA, surface pitting is likely to
occur as the lumps on or near the surface deteriorate under rolling or traf-
fic. If the total percent of day lumps or friable particles in the HMA is very
low (less than 1 percent), .the surface problems will be minimal, and per-
formance will be satisfactory. Some pitting and popouts may occur, but
will
--~o_th.e.r.:wi.s_e_th_e_p_erformance not be adversel;i~affc=e=ct=e=d~._____________ ..J._
In this test for day lumps and friable particles, aggregates are first
washed and dried to remove material passing the No. 200 (75 µm) sieve
and then separated into different sizes. Test samples are weighed and
soaked in water for 24 hours. Any particles that can be broken with the fin-
il
gers after soaking and removed by wet seiving are classified as clay lumps
or friable particles, and the percent of this material is calculated by weight
of the total test sample. Specifications normally limit the amount of clay
lumps and friable particles to a maximum of 1 percent.
Plasticity Index. The Plasticity Index (PI) (ASTM D43 I 8) is the dif-
ference between the liquid limit and the plastic limit of the material pass-
ing the No. 40 (0.42 mm) sieve. PI is a measure of the degree of plasticity
of fines (material passing the No. 200 sieve) and can indirectly indicate the
amount and type of plastic fines. ASTM D1073 and D242 limit the PI of
the aggregate fraction passing the No. 40 sieve of both the fine aggregate
and the mineral filler to four or less
Specific Gravity
The specific gravity of an aggregate is useful in making weight-volume

conversions and in calculating the void content in a compacted HMA. By
definition, the specific gravity of an aggregate is the ratio of the weight of
a unit volume of the material to the weight of an equal volume of water at
approximately 23°C (73.4°F). One equation for specific gravity that is
often used is:
weight
specific graviry =
volume(unit weight of water)
When working in the metric system the unit weight of water is 1.0 gram
per ml. Hence the equation for specific gravity becomes:
weight
specific graviry =
volume
_ _ _ _w.he.n_tb_e_w.eigh_t is in grams and the volume is in ml. The!:<:; .tJe four dif-
ferent aggregate specific gravities used for HMA based on the method used
to define the volume of the aggregate particles:
1. Apparent specific gravity;

2. Bulk specific gravity;

3. Effective specific gravity; and
4. Bulk impregnated specific gravity.
Referring to Figure 3-9, the apparent specific gravity includes only the
volume of the aggregate particle. It does not include the volume of any
pores or capillaries that become filled with water after a 24-hour soaking.
The bulk specific gravity includes the overall volume of the aggregate par-
ticle, as well as the volume of the pores that become filled with water after
a 24-hour soaking. The effective and bulk-impregnated specific gravity
includes the overall volume of the· aggregate, plus the pores that become
filled with water after a 24-hour soaking, min4~ the volume of the larger
pores that absorb asphalt. The volume of solids ~) contains internal pores
that are impermeable to both water and asphalt. Since the volume of the
impermeable pores cannot be accurately determined, for all practical pur-
poses they are considered as part of the volume of the solids. Reducing the
aggregate size by crushing exposes some of these internal voids and thus
increases the apparent specific gravity of the aggregate.
Air voids in the compacted HMA appears throughout the mix as small
pockets of air between the asphalt coated aggregate particles. The choice of
an aggregate specific gravity to be used in asphalt mix calculations can have
a substantial effect upon the calculated amount of air voids in the com-
pacted HMA. The effective specific gravity of the aggregate in the mixture
depe1J.ds upon the degree to which the aggregate absorbs asphalt. When the
apparent specific gravity is used in the calculations, the voids between the
asphalt filrp. and the aggregate (pores not filled with asphalt) are included
in the calc'½llated voids. If bulk specific gravity is used, it is assumed that no
asphalt is absorbed by any of the water permeable pores; hence, the mea-
sured voids include the voids between the aggregate coated particles minus
the volume of the absorbed asphalt. Except in rare cases, neither the bulk
or apparent specific gravity provides a correct measure of air voids in the
HMA mixture. The concept of effective specific gravity more nearly
describes the true case for asphalt absorption when calculating the voids in
a compacted asphalt-paving mix. In this case, the calculated voids are truly
those voids between the asphalt coated particles and do not include the
----vo_1._,.ds in ti:ie pores between--rhe-asphal-r-a-nd-aggrega:1:e:-T-he-e:ff-e-c-t-ive-speci.f-~-----
ic gravity is equal to or higher than the bulk specific gravity and equal to
or lower than the apparent specific gravity. When the aggregate absorption
is zero, all four types of aggregate specific gravity are theoretically equal.
The bulk and apparent specific gravities of coarse and fine aggregates
can be determined by ASTM C127 and Cl28, respectively. The effective

specific gravity can be calculated from the theoretical maximum specific
gravity (Rice specific gravity) using ASTM D2041. The bulk-impregnat-
ed specific gravity is used by the U.S. Corps of Engineers to design and
control HMA when using aggregates with a water absorption greater than
2 1/2 percent. The bulk-impregnated specific gravity is similar to the effec-
tive specific gravity in concept but involves immersion of aggregate in
asphalt, whereas the Rice specific gravity is obtained by testing the actual
HMA mixture. A further discussion on computation of voids is presented
in Chapter 4.
Specific Gravity and Absorption of Coarse Aggregate. Equipment
and procedures for determining the specific gravity and absorption of
coarse aggregates are outlined in AASHTO T85 and ASTM C127. The
method is briefly outlined below:
Oven-dry weight ot aggregate = W 3

s
Unit weight of water, "w" I g/cm
IF or simplicity, metric measurements are typically used in these calculations.)
Vs = Volume of solids
Vpp = Volume of water permeable pores
Vop = Volume cif pores absorbing asphalt
Vpp - Vop = Volume of waler permeable pores not absorbing asphalt
Ws
Apparent specific gravity
Bulk specific gravity
Effective specific gravity = Gse =

IVs+ Vpp-Vapl"w
Figure 3-9. Relationship Between the Different Specific Gravities of

an Aggregate Particle
Approximately 5 kg of thoroughly_ washed aggregate retained on a No. 4

(4.75 mm) sieve is oven-dried to a constant weight. The dried sample is
then immersed in water for 24 hours. The aggregate is removed from the
water, drained, and surface-dried until all visible films of water are
removed. The surfaces will still appear damp. The weight of the sample in
the surface-dry condition is then obtained and recorded as B. The saturat-
ed surface-dry sample is placed in a wire basket, submerged in water, and
the submerged weight determined and recorded as C. The sample is,then
removed from the water, drained and placed in an oven and dried to a con-
stant weight. The oven-dried weight is recorded as A. The specific gravity
and absorption are calculated as follows: ·
Apparent specific graviry, Gsa = A (1)

A-C
Bulk specific graviry, Gsb = ~ (2)

B-C.
Absorption, % = (B-Ai 1OO (3)
where,
A= oven-dried weight of aggregate, g;
B= saturated surface-dry (SSD) weight of aggregate, g; and
C= submerged weight of aggregate in water, g.
Specific Gravity and Absorption of Fine Aggregate. Equipment and

procedures for determining the specific gravity and absorption of fine
aggregates are described in MSHTO T84 and ASTM Cl28. The proce-
dure is briefly outlined below:
I
T~e flask (pycnometer) to be used to measure specific gravity is filled
,ii with water and the weight recorded as B. Approximately 1,000 g of fine
'I
aggregate is oven-dried to a constant weight. The dried sample is then
I immersed in water for 24 hours. The fine aggregate is spread on a dean
flat surface and exposed to a gently movmg current of warm air until a , - - - - -
saturated surface dry condition is achieved. A saturated surface-dry
condition is reached at the moisture content at which the lightly com-
pacted material in a cone first slumps when the cone is removed. The
aggregate has some cohesion as long as there is surface moisture but has
no cohesion when the surface moisture evaporates; hence, the fine

aggregate slumps when the cone is removed. The cone test must be
repeated several times to ensure that the highest moisture content (sat-
uration) at which the aggregate will slump is determined.
Approximately a 500 g sample of the saturated surface dry material is
placed in the flask and the actual weight of the sample recorded.as D.
The flask is then filled with water, using a specified procedure, and the
weight recorded as C. The fine aggregate is removed from the flask,
oven-dried to a constant weight, and the weight recorded as A. Specific
gravity and absorption are calculated as follows:
Apparent specific gravit:y, Gsa = B+f- C

(4)
Bulk specific gravit:y, Gsb = A (5)

B+D-C
Absorption, % = D - A x 100 (6)
A
where,
A = weight of oven-dry sample, g;
B = weight of flask (pycnometer) filled with water, g;
C = weight of flask (pycnometer) with specimen and water to
calibration mark, g; and
D = saturated surface dry weight (500 + 10 grams).
When the sample is tested in separate size fractions (for instance,

coarse and fine), the average specific gravity value can be computed as the
weighted average value of the various fractions using the following equa-
tion:
G =
(7)
where,
G = average specific gravity;
- - - - - - - - ~ G1-,-G.z.,---¥--5:P.ecific_gJ:aY.icy. __v.aluedorfraction 1, 2, ... n; and
P 1, P2 , ••• Pn= weight percentages of fraction I, 2, ... n.
This same equation can be used to calculate the average specific grav-
ity of an aggregate blend in which case G 1, G 2 , •.. Gn are the specific gravi-
ties of each of the aggregates I, 2, ... n and P1, P2 , ••• Pn are the weight per-
centages of each aggregate 1,2, ... n in the blend. This equation can be used
to determine average bulk specific gravity or average apparent specific grav-
ity, but the specific gravity type must be the same for all aggregates being
blended.
Example 3-1. The apparent specific gravity of an aggregate is 2.750

and the water absorption is 2.0 percent. Determine the bulk specific grav-
ity.
Recall that
G = dry weight of aggregate

• volume of aggregate
when the metric system is used and
G = dry weight of aggregate

"" ·volume of aggregate & external pores
where,·
G,a = apparent specific gravity; and
G,b = bulk specific gravity.
The volume of the external pores is equal to the volume of absorbed water
when the aggregate is in SSD condition. Assume for this example that the
weight of dry aggregate is I 00 grams. Then ·
volume of water (ml) = weight of water (grams)
water content= weight of water

dry weight of solids
The water content in the saturated surface dry condition is 2.0 percent as
_ ___,given in the problem. The aggregate must be in the SSD condition to mea-
sure all of the external pores
weight of water in SSD condition= 0.02 (100) = 2 grams

volume of water in SSD condition = 2 ml, since 1 gram of water
is 1 ml volume
Hence, the volume of external voids is 2 ml. Therefore
volume of aggregate- dry weight of aggregate= 100 =36.364m/

. G.,, 2.750
It then follows that

G _ 100 =____!QQ_
.- 36.364 + 2.000 38.364
G.=2.607
Example 3-2. Two aggregates will be combined in equal proportions

by weight to meet a HMA specification. The bulk specific gravity of aggre-
gate #1 is 2.500, and aggregate #2 is 3.000. Determine the bulk specific
gravity of the blend.
This problem is worked two ways. The equation for determining the aver-
age specific gravity of a blend of aggregate is:
G= P1+Pz
Pi Pz
-+-
Gl Gz
Substituting into the equation above
G =0.500 + 0.500
0.500 + 0.500
2.500 3.000
Simplifying
1 1
G=----=--
0.200 + .167 .367
------Hefl:€e--------------
G = 2.725
The problem will now be worked by simply dealing with weights and
volumes. The two aggregates are combined by equal proportions by
weight. Assume the weight of aggregate # 1 is 100 grams. Hence, the
weight of aggregate #2 is also 100 grams.
The volume of aggregate #l is determined as shown below:
.~ ~• weight of aggregate 1
volume o, aggreg...e 1 =
specific gravity of aggregate 1
100
=--
2.500
= 40.000
The volume of aggregate #2 is determined in a similar manner:
volume of aggregate 2. = weight of aggregate 2

specific gravity of aggregate 2
100
=--
3.000
= 33.333
Hence, the total weight of the blended aggregate is 200 grams and the total
volume of the blended aggregate is 73.333 ml. It follows that
G . = weight = 200
.sb volume 73.333
and
The slight difference in answers is caused by rounding errors. Notice that

averaging the two specific gravities does not provide the correct answer.
The average of the two specific gravities is 2.75 which is not the correct
specific gravity of the blend.
Size and Gradation
Desired Gradation Curves. Aggregate gradation is the distribution of

particle sizes expressed as a percent of the total weight. The gradation as a
percent of the total volume is of most importance, but expressing grada-
tion as a percent by weight is much easier and is standard practice. The gra-
dation by volume and weight is approximately the same as long as the spe-
cific gravities of the various aggregates being used are approximately equal ..
If there are large differences in the specific gravity of aggregates being used
for a particular mix, then the gradation should be determined as a percent
of total volume. Gradation is determined by sieve analysis, that is, by pass-
ing the material through.a series of sieves stacked with progressively small-
er openings from top to bottom, and weighing the material retained on
each sieve. The gradation of an aggregate is normally expressed as total per-
cent passing various sieve sizes.
Gradation of an aggregate can be graphically represented by a grada-
tion curve for which the ordinate is the total percent by weight passing a
given size on an arithmetic scale, while the abscissa is the particle size plot-
ted to a logarithmic scale. Sieves typically used for sieve analysis and gra-
. dation specifications for HMA are: 2 inches, 1 1/2 inches, 1 inch, 3/4 inch,
1/2 inch, 3/8 inch, No. 4, No. 8, No. 16, No. 30, No. 50, No. 100, and
No. 200 (50.8 mm, 38 mm, 25.4 mm, 19 mm, 12.5 mm, 9.5 mm, 4.75
mm, 2.36 mm, 1.18 mm, 0.6 mm, 0.3 mm, 0.15 mm, and 0.075 mm,
respectively). A 3/8-inch sieve has openings equal to 3/8 inch. A No. 8
sieve has 8 openings per inch. The openings in a No. 8 sieve are smaller
than 1/8 inch since the diameter of the wire must be considered in deter-
mining the opening size. The sizes of successive sieves usually differ by a
factor of approximately 2; therefore, when plotted on a logarithmic scale,
the distances between adjacent sieve sizes are usually about equal.
Aggregate gradations are sometimes described as dense or well-graded,
uniformly-graded (open), and gap-graded as shown in Figure 3-10.
Gradation is perhaps the most important property of an aggregate. It
affects almost all the important properties of a HMA, including stiffness,
----~tabili-ty,-d1:1:r~Hty,per-ffl:eabi-l-iry,we-r-k-abi!-i:ty,f-atigue-reslstance, frictional
resistance, and resistance to moisture damage. Therefore, gradation is a pri-
mary consideration in asphalt mix design, and the specifications used by
most states place limits on the aggregate gradations that can be used in
HMA.
Theoretically, it would seem reasonable that the best gradation for

HMA is one that gives the densest particle packing. The gradation having
maximum density provides increased stability through increased interpar-
ticle contacts and reduced voids in the mineral aggregate. However, there
must be sufficient air void space to permit enough asphalt cement to be
incorporated to ensure durability, while still leaving some air space in the
mixture to avoid bleeding and/ or rutting. A tightly packed aggregate (low
voids in mineral aggregate) also results in a mixture that is more sensitive
to slight changes in asphalt content.
Numerous investigators have proposed ideal gradations for maximum
density. One of the best known of these gradations is Fuller's curve pro-
posed by Fuller and Thompson (J.5). The equation for Fuller's maximum
density curve is:
P = 100 (d/D) 0 (8)
where d is the diameter of the sieve size in question, P is the total percent
passing or finer than the sieve, and D is the maximum size of the aggre-
gate. Studies by Fuller and Thompson showed that a maximum density can
be obtained for an aggregate when n=0.5.
200
100
100 00 30 ,. ." I
,,
I
,-- V
~
I I 90
90
~
,
/
I I :
I :
80
~ , II ;
I: I I :
70 70
I
/ :
"'ti I / I !
~ so~
0. I / I j
...z ,_ __ , I/
/"'4 !'--Uniformly j
~
a::
,0 Gop-groded> _
/ I
graded
""
-- --
UJ
0.
;l 40~
/ I
... ,,v I
I
'
4
...
0
I / I
3o~ I : 3
Well-graded or / I
20
10
~
~ I
I Dense grad~~
-~ I/
.I
I
I
:
20
0~
200
----
SIEVE NO.
100 00 30 ,. . ~,-:-
i" t" ¼" It"
SQUARE OPENINGS
0
Figure 3-10. Typical Terms Used to Identify Aggregate Gradations

Chapter_ 3-Aggregates 155
In the early 1960s, the Federal Highway Administration introduced an

aggregate grading chart which is based on the Fuller gradation but uses a
0.45 exponent in the equation. This chart is very convenient for deter-
mining the maximum density line and for adjusting aggregate gradings.
Using this chart, the maximum density lines can be easily obtained by
drawing a straight line from the origin at the lower left of the chart to the
actual percentage point of the nominal maximum size. The nominal max-
imum size is defined as the largest sieve size in the specification upon
which any material is retained. Figure 3-11 shows a family of maximum
density lines plotted on this chart. FHWA recommends this chart be used
as part of the HMA design process.
Gradations of maximurri density may not provide sufficient voids in
the aggregate for enough asphalt cement to provide adequate film thickness
for maximum durability without bleeding. In such cases deviations from
the maximum density curves are necessary in order to increase the total
voids in the mineral aggregate (VMA). Minimum VMA requirements have
been suggested an:d vary with the maximum nominal aggregate size (see
Chapter 4). Most asphalt paving agencies prefer that the gradation be
approximately parallel to the maximum density grading, but that it be off-
set a few percentage points either above or below the line. The two lines
will intersect at the nominal maximum size and be a few percentage points
different on the middle sieves. The requirement for minimum VMA is
necessary to ensure that there are sufficient voids in the aggregate to allow
enough asphalt cement to be added to provide a durable mix and sufficient
air voids to maintain stability. The Superpave mix design developed by the
Strategic Highway Research Program (SHRP) requires a selected number
of control points on the gradation chart (see Chapter 4).
Maximum Aggregate Size. The maximum particle size in a mixture is

important to ensure good performance. If the maximum particle size is too
small, the mix may be unstable; if it is too large, workability and segrega-
- - - - - ~ ·on..ma.y-6.e_p.r.oblemsJhere_ar_e__tw._o_designations for maximum particle
size (ASTM C125):
1. Maximum size, designated as the smallest sieve through which 100

percent of the aggregate sample particles pass; and
2. Nominal maximum size, designated as the largest sieve that retains

some of the aggregate particles, but generally not more than 10 per-
cent.
FEDERAL HIGHWAY ADMINISTRATION 0.45 POWER GRADATION CHART

SIEVE SIZES RAISED TO 0.45 POWER
100
100
.., 00
00 00
"' "'
00
,., ,.,
..,
:,0 :,0
zo zo
10 10
0
o
••
No.200JOO .50 JO ic 11211
.
'"
SIEVE SIZES
Figure 3-11. Maximum Density Lines for FHWA 0.45 Power

Gradation Chart
Mix designations used in specifications typically use the nominal max-

imum -size of aggregate. For example, in ASTM specification D3515, a
blend having nominal maximum aggregate size of 1 inch (25.4 mm) is
required to have 100 percent of the material passing the 1 1/2-inch
(38 mm) sieve and 90-100 percent passing the 1-inch sieve (25.4 mm).
Superpave mix design system uses the following aggregate size defini-
tions:
Maximum Size: One sieve size larger than the nominal

. .
max1mum size.
Nominal Maximum Size: One sieve size larger than the first sieve to
retain more than 10 percent.
- --- -------------------------------------
The maximum density line is obtained in Superpave by connecting the
origin at the lower left of the 0.45 power gradation chart to the maximum
aggregate size at the upper right of the chart (1§).
From a construction standpoint, the maximum aggregate size is nor-
mally limited to about one-half of the lift thickness. In recent years, there
has been an increase in the use of large stone mixes to minimize rutting
potential of HMA. The introduction oflarger stone sizes increases the vol-
ume concentration of the aggregate. This factor contributes to a reduction
of both design asphalt content and cost of the mix. These large stone mix-
tures are more resistant to rutting than the smaller aggregate size mixtures.
However, the use of a maximum aggregate size greater than 1 inch (25.4
mm) often results in harsh mixes that tend to segregate during placement.
Therefore, special attention must be given to mix design, mat thickness,
material handling, mixing, and paving procedures when these larger max-
imum aggregate size mixtures are used.
Typical Specification Bands and Tolerances. Aggregate gradation

specifications for HMA have been developed through accumulated field
experiences. In many cases they are established by trial and error to reflect
local conditions. Most specifications for HMA require well or dense-grad-
ed aggregate gradations with the middle portion of the curves approxi-
mately parallel to the maximum density curves. Table 3-7 contains the gra-
dation limits recommended in ASTM D 3515 and Table 3-8 gives the tol-
erances. In practice, most state specifications do not allow as much varia-
tion in gradation as that shown in Tables 3-7 and 3-8. The gradation band
from ASTM D3515 for 3/4-inch (19 mm) nominal maximum size mix-
ture is plotted on a FHWA gradation chart in Figure 3-12. For the most
part, the maximum density curve for a 3/4-inch (19 mm) maximum size
aggregate fits inside these limits. Some guidance for developing a gradation
band and potential problem areas are shown in Figure 3-13 UZ).
Another problem often caused by poor aggregate gradation is tender
mixes. These mixtures cannot be compacted in the normal manner because
they are slow in developing sufficient stability to withstand the weight of
the compaction equipment. Figure 3-14 illustrates a typical gradation pat-
tern for these mixes. Most notable is the hump in the curve near the No.
40 sieve and the flat slope between the No. 40 and the No. 8 sieves. The
most likely cause of an aggregate blend having this shape curve is excessive
use of poorly-graded natural sands. There are other causes of tender mix-
tures which are discussed in Chapter 6.
Sieve Analysis (ASTM C136 and Cl 17). The gradation of an aggre-

_ _ _ ____,gate is determined by a sieve anylysis. Standard procedures for a dry sieve
analysis are given in ASTM C136, and for a washed-sieve analysis for
determining the amount of material passing the No. 200 (75 µm) sieve,
the procedures are in ASTM C 117. The washed sieve analysis is a more
accurate measure of the true gradation, but the dry method is faster and is
.often used to estimate the actual gradation. When the dry method is used,
the measured amount of material passing the No. 200 (75 µm) sieve is
most likely significantly lower than the actual amount in the aggregate
mixture. For clean aggregate, a dry sieve analysis may be accurate enough
since the amount of material passing the No. 200 (75 µm) sieve is low.
Example 3-3. Assume that an aggregate blend was split into two sam-
ples. Determine the percent passing each sieve size using the dry and
washed gradation methods. A dry gradation was conducted on the first
sample which weighed 2012 grams (Table 3-9). A washed gradation was
performed on the second sample which weighed 2074 grams. The weights
and calculations for the washed gradation shown in Table 3-10.
The steps in a dry sieve analysis include splitting the sample to the cor-
rect size, determining the total dry weight, pouring it over a set of sieves
and shaking, weighing the material collected on each sieve size, determin-
ing the percent retained on each sieve, and determining the percent pass-
ing each sieve. When a washed sieve analysis is conducted to obtain the
washed gradation, the material is washed over a No. 200 (75 µm) sieve to
remove the dust. The material is then dried, weighed, poured over a nest
of sieves, vibrated, and the amount retained on each sieve determined.
0.45 Power Gradation Chart
1:11
C
j sot---+-+-+--+--+--+----r--1---.--+.+'-----+----leo
:.
i
i 40
-- -·---··--·---·- '~.'iot~;;o--.~oi=:f..re==i's=====•~==::;3;;';;,.-::;;;,====a..l,•.::;;_,~:::..=--=-:i,_g__________
Sieve Sizes
Figure 3-12. ASTM D 3515 Specification for 3/4-inch Nominal

Maximum Size Aggregate Plotted on the 0.45 Power
Chart
~
()
::,-
Dense Mixtures -
CT'
n
~
~
&
w
Sieve S ,:e Mix Designation and Nominal Maximum Sire or Aggregate
I
>I k~
2 inch 1½ inch 1 inch %Inch ½ Inch "Inch
No.4
(4.75 mm) No.8
No.16
(1.18mm) ~ ere,
~
~
(50 mm) (37.5 mm) (25.0 mm) (19.0 mm) (12.5 mm) (9.5 mm) (Sand Asphalt) (2.36 mm) (Sheet Asphalt)
""'
Grading or Total Aggregate (Coarse Plus Fine Plus Filler IC Required) t,
Amounts Finer than Each Laboratory Sieve (Square Opening), Percent by Weight ~
2½ inch (63 inm) 100 - - - - - - - - ....

VI
VI
2 inch (50 ~m) 90 to 100 100 - - - - - - - 6;l
1½ inch (37 5 mm) - 90 to 100 100 - - - - - - s..
1 inch (25.0 mm) 60 to 80 - 90 to 100 100 - - - - - a.
0
\ inch (19.~ mm) - 56 to 80 - 90 to 100 100 - - - - ::s
I
½ inch (12.~ mm) 35 to 65 - 56 to 80 - 90 to 100 100 - - - ~
n
I
% inch (9.5 imm) - - - 56 to 80 - 90 to 100 100· - -
....n::r,
! n
No. 4 (4.75 :mm) 17 to 47 23 to 53 29 to 59 35 to 65 44 to 74 55 to 85 80 to 100 - 100 ....0~
No. 8 (2.36 mm) 10 to 36 15 to 41 19 to 45 23 to 49 28 to 58 32 to 67 65 to 100 - 95 to 100 ::s
No. 16 (1.18 mm) - - - - - - 40 to 80 - 85 to 100 ,.,g-i
No. 30 (600 -m) - - - - - - 25 to 65 - 70 to 95 t,
No. 50 (300 -m) 3 to 15 4 to 16 5 to 17 5 to 19 5 to 21 7 to 23 7 to 40 - 45 to 75 e
n
- - - - - - -
a·~
No. 100 (150 -m) 3 to 20 20 to 40
No. 200 (75 -m) Oto 5 0106 1 to 7 2 to 8 2 to 10 2 to 10 2 to 10 - 9 to 20
~ ....01
\C)
"'
Sieve opening, In •
~g ......
.,,.,
"'"'0
q
00
00
t-Ol
qq
100°
00 00
I J I/
90
80
:~~::o~e~ol: tw:lily lo produce, I I I
,/I I
-- - IP --
g' 70
-.; -'V /
i 60
Crl!lcal m!xlures ~ become V
~ 50 readily unstabla with silght
------
excess of asphalt or waler / V
.~ 40
a.
30 --
L--- i---
~ .4"ar~h mbcfw;e Inclined lo
segregate, nol critical
20 '---;;;, --- ~
i:;.. i:::-
, L I I
Porous· mixture, ~: <lack te11slle strength
when mixed with fuel oil, best
I I
10
r:-- , r■sull1 with heavy bilu~n I
0
200 100 80 50 40 30 20 16 108
Sieve number
Figure 3-13. Gradation Limits for HMA and Expected Problems

when Outside the Limits (12)
Aggregate Blending to Meet Specifications. For various reasons,

mostly associated with achieving maximum density and desired void prop-
erties, certain desirable gradation limits are usually required of aggregates
for HMA. Because it is unlikely that a single natural or quarried material
will meet these specifications, two or more aggregates of different grada-
tions are typically blended to meet specifications limits. Aggregates are also
separated into sizes to improve handling characteristics. Mixing of coarse
and fine aggregate in one stockpile results in segregation; hence, aggregates
should be separated into sizes, for example, }/4 inch-3/8 inch
Table 3-8. ASTM D 3515 Gradation Tolerances

Sieve Size Tolerances, %
1/2 inch (12.5 mm) and larger .± 8
3/8 inch (9.5 mm) + 7
No. 4 (4.75 mm) + 7
No. 8 (2.36 mm) +6
No. 16 (1.18 mm) .± 6
'
No. 30 (600 µm) .± 5
~50-(3U0 µmJ ·- +:,
No. 200 (75 µm) .± 3
Bitumen content, weight % of total mixture .± 0.5
Table 3-9. Computation of Dry Aggregate Gradation
Sieve Size Weight Retained Percent Retained Percent Passing

3/4 inch 0.0 0.0 100.0
1/2 inch 86.5 4.3 95.7
3/8 inch 154.9 7.7 88.0
No.4 253.5 12.6 75.4
No. 8 269.6 13.4 62.0
No. 16 344.1 17.1 44.9
No. 30 380.3 18.9 26.0
No. 50 323.9 16.1 9.9
No. 100 158.9 7.9 2.0
No. 200 22.1 1.1 0.9
Pan 18.2 0.9
Total 2012.0 100.0
(19 mm-9.5 mm), 3/8 inch-No. 4 (9.5 mm-4.75 mm), and minus 4
(4.75 mm) prior to hauling and stockpiling. Another reason for blending
aggregates is that it is often more economical to combine naturally occur-
ring and processed materials to meet specifications than to use all processed
materials.
0.45 Power Gradation Chart
80
f
Ill
60
...
a.
e
C
40
:.
0~__,_....._.,_..,__,__ __,__ __.__ _ __.__ _ _ _....,__ ____.0

0 No. 200 JOO ~0 '30 16 8 4 3/8 In. 314 In. I In.
40
Sieve Sizes
Figure 3-14. Characteristics of a Tender Mix on the 0.45 Power
Gradation Chart
Table 3-10. Computation ofWashed Aggregate Gradation
Sieve Size Weight Retained Percent Retained Percent Passing
3/4 inch 0 0 100
1/2 inch 89.2 4.3 95.7

3/8 inch 159.7 7.7 88.0
No.4 238.5 11.5 76.5
No. 8 311.l 15.0 61.5
No.16 300.7 14.5 47.0
No. 30 398.2 19.2 27.8
No. SO 327.7 15.8 12.0
No.100 149.3 7.2 ,_ 4.8
No. 200 16.6 0.8 4.0
Pan 24.9 1.2
Washed Out 58.1 2.8
Total 2074.0 100.0
The nature of particle size distribution can be examined by graphical-

ly representing the gradation by (a) a cumulative percent passing on a semi-
log scale, or (b) the cumulative percent passing versus the sieve sizes raised
to the 0.45 power. Both methods are commonly used by engineers to help
select the best aggregate blend.
If the specifications are expressed in terms of the total percent passing
each sieve, they can be plotted as bands or envelopes (Figure 3-15). A
method of transforming a passing-retained specification to an approximate
equivalent total percent passing specification was described by Dalhouse
(1..i). The transformation enables the passing-retained specification to be
plotted on the, total percent passing chart, making visual examinations and
comparisons of different specifications possible. However, in this chapter
all gradations and specifications are expressed on a total percent passing
basis, unless otherwise stated.
A large number of blending methods (techniques of determining rela-
tive proportions of various aggregates to obtain a desired gradation) have
---b~~.0.-ci~v~l0p@El-s1-a-G@--ffl{l-SUgg~s1=i-0:0.-0f-1=he-m~-murn-cl:ensi-ey-·e1:1.rVes-by
~uller and Thompson (15). The suitability of these methods depends on
the types of specification and number of aggregates involved, the experi-
ence of the individual, and the major emphasis of the blending (closeness
to the desired gradation or economics).
Regardless of the number of aggregates or blending techniques select-

ed, the basic formula expressing the combination is
p = Aa + Bb + Cc + ... (9)
where,
p = the percent of material passing a given sieve for the
combined aggregates A, B, C, .....
A, B, C, ... = the percent of material passing a given sieve for each
aggregate A, B, C, ...
a, b, c, ... = proportions (decimal fractions) of aggregates A, B,
C, ... to be used in the blend, a+ b + c... = 1.00
It is desirable, no matter which method is used, to first plot the gra-

dations of the aggregates to be blended and the specification limits on a
gradation chart before actual blending is attempted. From these plots, deci-
sions can be made prior to any calculation on (a) whether a blend can be
found using the available aggregates to meet the specification limits, (b)
where the critical sieve sizes are, an (c) the approximate trial proportions to
be selected. These decisions can be made based on the following simple
facts and the gradations shown in Figure 3-15:
100
I
90 I
I
I
80
t!)
z 70
in
(J)
< 60
11..
I
E
I-
z ✓
ILi ✓
u
a:;
✓
✓
ILi
11.. 30 /A
✓
--- --
_../
10
o._..=-=;:;;;;L~--1L...-~....1....---1..--1_---1...J.....----1..--L..---'--__.__ __,
200 100 80 50 40 30 20 16 108 4 3/8 3/4 1-1/2 3 In.
SIEVE NUMBER
- -0 ~Jt---""-l-49---0.42-0-CUl~l-----2..00-4~7..6--L2.7 - . 25.4 76.1
·'\"T-"'• I I I I I I I I
SIEVE OPENING, mm
Figure 3-15. Examination of Aggregate Gradings with Specification

Band
1. The gradation curves for all possible combinations of aggregates A

and B fall between curves A and B. It is impossible to blend aggre-
gates C and B to meet the specification regardless of the method
used.
2. If two curves cross at any point (Band D), the grading curves for
all possible combinations pass through that point.
3. The curve for a blend containing more of aggregate A than B is clos-
er to curve A than B and vice versa.
The most common method of determining the proportions of aggre-

gate to use to meet specification requirements is the trial-and-error
method. As the name implies, a trial blend is selected (aided by experience
and plots of individual gradation curves and 'specification limits) and cal-
culations made using Equation 9 to determine the percent passing each
sieve size for the blend. The grading that is calculated from this trial is
compared with the specification requirements. Adjustments are made for
the second trial blend and the calculations repeated for the critical sieves
until a satisfactory or optimum blend is obtained. This method, guided by
a certain amount of reasoning, mathematics, and experience, is the most
widely used method and is the easiest procedure to determine a satisfacto-
ry blend.
The trial-and-error method involves the following steps:
1. Selecting critical sieves for the aggregates in the blend;

2. Determining an initial set of proportions a, b, c, etc., .which will
meet the specification requirements for the critical sieves;
3. Checking the calculated blend using the proportions determined
for all sieves in the specification requirements; and
4. Adjusting the proportions, as necessary, to ensure that the percent-
ages for all sieves are within specification limits.
a
Example 3-4. Determine blend of the two aggregates shown in
Table 3-11 which will meet specification requirements using the
trial-and-error method.
Examination of the grading indicates that it is possible to find a blend
that falls within the specification limits, possibly a 50-50 blend because of
the relative distance of the curves to the center of the band (desired grada-
----=t1:on)-:-Tlie1irstffial blend can be determmed more mtelhgently 1f certam
critical sieve sizes are selected. By inspecting the gradations, it is seen that
all material retained on the 3/8-inch or 9.5 mm sieve (100-80 = 20 per-
cent desired) has to come from aggregate A, and all material finer than the
No. 30 (0.6 mm) sieve must be furnished by aggregate B. With regard to
~ Q
~
~ ~
w
'f,_. i
,_.
Sieve Size
r$" I "'
('Jq
Aggregate Percent 3/4 inch 1/2 inch 3/8 inch No.4 No.8 No.30 No.50 No.100 No. 200 0
'"1-'j
Used
Aggregate A 100 100 90 59 16 3 0 0 0 0 ~
0
Aggregate B
50 percent A
50 percent B
100
50
50
100
50
50
100
45
50
100
29.5
50
96
8
48
82
1.5
41
51
0
25.5
36
0
18
21
0
10.5
9
0
4.5
f~
~
Blend 100 100 95 79.5 56 42.5 25.5 18 10.5 4.5
Desired 100 100 90 80 60 43 24 18 12 7
Specification 100 80-100 70-90 . 50-70 35-50 18-29 13-23 8-16 4-10
,_.
8:
aggregate A, because 100-59 = 41 percent is retained on the 3/8-inch

(9.5 mm) sieve from A, the percentage needed from A to retain 20 percent
on this sieve (specification median) is A= 20/41 = 0.49; the percentage of
aggregate B will be 1-0.49 = 0.51. With regard to aggregate B, because
there is 51 percent passing the No. 30 (0.6 mm), the percentage of aggre-
gate B required to arrive at the desired 24 percent passing this sieve is B =
24/51 = 0.47; the percent of aggregate A for this blend will be 1-0.47 =
0.53. Hence, analysis of the two critical sieves produces very similar results.
Evaluation of one of the critical sieves shows that the percent of aggre-
gate A should be 49 percent and the other critical sieve shows that A should
be 53 percent. For the first trial use 50 percent of aggregate A and 50 per-
cent of aggregate B.
A form that is easy to use is shown in Table 3-11. Multiply 0.50 times
the percent passing each sieve size for aggregate A and aggregate B. Add the
two results to get the gradation of the blend. The resulting gradation is
compared to the desired one and it appears that the use of 50 percent A
and 50 percent B results in a gradation very close to the desired gradation.
Hence, for this example, the proportions should be 50 percent A and 50
percent B. The percentage of these two aggregates could be changed slight-
ly, if desired, without having an adverse effect on the blend. For instance,
if aggregate A costs more than aggregate B, the percentages could be
changed to slightly increase the amount of aggregate B and decrease the
amount of aggregate A, thus reducing the overall cost of the mix. The pro-
ject specification will still be met but with a less expensive mix.
Example 3-5. Given the gradat'.ion of aggregates A, B, and C, determine

the required percent of each to result in a blend meeting the required spec-
ification requirements (see Table J-12).
The critical sieves in this case would be the No. 4 (4.75 mm) and No.
30 (0.6 mm). The No. 200 (75 µm) sieve is also critical since aggregate C
has 88 percent passing the No. 200 (75 µm) sieve and the specifications
only allow up to 11 percent. The desired 52 percent of material larger than
the No. 4 (4.75 mm) sieve must come from aggregate A. Hence, the per-
cent of aggregate A to use for the first trial should be:
percent ofA =~ = 64%

81
Now looking at the No. 30 (0.6 mm) sieve it is apparent that 75 per-
cent of the material larger than a No. 30 (0.6 mm) sieve must come from
materials A and B. Since 75 percent is the total desired to be retained on
the No. 30 (0.6 mm) and 0.64(95) is that portion provided by aggregate A.
percent ofB = 75-0.64 (95)

= 75 -61
= 14
Based on these calculations, it appears that the best first estimate would be
64 percent of aggregate A, 14 percent of aggregate B, and 22 percent of
aggregate C. It is obvious, however, that using 22 percent of aggregate C
would produce too much material passing the No. 200 (75 µm) sieve. The
percent of aggregate C should probably be no more than 8 percent, since
that aggregate alone will produce approximately 7 percent passing the
Table 3-12. Blending of Three Aggregates
Aggregate
Median of
Sieve Size A B C Specification Specifications
1 inch 100 100 100 94 to 100 97
½ inch 63 100 100 70 to 85 78
No: 4 19 100 100 40 to 55 48
'No. 8 8 93 JOO 30 to 42 36
No. 30 5 55 100 20 to 30 25
No. 100 3 36 97 12 to 22 17
No. 200 2 3 88 5 to 11 8
No. 200 (75 µm) sieve. Based on this analysis, the first estimate should be
71 percent of aggregate A, 21 percent of aggregate B, and 8 percent of
aggregate C. The calculations are shown in Table 3-13.
The estimated percentages of 71 percent A, 21 percent B, and 8 per-
cent C produce a blend that meets the specification requirements.
However, the blend can be improved slightly to get it closer to the desired
gradation. To improve the blend, use 2 percent less of aggregate C and 5
percent less of aggregate A and, hence, 7 percent more of aggregate B. For
each 1 percent that two aggregates are changed the resulting blend will
change by the difference between the percent passing for any given sieve
size for the two aggregates times 1 percent. For instance, increasing aggre-
-----g~a=te-B-oy:, percen:-r-.rn-d~d-ec::re"lfsin:g-.rggreg:i:te-Aby 5 percent will result in
a change of (.81) 5 or 4.0 on the No. 4 sieve.
After making the calculations (Table 3-13) it can be seen that the
amount of each aggregate to use is 66 percent A, 28 percent B, and 6 per-
cent C.
The computations required by the trial-and-error method shown in

the example were carried out manually to show the steps. However, su.:'-
cessive trials can be done easily and quickly by using personal computers
and electronic spreadsheet software. If desired, the gradations of the blends
as well as the specification band can be plotted and examined using the
graphic capabilities of a typical personal computer.
Table 3-13. Calculations for Blending Three Aggregates
Sieve Size
Aggregate Percent 1 Inch 1/2 No.4 No.8 No.30 No.100 No.200
Used Inch
A 71 71 44.7 13.5 5:7 3.6 2.1 1.4
B 21 21 21 21 19.5 11.6 7.6 0.6
C 8 8 8 8 8 8 7.8 7.0
Blend 100 73.7 42.5 33.2 23.2 17.5 9.0
Desired 97 78 48 36 25 17 8.0
Specification 94-100 70-85 40-55 30-42 20-30 12-22 5-11
A 66 66 41.6 12.5 5.3 3.2 2.0 1.3
B 28 28 28 28 26.0 15.4 10.1 .08
C 6 6 6 6 6 6 5.8 5.3
'~
Blend 100 100 75.6 46.5 37.3 24.6 17.9 7.4
There are several graphical methods that have been used to determine
the percent of each aggregate to use in a blend. These methods are gener-
ally not used since the trial-and-error method is quicker and easier in most
cases. These methods are especially difficult to use when more than two
aggregates are being blended, which is normally the case for HMA. For
more information on graphical methods, additional references should be
reviewed (§, 12.,_ 21, 22). ·
Computer programs have been developed to determine the optimum
blend for a specified number of aggregates. None of these programs have
been very successful since an experienced engineer can determine the
desired proportion of up to four aggregates by the trial-and-error method
- - - - 5 y tne time ne could mput the data mto tlie computer. Another probli-::ec:::m=------
with a computer method is that it is difficult to quantitively describe the
optimum blend. For example, the significance of a 2 percent deviation
from the desired gradation is not as critical for a No. 4 (4.75 mm) sieve as
for a number 200 (75 µm) sieve. Many factors must be considered when
evaluating the importance of deviations from the desired blend. Some

engineering judgement must be used, and one program cannot be devel-
oped that includes all the factors that must be considered when blending
aggregates. For more information on computer methods, additional refer-
ences should be reviewed (2Q, 22, 22). The most common use of comput-
ers in aggregate blending is in the trial-and-error method (using the engi-
neer's judgement for the best blend) and in plotting the specifications and
trial blends. Computer programs can help to factor in cost considerations
(2d,2J,2QJ.
Surface Area
One important property that can be computed from the aggregate gra-
dation is the surface area. The aggregate surface area is important since it
affects 'the amount of asphalt needed to coat the aggregate. Dense-graded
asphalt mixtures are usually designed to contain a desired amount of air
voids; hence, the aggregate surface area is not a design factor. It is possible
to increase the surface area of an aggregate and at the same time reduce the
optimum asphalt content. One way to do this is by increasing the dust
content of a mixture. Asphalt mixes that have high surface area and low
optimum asphalt content are undesirable because these mixes will have a
thin asphalt film on the aggregate and will probably not have adequate
durability. A further discussion of asphalt film thickness is provided in
Chapter 4.
The most common method. used for estimating the aggregate surface
area involves calculations made from the aggregate gradation. The surface
area is determined by multiplying surface area factors by the percentage
passing the various sieve sizes. These factors can be calculated by assuming
an aggregate specific gravity and assuming all particles are rounded or cube
shaped. The surface area per pound of aggregate can then be determined
for each aggregate size.
The surface area factors for various sieve sizes are shown in Table 3-14.
The surface area is determined by multiplying the factor by the percentage
passing for each sieve size and totalling for all sieve sizes. The units for these
calculations are square feet per pound. All surface area factors for all sieve
sizes shown in Table 3-14 must be used. The surface area is only slightly
------affectecl: by the percent passmg the larger sieve sizes-oiifis significantly
affected by the percent passing the smaller sieve sizes.
One of the primary reasons for estimating the surface area is to deter-
mine the asphalt film thickness. This is an estimated value, but it does
allow comparisons to be made for various mixtures.
Table 3-14. Surface Area Factors

Sieve Size Surface Area Factors
Percent Passing Maximum Sieve Size 2

Percent Passing No. 4 2
Example 3-6. An aspl;talt mixture has an ~gregate gradation as shown

below. Determine the surface area per pound of aggregate.
Sieve Percent
Size Passing
3/4 inch 100
1/2 inch 88
3/8 inch 72
No.4" 60
No.8 47
No. 16 35
No. 30 25
No. 50 20
No. 100 10
No. 200 5
Using the surface area factors and the percent passing each sieve shown
above, the surface area (SA) is calculated below:
. . I
SA= 2(100) + 2(.60) + 4(.47) + 8(.35) + 14(.25) + 30(.20) + 60(.10)
+ 160(.05)
Simplifying the results,
SA = 31.38 square feet of surface area per pound of aggregate.
An additional discussion of aggregate is provided in the National Stone

---~As_s_o-c1_a_t1_on-Aggregate HandooolcQZ'J.
REFERENCES
1. A Handbook of Crushi~g, Bulletin 4030, Pennsylvania Crusher

Corporation, Broomall, Pennsylvania, 1974.
2. American Society for Testing and Materials, Volume 0402, ASTM C
294 Standard Descriptive Nomenclature of Constituents of Natural
Mineral Aggregates, 1989.
3. The Identification of Rock Types. Bureau of Public Roads, 1960.
4. State-of-the-Art: Effect of Water on Bitumen-Aggregate Mixtures.
Highway Research Board, Special Report 98, 1968.
5. Cordon, WA. Properties, Evaluation, and Control of Engineering
Materials, McGraw-Hill, N.Y., 1979.
6. McLaughlin, J .F. et al. Distribution, Production, and Engineering
Characteristics of Aggregates. Highway Engineering Handbook, ed. by
K.B. Woods, McGraw-Hill, N.Y., 1960.
7. Mertens, E.W and J.J. Borgfeldt. Cationic Asphalt Emulsions.
Bituminous Materials: Asphalts, Tars, and Pitches, ed. by A.J. Hoiberg,
Vol. 2, Part 1, Interscience Publishers, John Wiley & Sons, N.Y., 1965.
8. Petersen, J.C. Chemical Composition of Asphalt as Related to Asphalt
Durability: State-of-the-Art. Transportation Research Board,
Transportation Research Record 999, 1984.
9. Marek, C.R., H. Herrin, C.L. Kesler, and E.J. Barenberg. Promising
Replacements for Conventional Aggregates for Highway Use. Highway
Research Board, NCHRP Report 135, 1972.
10. McLeod, N.W and J.K. Davidson. Particle Index Evaluation of
Aggregates for Asphalt Paving Mixtures. Proceedings Association of
Asphalt Paving Technologists, Volume 50, 1981.
11. Asphalt Concrete Mix Design and Field Control. FHWA, Technical
AdvisoryT 5040.27, 1988.
12. Kandhal, P.S., M.A. Khatri, and J.B. Motter. Evaluation of Particle
Shape and Texture of Mineral Aggregates and Their Blends. Asphalt
PavingTechnology, Vol. 61, 1992.
13. Brown, E.R. and S.A. Cross. A National Study of Rutting in Asphalt
Pavement. Asphalt Paving Technology, Vol. 61, 1992.
14. Lee, D.Y. The Relationship Between Physical and Chemical Properties
-----~oLAggr.egates_and.TheiL.As.phalt...Ahs.o..rp.tio.n.J>r.oceedings.. Association
of Asphalt Paving Technologists, Volume 38, 19r69.
15. Fuller, WB. and S.E. Thompson. The Laws of Proportioning
Concrete. Journal of Transportation Division, American Society of
Civil Engineers, Vol. 59, 1907.
16. McGennis, R.B., R.M. Anderson, T.W Kennedy, and M. Solaimanian.

Background of Superpave Asphalt Mixture Design & Analysis.
FHWA, Report No. FHWA-SA-95-003, November 1994.
17. Hveem, F.N. Gradation of Mineral Aggregate for Dense-Graded
Bituminous Mixtures. Proceedings Association of Asphalt Paving
Technologists, Volume 11, 1940.
18. Dalhouse, J.B. Plotting Aggregate Gradation Specifications for
Bituminous Concrete. Public Roads, Vol. 27, No. 7, 1953.
19. Lee, D.Y. Review of Aggregate Blending Techniques. Highway
Research Board, Highway Research Record 441, 1973.
20. Lutter, B.E. and T.D. White. Computer Method for Aggregate
Blending. U.S. Army Engineers Waterways Experiment Station,
Instruction Report S-70-5, 1970, "-
21. Rothfuchs, G. Graphical Determination of the Proportioning of the
Various Aggregates Required to Produce a Mix of a Given Grading.
Betonstrasse, Vol. 14, No. 1, 1939.
22. Mackintosh, C.S. Blending of Aggregates for a Premix Carpet.
Transvaal Roads Department, Pretoria, 1959.
23. Nuemann, D.L. Mathematical Method for Blending Aggregates.
Journal of Construction Division, American Society of Civil
Engineers, Vol. 90, No. 2, 1964.
24. Ritter, J.B. and L,R. Shafter. Blending Natural Earth Deposits for
Least Cost. Journal of Construction Division, American Society of
Civil Engineers, Vol. '87, No. 1, 1961.
25. Easa, S.M. Trade-Off of Gradation and Cost Requirements in
Aggregate Blending. Cement, Concrete, and Aggregates, Vol. 7, No. 1,
1985.
26. Miller, L.J. and D.C. Bean. Selection of Aggregate Components by
Computer fo1_;! Optimum Mix Design. Proceedings Canadian Technical
Asphalt Association, Vol. 25, 1980.
27. Barksdale, R.D. (Editor). The Aggregate Handbook. National Stone
Association, Washington, DC, 1991.
CHAPTER·4. HOT MIXASPHALT
MIXTURE DESIGN METHODOLOGY
INTRODUCTION
In 1995, a few states began to use the Superpave design procedures for
a small number of projects. It is anticipated that within the next 2-5 years
most HMA projects will be designed using the Superpave concept.
Most HMA produced during the 50 years between the 1940s and mid
1990s were designed using the Marshall or Hveem methods. According to
a survey in 1984; approximately 75 percent of the state highway depart-
ments used some variation of the Marshall method while the remaining 25
percent used some variation of the Hveem method (1). Some states used
both methods ori the same materials to get a better understanding of the
characteristics of the mixtures. It is important to recognize that all three
methods are designed to help the asphalt technologist choose an appropri-
ate asphalt content with which to begin field construction. It is important
to understand that if the service conditions for which a mixture is being
designed are different from those for which the design method was devel-
oped, the mixture design may not be adequate for service even though it is
designed according to the method. It is incumbent upon the user of each
of these methods to understand the limitations of that design method to
ensure that the method is used appropriately. To enhance that understand-
ing, these three mixture design methods are discussed in detail in this chap-
ter. The design of only dense-graded HMA mixtures is discussed.
HISTORY OF MIX DESIGN _____ _
Crawford has traced the history of asphalt mix design dating back to
the 1860s and has given a very interesting account of mix design develop-
ment in Reference 2_.
Tar was used as a binder in the first bituminous pavements placed

mostly in Washington, D.C. in about 1868 and 1873. The importance of
aggregate proportioning was not understood at that time, and the mixing
process was not mechanized. These tar mixes did not fare well as surface
mixes. However, their performance as bases was satisfactory. Tar was also in
competition with Trinidad Lake asphalt during that period.
Clifford Richardson, considered as the great American asphalt tech-
nologist, obtained and analyzed the samples of those earlier bituminous
bases in 1906. He understood the importance of material selection, espe-
cially the significant role of fine aggregate fractions. He studied and docu-
mented the important principles ofHMA design including the voids in the
mineral aggregate (VMA) and air void content. Richardson published a
book, "The Modern Asphalt Pavement," in 1905 (3).
The Pat Test
Richardson described the "Pat Test" in his book. It can probably be

considered as the first mix design test to determine the optimum binder
content of a HMA mix. It was applied to sheet asphalt (a hot sand asphalt
mix) used as surface mix. Pat samples of sheet asphalt containing different
amounts of binder were pressed (compacted) against a brown manila
paper, and the residual stains on the paper were assessed visually. A heavy
stain indicated too much binder, a light stain indicated too little binder,
and a medium stain indicated "just right" or an optimum binder content.
Richardson provided comparative illustrations of these stains in his book.
The Pat Test was still being used in the early 1920s for fine-graded mixes
and for the fine aggregate portion of coarse mixes. Richardson recom-
mended a top size of about one inch or larger for the aggregate in
dense-graded HMA base mixes for heavy traffic. However, he used the
visual assessment to establish the optimum binder content.
Large Stone Mixes
Frederick Warren obtained a patent for the so-called Bitulithic pave-

ment at the turn of this century. The Bitulithic pavement allowed larger
aggregate size up to three inches. Bitulithic pavements containing 3/4 inch
(19 mm) and 11/4 inch (32 mm)__maximum-agg.i;~gaEcHiz-es-have-been-doc=-- - - -
umented in Rhode Island and Texas, respectively. Larger top size mixes
were placed in other states. These large stone mixes performed better than
thick sheet asphalt which had a tendency to rut when subjected to traffic
Chapter 4-Hot Mix Asphalt Mixture Design Methodology 175
consisting of slow-moving carts with narrow, steel-rimmed wheels or

solid-tired trucks in the 1900s.
The City of Topeka, Kansas, was sued by Warren Brothers in 1909 for
alleged patent infringements. This resulted in the "Topeka decree" of 1910
which, among other factors, stated that the Bitulithic patent would not be
infringed when the top size of the aggregate was less than one-half inch.
The "Topeka decree" encouraged an extensive use of fine-grained mixes
throughout the U.S.A. These mixes were easier to shovel and rake than the
large stone mixes, provided smoother surfaces, and were more suitable for
mechanized paving operations. Fortunately, these fine-grained mixes ben-
efited from the development and usage of pneumatic tires on heavy vehi-
cles. Pneumatic tires had significantly lower contact pressures than solid
tires and steel-rimmed wheels. Although the Bitulithic patent expired in
1920, the fine-grained mixes continued to be used extensively and per-
formed reasonably well. However, in recent years tire designs have
changed, tire pressures have increased significantly, and traffic load and
volume have also increased substantially. This has led to some premature
rutting of heavy duty pavements. Therefore, there is a trend to return to
the use of large stone and high stone content mixes to minimize rutting.
Dense Gradation
Roy Green, an associate professor at the Agricultural and Mechanical

College of Texas, now Texas A&M University, studied several Bitulithic
pavements in Texas, especially the fine aggregate fractions of Bitulithic
mixes. He developed a procedure for obtaining an ideal dense gradation for
these mixes. He simply connected the percent passing the No. 200
(0.075 mm) sieve with 100 percent passing the top size aggregate sieve on
a gradation chart and plotted the intermediate sieve positions on this
straight line. This procedure facilitated comparison of gradations of differ-
ent mixes. It is interesting to note that Green developed this method 42
years before Goode and Lufsey developed their 0.45 power gradation chart.
Hubbard-Field Method
A: test metnocl-fur determmmg the optimum aspli.a!Ccontent of sheet

asphalt surfaces and sand asphalt bases was devised by Hubbard and Field
in the middle 1920s. The test consisted of determining the maximum load
developed as a specimen 2 inches (50.8 mm) in diameter by 1 inch
(25.4 mm) high was forced through a 1.75-inch (44.4 mm) diameter stan-
dard orifice. The load was reported as the stability value. Stability numbers
corresponding to various asphalt contents were plotted and the optimum
binder content determined. This was probably the first attempt to quanti-
fy empirical mix stability values at various asphalt contents. The method
was widely accepted and survived in some states for a long time. The test
was modified in the mid 1950s to test six-inch (150 mm) diameter speci-
mens to accommodate mixes containing coarse aggregate up to 3/4 inch
(19 mm) maximum size. However, the modified version was used for a
very short time because the Marshall test had started to gain popularity
during that period.
Hveem Mix Design Method
Reference 1. contains a detailed account of the evolution of the Hveem

method. As a resident engineer in California, Francis Hveem began to
work with "oil mixes" during the late 1920s. Oil mixes, which were a com-
bination of fairly good quality gravel and slow-curing liquid asphalts, were
being used in California to obtain an intermediate _type surfacing for mod-
erate traffic conditions. When mechanical pavers were introduced in 1937,
it became possible to place mixes with heavier grades of asphalt cements.
Hveem noticed that there was a relationship between the gradation of the
mineral aggregate and the amount of oil required to maintain a consistent
color and appearance of the mix. He subsequently found a method for cal-
culating surface area of aggregate gradations developed by a Canadian engi-
neer, L.N. Edwards. Refinements in the method occurred as Hveem real-
ized that the film thickness on a particle decreased as the diameter of the
particle decreased. The kerosene equivalent test was developed to take into
account differences in oil requirements as the absorption and surface
roughness of aggregates varied.
Hveem realized that having the proper oil content did not guarantee
good performance relative to rutting; therefore, another test was needed to
evaluate the stability, or the ability to resist the shear forces applied by
wheel loads. This led to the development of the Hveem stabilonieter. A
specimen of 4-inch (101.6 mm) diameter and 2½-inch (63.5 mm) height
is subjected to a vertical load on the circular surface and the amount of ver-
tical load that is transmitted horizontally is measured. The circumferential
-----.p=e...:rimeter of th:e specimen is restrained by a neoprene diaphragm and 1s
surrounded by an oil reservoir to simulate field loading conditions.
Empirical stability numbers are obtained at various asphalt contents.
Specimens are prepared with a kneading compactor, which is a hydraulic
device that applies pressure to the specimen through a hydraulically oper-
ated tamper foot. The foot is raised after a specified pressure is sustained
by the specimen, the base rotates 1/6 of a revolution, the tamper foot is
automatically lowered, and the specified pressure is applied again around
the perimeter of the specimen. The area of the tamper foot is one-fourth
the cross sectional area of the specimen.
A second mechanical test device called a cohesiometer was developed
along with the Hveem stabilometer. It was designed to measure the cohe-
sive strength across the diameter of a compacted specimen on which the
stability test had already been conducted. The specimen is placed in the
cohesiometer, the specimen is secured, and the load is applied by lead shots
flowing from a reservoir into a bucket at the end of the moment arm
(details are given in ASTM D1560). Load is applied until the specimen
fails, and at that time the shot supply is automatically shut off. The total
applied force is determined and the cohesion calculated by a formula. This
test was aimed at measuring a tensile property of the oil mixes that could
be related to a minimum level to preclude ravelling of surface mixes under
tractive forces. This test proved to be of little value in characterizing HMA
surfacings. HMA surfacings were made with asphalt cement and always
had cohesion values large enough to prevent ravelling. Therefore, when oil
mixes were replaced by HMA surfacings after World War II, the cohe-
siometer test served no real purpose and it gradually fell out of favor.
By 1959 the Hveem stabilometer procedure for design ofHMA mixes
had evolved into its final form. Several states, primarily in the West, adopt-
ed the Hveem procedure. However, many states modified various elements
in the procedure.
Marshall Mix Design Method
A detailed account of the development of the Marshall method is given

in References 5. and Q.
The earliest version of the Marshall mix design method was developed
at the Mississippi Highway Department by Bruce Marshall around 1939.
The Corps of Engineers Waterways Experiment Station (WES) began a
study in 1943 to develop a simple portable apparatus for designing asphalt
mixtures for airfield pavements. This was primarily prompted by the
-----'·-creas-e-i1n:i:rcrafr-wh-ed-I-mrds-dmrng-Wo-rtd-War-II. Tire pressures and,
therefore, contact pressures also increased, warranting a suitable mix design
method to cope with increased tire pressures. The Corps began experi-
menting with Bruce Marshall's apparatus and embarked on a series of lab-
oratory and field experiments. Mixtures were designed in the laboratory
using a variety of compactive efforts in an attempt to produce densities in

the laboratory which were similar to those achieved in the field under con-
struction and simulated aircraft loadings. The laboratory compactive
efforts involved different drop hammer weights; different combinations of
numbers of blows per side and/ or compressive loads; different compactor
foot designs including both size and shape; and different mold base shapes
and materials.
One goal of the laboratory compaction study was to adopt a sample
preparation procedure that would involve minimum effort and time but
would provide a basis for selecting the proper optimum asphalt content. It
was also desirable to develop a method that was portable and which could
be taken to the field for quality control purposes. Based on the results of
this study, a IO-pound (4.54 kg) hammer with a 37/s-inch (98.4 mm)
diameter foot was selected, and a compaction effort of 50 blows per side of
the specimen was adopted as a standard.
Test sections trafficked with loads and tire pressures of 15,000 lbs (50
psi), 37,000 lbs (110 psi), and 60,000 lbs (110 psi) [66.7 kN (345 kPa),
164.4 kN (759 kPa), and 266.7 kN (1379 kPa)] for up to 3500 coverages
showed no significant increase in density after traffic, regardless of wheel
load. Using the results from tests conducted on field-secured specimens,
the following limits were established as the criteria for mixture design:
· 50 Blow Marshall Criteria for Surface Mixes (i)
Asphalt Content
Selection Limiting Criteria
Test Property
HMA Sand Asphalt HMA Sand Asphalt
Stability, lbs. (N) Peak . Peak 500 (2220) min 500 (2220) min
Flow, 0.01 inch (0.25 mm) - 20 max 20 max
Unit Weight Total
Mix, pcf (g/ml) Peak Peak
Voids Total Mix, % 4 6 3-5 5-7
Voids Filled with
Asphalt,% 80 70 75-85 65-75
During the late 1940s and early 1950s, aircraft size and weight con-
!-----tinued to increase-with-the-resul:t--that-rire-pressures-were-i.ncreased-ro-2-0 - · ~ - - - - -
psi (1380 kPa) in order to keep aircraft gear and tire sizes about the same.
This increase in tire pressure significantly increased the near surface com-
pactive effort of traffic. The WES began a study to determine what labo-
ratory compactive effort was required to correlate with traffic-induced
densities and to modify the low pressure mix design and quality control
criteria. The original full-scale test sections were trafficked by 30,000 lb
(133.3 kN) wheel loads on 200 psi (1380 kPa) tires and the pavement per-
formance monitored.
Even though the surface density continued to increase with the appli-
cation of the high pressure tire loads, the surface rutting measurements
were confounded as a result of a failure in the asphalt binder course under-
neath the surface. Using the data from both the field study and the labo-
ratory compaction study, the following criteria were adopted for the high
pressure tires (200 psi or 1380 kPa) which included adjustments in stabili-
ty and flow based on test data from the trafficked surface as well as a mod-
ification to the laboratory compaction effort:
Initial High Pressure Tire Criteria

75 Blow Marshall Criteria for HMA Surface Mixes Ci)
Selection of
Test Property Asphalt Content Limiting Criteria
Stability, lbs. (N) Peak 1000 (4440) min

Flow, 0.01 inch (0.25 mm) 16 max
Unit Weight Aggregate Only, pcf (g/ml) Peak
Unit Weight Total Mix, pcf (g/ml) Peak
Voids Total Mix, % 5 4-6
Voids Filled with Asphalt, % 80 75-82
In addition to the development of the high tire pressure criteria in the

1950s, WES noted performance problems, especially rutting, on a number
of projects. Investigations revealed that large quantities of natural sand
were present i:n those mixes experiencing rutting problems. Attempts to
control the use 'of natural sands resulted in raising the Marshall stability to
1800 lb (8000 N) and subsequently a limitation on the allowable amount
of natural sand that could be used in airfield pavements. The initial limit
on natural sand was 10 percent of the total aggregate, but this was later
raised to 15 percent.
In summarizing the development and evolution of the Marshall
-----method,'Whi-1:e-E:2)-e0neluded-that-two-variables cstand out in the design
0
and performance of HMA: asphalt content and density. In the field it is the
highest satisfactory asphalt content at a density achieved under traffic that
is significant. In the laboratory the important feature is selecting a com-
paction procedure that represents traffic-induced density and then select-
ing response properties that can be averaged to yield an asphalt content

that will produce satisfactory performance.
In reviewing the development of the Marshall method, it is apparent
that a controlling factor is the correspondence between the density
achieved in the field under traffic and that produced in the laboratory with
the specified compactive effort. When field conditions changed due to
changes in traffic, WES monitored the field conditions and adjusted the
laboratory procedures to produce densities similar to those being achieved
in the field under traffic. Therefore the driving force for change in the
Marshall procedure was changes in the field density brought about by
changes in traffic.
Applying this principle of empirical de~!gn to highway practice, shows
the importance of determining density induced by traffic before making
changes in laboratory procedures. Only by knowing field conditions can
proper adjustments be made in the laboratory since the laboratory condi-
tions must replicate field conditions.
Figure 4-1 shows the states that used various mix design methods
according to a 1984 survey (Jj. Notice that 38 states used some version of
the Marshall method while 10 states used the Hveem method.
ND
IIT
SD
'If'(
~HVEEM
C]MARSHALL
~ TEXAS
-GRADATION
Figure 4-1. HMA Mixture Design Methods Used by States in the

USA (J)
Chapter 4--Hot Mix Asphalt Mixture Design Methodology 181
Kandhal and Koehler (1) discuss a number of factors that significantly

affect the test results obtained from specimens prepared with different
types of hammers and tested with breaking heads with different bevel
geometry on the lower inside corner of the breaking head segment. Because
of the different methods for selecting optimum asphalt content used by the
different states, it is anticipated that a significant variation would be
obtained among the states if a common material were compacted and test-
ed and the optimum asphalt content determined.
Superpave Mix Design Method
A detailed description of the Superpave mix design method is provid-

ed in reference (Z).
The Superpave mix design method is a direct result of the Strategic
Highway Research Program. The primary purpose of this program was to
develop a rational mix design procedure. The Superpave gyratory com-
pactor (SGC) was selected for compaction because it orients the aggregate
particles in a way similar to that observed in the field.
The mix design method uses volumetrics to select the optimum
asphalt content. Mixtures are compacted to provide a density in the labo-
ratory equal to the estimated density in the field after various levels of traf-
fic. The number of gyrations are varied to simulate anticipated traffic.
The air voids at NinitiaJ, Ndesign' and Nmaxirnum are measured to evaluate
the mixture quality. The mixture should have at least 11 percent air voids
at NinitiaI, 4 percent air voids at Ndesign' and at least 2 percent air voids at
Nmaximum. The number of revolutions for Ninitial> Ndesign' and Nmaximum are
defined later in this chapter.

This mixture design method will be implemented by State DOTs in
the late 1990s. It is anticipated that some changes in the mix design pro-
cedure will qccur during the first few years of its existence as more research
is accomplished.
OBJECTIVES AND ELEMENTS OF MIX DESIGN
Objectives
HMA mix design should be developed with the following objectives in

mind.
Resistance to Permanent Deformation. The mix should not distort or

displace when subjected to traffic. The resistance to permanent deforma-

tion (or rutting) becomes critical at elevated temperatures during hot sum-
mer months when the viscosity of the asphalt cement binder is low and the
traffic load is primarily carried by the mineral aggregate structure.
Resistance to permanent deformation is controlled by selecting quality
aggregates with proper gradation and selecting the asphalt content so that
adequate voids exist in the mix. Test methods such as static or dynamic
creep discussed in Chapter 5 may be more suitable for predicting perma-
nent deformation.
Fatigue Resistance. The mix should not crack when subjected to repeated
loads over a period of-time. Complex repeated load (either constant stress
or constant strain controlled) tests discussed in Chapter 5 can be run to
estimate the number of cycles to cause failure (cracking) in HMA mixtures.
Resistance to Low Temperature Cracking. This mix property is impor-

tant in cold regions which experience low ambient temperatures. As dis-
cussed in Chapter 2, low temperature cracking of HMA pavement is pri-
marily influenced by the low temperature properties of the asphalt cement
binder. Therefore, a proper selection of binder minimizes this problem.
Durability. The mix must contain sufficient asphalt cement to ensure an

adequate film thickness around the aggregate particles, thus minimizing
asphalt cement hardening or aging during production and in service. The
compacted mix should not have very high air voids (increased permeabili-
ty) which accelerate the aging process.
Resistance to Moisture Induced Damage. Some HMA mixes, when sub-

jected to moisture or water lose adhesion between the aggregate surface and
asphalt cement binder. Aggregate properties are primarily responsible for
this phenomenon, although some asphalt cements are more prone to mois-
ture damage (stripping) than others. If a HMA mix is prone to stripping,
then antistripping agents should be used. Making the rnix impermeable to
water also minimizes. the problem. Moisture susceptibility of HMA mixes
is discussed later.
11-----S'kid- Resistance;-'Fhis-requiremencis'-orrly-a:ppticable-to-stirface-mixe-s
1
which must be designed to provide sufficient resistance to skidding to per-

mit normal turning and braking movements to occur. Aggregate charac-
teristics such as texture, shape, size, and resistance to polish are primarily
responsible for skid resistance. However, the mix should also not contain
r
Chaprer 4--Hot Mix Asphalt Mixture Design Methodology 183
'1
\
PERCENT PASSING
8
- .
0 0
co g i [j! 0
\
\
\
f\ \
I\ I\
i\
\ \
I\
I\ f\
I\ I\ I\
ffi
3
0
a. \ I\
~~ \ \
0
~ ~ I\ i\
0
l- - ;;;~ [3
N
a I'\. '
[", I'\.
\ i\ \ cij
~ I\ [\ w
<
0::
~ ~
\ \ >
w
en
~ I'\ I\
I'\ I\ [\ I\
[3
N ~
' f" I"\ I'\ \ I\' I\I\
(/)
w
Gi "' r---,.
1'-,,
f" I'\
cij 1'.
1'-,, I'\ I'\. I\f\ i\
I"\~ \
" t---- r-.... t---- '
~
r---,. ' ~I'\ I\ [\\
--t--..
r--.. r--..
t----
r--..
'r---.
I'---
~
I"--
'~ "' "~
r---,. l\L'\. ~
\
0
"l
I"-- " r---,. "~ ~ ~

r-- i----: !
'r---.
-
0
r--.. ~
r-- 'r---. ', ~ ~ i
0
~
r--.. ............ I"-..
....... i::::: '"~ ~~
~
~\ !
0
.,0
"' 0
0
·-------+--------9NISS~d-J.N3::>!:13d- - -
Figure 4--2. Recommended FHWA Maximum Density Curves Based

on the 0.45 Power Gradation Chart
0
0 0 0 g 0 0 0 0
N
0 0
Ol CD CD
"' "' N
~
-
~ z ,,,,
~
0
\ I\ /<f.
i==
\ l;- <f.
I'. Cl
<f.
I'\ f\ ' a::
~~ V
(!)
"" ' '\

~
~
:i-
;;;
~
' )\
~
I
......... ......... a) 17I/ ~ f\.. ~
•W
~N
Cf)
u " .\ k' .w
~>
ci5
"[\_~ ~ ~
"\ ' "'w
ci5
'- st
~~
' I\.
'
CD
"~ ~ I',.,_
"'
I'\ ~ 0
p
' \.
, .~
.,0
1::-.;
' .~ 0
~
-·- -~--
0
' 0 0 0 0 0
0
r-
Ol
"' "' "'
E>NISS'vd 1N3:::>H3d
Figure 4-3. Three Common Methods of Plotting Maximum Density
Lines on 0.45 Power Gradation Chart
Chapter 4-Hot Mix Asphalt Mixture Design Methodology ' 185
too much asphalt cement binder to cause the HMA mix to flush and cre-
ate a slippery surface. Skid resistance is discussed in detail in Chapter 8. It
is also discussed briefly later in this chapter.
Workability. The mix must be capable of being placed and compacted

with reasonable effort. No test method is currently available to quantify
workability during the laboratory mix design phase. Workability problems
are most frequently discovered during the paving operations. However,
suitable adjustments to the mix design can be made quickly to overcome
the workability problems.
Gradation and Maximum Density Curves
Blending of aggregates is discussed in Chapter 3. Grading curves are

very helpful in making necessary adjustments to the mix design such as
VMA (voids in the mineral aggregate). Maximum density curves, as men-
tioned in Chapter 3, are particularly suitable for this purpose. In 1962
Goode and Lufsey proposed that aggregates should be graded using a
mathematical concept of packing the void space between aggregates of
large diameter with aggregates of smaller diameter (ff). Nijboer cites the
work of Andreasen in determining the grain picture of an aggregate con-
taining particles of all sizes to produce the most efficient arrangement of
packing spheres in a confined space (2). The mathematical relationship
describing the maximum density gradation to produce the densest packing
of particles based on Fuller's curve is given in Chapter 3. However, Goode
and Lufsey (fl) noted that if the gradation corresponding to the exponent
of 0.5 is used as proposed by Fuller, then the VMA may be too low to
ensure both sufficient air void content and sufficient asphalt cement for
durability without encountering field stability problems. Therefore, rec-
ommendations in the FHWA Advisory (JJ!) include the suggested use of
the 0.45 power curve to evaluate and adjust aggregate gradations. The most
convenient way to evaluate gradation is to plot the combined gradation on.
a plot on which the FHWA 0.45 power gradation line has been graphed
such as shown in Figure 4-2.
There has been some confusion on how the various agencies prepare
------,.he-plots-representing-the-0;4§-power-curve. In referring to the original
Goode and Lufsey (ff) paper, it is not clear in the paper how the maximum
particle size D or the minimum particle size is determined and plotted. ·
There appear to be three prevalent methods for plotting the maximum
density lines.
The solid line in Figure 4-3 shows the gradation of a binder course
mix (as an example):
Sieve Size Percent Passing

1½ inch (37.5 mm) 100
1 inch (25 mm) 90
1/2 inch (12.5 mm) 61
These gradation data are used three different ways to produce the maxi-
mum density lines labeled A, B, and C in figure 4-3. The construction of
each line is described below.
Maximum Density Line ''N.'-obtained by drawing a straight line from

the origin at the lower left of the chart to the maximum (top) particle size
(1 ½ inch (37.5 mm) in the example) at the top.
Maximum Density Line "B"-obtained by drawing a straight line from

the origin to the nominal maximum particle size (1 inch (25 mm) in the
example) at the top. The nominal maximum particle size is defined as the
largest sieve size listed in the applicable specification upon which any mate-
rial is retained.
Maximum Density Line "C"-obtained by drawing a straight line from

the origin to the percentage point plotted for the largest sieve with mater-
ial retained (90 percent passing point for 1 inch (25 mm) sieve in the
example).
The Maximum Density Line "C" was recommended by Goode and

Lufsey of the then Bureau of Public Roads (now the FHWA) who devel-
oped this gradation chart in the early 1960s. Their preference is stated in
the discussion of the paper which was reported in the Proceedings of the
Association of Asphalt Paving Technologists (Volume 31, 1962). Line "C"
is considered to be the most realistic maximum density line. It is very
important to have consensus among designers on this issue, especially since
··~---some state specifrcationnequire-tlrarthe mix-graciation be below tn1,-e~m~ax~-- - - - -
imum density line. The 0.45 power curve for maximum density line
should be used only as a guide and should not be incorporated into speci-
fications that require that all gradations be on, above, or below this line.
Satisfactory mixes can be produced under all three conditions. However, a
gradation in close proximity to the maximum density line generally gives

low VMA, resulting in a less durable mix that is more sensitive to slight
variations in asphalt content during production. To avoid these difficulties,
gradations are usually shifted away from the maximum density line.
The following general recommendations should be considered in mix-
ture design:
1. If the fine aggregate is predominately round natural sand, it is usu-

ally preferred to stay below the maximum density line to minimize
the amount and the effect of the sand.
2. If crushed or very angular fine aggregate is used, the gradation may
be above the maximum density line. However, these mixes may not
be economical since the voids must be filled with asphalt cement in
order to meet the air voids criteria in mix design.
3. It is undesirable to have a hump in the gradation in the No. 30-50
size range, but this should especially be avoided when natural sands
predominate in the fine aggregate. Humps in this size range tend to
produce mixes which are tender, i.e., they exhibit high plastic dis-
tortion under normal compaction equipment immediately after lay-
down.
These preceding recommendations should be considered especially when

the mixture being placed is giving problems during laydown or com-
paction or experiencing appreciable rutting during service.
Asphalt Film Thickness
Consideration of film thickness is not directly considered as a part of

mixture design. However, in analysis of mixtures which have exhibited pre-
mature distress, it may be desirable to check film thickness. Therefore, a
procedure for calculating film thickness is included in this section.
The thickness of the asphalt cement film around a particular aggregate
is a function of the diameter of the aggregate and the percent asphalt
cement in the mixture. Hveem found that his initial assumption, that each
aggregate particle needed to be covered with the same optimum film thick-
-----'-"'ess,_w:~s_w..r.ong. Rather, Hveem realized thanh~J'ilm thickness on a par-
ticle decreases as the diameter of the particle decreases. The current tech-
nique for calculating film thickness is based on the surface area factors
mentioned in the section on the Hveem method of mix design. The
asphalt cement film thickness is calculated using the following formula:
Tp =
vase
SA x W (1,000
)
where,
TF = Average film thickness, microns;
V asp = Effective volume of asphalt cement (liters);
SA = Surface area of the aggregate (m2 per kg of
aggregate); and
W = Weight of aggregate (kg).
The volume of asphalt cement, V asp, is the amount of material avail-

able for coating aggregates and, therefore, must have the weight of asphalt
cement absorbed by the aggregate subtracted from the total asphalt cement
added to the mix. The calculated film thickness is an average film thickness
which has been generally correlated with durability. If the asphalt cement
film is too thin, air which enters the compacted HMA can more rapidly
oxidize these thin films, causing the HMA to become brittle and to fail
prematurely by cracking. Additionally, if the aggregates are susceptible to
water damage, thin films are more easily and rapidly penetrated by water
than thick ones producing the typical manifestations of water damage: rut-
ting, shoving, ravelling, and bleeding.
Example 4-J. The surface area of an aggregate has been determined,

as discussed in Chapter 3, to be 35.2 square feet per pound (7.23 m 2/kg).
The bulk specific gravity of the aggregate was measured to be 2.680 and
the effective specific gravity was determined from the theoretical maxi-
mum density to be 2.700.,Determine the asphalt film thickness in microns
if the asphalt cement specific gravity is 1.030. The asphalt content is 5.0
percent and the bulk density of the mixture is 146 pcf (2.340 g/ml).
Recall that
vase
SA X w (1,000)
The effective volume of asphalt cement is equal to the total volume

minus the absorbed volume. For a one liter sample the
- - ~ - --Total·volumeofasphalnemenr=-1.e14 Lw.e:_igh;t.x_aspl2aluo.J:J:t.m.:t._ _ _ _ _ _ _ _ _ _ _,_
Sp grx'Jw
2340 g x 0.05 = 113 _6 ml

1.03 (1)
Asphalt absorbed, Pb (by weight ofaggregate) = 100 G,, - G,b Gb

G,b G,,
100 2.700 - 2.680 1.030

(2.680) (2.700)
0.2847%
0.002847
Weight ofabsorbed asphalt 0.002847 (weight ofaggregate)
0.002847 (2340 x 0.95) = 6.33 grams
Volume ofabsorbed asphalt 6.33 6.15 ml

1.03 (1)
Effective volume ofasphalt = 113.6 - 6.2 = 107.4 ml
Now substituting into the equation for film thickness,
vasp
SA X w (1,000)
107.4 (1,000)
7.23 (2340 X 0.95)
Tp = 6.7 microns
Laboratory Compaction
Based on the discussion of the evolution of both the Marshall and the
Hveem methods of mixture design, it is apparent that the laboratory com-
paction methods included in the initial methods were referenced to field
conditions. The reference field conditions were not based on the as con-
structed densities but on the density achieved after the pavements had sus-
----~t.ai.n~_d_a_.r_e_a_sonable level of traffic;, Th_e_S_up_e.rp_ave methQdis also designed
to provide a density equal to that after some amount of traffic.
Therefore, it is clear that if the engineer intends to design a mixture in
the laboratory that will sustain the loads imposed on it by traffic, that lab-
oratory mixture should be compacted to the same general density (void
content) to which the traffic will finally compact it under service condi-
tions. Otherwise, the traffic levels may be so high in service that the mix-
ture may be compacted to such a dense state that the void content is too
.low and instability results. Alternatively, the traffic levels may be so light
that the compaction energy used in the laboratory produces a density too
high for the actual traffic level. The result is too little asphalt cement and
a mixture that has low durability in the field. This second type of mix does
not exhibit instability in the field, but the film thicknesses are too thin for
good durability under light traffic. It is apparent then that the selection of
the laboratory compaction level is critical in designing a mixture that will
provide adequate field performance.
Figure 4-4 illustrates the type of relationship that exists between air
void content and level of compaction in th:-e laboratory and the time the
same mix is subjected to field traffic as reported by Arena (11). Notice in
the plot that as the number of 'revolutions of the gyratory compactor
increases, the air voids go down. In the gyratory compaction device, a com-
pressive load is applied to a specimen while the mold containing the spec-
imen is tilted from side to side in a gyrating action by a rotating flanged
9
MIX NO. 2
ASPHALT 5.8%
8
100 PSI
•
Cl)
0
0
>
I- 6
z
w
7
\-L •
~ 5 BLOWS, PLANT CMECHJ
u
c::
w •
a.. 5
4,
0
·--.
10 20 30 40 50 60 70
NUMBER OF GYRATIONS
Figure 4-4. Comparison of Void Contents Derived from Laboratory

Compacted Specimens and Field Measurements (1J)
Chapter 4--'-Hot Mix Asphalt Mixture Design Methodology 191
mold chuck (see Figure 4-5). Notice also in Figure 4-4 that several refer-
ence points are noted on the curve. The first point, "ORIGINAL," corre-
sponds to the original as constructed density in the field which is normal-
ly specified as some percentage oflaboratory density; the second point, "75
BLOWS PLANT (MECH)," corresponds to the air void level achieved
from the field mixed material that was compacted in the field laboratory
using 75 blows from a mechanical Marshall compactor; and the next 3
points, "6 MONTHS," "15 MONTHS," and "36 MONTHS," represent
air void contents determined from cores taken from the wheel paths after
the mixture was subjected to traffic for those respective periods. While the
Marshall compaction effort in the field laboratory corresponded to an air
void content of slightly less than 6 percent, the mixture appeared to stabi-
lize under traffic at an air void content between 3 and 4 percent. Notice
also that the gyratory compactor is capable of achieving air void contents
much lower than that achieved by mechanical Marshall hammer compac-
tion. The compaction curve shown in Figure 4-4 applies only to the spe-
cific mix investigated and cannot be considered as a typical plot for all
mixes.
If surveys show that the laboratory compaction method consistently
produces densities substantially lower or higher than those achieved in the
field under traffic, the specifying agency should consider changing the lab-
oratory compaction methodology or blows (compaction effort) in order to
achieve laboratory densities which are consistent with those being pro-
duced by traffic.
Fixed Upper Rom
Fixed Roller
Gyrolory Angle
Adjustment Knob
Figure 4-5. Cross-section of Gyratory Compactor Showing

Compression and Shear Force Mechanisms
Weight Volume Relationships J
A basic understanding of weight-volume relationships of compacted

HMA is important from both a mixture design standpoint and from a field
construction standpoint. It is important to understand that mix design is a
volumetric process whose purpose is to determine the volume of asphalt
cement and aggregates required to produce a mixture with the desired
properties. However, measurements in the laboratory or field of the vol-
ume of aggregates and asphalt cement are very difficult and totally imprac-
tical. Therefore, to simplify the measurement problem, weights are used
instead of volume, and the specific gravity is used to convert from weight
to volume. Two very important parameters are voids in total mix (VTM),
or air voids in the compacted HMA, and ,:yolume of voids in the mineral
aggregates (VMA). Figure 4-6 illustrates these volumes in a simplistic for-
mat. Unfortunately, accurate calculation of these volumes is compounded
by the partial absorption of asphalt cement into the aggregate. If asphalt
cement is not absorbed into the aggregate, the calculations are relatively
straightforward in that the bulk specific gravity of the aggregate can be
used to calculate the volume of aggregate. If asphalt cement absorption is
identical to water absorption as defined by ASTM C127 and C128, the
calculations are relatively straightforward in that the apparent specific grav-
ity of the aggregate can be used to calculate the volume of aggregate~ Since
almost all mixtures have partial asphalt absorption, the calculations are
more cumbersome as explained later.
Definitions. A more complete understanding of weight-volume rela-

tionships must start with definitions. The definitions given below are con-
sistent with those advanced by ASTM, AASHTO, and the Asphalt
Institute.
Voids in the Mineral Aggregate NMA)-The volume of intergranular

void space between the aggregate particles of a compacted paving mixture
that includes the air voids and volume of the asphalt not absorbed into the
aggregates (see Figures 4-6 and 4-7).
Voids in Total Mix (VTM)-The total volume of the small pockets of

air between the coated aggregate particles throughout a compacted paving
H---_ _ _mixture, exeressed as a percent of the b_ulk_~olum..e_oLthe_co.mp.acte.' - L - - - - - -
paving mixture (see Figures 4-6 and 4-7).
Voids Filled with Asphalt Cement (VFA)-The percent of the volume

of the VMA that is filled with asphalt cement (see Figure 4-6).
VT COMPACTED WT
SPECIMEN
Vy Air Voids 0
VEAc Effective Asphalt Cement WEAC

VAAC Absorbed Asphalt Cement WAAC
VA"" Aggregate WArt.rt.
V,
Ji'7M = __! X 100
VT
W.4gg
VAa = ----=---
Sp Gr.4gg x Yw
' V,
YFA = .&tC V x 100
V.&(C + y
where,
VT = Total volume of compacted specimen;
Vv = Volume of air voids;
VEAC = Volume of effective asphalt cement;
VMC = Volume of absorbed asphalt cement;
VAss = Volume of aggregate;
WT = Total weight of compacted specimen;
WEAc = Weight of effective asphalt cement;
WMc = Weight of absorbed asphalt cement;
WAss = Weight of aggregate;
VTM = Voids in total mix, percent;
VMA = Voids in mineral aggregate, percent;
VFA = Voids filled with asphalt, percent;
AC = Asphalt content by weight of mix, percent;
'Yw = Unit weight of water; and
Sp Gr = Specific gravity.
Figure 4-6. Weight-Volume Relationships for Hot Mix Asphalt
In order to calculate VTM, VMA, or VFA, it is necessary to have data

on the bulk specific gravity of the compacted HMA (ASTM Dl 188 or
D2726), the appropriate specific gravity values of the individual compo-
nents of the mixture, and the asphalt content.
absorbed asphalts
Impermeable
voids
effective asphalt
water
permeable voids
voids
voids in mineral aggregate = air voids plus volume of effective asphalt

EXAMPLE• ASSUME BULK SPECIFIC GRAVITY- OF COMPACTED MIX = 2.250
VOLUME OF AIR VOIDS = 4%
ASPHALT CONTENT = 6.2%
ABSORPTION = 2%
EFFECTIVE ASPHALT CONTENT = 4.2%
2 250 042
VOLUME OF EFFECTIVE ASPHALT CONTENT = " • x • x I 00 = 9.3%
1.02
VOLUME OF VOIDS IN MINERAL AGGREGATE = 4% + 9.3% = 13.3%
Figure 4-7. Diagram Illustrating Air Voids and Voids in Mineral

Aggregate (VMA) (llJ
impermeable voids
voids
apparent = weight of oven dry aggregate

specific gravity volume of aggregate plus volume of impermeable .voids
VOLUME IS DETERMINED BY DISPLACEMEI-ILIILWAT.Ea-WI-T-1=1-0.V-EN-DR¥-AGGREGA-T-E-OR'------

---.-tJSING_A_SATIJRATED-S0RFACE--DRY. AGGREGATE AND THEN SUBTRACTING THE
VOLUME .OF WATER EOUAL TO THE AGGREGATE WATER ABSORPTION
Figure 4-8. Diagram Illustrating Apparent Specific Gravity (llJ

impermeable voids
bulk (dry) = weight of oven dry aggregate

specific gravity volume of aggregate plus volume of impermeable voids
and permeable voids
VOLUME IS DETERMINED BY DISPLACEMENT IN WATER WITH

SATURATED SURFACE DRY AGGREGATE
Figure 4-9. Diagram Illustrating Bulle (Dry) Specific Gravity (11)
impermeable voids
permeable voids
weight of oven dry aggregate plus weight of waler

bulk (saturated surface dry) _ in permeable voids
specific gravity - volume of aggregate plus volume of impermeable
voids and permeable voids
--------------,------------
VOLUME IS DETERMINED BY DISPLACEMENT IN WATER WITH
SATURATED SURFACE DRY AGGREGATE·
Figure 4-10. Diagram Illustrating Bulle (SSD) Specific Gravity (11)

In order to facilitate further discussions relative to voids, definitions

are needed for the methods used for measuring specific gravity of the
aggregates and their relevancy to voids considerations. Specific gravity is
needed in voids calculations to convert from weight measurements to vol-
ume determinations. Apparent and bulk specific gravities of the mineral
aggregates are discussed in detail in Chapter 3. Their definitions only are
given here.
Apparent Specific Gravity (ASTM CI27 and Cl 82)-The ratio of the

oven dry weight in air of a unit volume of an impermeable material at a
stated temperature to the weight of an equal volume of gas-free distilled
water at a stated temperature (see Figure 4-8). Apparent specific gravity is
normally used for weight to volume calculations of the mineral filler only,
since bulk specific gravity values of this fraction are very difficult to obtain.
Bulle Specific Gravity, Dry (ASTM C127 and C128)-The ratio of the
oven dry weight in air of a unit volume of a permeable material (including
both permeable and impermeable voids normal for the material) at a stat-
ed temperature to the weight of an equal volume of gas-free distilled water
at a stated temperature (see Figure 4-9).
Bulk Specific Gravity; SSD (ASTM C127 and C128)-The ratio of

the SSD weight in air of a unit volume of a permeable material (including
both permeable and impermeable voids normal for the material) at a stat-
ed temperature to the weight of an equal volume of gas-free distilled water
at a stated temperature (see Figure 4-10). This specific gravity is seldom
used in the design of HMA mixes.
Effective Specific Gravity of an Aggregate-The ratio of the oven dry

weight in air of a unit volume of a permeable material (excluding voids per-
meable to asphalt) at a stated temperature to the weight of an equal vol-
ume of gas-free distilled water at a stated temperature (see Figure 4-11).
Bulk Specific Gravity, Compacted Asphalt Concrete (ASTM D1189

and D2726}-The ratio of the weight in air of a unit volume of a com-
pasted spe_cimen of HMA (including_.1.2ermeable_"\:.O.ids)_at...a.statecltempe:1.,-o.------
----
ature to the weight of an equal volume of gas-free distilled water at a stat-
ed temperature. The value is used to determine weight per unit volume of
the compacted mixture. The bulk specific gravity (GmJ of a compacted
asphalt mixture is equal to:
WssD - Wsub
where,
WO = Dry weight, grams;
W 550 ,,; Saturated surface dry weight, grams; and
Wsub = Saturated surface dry weight submerged m water,
grams.
Theoretical Maximum Specific Gravity of Bituminous Paving Mix-

tures (ASTM D2041)-The ratio of the weight in air of a unit volume of
an uncompacted bituminous paving mixture at a stated temperature to the
weight of an equal volume of gas-free distilled water at a stated tempera-
ture (see Figure 4-12). It is also called Rice Specific Gravity (after James
Rice who developed the test procedure) or theoretical maximum density
(TMD).
impermeable voids
permeable voids not filled

with asphalt
permeable voids filled with asphalt
,effective specific _ weight of oven dry aggregate

gravity of aggregate - volume of aggregate plus volume of impermeable voids
plus volume of waler permeable voids less volume of
absorbed asphalt
EFFECTIVE SPECIFIC GRAVITY DETERMINED BY PROCEDURES OUTLINED IN .ASTM D2O41.
BULK IMPREGNATED SPECIFIC GRAVITY OF CORPS OF ENGINEERS USUALLY
INDICATES MORE ASPHALT ABSORPTION AND HIGHER EFFECTIVE SPECIFIC GRAVITY.
EFFECTIVE SPECIFIC GRAVITY SHOULD FALL BETWEEN APPARENT AND BULK
IDRYI SPECIFIC GRAVITY.
Figure 4-11. Diagram Illustrating Effective Specific Gravity of

Aggregate (12)
-,
i
volume of aggregate
impermeable
voids volume of voids not filled with
asphalt
volume of voids filled with asphalt
maximum theoretical specific = weight of aggregate plus weight of asphalt

gravity of mixture volume of aggregate plus volume of
impermeqble voids plus volume of
permeable voids not filled with asphalt
plus total volume of asphalt
GMM = _ _P_M_M_ _ PMM = TOTAL WEIGHT OF MIX
PS = WEIGHT OF AGGREGATE
~ + PB
GSE GB PB = WEIGHT OF ASPHALT
GSE = EFFECTIVE SPECIFIC GRAVITY OF AGGREGATE COATED
WITH ASPHALT
GB = SPECIFIC GRAVITY OF ASPHALT
Figure 4--12. Diagram Illustrating Maximum Theoretical Specific

Gravity of Mixture (12)
A review of the above specific gravity definitions indicates the following:
1. The difference between apparent and bulk specific gravity (dry) is

the volume of the aggregate used in the calculations. The difference
between these volumes is equal to the volume of the absorbed water
in the permeable voids (difference between SSD and oven dry
weight of aggregate when weighed in grams). Both specific gravi-
ties use the oven dry weight of the aggregate.
2. The difference in calculations between bulk (dry) and bulk (SSD)
specific gravity is the weight of the aggregate. The volume of the
aggregate is identical for both specific gravities. The difference in
the weights is equal to the absorbed water in the permeable voids
(difference between SSD and oven dry weights of aggregate).
3. The differences in c.ak_ulations_am.ong-apparent,-bulk-(-d13/-)-anu-- - - - - -
effective specific gravity is the volume of the aggregate. All three
specific gravities use oven dry weights of aggregate.
4. The difference in calculations between the compacted bulk specif-
ic gravity and maximum theoretical specific gravity is the volume
as the weights are identical. The difference in the volumes is that

associated with the volume of the air in the compacted mixture.
5. The measured values of the above specific gravities can be conve-
niently checked to a first approximation by realizing the following:
(a) The apparent specific gravity will always be equal to or larger
than the effective specific gravity which will always be equal to
or larger than the bulk (dry) specific gravity.
(b) The bulk (SSD) specific gravity will always be equal to or larg-
er than the bulk (dry) specific gravity.
(c) The maximum theoretical specific gravity will always be equal
to or larger than the compacted bulk specific gravity of the
HMA mixture.
(d) The aggregate specific gravity [apparent, effective, bulk (dry),
bulk (SSD)] will always be larger than the maximum theoreti-
cal specific gravity of the HMA mixture.
Calculations. Calculations associated with weight-volume relationships

for HMA mixtures can be found in considerable detail in Asphalt Institute
publications Cl.3). These calculations are summarized below.
Bulk Specific Gravity of Combined Aggregates-Most HMAs contain

several different aggregates (crushed limestone, lightweight aggregate, grav-
el, field sand, screenings, etc.) which are combined to meet the desired gra-
dation. Usually these aggregates have different specific gravities which need
to be combined for determining weight-volume relationships of the mix-
tures. The equation used to determine the specific gravity of the combined
aggregate is given in Chapter 3.
The bulk specific gravity of mineral filler is difficult to determine accu-
rately at the present time. However, apparent specific gravity of the filler
can be used and the error is negligible.
Volume of Voids in Mineral Aggregate (VMA)-As noted above, the

volume of voids in the mineral aggregate VMA is an important factor for
mixture design.
The formula for VMA can be developed as follows based on the
weight-volume relat10nships shown m Figure Lf-6. It is recommended that
the bulk specific gravity of aggregate be used for calculating VMA.
simplifying,
V (bulk)
VMA = 100 -' ..tu (100)
VT
substituting volume with weight divided by specific gravity gives
W. IG
VMA = 100 - AU sb 100
WJGm1,
substituting,
and simplifying
(1)
where,
VAgg = Volume of aggregate;
VT= Total volume of compacted mixture;
WAgg = Weight of aggregate;
WT= Total weight of mixture;
G.b = Bulk specific gravity- of aggregate;
Gmb = Bulk specific gravity of compacted mixture; and
Pb= Asphalt content.
Air Voids in Compacted Mixture (VfM)-The formula for calculat-

ing the percent air voids can be developed as follows:
By definition,
Vy
Jl7M = - X 100
VT
substituting,
multiply the numerator and denominator by WT and simplify,
YIM = 1 - w:T /V.T ) 100

( WT /(VEAC + V.tu>
substituting,
(2)
where,
VTM = Voids in total mix;
Vv = Volume of air voids;
Vy = Total volume of compacted specimen;
VEAc = Volume of effective asphalt cement;
VAgg = Volume of aggregate (bulk);
Wy = Total weight of compacted specimen;
Gmb = Bulk specific gravity of compacted specimen; and
Gmm = Maximum theoretical specific gravity of mixture.
Voids Filled with Asphalt NFA)-VFA is simply the percentage of

VMA filled with asphalt cement. The following formula is used to calcu-
late the VFA:
VFA = Vi\l4 - YIM x 100

™4
Effective Specific Gravity of Aggregate-The effective specific gravity

(Gse) is calculated from the maximum specific gravity (Gmm) of the mix-
ture. Tests for Gmm are run on voidless loose mixtures. The formula for Gse
can be derived as follows:
G,, =
since the effective specific gravity (G,J is equal to the weight of aggregate
-----Jivtded-by-rhreffective-voforne··ohhe-a:ggrega:te-(V:e.f!}.The effective vol-
ume is the volume of the aggregate plus the external voids minus the exter-
nal voids filled with asphalt cement. In the test for maximum specific g~av-
ity (Gmm), the total volume of the voidless mix is measured and the volume
of asphalt cement is determined from the weight of asphalt cement and the
202 Hot Mix Asphalt Materials, Mixture Design and C(i)nstruction
specific gravity of asphalt cement. The effective volume of the aggregate is

determined by subtracting the volume of asphalt cement from the total
volume.
Hence,
G,, =
substituting,
G = WT - Pb WT
se WT
-- ---
w;..c
G_ Gb
simplifying,
Gu =
where,
w.gg = Weight of aggregate;
VEff = Effective volume of aggregate;
VAC = Volume of total asphalt cement;
Wr = Total weight of mixture;
Vrv = Total volume of voidless mixture;
Pb = Asphalt content;
WAc = Weight of total asphalt cement;
Gb = Specific gravity of asphalt cement; and
Gmm = Maximum specific gravity of mixture.
Maximum Specific Gravity of Mixtures with Different Asphalt Con-

tents-Normally, the maximum specific gravity of the HMA mix. is
obtained at one asphalt content (preferably near the optimum). If the
effective specific gravity of an aggregate is obtained by Equation 3 after
determining the maximum specific gravity of the HMA mix at that asphalt
content, the maximum specific gravity of the HMA mix at other asphalt
. -·--· contents_can_b.e_calculateclas_follo:w:s_Note-that-these-valueS-ai:e-need{:d-to•- - - - - -
calculate the VTM for each asphalt content.
Substituting,
simplifying,
where,
Wagg = Weight of aggregate;
WAc = Weight of total asphalt cement;
VAc = Volume of asphalt cement;
WT= Total weight of mix;
Gse = Effective specific gravity of aggregate;
Gb = Specific gravity of asphalt cement; and
Pb= Asphalt content, percent by weight of mix.
Asphalt Absorption-The percentage of asphalt absorbed by the min-

eral aggregate is usually expressed by the weight of aggregate rather than by
weight of total mix. The equation to calculate asphalt absorption can be
derived as follows. The equation for absorbed asphalt is
pba = ( ; : ) 100
substituting, weight = volume x specific gravity
Pba = (VM;:J) 100
The volume of absorbed asphalt is the difference between the bulk volume
- - - - - - t i :£.-ihe-aggregate-and-the-effecti:v:e-volume.-Hence----- -.--·
Substituting, volume= weight/ specific gravity
simplifying,
where,
Pba = Absorbed asphalt by weight of aggregate;
W MC = Weight of absorbed asphalt;
W Agg = Weight of aggregate;
VMC = Volume of absorbed asphalt;
Vbulk = Bulk volume of aggregate;
Gb = Specific gravity of asphalt cement;
G,e = Effective specific gravity of aggregate; and
G,b = Bulk specific gravity of aggregate.
MARSHALL-MIXTURE DESIGN METHOD

The overall procedure for mixture design always begins with accep-
tance tests performed on the aggregates and asphalt cement considered for
the design. If the acceptance tests on the aggregates and asphalt cement
pass, the procedure included in ASTM D1559, Resistance to Plastic Flow
of Bituminous Mixtures Using the Marshall Apparatus, can be performed
on the mixture. The procedure that follows is appropriate only for paving
grade asphalt cements used with maximum size aggregates not exceeding 1
inch (25 mm). Since the mold diameter to maximum particle size ratio
should exceed at least 4, a I-inch (25 mm) diameter particle is the largest
aggregate size permitted in the 4-inch (101.6 mm) Marshall mold. For sit-
uations where mixtures have particle sizes that exceed 1 inch (25 mm), the
reader is referred to Chapter 7 and the discussion of a modification to the
------ -Marshall-proeeclure-whieh-uses--6-ineh-(-l-5z.-;4-mmJ--molds:-'Fhe-majo,,,____----
steps involved in performing a mixture design begin with source accep-
tance tests in Steps A and B; specimen preparation in Step C; Rice specif-
ic gravity in Step D; stability and flow tests in Step E; density and voids
analysis in Step F; and analyzing test results in Step G.
STEP A. AGGREGATE EVALUATION
A-1. Determine acceptability of aggregate for use in HMA construction;

tests often performed include Los Angeles· abrasion, sulfate soundness,
sand equivalent, presence of deleterious substances, polishing, crushed face
count, and flat and elongated particle count.
A-2. If material is acceptable in Step A-1, then perform other required

aggregate tests: gradation, specific gravity, and absorption.
A-3. Perform blending calculations as described in Chapter 3. As a sug-

gested beginning· point, plot the specification midband gradation on a
FHWA 0.45 power gradation chart and see how close the midband grada-
tion comes to the maximum density line. If it is too close, the VMA is like-
ly to be too low. The target gradation should be adjusted to deviate from
the maximum density line, especially on the No. 8 (2.36 mm) sieve. VMA
increases as the gradation lines move away from the maximum density line
plot either up or down. A good beginning point for estimating percentages
of aggregates in a blend can be made by using the target percentage pass-
ing the No. 8 (2.36 mm) and No. 200 (0.075 mm) sieves and estimating
the quantities of each aggregate required to meet those two percentages. A
full gradation check must then be made to check that the percentages pass-
ing for other sieve sizes are met. Aggregate blending is discussed in detail
in Chapter 3.
A-4. Prepare a specimen weigh-,-out table by multiplying the percent

aggregate retained between sieves times an aggregate weight of approxi-
mately 1150 g, then determine the cumulative weights starting with the
material passing the No. 200 (0.075 mm) sieve.
!STEP B. ASPHALT CEMENT EVALUATIONj
B-1. Determine appropriate asphalt cement grade for type and geograph-
ical location of mixture being designed.
B-2. Verify that specification properties are acceptable.
B-3. Determine asphalt cement specific gravity and plot viscosity data on
a temperature-viscosity plot.
B-4. Determine the ranges of mixing and compaction temperit~res from

the temperature-viscosity plot:
1. Mixing temperature should be selected to provide a viscosity o.f 170

± 20 centistokes. ,__,
2. Compaction temperature should be selected to provide a viscosity
of 280 ± 30 centistokes.
!STEP C. PREPARATION OF MARSHALL SPECIMENS!
Prepare Marshall Specimens using the procedures given in Steps C-1 thru
C-7 for individual specimens. As an alternate, HMA for a desired asphalt
content can be prepared in bulk to produce the required number of com-
pacted Marshall specimens and to provide a sample for determination of
maximum (Rice) specific gravity.
C-1. Dry and sieve aggregates into sizes (preferably individual sizes) and
store in dean sealable containers. Separate enough material to make 18
specimens of approximately 1150 g each. Minimum aggregate and asphalt
cement requirements to prepare one series of test specimens of a given gra-
dation are 50 pounds (25 kg) and one gallon (4 liters), respectively.
C-2. Weigh out aggregate for 18 specimens placing each in a separate c;:on-
tainer and heat to mixing temperature determined in Step B-4. However,
the total aggregate weight should be determined as discussed in Step C-3.
C-3. It is generally desirable to prepare a trial specimen prior to preparing

all aggregate batches. Measure the height of the trial specimen (h 1) and
check against height requirements for Marshall specimens: 2.50 inch (63.5
mm) ± 0.20 inch (5.1 mm). If the specimen is outside this range, adjust
quantity of aggregate included in a specimen using the following formula:
where,
Q = weight of aggregate to produce a specimen 2.5 inches

(63.5 mm) tall, g; and
hi = height of trial specimen, inches (mm).
C-4. Heat sufficient asphalt cement to prepare a total of 18 specimens.

Three compacted specimens each should be prepared at five different
asphalt contents. Asphalt contents should be selected at 0.5 percent incre-
ments with at least two asphalt contents above "optimum" and at least two
below "optimum." See appropriate specifications for a guide on approxi-
mate "optimum" asphalt content or the estimate of optimum can be based
on experience. Three loose mixture specimens should be made near the
optimum asphalt content to measure Rice specific gravity or theoretical
maximum density (TMD). (Note: Some agencies require the Rice specific
gravity test at all 5 asphalt contents. However, the precision of the test is
best when the mixture is close to the optimum asphalt content. It is prefer-
able to measure the Rice specific gravity in triplicate. After averaging the
results from these three tests and calculating the effective specific gravity of
the aggregate, the maximum specific gravity for the other four asphalt con-
tents can be calculated using the formula given earlier.)
C-5. Review appropriate specifications to determine number of blows/side

and type of compaction equipment required for compaction of Marshall
specimens.
C-6. Remove the hot aggregate, place it on a scale, and add the proper
weight of asphalt cement to obtain the desired asphalt content.
C-7. Mix asphalt cement and aggregate until all the aggregate is coated. It
is helpful to work on a heated table. Mixing can be by hand, but a mechan-
ical mixer is preferred.
C-8. Check temperature of freshly mixed material; if it is above the com-

paction temperature, allow it to cool to compaction temperature; if it is
below compaction temperature, discard the material and make a new mix.
----------1-Note:-Some-st-a:tes-per-mi:t-mixes-te-be-plaeed-baek-in the oven to be reheat-
ed; some states use an oven curing time to better simulate what happens to
the HMA mix in the field. This curing time is especially important for
aggregates with high absorption since the asphalt absorbed into the aggre-
gate increases with time.)
C-9. Place a paper disc into an assembled, preheated Marshall mold and
pour in loose HMA. Check the temperature. Spade the mixture with a
heated spatula or trowel 15 times around the perimeter and 10 times over
the interior. Remove the collar and mound materials inside the mold so
that the middle is slightly higher than the edges. Attach the mold and base
plate to the pedestal. Place the preheated hammer into the mold, ~d apply
the appropriate number of blows to the top side of the specimen.
C-10. Remove the mold from the base plate. Place a paper disc on top of
the specimen and rotate the mold 180° so that the top surface is ori bot-
tom. Replace the mold collar and attach the mold and base plate to the
pedestal. Place the hammer in the mold _and apply the same number of
blows to the opposite side of the specimen~ .
C-11. Remove the paper filters from the top and bottom of specimens.
Cool the specimens and extrude from the mold using a hydraulic jack.
Place identification marks on each specimen with an alphanumeric code
using a grease pencil. Allow specimens to sit at room temperature
overnight before further testing.
C-12. Determine the bulk specific gravity for each specimen by weighing
in air. Submerge the samples in water and ailow to saturate prior to getting
submerged weight in SSD condition. Remove the sample and weigh in air
in saturated surface dry condition. This test is conducted in accordance
with AASHTO T166.
C-13. Measure the Rice specific gravity on the loose HMA mi?c samples in
accordance with AASHTO T209 (ASTM D2041). This specific gravity is
required for voids analysis. If the mix contains absorptive aggregate it is
recommended to place the loose mix in an oven (maintained at the mix
temperature) for 4 hours so that the asphalt cement binder is completely
absorbed by the aggregate prior to testing. Keep the mix in a covered con-
tainer while in the oven. If the test is run in triplicate on the mix contain-
ing near optimum asphalt content, average the three results, calculate the
!: effective specific gravity of the aggregate, and then calculate the Rice spe-
.~___cific gravity _or__.TMD_for__.the-remaining-m.ixes-with-d-iffa.r-e-m-asphal:t-ron-- - - - - -
f tents. If one TMD test is conducted on each mix containing a different
I asphalt content, then calculate the effective specific gravity of aggregate in
1
each case. Calculate the average effective specific gravity, and then calculate
the TMD values using the· average for all five mixes.
r
I Chapter 4-Hot Mix Asphalt Mixture Design Methodology 209
I STEP D. DENSITY AND VOIDS ANALYSIS
Refer to the section on weight volume relationships in this chapter to

complete the calculations in the following steps:
D-1. For each specimen, use the bulk specific gravity (Gmb) from Step
C-12 and the Rice Specific gravity (Gmm) from C-13 to calculate the per-
cent voids or VTM.
VIM = (1 - !:) 100
D-2. Calculate the density of each Marshall specimen as follows:
Densi-ty (g!ml) = Bulk Specific Gravi-ty (Gmt,) x (Ow)
D-3. Calculate the VMA for each Marshall specimen using the bulk spe-
cific gravity of the aggregate (G, 6) from Step A-2, the bulk specific gravity
of the compacted mix (Gm6) from Step C-12, and the asphalt content by
weight of total mix (Pb):
D-4. Calculate the VFA (voids filled with asphalt) for each Marshall spec-
imen using the VTM and VMA as follows:
ISTEP E. MARSHALL STABILITY AND FLOW TEST I

- - - - -....
~l-;-Hea-r-tlre-water-bath-ro-I-40-0-F-E60°GJ-and-place specimens to be test-
ed in the bath for at least 30 but not more than 40 minutes. Place speci-
mens in the bath in a staggered manner to ensure that all specimens have
been heated for the same length of time before testing. Use a waterbath
large enough to hold all specimens prepared for the mixture design.
E-2. After heating for the required amount of time, remove a specimen
from the bath, pat with towel to remove excess water, and _quickly place in ',"-"
the Marshall testing head.
E-3. Bring the loading ram into contact with the testing head. Zero the
pens if using a load-deformation recorder or zero flow gauge, and place the
gauge on the rod of the testing head~ Apply the load at 2 inches/minute
(50.8 mm/minute) until maximum load is reached. When load just begins
to decrease, remove the flow meter, stop ram movement, and record the
stability (maximum load) in lbs (Newtons) and flow in 0.01 inches
(0.25 mm). Testing should be completed within 1 minute from the time
the specimen is removed from the bath.
E-4. Repeat Steps D-2 and D-3 until all specimens have been tested being
careful to check that:
1. The total elapsed time between removing a specimen from the

water bath and the maximum load applied is less than 60 seconds.
2. The total time in the water bath for each set of 3 specimens is the
same and is between 30 and 40 minutes.
STEP F. TABULATING AND PLOTTING TEST RESULTS
F-1. Tabulate the results from testing, correct the stability values for spec-
imen height (ASTM D1559), and calculate the average of each set of 3
specimens.
F-2. Prepare the following plots:
• Asphalt content versus density (or unit weight);

• Asphalt content versus Marshall stability;
• Asphalt content versus flow;
• Asphalt content versus air voids (or VTM);
• Asphalt content versus VMA; and
- - - - --~-Asphalt-content-versus-V-f.A-.- - - - - - - - - - - - - - - - - - - -
F-3. Review the plots for the following trends:
• Stability versus asphalt content can follow two trends:

r Chapter 4-Hot Mix Asphalt Mixture Design Methodology
(a) Stability increases with increasing asphalt content, reaches a

211
peak, and then decreases.

(b) Stability decreases with increasing asphalt content and does not
show a peak. This curve is common for some recycled HMA
mixtures.
• Flow should increase with increasing asphalt content.
• Density increases with increasing asphalt content, reaches a peak,
and then decreases. Peak density usually occurs at a higher asphalt
content than peak stability.
• Percent air voids should decrease with increasing asphalt content.
• Percent VMA decreases with increasing asphalt content, reaches a
minimum, and then increases.
• Percent VFA increases with increasing asphalt content.
lsTEP G;· OPTIMUM ASPHALT CONTENT DETERMINATION!
G-1. The following two methods are commonly used to determine the
optimum asphalt content from the plots:
Method I-NAPA Procedure contained in TAS 14 (11)
1. Determine the asphalt content which corresponds to the spec-

ification's median air void content (4 percent typically). This is the
optimum asphalt content. ,
2. Determine the following properties at this optimum asphalt content
by referring to the plots:
• Marshall stability;
• Flow;
•VMA; and
•VFA.
3. Compare each of these values against the specification values and if
all are within the specification, then the preceding optimum asphalt
content is satisfactory. If any of these properties is outside the spec-
ification range, the mixture should be redesigned.
Method 2-Asphalt Institute Method in MS-2 (1.3.)
1. Determine: (a) Asphalt content at maximum stability

(b) Asphalt content at maximum density
(c) Asphalt content at mid point of specified air void
range (4 percent typically)
2. Average the three asphalt contents selected above.

3. For the average asphalt content, go to the plotted curves and deter-
mine the following properties:
,I
• Stability;
• Flow;
• Air voids; and
•VMA.
4. Compare values from Step 3 with criteria for acceptability given in
I!
Table 4-1 and Figure 4-13.
G-2. Failure to meet the criteria for any property requires that special
action be taken to determine if the mix can be used in construction. Local
experience with a mix which fails to meet'particular criteria may verify that
the mix performs satisfactorily. If that is the case, the criteria ~ay be mod- .
ified only for those mixes that serve satisfactorily under specified traffic and
environments. If the VMA. criteria are not met, the aggregate gradation
must be modified and the mixture redesigned.
G-3. The laboratory developed mixture design forms the basis for the ini-
tial job mix formula QMF). The initial JMF must often be adjusted early
I, in the construction project because of slight differences between the labo-
ratory supplied aggregates and those being used in the field. Often, the
asphalt content developed in the laboratory is too high for field construc-
tion and must be reduced a few tenths of a percent, especially if the VTM
of the produced mix is low. Construction experience during the first few
days should be carefully reviewed and the JMF modified to provide a mix
that can be satisfactorily mixed, placed, and compacted. Ensure that· the
produced mix satisfies all Marshall parameters.
An example of Marshall mixture design is given later in this chapter.
11:
! :
Marshall Criteria
As mentioned earlier, the criteria used by the various states that design
I
mixes using the Marshall method vary considerably. However, in order to
I:
include all the criteria in this discussion, the reader must understand that
....jl.,._':---=n=ot~al=l these criteria ar.e us.e_d_b.y_alLsp.ecify.ing-ageR-G-ies.-GeI1~-al-ly-m-0si:--------
' states include measurement of the following properties in a Marshall

mixture design: a minimum amount of voids in the mineral aggregates
(VMA), a range of acceptable air void contents, a minimum stability, and
a range or minimum flow value. Some states also add a requirement that
---- ------ -~ -- ---- -~ ~
~
()
::,-
~
-t
CT'
n i...,
t
~ ::i:::
~
Marshall Traffic
~
t,:j ~ ,
Method
Light Medium Heavy ~ ~ >
Mix
"' =-
n "' .r,
::,-
Criteria Minimum Maximum Minimum Maximum Minimum Maximum ~~ ~
Compaction,
No. of
blows/side
35 50 75
~
~-~ Iti
l"'I
n ~-"'
:::s
Stability, lb (N) 750 (3333) --- 1200 (5333) --- 1800 (8000) --- t::J ~
n
"'.... 9-
0
Flow ~
0.
g_
0.01 inches 0
(IQ
(0.25 mm) 8 18 8 16 8 14 n....

l"'I
'<
....
n
Air Voids,% 3 5 3 5 3 5 ....
l"'I
~
Voids in 8"'
l"'I
Mineral (Graphically shown in Figure 4-13)
Aggregate Vl
=l"'I
~
n
n
"'
~
....,_.
N
i:i..
Mllllmetres1,2
co
(/J
"l
rn
,.._ U'l U'l
ru
0
oi
0
ui
U'l
.,:
,., 0 ,.,
(\J .,f oi (\J U'l (/J
4◊- 40
30 30
Normally
~i1t 0
l
'\-
Acceptable
VMA
~j~arp°:c1flc
Gravity ot Aggregate
<i.
:i 20
>
E
E
~
Deficient In
Either Asphalt
or Air Voids
10 10
No.16 No.8 No.4 318In. ll2in. 314in. lln. I ½Jn. 2In. 2½Jn.
Sieve Slze 1r2
'standard Speclflcallon lor Wire Clolh Sieves /or Testing Purposes, ASTM Designation Ell
IAASHT0 Deslgnallon M92J.
2For processed aggregate, the nominal maximum particle size Is the largest sieve size listed In. the
applicable speclflcallon upon which any material Is retained.
3Mixture in the I% tolerance band shall be permitted only when experience lndlcales that lhe mix-
ture wlll perform satisfactorily and when all other criteria are met.
Figure 4-13 Relationship between Voids in Mineral Aggregates and

Maximum Sieve Sizes with Criteria Superimposed (13)
the percent of voids filled with asphalt (VFA) be within a specified range.
Each of these criteria is discussed below.
Voids in the Mineral Aggregate. VMA is the total volume of voids

within the mass of the compacted aggregate. This total amount of voids
significantly affects the performance of a mixture because .if the VMA is
-:,, too small, the mix may suffer durability problems, and if the VMA is to.o
5;1 large, the mix may show stability problems and be uneconomical to pro-
duce. The VMA has two co'mponents:the volume of the voids that is f i l l e d ~
-;rrh asphalt-!a'nd the volume of voids remaining after compaction that is
available for thermal expansio~ of ~e asphalt ce~~nt during ~ot weathe(-'
1
---'.I'he-v:olume-0£-asphalt-cement-1s-cF1-t1Gal-fer-0,1:1-Fa-0il+ty-0-:f-1:he-m-B1::1:1:1:re-c-T'l1:1s .
asphalt cement volume along with the aggregate gradation determines the
thickness of the asphalt film around each aggregate particle. Without ade-
quate film thickness, the asphalt cement can be oxidized faster, the films
are more easily penetrated by water, and the tensile strength of the mixture
is adversely affected. Figure 4-13 shows the recommendation for VMA

proposed by the Asphalt Institute (13). Notice that as the maximum par-
ticle size decreases, the minimum VMA increases. This occurs because the
total void space between small particles is greater than that between large
particles. Therefore, the optimum asphalt content for a dense-graded
aggregate mixture with a maximum aggregate size of 2 inches (50 mm)
may be as low as 3.0 to 3.5 percent while for a 3/8 inch (9.5 mm) maxi-
mum aggregate size the asphalt content may be as high as 7.0 to 7.5 per-
cent. The VMA for any given mix must be sufficiently high to ensure that
there is room for the asphalt cement plus the required air voids. Notice that
if a specifying agency has both a VMA and a minimum air void content
criteria in the mix design procedure, they are also specifying VFA, because
VFA is calculated from VMA and air void content. Many specifying agen-
cies rigidly apply VMA specifications. They should recognize that any
VMA specification should be supported by field performance data. Foster
(J.5J has questioned the data from field performance studies conducted by
McLeod in the 1950s on which the minimum VMA recommendations are
based.
Voids in Total Mix (VTM). The VTM in the compacted dense-grad-

ed HMA specimen at optimum asphalt content is suggested by most agen-
cies to lie between 3 and 5 percent. There are a number of reasons for rec-
ommending this void content range. However, this air void content is for
laboratory compacted samples and should not be confused with field com-
pacted samples. This void content must be approached during construc-
tion through the application of compactive effort and not by adding
asphalt cement to fill up the voids. HMA pavement layers transfer the load
from the surface to underlying layers through intergranular contact and
resistance to flow of the binder matrix; therefore, high shear resistance
must be developed in the HMA layers if adequate performance is to be
achieved. This high shear resistance must be present to prevent additional
compaction under traffic which could result in rutting in the wheel paths
or flushing and bleeding of the asphalt cement at the surface.
In addition, the dense-graded HMA wearing course must provide a
surface that is relatively impermeable to both air and water. If the in-place
air void content is only slightly higher than the 3 to 5 percent range, both
the air and wateCfferrrfeabitities-should-be-quite low-because the voids
which are present are not interconnected but rather are isolated individual
voids within the aggregate and asphalt mass. Low air void contents mini-
mize the aging of the asphalt cement films within the aggregate mass and
also minimize the possibility that water can get into the mix, penetrate the
thin asphalt cement film, and strip the asphalt cement off the aggregates.
It is very important that the HMA be compacted to a laboratory density
that approximates the ultimate density achieved under traffic and at the
same time have an air void content in the 3 to 5 percent range. The
in-place air void content should initially be slightly higher than 3 to 5 per-
cent to allow for some additional compaction.
Density. The magnitude of the density achieved during compaction in

the laboratory is not so important. What is important is how close the den-
sity achieved in the laboratory is to the density achieved in the field after
several years of traffic. Remember that in the discussion of the develop-
ment of the Marshall, Hveem, and Supef_pave methods, the developers of
the laboratory compaction methods applied enough compactive effort to
provide laboratory densities which were comparable to field densities
reached for the same materials serving traffic and giving good field perfor-
mance. Many states now specify the field density requirement as a percent
J of the theoretical maximum density, hence density can be achieved by
increased.,:o~ction, by increased as~t content,. by increased f i l l ~
tent, or any method that reduces voids. Tlieretoreit is important how the
I
density is achieved, whether by applying compactive effort to force the par-
ticles closer together or by filling the voids with asphalt cement or filler.
Either of these methods increases density (as a percent of theoretical max-
imum density), but satisfactory compaction effort on a properly designed
mixture produces a mixture with high shear strength, while modifying the
mixture to reduce in-place voids will provide a mixture with low shear
strength with a tendency for high permanent deformation under heavy ver-
tical loads or large acceleration or deceleration forces such as at intersec-
tions.
Since density is usually expressed as a percent of theoretical maximum
J density, it is easy to see that anything that decreases the in-place air voids
will increase the p e ~ t y - Notice that increasing t h ~ t e n t
and the amount of dust in a mix both increase the percent density but do
not necessarily have a positive effect on performance. This is one reason
that compaction of Marhsall specimens using field produced materials is
important. Increased density due to increased compactive effort will
increase shear resistance and improve performance assuming that there is
adequate ·asphal:t-cetn·e:n:ca:va:iia:Me-to prevent durao1hty problems and not
too much asphalt cement to cause permanent deformation problems.
In the Marshall mix design procedure, the density varies with asphalt
content. Density increases initially as the asphalt content increases because
the hot asphalt cement lubricates the particles allowing the compactive
I
r
I
effort to force them closer together. The density reaches a peak and then
begins to decrease because the additional asphalt cement produces thicker
films around the individual aggregates, thereby pushing the aggregate par-
ticles further apart and resulting in lower density.
Stability. Marshall stability is defined as the maximum load carried by

a compacted specimen tested at 140°F (60°C) at a loading rate of 2 inch-
es/ minute (50.8 mm/minute). This stability is generally a measure of the
mass viscosity of the aggregate-asphalt cement mixture and is affected sig-
~ificantly by the angle of internal friction of the aggregate and the visc.£_Si-
ty of the asphalt cement .at 140°F (60°C). Since the viscosity at l40°F
(60°C) is primarily affected by the asphalt cement grade, one of the easiest
ways to increase the stability of an aggregate mixture is to change to a high-
er viscosity grade of asphalt cement. One may also significantly affect the
Marshall stability of a mix by selecting a more angular aggregate; i.e.,
crushed angular aggregates of the same gradation yield higher stability
mixes than do aggregates that are rounded or subrounded. It is possible
then for the designer to specify certain asphalt cement and aggregate char-
acteristics to produce mixes with either higher or lower stabilities.
Anything that increases the viscosity of the asphalt cement increases the
Marshall stability. One may add a small amount of very fine dust which
may have the effect of making the asphalt cement/ dust mixture act as a
more viscous binder thus increasing the Marshall stability. However, if the
dust is extremely fine, it may extend the asphalt cement, making it act like
a higher asphalt content, and lower the stability, depending on the shape
of the Marshall stability versus asphalt content curve. One must then care-
fully and completely analyze the effect of changes in components on the
properties of a HMA material.
Marshall stability and field stability are not necessarily related.
Practitioners often assume that if a stability problem (rutting) occurs in the
field, it can be solved by increasing the minimum acceptable Marshall sta-
bility used in design. Experience in the field has demonstrated that this
solution does not work. The basic premise for such a solution is: Marshall J
stability is a surrogate measure for .field stability. Since the stability of a /
rrrixru:re m the field is affected by the ambient temperature, types of load-J \,,
ing, rate of loading, tire contact. pressure, and num.erous mixture proper-
ties, it is difficult to accept the above premise since the laboratory Marshall
stability only takes mto accounco-ne~subserofthe-conditions mentioned
above. Therefore, if field stability is aproblem, a comprehensive study -._
sno'uld be conducted to determine which of the mixture characteristics
appear to be causing the problem rather than assuming that higher labora-
tory stability will solve the problem.
1
The primary use of Marshall stability is in evaluating the change in sta-
bility with increasing asphalt content to aid in selection of the optimum
asphalt content. The Marshall stability is also useful in measuring the con-
sistency of a plant produced HMA. Because there are large variations in the
stability test values, i.e:, the precision of the test is very poor, ·using the
Marshall stability as a specification requirement must be done carefully. In
fact, ASTM Dl559 contains no precision statement even though attempts
to gather data for a precision statement have been made periodically for
many years.
Flow. The flow is measured at the same time as the Marshall stability.
The flow is equal to the vertical deformation of the sample (measured from
start of loading to the point at which stabil1ty begins to decrease) in hun-
dredths of an inch. High flow values generally indicate a glastic ~ix that
will experience permanent deformation under traffic, whereas low flow val-
~
ues may indicate a mix with higher than normal voids and insufficient
asphalt for_cl!:!I~bility and one that may experience pre~ure crac}ung due
to mixture b r i ~ d g g ~ life of tb.~,~~nt.
For 75-blow Marshall designs that are used on interstates or high vol-
ume roads or airfields, the flow value is usually specified to be in the range
of 8-16. For 50-blow Marshall designs, used on lower traffic volume
roads:the flow is sometimes allowed to be as high as 20.
Percent Voids filled with Asphalt. There are a number of states that
include percent voids filled with asphalt cement (percent VFA) in addition
to the criteria listed above. If a specifying agency includes a VMA require-
ment and exercises air void control during construction, percent VFA is a
redundant requirement for dense-graded HMA. Most states that include
percent VFA requirements generally specify that the VFA range from 7(}JQ_
85 percent.
HVEEM MIXTURE DESIGN METHOD
The basic philosophy of the Hveem method of mix design was sum-
marized by Vallerga and Lovering (1) as containing the following elements:
- - - ·---1-;-lt-should-provide-suffrcienr-asphaltcement for aggregate absorption

and to produce an optimum film of asphalt cement on the aggre-
gate as determined by the surface area method.
2. It should produce a compacted aggregate-asphalt cement mixture
with sufficient stability to resist traffic.
3. It should contain enough asphalt cement for durability from weath-

ering including effects of oxidation and moisture susceptibility.
These elements were combined into a summary statement which forms the
basis for selecting the optimum asphalt content in the Hveem method: Use
a dense, well-graded aggregate with high internal friction without an
excess of fines and add as much asphalt cement as the mixture will tolerate
without losing stability. At least 3 percent air void content is desired in the
Hveem mix design.
Just as with the Marshall Method discussed previously, the procedure
is presented in steps to show how the various elements of the design pro-
cedure fit together. Steps A and B are the same as in the Marshall method
and will not be repeated here; the Centrifuge Kerosene Equivalent test for
the fine aggregate is outlined in Step C; the surface capacity test for the
coarse aggregate is outlined in Step D; the procedure for estimating the
optimum asphalt content is outlined in Step E; stabilometer specimen
preparation is outlined in Step F; calibration of the displacement of the
Hveem stabilometer is contained in Step G; the procedure for the sta-
bilometer test is outlined in Step H; procedures for voids and density
analysis are outlined in Step I; and Step J includes procedures for analyz-
ing Hveem mix design test results to determine the optimum asphalt con-
tent. An example of Hveem mix design is given later in the chapter.
Details of specimen preparation with the California Kneading
Compactor and testing by the Hveem apparatus are given in ASTM
Dl561 and ASTM D1560, respectively.
The purpose of the tests and procedures in Steps C, D, and Eis to esti-
mate the optimum asphalt content at which to compact Hveem test spec-
imens. If previous information on aggregates is available to provide guid-
ance on asphalt content, go directly to step F.
I STEP C. CENTRIFUGE KEROSENE EQUIVALENT

C-1. Calculate the surface area of aggregate gradation selected using the
following table and as discussed in Chapter 3.
C-2. Place 100 g of dry material passing a No. 4 (4.75 mm) sieve in a cen-
_ _ _ _ __.,i.:ifuge_cup_wi.th__s.cr.een_andlilter_pap_er_and weigh (W0 ). _ ..
C-3. Place the cup in kerosene, submerging the aggregate, and leave until
the aggregate is saturated.
C-4. Place the cup in the centrifuge and run at 200 g's for 2 minutes.
Seive Size Percent Passing* Surface Area Factor Surface Area ft2/Jb (m2/kg)
(mm) (SAF) Percent Passing x SAF/ 100
maximum 100 2 (0.4) 2.00
size
No. 4 (4.75) 2 (0.4)
No. 8 (2.36) 4 (0.8)
No. 16 (1.18) 8 (1.6)
No. 30 (o.600) 14 (2.9)
No. 50 (0.300) 30 (6.2)
No. 100 (0.150) 60 (12.4)
No. 200 (0.075) 160 (33.0)
- SA; L ft 2 /lb (m2/kg)
*Percent passing fi:om blended aggregate for selected combination of aggregates.
C-5. Stop the centrifuge, remove the cup and weigh, and record the
weight of the aggregate and absorbed kerosene as Ww. Calculate CKE as
percentage of kerosene retained by aggregate:
C-6. Obtain specific gravity (SG) of fine aggregate from Step A. If the SG
is greater than 2.70 or less than 2.60, correct the CKE using the following
relationship:
' SG
CKECORRECTED = CKE X --
2.6S
I STEP D. SURFACE CAPACITY OF COARSE AGGREGATE I

D-1. Place 100 g of dry aggregate retained between the No. 4 (4.75 mm)
and 3/8-inch (9.5 mm) sieves in a metal funnel and weigh, Wn.
---"""'D-2._Immerse the funnel in SAE_N_o_._l0-1ub.ricacing-oiLat-toom-temper.a-~------

ture for 5 minutes.
D-3. Drain at room temperature for 2 minutes. Place the funnel and sam-
ple in an oven and drain at 140°F (60°C) for 15 minutes.
,...--
i
p--4. Remove from the oven, cool, and weigh (Ww). Calculate the amount
of oil retained as a percent of dry aggregate:
Ww- WD
Percent Oil Retained = - - - - x 100
WD
D-5. Obtain coarse aggregate specific gravity (SGcoursJ from Step A. If the
SG is greater than 2.70 or less than 2.60, correct percent oil retained using
the following relationship:
Percent Oil Reta(nedcorrecud = Percent Oil Retained X SG

2.65
STEP E. ESTIMATED OPTIMUM ASPHALT CONTENT
E-1. Using the CKE from Step C-6, enter the chart in Figure 4-14 to
determine the surface constant, Kr, for fine material.
E-2. Using the percentage oil retained from Step D-5, enter the chart in
Figure 4-15 to determine the surface constant, Kc, for coarse material.
E-3. Using the surface area of aggregate (Step C-:-1), the percent coarse
aggregate (percent retained above No. 4 sieve), and the absolute value of
K:-Kf (Steps E-2 & E-1), determine the correction to Kr from the chart
in Figure 4-16.
E-4. Calculate K.n as follows:
K.n = Kr + Correction to Kr
Correction to Kr is:
negative if Kc-Kr is negative
positive if Kc-Kr is positive
E-5. Calculate the Approximate BituI?en Ratio (ABR) for a mix that uses
cutbacks, i.e., an oil mix (cutbacks· RC-250, MC-250 and SC-250).
~-----Note:..A-cor:rection-wi1Lbe...made.Jor_the_viscosicy-_o£paving grade asphalt
cements. Using SA from Step C-1, Average Aggregate SG from Step A, Km
from Step E-4, and the chart in Figure 4-17, determine oil ratio if oil was
a cutback. If oil is a paving grade asphalt, go to Step E-6 for a viscosity cor-
rection, otherwise stop.
CHART FOR DETERMINING Kt FROM C.K.E.
~,,, 3,0
' '~ ' ',,, ~" ,~,

~
~~" UI
'~ '~ ' ' '

~ '2.0
;z
,'I-A ~
l.ll
'~ 1
0
'~ll'>_, ,....,~
~<i> .
u
u
~~~ ' ~--:-.,.-.
~,~~ l'>I'"
1.0
.9 ~
cil
.8
I' "~'!, N. ,~r-.~~
r,.. ~ .7
r-.,••'l ,, r-;
r,<"o ~,,,r,.. I:: ' .6
.ll
20
1B '-
16 '
tj 14
, 12
a.
~ 10
,2 9
.., B
7
~
s
6
ll
4
~ 3
C.K.E. Corrected ~ C.K.E. • 5P, gr, llne

2.65
NOTE• Do not confwae !his correction
lo C.K.E. wllb Iba! used In Fig, IV-5
~-----m•'------ff-&---------------------7-------
•SUrface area, kg • 0.204816 T Courfuy of Calfornla Deparfmenl of Tr-portallon.
Figure 4-14. Chart for Determining Surface Constant for Fine

Material, Kt, from C.K.E., Hveem Method of Design
3.0
/
2.8
2.6
/
/
2.4
2.2 /
,:22.0
/
1.8
/
I-
z<( /,,
1.6
I-
en
z
0 1.4
u V
/
w
u 1.2 /
/
<(
LL.
a::
::>
en 1.0
V
.8
/
/
/
1.5 2 3 4 5 6 7 8 9
PER CENT OIL RETAINED - CORRECTED FOR SP. GR. OF AGGREGATE
Material Used• .Aggregate - Passing 9.5 mm (3/8•>, Rel 4.75mm (#4) Sieve
Oil - SAE 10
% Oil Rel. Corrected .. % Oil· Rel. " sp. gr. of Coarse Aggregate
2.65
Courtesy on;aaforntaDeparlment -01 Tran1portatton.
Figure 4-15. Chart for Determining Surface Constant for Coarse

Material, Kc, from Coarse Aggregate Absorption,
Hveem Method of Design
CHART FOR COMBINING Kr AND Kc TO DETERMINE Km
II <Kc - Kr> 11 neg., corr. la neg.

II <Kc - Kr> I• pos., corr. Is pos.
0.8
Km • Kt • c~rr. to Kr
0.:5
0.4
~
0.3
.2
C
(.0 ,g
1i5 0.2 f"
st. a
-u
0
C\J
0
..E I"'
.,., OJ
300 ...+-+--.l--!-,-.+-.+-,,-+,-+--.-1,
200 f-+-+--"tr-l--\-+-'~...+--1-1-\....JI-I
i0.
;;: 100~-1--1--14-+--Y--li
,.. 90·1-'--~l-l-++-4--l
3 eol-'---1-1-H+-:-!\
! 70•1----+......l--l+-
f eo,1-----1-1--1---+-+-'l-+-++-.,....+--r-++-1
"'
8' 50 l-'----'\-1--l't-+-.li+-'r-\l--\-'f-l-'H
: 401----+-I---H--l-_....._ _ _1-+l+•U
.
ls
'° 30
]
~
-Cour-tesy-ot-Calltomla-Departmenl-oHranspo)'latto
Figure 4-16. Chart for Combining Kr and K.: to Determine Surface

Constant for Combined Aggregate, K,,,, Hveem Method
of Design
-,
"'!j
Q
1't :i
l 100
D
I
.g
!'i
....
:-,l
.: 90
~
.S
i
BO
70
60 /LI
,
, ,, ,,
I
/
CHART FOR COMPUTING APPROXIMATE
BITUMEN RA TIO CABRI FOR DENSE
GRADED BITUMINOUS MIXTURES
t
~
• ,.:J~LY
I' I'/ ~
~Q ; 50 'l,!6
v ><
gs //"'?
~/'II~
140 _.,..,l,,,j
.i-"17:~ ►
~
::r'
§ 30
~
g· 8"> j
r/ ~ ~ ,I V
7, V //
;~
i:: "'I
"'s= vv
C':) ~ ~ [/ 1/V
a·! 15
~~ [/ V
I/
v v 8
c-:.l· p«~ ~ '.t-rf.i.':',1!1!".. ti
~
~-
a a.=
20
10
I/ ~-:,,~~~.~":.-~
I/[/ [/ I5
~:!:=t =~:.1~r:1,t;1 ~
w
::c: (Jq I/[/ ,_
lfnewllh1calo8•ollrcllo. e-
0
~ 0
l'I)
9 -
....
I/
) .
'M('_,,,
~p
~
-
,,. ~-
V..,.. -::;; ;.,
....,..,
_....._
10
9
8
Can 2.. OIYen aurfaoe area,
of aggrogala.
Flnd
1p.
mace - .. ,calo D, ProCNd

gr. and km C 100
90
80
:
~ I
g,
0
~
.,,,. v n )_,,, ~,..... ..-:::v ,.,,,. 7 ~
::."';~::,~~- ~~...,::,.1~. 70
;.~~s·
~f,:;: ~
.,,,,.
V ...... ::::::,,I.,' :,,:v V / .,,,,. 6,g
5
f:~::L"'tl:, loolKfdtil.hen horlzonlallr
60
!
~
.,,,,.
f:::C: ~;:::: t:-t:::: ~v vi:,,,,-
0:
Is
0Q ratio • lbs. of oll 1,er I00 lb.- 40
!
0 4 ~
8"> :::::~ [::: ~~ ~,.,,,.v- Vv
::;;-i;.,
ri :l,9f!l !1 ~~2~g '/.\'c~;~lyandt
0
,:l..
~--i.,,vv
0 "'I
~
t:::- t:= V t::J-6 V
3
:~-!~ ro~ct~·:-1t~~r.r•• 30
f.
~~
'"">
povlng osphalls. Fig. IV 6
t::, t::, _v V vv
fl l'I)
.... = 2 V v
l•J 1-J 8 2
20
!
~ 7 1.5 1.0 0.5 0.5
I
OJ AOJ 0.5 0.5
m2 112
~
~
'" !11 rl• IlI I II Ij 11I 11I I I I I I I I •I I I • l l I • I • E (Corr- I ~~r• kg •
I
"'ib'"
l
I O........., •St&rface Area, 0.204816
1.8 1,8 2.0 2.2 2.4 2.8 2.8 3.0 3J!
c 10
Sp, Gr, of Aggrogqfa
••Kg. al oil' per 45.36 kg, N
,:l..
~
CHART FOR CORRECTING BITUMEN
Jorj
I N
N
1'
n
REQUIREMENT DUE TO INCREASING VISCOSITY
•When viscosity grades of aslj!1alt cement

°'
t (I) 121
In Tables I and 2 of AASH O Sfceclflca-
lion M 226 and ASTM 8/eclficat on
"""'
po Llqllld Paving Grades CAR)
D 3381 ore used llnsteo of AR grades
In Tobie 3l, substitutions on Scale A may
D Asphalts
be made as follows• rl
o~n 2 .----:::i
8 ~ 8
0 4
~ 2~ ~
0
n =r
'"I') 0 A AC 2.5 for AR 1000
t, ~ e; (\I m !!! AC 5 for AR 2000
10 for AR 4000 t-13
;, el ... 3
Olis and Asphalts•
AC
AC 20 for AR 8000
~
=0
... !:!l'I
Arr~1e• according lo Absolute Viscosity
AC 40 for AR 16000
aq '"'
Sn l
~
4 0) S MC or RC
121 A • Asphalt Residue
12
:I:
"' 0 ::,
er ~
n '"' ILi 5 II
g-;
0 '"'
"' n
... .,
~ e.
C
lo .f
;g 8
;,.,
►
o· =
'"l')Cl(l
e~
~l,
8 10
l~
"'
'Cl
::,'
g, PROCEDURE
~
'O ... ::, 8
CJ)
7 9 i~ .,~
E-.o
[i
"' txi
,.. n
:61
~! 8
BY MEANS OF A STRAIGHTEDGE
CONNECT THE POINT ON SCALE AF
WHICH REPRESENTS THE GRADE 0
!/
8
:E-!
l!ID
t
F
ct
::i.
>
n= ... BITUMEN TO BE USED WITH THE
POINT ON SCALE B REPRESENTING
§
n
a ,.c~ ""
:§ 9
THE SURFACE AREA OF THE AGGRE-
GA TE. THROUGH THE POINT OF
7 ~2
g 0:: INTERSECTION ON LINE C PLACE A ...
C>
~ ij, 5 10
STRAIGHTEDGE CONNECTING WITH
THE PREVJOUSLY DETERMINED OIL B C1
C>
::C: f!l.
;a an RATIO VALUE ON SCALED, THE
INTERSECTION OF THE STRAIGHT
(IQ
.,::,::,
aa II EDGE WITH SCALE E REPRESENTS
THE BITUMEN RA TIO CORRECTED
FOR VISCOSITY OF THE BITUMEN.
~ B
5
p..
(')
~ t, 4 0
12 ::,
E "'
n i::
S- n m2 112
..Surface area,-• 0,204816b
a'
0
0 ... kg I ,:t.
i:i..o 0
::,
r
i
E-6. Correct for viscosity of paving grade asphalt cement. Using asphalt
cement viscosity, surface area of aggregate, oil ratio from E-5, and chart in
Figure 4-18, determine the viscosity corrected Bitumen Ratio.
E-7. Bitumen Ratio is lbs (kg) of oil per 100 lbs (kg) of aggregate.
Therefore, asphalt content (AC) by weight of total mix is calculated as fol-
lows:
0
cerce
nt AC
=-------
Bitumen Ratio
100 + Bitumen Ratio
X 100
ISTEP F. PREPARATION OF STABILOMETER SPECIMENS I

F-1. Compute batch weights using the same procedures as described in
Step C of Marshall mix design as well as adjusting specimen height as
described therein.
F-2. Select mixing temperature as a function of asphalt cement grade using

the following table.
Grade Temperature Range, °F(0 C)

AC-2.5, AR 1000, 200-300 Pen 225-250 (107-121)
AC-5, AR 2000, 120-150 Pen 250-275 (121-135)
AC-10, AR 4000, 85-100 Pen 275-300 (135-149)
AC-20, AR 8000, 60-70 Pen 300-325 (149-163)
AC-40, AR 16000, 40-50 Pen 300-325 (149-163)
F-3. Heat the aggregates and the asphalt cement to the required tempera-
ture, take the aggregate out of the oven, form a crater in the aggregates, and
pour in the required amount of asphalt cement. Note: prepare specimens
at asphalt contents corresponding to:
l. Value determined from CKE procedure;

2. 2 percent above the CKE amount at 0.5 percent increments;
and
---------"\---1-pereertt-belfrw-che-GKE-a-m0um-0y-Q.§-pereent.- -- -
At least one set of specimens should have enough asphalt cement to flush
after compaction. Flushing occurs when asphalt cement is forced to the
outside resulting in a shiny surface.
F-4 Mix aggregates until coated. Maintain the temperature by mixing on

a hot plate; do not overheat. Mixing may be by hand or mechanical mixer;
however, a mechanical mixer is preferred. After mixing, put the mix in a
flat pan, and cure in an oven for 15 hours at 140° ± 5°F (60° ± 3°C).
F-5. Compact the specimen using a kneading compactor. Preheat the

compaction molds to the approximate compaction temperature, and pre-
heat the compactor foot to avoid sticking of the mix. Use a variable trans-
former to adjust heat on the compactor foot.
F-6. Place the mold in a mold holder, place a steel shim under the edge of
the mold, tighten the positioning screw, and insert a paper disc to cover the
base plate.
F-7. Spread the preheated mix in the feeder trough, transfer halfof the mix
into a compaction mold and rod the mix 20 times in center of the mold
and 20 -times around the edge using a bullet nose steel rod. Add remainder
of the mix and rod as before.
F-8. Place the mold assembly into position in a mechanical compactor and
apply semi-compaction tamping blows as follows:
1. Apply 10 to 50 blows, typically 20 is sufficient, at a foot pressure

of 250 psi (1725 kPa) to precompact the mix to support the full
foot pressure during compaction.
2. If the mix moves out from under the foot and is not densifying,
apply a 40,000 lb. (178 kN) static load using the double plunger
method (i.e., base plate is not touching bottom of load platen).
Apply the load at 0.05 inches/minute (1.3 mm/minute) and hold
for 30 ± 5 seconds.
F-9. Remove the shim, release the mold tightening screw, increase the foot
pressure to 500 psi (3450 kPa), and apply 150 tamping blows. Observe-the
surface of each specimen and note each specimen that flushes. (Continue
adding asphalt cement until at least 1 specimen has a moderate or heavily
flushed surface.) Flushing can be ranked as follows:
~-------1-;-Slighra-if-surface-has-a-s-I-ighrsheen.
2. Moderate-if paper sticks to the surface but no surface distortion is
observed.
3. Heavy-if there is free asphalt cement puddled on the surface or if
surface distortion is observed.
Chapter 4-Hot Mix Asphalt Mixture. Design Methodology 229
F-10. Place the mold and specimen in an oven at 140°F (60°C) for 1½
hours; then apply a levelling load of 1000 psi (6900 kPa) at a rate of 0.25
inches/minute (6.4 mm/minute), releasing immediately once 1000 psi
(6900 kPa) is reached. Allow the specimen to cool in the mold.
I STEP G. CALIBRATION OF HVEEM STABILOMETER I

Variations in displacement are caused by air within the fluid and the fail-
ure of the diaphragm to fit tightly against the specimen after the initial
5 psi (34.5 kPa) pressure is applied. This calibration procedure is provided
to account for these variations.
G-1. Turn the stabilometer pump handle counterclockwise several turns to

retract the rubber diaphragm, place the stabilometer on its. base, secure the
set screw, and put the initial displacement cylinder ("dummy specimen")
into the stabilometer cell.
G-2. Place the stabilometer on a compression machine and apply 1000 lbs
(4.45 kN) to the top of a dummy specimen.
G-3. Open the fluid valve and rotate the pump handle clockwise until the
pressure is 5 psi (34.5 kPa) (tap gauge gently to ensure that the needle valve
is operating properly). NOTE: If the pressure exceeds 5 psi (34.5 kPa),
reduce the pressure by turning counterclockwise to below 5 psi (34.5 kPa)
and then turning clockwise to achieve 5 psi (34.5 kPa). Adjust the dis-
placement dial gauge to zero.
G-4. Rotate the pump handle clockwise in a continuous motion until the
pressure increases from 5 to 100 psi (34.5 to 690 kPa). Read the dial indi-
cator and record the initial displacement (0.1 inch (2.5 mm) dial reading
= 1 turn displacement). ·
G-5. Standard displacement = 2.00 ± 0.05 turns of displacement. If the

measured displacement is outside the tolerances, adjust air inside the sta-
------1~-ilometer:
I. If the displacement from G-4 > 2.05 turns, remove some air from
the system using the bleeder valve.
2. If the displacement from G-4 < 1.95 turns, add some air to the sys-
tem using the rubber bulb provided for this purpose.
230 Hot Mix Asphalt Materials,. Mixture Design and Construction
G-6. Repeat Steps G-2 through G-5 until the displacement is within tol-
erances.
STEP H. HVEEM STABILOMETER TEST
H-1. Recheck the initial displacement of the stabilometer before each

series of tests.
H-2. Place specimens in molds in an oven at 140°F (60°C) for at least 1

hour before testing.
H-3. Set the compression machine head speed at 0.05 inch/minute (1.3
mm/minute).
H-4. Place the stabilometer on a base and adjust the stabilometer base so
that the distance from the bottom of the upper tapered ring on the sta-
bilometer to the top of the base is 3.5 inches (88.9 mm).
H-5. Remove the mold and specimen from the oven and place over the
top of the stabilometer. Using the plunger, hand lever, and fulcrum, force
the specimen from the mold into the stabilometer. Be sure the specimen is
in straight and firmly seated. Place the follower on top of the specimen and
position the entire assembly in a compression machine for testing.
H-6. Using the displacement pump, raise the lateral pressure on the spec-
imen to 5 psi (34.5 kPa) initial pressure, tap the needle valve to ensure an
accurate reading, and then close the displacement pump valve.
H-7. Apply the test load at a head speed of 0.05 inch/minute (1.3
mm/minute). Record the test gauge readings at 500, 1000, and at 1000 lb
(2.22, 4.45, and at 4.45 kN) increments up to 6000 lbs (26.7 kN).
Immediately reduce the load to 1000 lbs (4.45 kN) as soon as 6000 lbs
(26. 7 kN) is reached.
H-8. Open the displacement pump valve and reset the pressure to 5 psi
- - - -·(ignore-the readmg om~he compression tester). Set ilie Ames dial gauge on ~----=c-------
the pump to zero using'the thumbscrew.
H-9. Turn the displacement pump handle clockwise rapidly (at 2 revolu-
tions/second) until the pressure test gauge reads 100 psi (690 kPa). Read
r
the Ames dial, record in 0.01 inch (0.25 mm) units (0.1 inch (2.5 mm)
equals 1 revolution or turn of the crank handle) as the displacement of the
specimen, D 2 •
H-10. Remove the compressive load, reduce the pressure on the test gauge
to zero, back the displacement pump past zero 3 turns counterclockw-ise,
remove the specimen from the stabilometer, and measure the height< in
inches.
H-11. Calculate the stabilometer value (S) as follows:
22.2
where,
Pv = vertical pressure (400 psi);

Ph =horizontal pressure gauge reading @ 400 psi
compressive load; and
D 2 = specimen displacement, from Step H-9.
If the specimen height is between 2.4 and 2.59 inches (61.0 and 65.8 mm),
no height correction is necessary. Otherwise, correct S for specimen height
using the chart in Figure 4-19.
H-12. Repeat Steps H-5 through H-11 until all specimens have been
tested.
STEP I. DENSITY AND VOIDS ANALYSIS
_ _ _ _ _l~-~l._D_e_tenni.n.e_the_hulk.sp_ecifi.c_gra..vig,_o£all.sp_ecimens_ and the Rice spe-

cific gravity of all mixes as per Step C of the Marshall mix design method.
1-2. Calculate the percent air voids (or VTM) and density of the compact-
ed specimens as per Step E of the Marshall mix design method.
CHART FOR CORRECTING ST ABILOMETER VALUES

TO SPECIMEN HEIGHT OF 64 mm (2.50 in.>
Height correction should be made using the table and
chart below.
Example• Overall height of 69 mm (2.74 in.) select
correctron curve "B". Stabilometer value uncorrected
= 35 Stabllometer value corrected = 38.
Overall Specimen Ht. "- Correction Curve
71 mm lo 76mm (2.80In. lo 3.00in.l A
66mm lo 70mm 12.60in. to 2. 79in.) B
61 mm to 65mm (2.40in. to 2.59in.) C
56mm lo 60mm 12.201n. to 2.391n) D
51 mm to 55mm 12.00ln. to 2.l 9in.) E
A B C D E
50
--g
0
CD
L
40
L
0 II
U 30
CD
:,
0
>
L
CD 20
"i
E
..2
:a
0
ci5 10
-0 - -- - - --
0 IO 20 30 40 50 60
Stabilometer Value Before Height Correction
Figure 4-19. Chart for Correcting Stabilometer Values to Effective
Specimen Height of 2.5 inches (64 mm)
STEP J. ANALYSIS OF TEST RESULTS
J-1. Tabulate results from testing and prepare plots of:

• Hveem stability versus asphalt content;
• Density versus asphalt content; and
• Percent air voids versus asphalt content.
J-2. Review trends of each plot for normal behavior:

• Hveem stability should decrease with increasing asphalt content.
This plot will not be similar to that from the Marshall test; in fact,
the plot may vary from being quite flat to quite steep.
• Density will generally increase with increasing asphalt content in a
dome shaped relationship; the curve may, however, not reach a peak.
• Percent air voids should decrease with increasing asphalt content.
J-3. This step is intended to determine optimum asphalt content.

Remember that the philosophy of the Hveem method of mix design is to
select the highest asphalt content to get the highest durability without
exceeding the point of minimum allowable stability. The "pyramid"
method of selecting a design asphalt content is based on this principle.
Construct a pyramid of boxes making the number of boxes in the base
equal to the number of asphalt contents at which specimens were prepared.
At each subsequent level reduce the number of boxes by 1. Determine the
optimum asphalt content using the following procedure:
1. Write the design series asphalt contents in each box with the lowest
on the left.
2. Referring to observations made in Step F-9, mark out the asphalt
contents in which moderate to heave surface flushing was observed.
3. Write the 3 highest remaining asphalt contents in the second level
of blocks in the pyramid.
4. For these 3 asphalt contents, check the Hveem stability and mark
out any which do not meet minimum stability.
------~S-:-Seleet-the-2-highest-asph-alt-content"S-that-meet-minimum stability
and record in the third level of the pyramid.
6. Referring to the plots prepared in Step J-1, select the highest
asphalt content that has at least 4 percent air voids and enter at the
top of the pyramid. This is the optimum asphalt content.
J-4. The laboratory developed mixture design forms the basis for the ini-
tial job mix formula QMF). The initial JMF must often be adjusted early
in the construction project because of slight differences between the labo-
ratory supplied aggregates and those being used in the field. The construc-
tion experience during the first few days should be carefully reviewed and
the JMF modified to provide a mixture that can be adequately mixed,
placed, and compacted.
SUPERPAVE MIX DESIGN METHOD
Background
The Hveem and Marshall methods of mix design have been used since
the 1940s and 1950s. These mix design methods have performed well for
many years but with more traffic and heavier loads it was determined in .
the early 1980s that an improved method of mix design was needed. A
method was needed that could be used to rationally design mixtures for
various traffic volumes, axle loads and environments. It was hoped that this
new mix design procedure would provide tests and models that could use
these test results to predict rutting, thermal cracking, and fatigue cracking.
In 1988, the Strategic Highway Research Program (SHRP) was initi-
ated with a primary goal of developing an improved mix design procedure.
(Z, 16, lZ ,18,) The SHRP program was completed in 1993. The effort
consisted of hundreds of researchers working on various parts of the
research program and it took a great deal of effort to bring all of the results
to a cohesive solution in five years. Even though the volumetric portion of
the mix design procedure is essentially complete, a lot of additional work
is on-going to verify portions of the procedure and to make modifications
where needed.
As directed the study primarily focused .on asphalt binder evaluation,
mix design, and mixture analysis. However, no part of the research effort
looked specifically at aggregates. Certainly, the mixture part of the study
included aggregates but there was no effort to evaluate aggregate separate
from the other mixture components.
Since aggregates are a very important part of the mixture the SHRP
----±•1esearc:hers-r:eali.zed-t:hat-·S0me-g1:1idanee-m-t1st-be-provided-:-'Fhe-g1:1:ida:neP------
that was eventually provided was based on a consensus approach by a
group of experts on aggregates. Recommendations were made for desired
aggregate properties for various traffic levels and for various depths below
the surface of the pavement.
At the conclusion of the SHRP Program in 1993, the resulting system

contained the following elements: new grading system for asphalt binder
(performance graded (_!'G) grading system discussed in Chapter 2), con-
sensus properti~ggregate, new_ mi:x:_a~8-ig~-pro-cedure, _ana:-~
analysisjirocedg.i:es.-This. system -is referred to as the Superior Perfor~ing
&phalt Pavement System (Superpave). After some fine tuning, much of
this information is being evaluated for rapid implementation by State
Departments of Transportation (DOTs). However, it will likely be some-
time before the mixture analysis procedures are finalized and published for
adoption.
Materials
The Superpave mix design procedures provide guidance on selection

of aggregates and asphalt binder. The asphalt binder selection has been dis-
cussed earlier in Chapter 2 and will not be discussed here. A discussion of
the aggregate consensus properties as recomme~ed in Superpave is pro-
vided below.
The aggregate properties that are specified as a result of the SHRP---p-ro-
grarn are the coarse and fine aggregate angularity, flat and elongated parti-
cles, and sand equivalent results. These tests that were adopted by SHRP
were primarily chosen from existing tests. Some of these test procedures
were already used by many state DOTs to help specify desired aggregate
quality. In some cases, Superpave modified the test procedures for aggre-
gates and in other cases the criteria was modified. Specific recommenda-
tions have been made by Superpave for all of these tests.
Specific guidance was provided for angularity of coarse and fine aggre-
gate. The angularity for coarse aggregate is determined by measuring the
percentage of coarse aggregate particles with fractured faces. A sample of
coarse aggregate is divided into 3 stacks. One stack contains all the parti-
cles with O fractured faces. The second stack contains all particles with 1
fractured face, and the third stack contains all particles with 2 or more frac-
tured faces. The percentage with one or more fractured faces and with two
or more fractured faces is then determined. One test meth9d that is often
used for this test is Pennsylvania DOT's test method No. 621
"Determining the Percentage of Crushed Fragments in Gravel". The
Superpave recommended criteria for coarse aggregate angularity is shown
. in Table 4-2.
Table 4--2. Requirements for Coarse Aggregate Angularity"'
Traffic Depth from Surface

ESALs (millions) < 100 mm > 100 mm
< 0.3 55/- -/-
<1 651- -/-
<3 75/- 50/-
< 10 85/80 60/-
< 30 95/90 80/75
< 100 100/100 95/90
> 100 100/100 100/100
*First number is minimum requirement for one or more fracrured faces and second number is mini-
mum requirement for two or more fractured faces.
The numbers shown in Table 4-2 are numbers that were developed
through concensus testing with a number of experts. These values are
believed to be acceptable for most applications of HMA. However, local
experience may indicate that the specifications should be different than
those values shown in Table 4-2. Local experience and materials should
always be considered when preparing project specifications.
The fine aggregate angularity is measured by determining the amount
of voids when the fine aggregate is poured into a cylinder as shown in
Figure 4-20. The fine aggregate is poured into the top end of the cylinder
and the amount of voids in the loose aggregate is determined. The higher
the amount of voids the more angular the aggregate. The test is often
referred to as the National Aggregate Association (NAA) flow test. The test
should be conducted in accordance with AASHTO TP 33 Method A "Test
Method for Uncompacted Void Content of Fine Aggregate (as Influenced
by Particle Shape, Surface Texture, & Grading)". The amount of uncom-
pacted voids in the cylinder can be calculated from the following equation:
I
uncompacte
TT. vot"ds -- V-VW/G,b
dTl'. X 100
~------where-----------
V = Volume of cylinder, milliliters

W = Weight of loose fine aggregate to fill cylinder, grams
G,b = bulk-specific gravity of fine aggregate
Figure 4-20. NAA Flow Test Set-Up
The recommended minimum requirements for the angularity of fine

aggregates for various traffic levels and depth from the surface, are shown
in Table 4-3. The criteria was established for the blended fine aggregate
and was not intended to be a requirement for each fine aggregate in the
blend.
Table 4-3. Requirements for Fine Aggregate Angularity
Traffic Depth from Surface

ESALs (millions) < 100 mm > 100 mm
< 0.3 - -
<1 40 -
<3 40 40
< 10 45 40
< 30 45 40
< 100 45 45
~ 100 45 45
The performance of HMA is related to the angularity of fine and

coarse aggregate. When the materials are more rounded, the shear strength
-------1·:s-redttced.--an:d-rtttting-of the mixture~is-more
~ - 1likelv -- to occur.
- The two pri-
mary sources of rounded aggregates are gravels and natural sands. Data was
developed from a rutting study conducted at NCAT that showed the rela-
tionship between aggregate angularity and rutting (12). This data was
presented in an empirical model as shown below. This model was shown to
be accurate when the air~ ~ . . . r e v ~ t plastic fuw.

This data compares favorably 'to the Superpave requirements shown in
Tables 4-2 and 4-3.
@ Y= (0.080- 0.000089 (CF)- 0.00152 (NAA)) x ✓ESALs

where Y = rut depth in mm
CF = Percentage of coarse aggregate particles with 2 or
more fractured faces
NM = NM flow test results for fine aggregates
Another aggregate property that is important for good performance is

the test to evaluate flat and elongated particles. When aggregates are flat
and elongated, they often tend to lie~flat when compacted resulting in a
mixture with low voids in the mineral aggregate (VMA). Flat and elongat-
ed aggregates also tend to breakdown more during compaction, resulting
in an in-place mixture that may be significantly different from the
designed mixture. In some cases, mixtures with an excessof Rat and elon-
gatecl~are d ~ :;'.'9..ft~~~ ~9-m.p_act. .
In the past aggregates were classified as flat and/or elongated. An
aggregate particle was considered to be elongated if it was significantly
longer than wide and it was considered to be flat if it was significantly
wider than thick. The Superpave method has combined these terms for flat
and elongated particles into one term. The longest dimension of each
aggregate piece is now compared to the smallest dimension of that piece to
get one number referred to as flat or elongated particles. According to
Superpave, an aggregate is considered to be flat or elongated i f ~ J
divi b thicknes exceeds five. The test procedure is provided in ASTM
D4791, "Flat or Elongate Particles in Coarse Aggregate". The criteria
shown in Table 4-4 has been recommended as a part of Superpave.
Table 4-4. Requirements for Flat or Elongated Particles
I Maximum Percentage of
I
TRAFFIC particles with length/thickness ·
I.
!
ESALs (millions) greater than 5
I
< 0.3 -
<1 -
<3 - - ···-~ ~-· -- 10
< 10 10
< 30 10
< 100 10
i
2 100 10
The final aggregate property that was evaluated by consensus analysis

was the clay content. If the amount of clay in a mixture is excessive, a num-
ber of problems can result. The most likely problem is the clay adhering to
the aggregate and preventing a good bond between the asphalt binder and
aggregate. "When this occurs stripping is likely to result in the mixture. To
minimize the possibility that this problem or other problems occur, some
maximum limit for the amount of clay is specified. The test used to mea-
sure the amount of clay is the sand equivalent test. This test can be con-
ducted in accordance with MSHTO T 176 or ASTM D 2419, "Plastic
Fines in Graded Aggregates and Soils by Use of the SancJ. Equivalent Test."
The requirements suggested by Superpave are shown in Table 4-5.
Table 4-5. Requirements for Sand Equivalent
TRAFFIC SAND EQUIVALENT

ESALs (millions) Minimum (Percent)
< 0.3 40
<1 40
<3 40
< 10 45
< 30 45
< 100 50
:2: 100 50
There are other aggregate properties besides the consensus properties

discussed above that are important to ensure that a quality aggregate is
used. Some of these additional properties (source properties)· include LA.
Abrasion, sulfate soundness, and absorption. These test procedures and cri-
teria were not modified by Superpave so they are not discussed in detail
under Superpave mix design.
Selection of Aggregate Blend
The applied loads must ultimately be carried by the aggregate. For sat-
isfactory performance an aggregate meeting the qualityrequirements dis-
-----,crrs~-e<::I-eadier with the proper grading must be used. The gradation of the
aggregate is important to ensure 1) the maximum aggregate size is not too
large or too small, 2) VMA requirements are met, and 3) a satisfactory
aggregate skeleton is obtained.
According to Superpave, the nominal maximum aggregate size is one
IO
N
Cl)
II-- ' ......
' 1·
r.q
N E
..- E
r.q cl
w
Cl)
s0
~
0\ a..
IO
-
II '<I"
~ ci
-
N
CZ)
IO
I'-;
'<I"
w
N
en
-~
w
~0 -
,• !---11-··!
(0
~
>
w
w
z N
(X)
"-:
..-
.,_~
0
0
' " " • , - - -.. MOMH-• (")
0 IO
..-
·•--• IO
r---
'-=!
0
0
0 0 0 0 0 0 0 0 0 0 0
0 Cl) GO r--- (0 IO '<I" (") N ..-
..-
DNISSVd .LN3::J1I3:d
Figure 4-21. Illustration of Gradation Requirements for 19 mm

Nominal Size.
r
sieve size larger than the ~ieve tC> __!e~ain.~(),re than 1_Q__pe~,:L.QLt~e
material. The maximum aggregate size is defined as one sieve larger than
the nominal maximum s:ze (Figure 4-21). If a 19 mm (3/4 inch) maxi-
mum aggregate size is specified, then 100 percent of the aggregate must
pass the 25 mm (1 inch) sieve size. At least 90-100 percent of the aggre-
gate must be finer than the nominal maximum aggregate size (19 mm or
3/4 inch). Less than 90% of the aggregate must pass the 12.5 mm (1/2
inch) sieve. Control points are also set on the 2.36 mm (No. 8) and the
0.075 mm (No. 200) sieve sizes. The aggregate gradation must be within
the control limits to meet the Superpave requirements.
/.
Another part of the grading specification is the restricted zone. This is
the most misunderstood portion of the grading requirements. The restrict-
ed zone is meant to be a guide to help ensure that too much natural sand
is not used in the__mixt.uE._and to help ensure that minimum V1v1A require-·
ments are met. However, there are aggregates that can pass through the
restricted zone that are well crushed and meet the minimum V1v1A require-
ments. Since the restricted zone is only a guide, aggregate blends that pass
through the restricted zone that do not use excessive amounts of rounded
aggregates and that meet the minimum V1v1A requirements should per-
form satisfactorily.
There is evidence to support the belief that mixtures closer to the low
end of the control limits (those going underneath the restricted zone) pro-
vide a better stone structure. Most Superpave designs are performed below
the restricted zone to provide a mixture that is believed to be more resis-
tant to rutting. Typical specifications for 9.5 mm nominal size aggregates
to 37.5 mm nominal size are shown in Tables (4-6 thru 4-10).
Table 4-6. 37.5 mm (1-1/2 inch) Nominal Size

SIEVE SIZE CONTROL POINTS RESTRICTED ZONE BOUNDARY
mm LOWER UPPER LOWER UPPER
50 100.0 100.0 - -
37.5 90.0 100.0 - -
25 - 90.0 - -
19 - - - -
12.5 - - - -
9.5 - - - -
4.75 - - 34.7 34.7
-----
2.% (5:-0 41.0 23.3 27.3
1.18 - - 15.5 21.5
0.60 - - 11.7 15.7
0.30 - - 10 10
0.15 - - - -
0.075 0.0 6.0 - -
Table 4-7. 25 mm (1 inch) Nominal Size

37.5 100.0 100.0 - -
25 90.0 100.0 - -
19 - 90.0 - -
12.5 - - - -
9.5 - - - -
4.75 - - 39.5 39.5
2.36 19.0 45.0 26.8 30.8
1.18 - - 18.1 24.1
0.60 - - 13.6 17.6
0.30 - - 11.4 11.4
0.15
0.075
-
1.0
-
7.0
- -
- -
-
Table 4-8. 19 mm (3/4 inch) Nominal Size

25 100.0 100.0 - -
19 90.0 100.0 - -
12.5 - 90.0 - -
9.5 - - - -
4.75 .- - - -
2.36 23.0 49.0 34.6 34.6
1.18 - - 22.3 28.3
0.60 - - 16.7 20.7
0.30 - - 13.7 13.7
0.15 - - - -
0.075 2.0 8.0 - -
Table 4-9. 12.5 mm (1/2 inch) Nominal Size

19 100.0 100.0 - -
12.5 90.0 100.0 - -
9.5 - 90.0 - -
4.75 - - - -
----· ·- -2.-36· 2-8,0- 5·8:~ :;9_~ 39:':
1.18 - - 25.6 31.6
0.60 - - 19.1 23.1
0.30 - - 15.5 15.5
0.15 - - - -
0.075 2.0 10.0 - -
Table 4-10. 9.5 mm (3/8 inch) Nominal Size

12.5 100.0 100.0 - -
9.5 90.0 100.0 - -
4.75 - 90.0 - -
2.36 32.0 67.0 47.2 47.2
1.18 - - 31.6 37.6
0.60 - - 23.5 27.5
0.30 - - 18.7 18.7
0.15 - - - -
0.075 2.0 10.0 - -
It is desirable to evaluate 2-3 aggregate gradations prior to performing

a complete mix design. This is done by preparing 2 HMA samples for each
of the 2-3 aggregate gradations. The estimated optimum asphalt content
is added to these samples and the samples compacted in the Superpave
gyratory compactor at the required setting. Several items can be evaluated
from compacting these trial samples. Some of these items include: What is
the approximate optimum asphalt content, will the mixture with the aggre-
gate blend being evaluated meet the VMA requirements, and will the gyra-
tory properties such as Ni and Nmax be met? After the initial samples have
been completed and evaluated the best aggregate blend can be selected for
conducting the mix design.
Figure 4-22. Photos of Superpave Gyratory Machines (2 manufacturers).

There are times that the aggregate structure regardless of gradation

l
does not meet requirements for VMA, Ni and other properties. This is a
good indication that the source of aggregate is not acceptable. For exam-
ple, if an aggregate is excessively soft, it will be difficult to impossible to
meet VMA requirements. If an aggregate has an excessive amount of
rounded material it will likely fail the Ni requirements.
Superpave Gyratory Compactor

A key component of the Superpave mix design procedure is the
Superpave gyratory compactor (Figure 4-22). This compaction device has
been designed to compact HMA samples to a density similar to that
obtained in the field under traffic. It also tends to orient the aggregate par-
ticles similar to that observed in the field.
The Superpave gyratory compactor was developed based on evaluatio_n
of existing gyratory machines. Although the Superpave gyratory has some
similarities to other gyrato.ry compactors, it is a unique piece of equipment.
There are three parameters that control the compaction effort on the
Superpave machine. These settings are ve~cal pressure, a~le of gyration,
and number of gyrations. For the Superpave procedure, the vertical pres-
sure is set at 600 kPa (87 psi) and the angle of gyration is set at 1.25°. The
gyrations are applied at a rate of 30 revolutions per minute. The number
of gyrations is varied to simulate traffic level.
The Superpave gyratory compactor can compact samples 150 mm (6
inches) in diameter or 100 mm (4 inches) in diameter. Currently, samples
are specified to be 150 mm (6 inches) but studies are underway to evalu-
ate the possibility of using 100 mm (4 inches) samples for some mix types.
The required number of gyrations is based on traffic level and average
design high air temperature. Mixtures that are exposed to higher tempera-
tures and higher traffic levels in the field will densify more and therefore
must be compacted in the laboratory to a higher density. This higher den-
sity is obtained in the laboratory by increasing the number of gyrations.
The specified number of gyrations for spedfied traffic levels and air
temperatures are shown in Table 4-11. These numbers were developed
with a limited amount of data and additional work is being performed to
verify that these numbers are correct and to simplify the system.
N;, Nd and Nm are three terms that need to be identified. Nd is N-
--design and-is-the number:-ef-gyr-a:ti-ons-required-ro-produce-a-derrsiry-i:rrth------
mix that is equivalent to the expected density in the field after the indicat-
ed amount of traffic. In the mix design process, an asphalt content is select-
ed that will provide 4 percent air voids when the mix is compacted to Nd
gyrations.
Table 4-11. Selection of Number of Gyrations for Superpave

Gyratory Compactor
Design AVERAGE DESIGN HIGH AIR TEMPERATURE

ESALs < 39°C 39-40°C 41-42°C 43-44°C
(millions) .N; Nd Nm N;· Nd Nm N; Nd Nm N, Nd Nm
<0.3 7 68 104 7 74 114 7 78 121 7 82 127
<1 7 76 117 7 83 129 7 88 138 8 93 146
<3 7 86 134 8 95 150 8 100 158 8 105 167
<10 il 96 152 8 106 169 8 113 181 9 119 192
<30 8 109 174 9 121 195 9 128 208' 9 135 220
<100 9 126 204 9 139 228 9 146 240 10 153 253
>100 9 143 233 10 158 262 10 165 275 10 172 288
Ni is N-initial and is a measure of mixture compactibility. Mixtures

that compact too quickly are believed to be tender during construction and
may be unstable when subjected to traffic. A mix that has 4 percent air
voids at Nd should have at least 11 percent air voids at Ni. Mixtures that
tend to fail this requirement are often finer mixtures and often tend to have
a large amount of natural sand.
Nm is the N-maximum and is the number of gyrations required to pro-
duce a density in the laboratory that should absolutely never be exceeded
in the field. Nd provides an estimate of the ultimate field density. Nm
provides a compacted density with some safety factor to ensure that the
mixture does not densify too much resulting in low in-place voids thus
leading to rutting. The air voids at Nm are required to be at least 2 percent.
Mixtures that have less than 2 percent air voids at Nm are believed to be
more susceptible to rutting than mixtures exceeding 2 percent air voids.
The number of gyrations for Nd was established based on laboratory
and field testing to compare in-place density with laboratory density for
various numbers of gyrations. Once Nd was established for each traffic level
and air temperature, the number of gyrations for Ni and Nm, was estab-
lished based on Nd. The equations for Ni and Nm are shown below:
Ni (NJ0.45
Nm (Nd)l.lO
Compaction of Samples
Samples of HMA are blended and compacted to produce specimen

sizes of 150 mm (6 inches) in diameter and 115 mm tall. These compact-
ed samples are evaluated to determine the volumetric properties as well as
the voids at Ni and Nm.
All samples for mix design and. volumetric determination are required
to be aged at 135°C (275°F) in an oven for 2 hours after mixing and prior
to compaction. This aging is needed to allow absorption of the asphalt
cement into the aggregate pores prior to compaction and cooling. This
absorption produces a mixture that is more representative of the mix in the
field. The sample is spread in a thin layer for the aging process. If the mix
is not allowed to age a lower optimum asphalt content will be selected
which could result in some loss in durability during the life of the pave-
ment.
During the design process all samples are compacted to Nm and the
voids are determined by machine calculations at Ni, Nd and Nm. The sam-
ple height is measured and recorded during the compaction process which
allows determination of mixture volume. 'Fhe sample weight is input ·into
the computer and the sample volume is calculated as it is compacted by
knowing the diameter (150 mm or 6 inches) and the sample height. The
estimated bulk density can be determined from these values. This estimat-
ed density is slightly incorrect, however, due to the surface voids on the
top, bottom, and sides of each specimen being considered as part of the
specimen volume. After each sample has been compacted to Nm it is
removed from the compactor and the actual bulk density is determined by
weighing in air and water. The actual density at Nm is compared to the cal-
culated density at Nm to get a correction factor. The calculated densities at
Ni and Nd are corrected based on this correction factor. For example,
the correction factor (CF) is calculated on samples compacted to Nm
gyrations as:
Actual density
CF=
Calculated density
The actual density at Ni and Nd can then be determined by:
Actual Density CF (Calculated Density)
Loose samples of mixture are also used to determine the theoretical

maximum density (TMD). By knowing the TMD and bulk density the
void level can be determined at the various compaction levels. The TMD
samples should be aged in the same way as the samples for compaction.
Selection of Optimum Asphalt Content
After the aggregates and asphalt binder have been determined to meet
the quality requirements and after the aggregate proportioning necessary to
produce a suitable aggregate structure has been established the optimum

asphalt content is determined. The optimum asphalt content has been
defined by Superpave as the asphalt content that produces 4 percent air
voids at Nd. Of course, this designed mixture must meet the requirements
for Ni and Nm and any other requirements that may be specified. After
designing the mixture it should be evaluated for water susceptibility using
one of the conventional test methods. Superpave currently recommends
that AASHTO T-283 be used.
The first step in selecting the optimum asphalt content is to determine
an estimate of the optimum asphalt content. A good estimate should be
available from the design aggregate blend trials.
After the optimum asphalt content has been estimated prepare three
samples each at 0.5% below estimated optimum, at estimated optimum, at
0.5% above estimated optimum, and at 1.0% above estimated optimum.
Mix each sample and place in the oven at 135°C (275°F) for the required
aging time. Compact each sample up to Nm gyrations based on design traf-
fic and performance temperature.
After each sample has been compacted, remove the sample from the
mold and measure bulk density at Nm. Determine the correction factor
from the actual bulk density and the calculated bulk density and correct
the calculated voids for each sample at Ni and Nd. Average the results for
the corrected percent ofTMD for the three samples and plot on a graph as
shown in Figure 4-23. The percent ofTMD is .determined by dividin the
bulk density by the TMD an multi.e_lying by 100 pe;_S,ent. Remember that
the percentage of air voids for each sample is equal to 100 minus percent
ofTMD.
100
92 ~ 4.S%AC
-6---- 5.0 %AC
-6- 5.S¾AC
84
IO 100 1000
Number of Gyrations
Figure 4-23. Plots for Determination of Optimum Asphalt Content

and Evaluation of Ni and Nm
After all of the data has been plotted as shown in Figure 4-23, pick the
asphalt content that will provide 4 percent voids at Nd. This is done by
drawing a vertical line at Nd and a horizontal line at 96 percentTMD. The
point at which these lines cross is the optimum asphalt content. This will
likely require some interpolation between the actual data points. It is satis-
factory to interpolate but one should not extrapolate. The asphalt content
selected to produce 96 percent ofTMD is the optimum asphalt content.
Now at Ni draw a vertical line up to the optimum asphalt content and then
horizontally over to determine percent of TMD. The percent of TMD
should be less than 89 percent. Now do the same thing at Nm. The num-
ber selected should be less than 98 percent. The requirements for Ni and
Nm are met for this example (Figure 4-23). Remember that the TMD is
different for different asphalt contents. \X&en the TMD is known at one
asphalt content, it can be calculated at, another asphalt content by deter-
mining the effective specific gravity of the aggregate and by making a foy
calculations as discussed elsewhere in this chapter. .
! If the air voids. are 4 percent at ~d, _greater ~an 11 percent at Ni and
I greater than 2 percent at Nm, then this is the optimum asphalt content. If
) the mixture fails to meet Ni or Nm then the aggregate grading or possibly
(' aggregate source will need to be changed. Two other criteria that must be
) satisfied are VMA and voids filled with asphalt (VFA). The VMA criteria
\ are shown in Table 4-12 and the VFA criteria are shown in Table 4-13.
Table 4-12. VMA Requirements
Nominal Maximum Aggregate Size, mm Minimum VMA Percent

9.5 15.0
12.5 14.0
19.0 13.0
25.0 12.0
37.5 11.0
Table 4-13. Voids Filled with Asphalt Requirements
ESALs (millions) VFAPercent

< 0.3 65-80
< 1 65-78
<.3 6.5-'.;U~
< 10 65-75
< 30 65-75
< 100 65-75
;?: 100 65-75
If the mix design meets all of the requirements except VFA, the asphalt
content may be able to be adjusted slightly to keep the air voids close to 4
percent and to bring the VFA within limits. The VFA is a somewhat redun-.
dant term since it is a function of air voids and VMA. From a practical
point of view, the volumetric quality of a HMA can be controlled by air
voids and VMA.
Two items need to be mentioned about Superpave mix designs.
Individuals that have experience designing mixtures with 100 mm (4 inch)
samples will find that much more material is needed for a 150 mm (6 inch)
design. It is estimated that approximately 4 times the material is needed for
a 150 mm Superpave design as that for a 100 mm (4 inch) Marshall or
Hveem design. Since the samples must be aged in an overi, one will also
realize that a substantial amount of oven space is needed. Depending on
the number of samples to prepare and compact, this may require that more
than one oven be available during the mix design process.
The Superpave mix design system is certainly a step in the right direc-
tion to improve the quality of HMA. This is a new process that will see
substantial changes within the next few years. Changes are expected with
any new technology and the Superpave system will almost certainly under-
go some modifications during the first few years of its existence.
OTHER CONSIDERATIONS
There are other considerations that should be discussed relative to var-
ious features of mixture design, including moisture susceptibility, recycled
mixture design, and skid resistance for surface mixtures. Each of these con-
siderations is discussed briefly below.
Moisture Susceptibility
A discussion on mixture design is not complete without some discus-

sion of moisture susceptibility or the deterioration of a HMA due to the
detrimental influences of moisture, called stripping. Stripping produces a
loss of strength through the weakening of the bond between the asphalt
cement and the aggregate. This loss of strength can be sudden and cata-
strophic where the asphalt peels off the aggregate, the cohesion of the mix-
~-----ture is lost, and distresses develop--rapkl:ly.--Sm:netirnes the asphalt cement
actually washes off the aggregate so that bare aggregate remains. The more
typical situation is that there is a gradual loss of strength over a period of
years which contributes to the development of rutting and shoving in the
.wheel paths.
To combat stripping, proper mixture design is absolutely essential;

however, it is possible for a properly designed mix to strip if field com-
paction produces void contents high enough to permit water to enter the
HMA layer. Therefore, one must evaluate each mixture to determine if it
is susceptible to water damage in the event that water penetrates into the
HMAlayer.
Numerous water susceptibility tests have been proposed and are being
used (2.Q). These tests are discussed in Chapter 5 - Characterization of
Asphalt Mixtures.
Hot Recycling
More and more mixtures being produced around the USA contain
some amount of reclaimed asphalt pavement (RAP) simply because this
material usually contains valuable aggregates and asphalt cement that can
be reused. Field Investigations (21) have shown no significant difference
between the recycled and virgin HMA pavements in terms of in-situ mate-
rial properties and pavement performance. If RAP material is not used in
a recycled mix, it is typically used.on shoulders, as bases, or is disposed of
in landfills. None of these uses take advantage of the monetary value of this
product. Additionally, with the growing use of milling machines to remove
unwanted pavement layers and badly distressed materials, there is a grow-
ing supply of RAP which in most states belongs to the contractor. The con-
tractor wants to use this material in the most cost-effective way and that is
in recycled HMA mixtures. However, any mixture containing RAP mate-
rials should be required to meet the same specifications as a virgin mixture.
Therefore, the contractor must carefully consider the quality and unifor-
mity of the RAP stockpiles to ensure that the HMA produced meets the
project specifications. Additionally, the HMA facility must be modified to
handle RAP materials and often special stockpiling procedures must be fol-
lowed to ensure that the material can be easily handled during production.
In order to incorporate a RAP material in a mix design, the gradation
of the RAP and the properties of the asphalt cement must be determined.
Reference 14 suggests that three samples be taken from the RAP stockpile
and an asphalt cement extraction process be performed on each sample.
The extraction process involves dissolving the asphalt cement off the aggre-
- - . gates·.-Once performed;-rhe-aggregate-gra:ctation andasphaltcontentcanb~e------
determined for each of the three samples. The RAP stockpile gradation is
the average of the three sample gradations, and the RAP asphalt content is
the average of the three sample asphalt contents. The sqlvent containing
the dissolved asphalt cement should be collected and an asphalt cement
Chapter 4--Hot Mix Asphalt Mixture Design Methodol?gy 251
recovery process performed (ASTM D1856) to recover the asphalt cement.

The viscosity and penetration of this recovered asphalt cement should be
· measured.
Using the average aggregate gradation of the RAP stockpile, a blend
with the available virgin aggregates is deter:mined. Milled RAP may con-
tain a high percentage of fines so care should be exercised to ensure that the
mixture meets all agency gradation specifications as well as VMA and any
guidelines on dust (fines) to asphalt ratio. The type of HMA facility deter-
mines the maximum amount of RAP which can typically be included in a
mixture and still maintain reasonably efficient production while meeting
emission and opacity requirements for exhaust stack emissions. Commonly
recommended maximum percentages on RAP material in a HMA under
ideal conditions are:
l. 70 percent of total mix for modified drum mixers, and

2. 50 percent of total mix for modified batch mixing facilities.
Typically, the maximum production values are in the range of 50 percent

RAP for drum mixers and 30 percent for batch facilities. These lower per-
centages permit higher total production rates with the high moisture con-
tents typical of RAP materials. Once the combined gradation and the
amount of RAP to be included in the mixture have. been established, the
grade of virgin asphalt or rejuvenating agent needed to produce the desired
asphalt cement properties is determined.
Some agencies use the viscosity of the recovered asphalt cement to
select a rejuvenating agent to mix with the RAP material to restore the
asphalt cement to its original viscosity. Viscosity blending charts used for
this purpose are contained in Chapter 7 on special mixtures including hot
mix recycling. This type of chart can be used for blends of the recycled
asphalt cement with either another paving grade asphalt cement or a reju-
venating agent.
The RAP content of the recycled mixture should be kept low enough,
or be lowered, so that the grade of asphalt cement being added as the new
liquid is no more than two grades softer than the asphalt cement grade
desired in the recycled mixture.
A second procedure for determining the grade of the virgin asphalt to
_ _ _ _ _..,. e-adJecLis..sug-gested-in.Refei::ence.-1.4___When_using up_,ro 30 percent RAP,
use an asphalt cement that is one grade softer than that desired in the final
recycled mixture. When using higher percentages, up to 60 p~rcent, use an
asphalt cement that is two grades softer or use a softening (rejuvenating)
agent.
Once the grade of the virgin asphalt to be added to the mixture has
been determined, the amount of RAP and virgin asphalt can be deter-
mined for each asphalt content (total) to be used in the mixture design.
The following formulas are given in Reference 14 to calculate these quan-
tities:
RAP - ( B X C ) (
100 - D
WT)
100
where,
RAP = weight of reclaimed pavement, g;
B =_ percent of all aggregate in total mix;
C = percent reclaimed aggregate in the aggregate blend;
D = percent asphalt cement in RAP;
Wr = total weight of test specimen, g;
Av = weight of virgin asphalt cement, g; and
A= percent asphalt cement in total mix.
In preparing the test specimens, several special items should be noted.

First, it is difficult to separate the RAP particles to meet the precise quan-
tities needed for individual specimen preparation; therefore, some labora-
tory technicians prepare a large batch of all the material needed for each
asphalt content. Individual test specimen quantities can be quartered out
of the large batch. The RAP material is heated separately and at a much
lower temperature (about 140°F (60°C)) than that needed for mixing and
compaction. The virgin aggregates are heated enough so that when mixed,
the resulting mix is within the required mixing temperature range. Heating
the RAP at a lower temperature prevents additional hardening of the RAP
asphalt cement. ,
The balance of the mixture methodology is exactly the same as is
required for mixtures of virgin materials. The recycled HMA should meet
all test procedures and criteria as required for virgin materials.
Frictional Resistance of Surface Mixes
One additional mixture design -consideration relates only to surface

mixes, i.e., those mixes that are designed to be in contact with the vehicu-
J
lar tires for the life of the project after construction is complete. These sur-
faces must not only be stable under acceleration, deceleration, and vertical
loads, but they must also provide sufficient resistance to sliding to permit
normal turning and braking movements to occur. This means that it is the
responsibility of the paving engineer to design a surface mixture so that the
necessary tire-pavement friction is attained even under the variable envi-
ronmental and vehicle operating conditions occurring in the field. The mix
designer must choose aggregates that will provide adequate frictional resis-
tance for all surface mixtures and ensure that the asphalt content of the mix
is not so high that bleeding occurs with subsequent loss in frictional re-
sistance.
Four aggregate characteristics should be evaluated in order to select a
material that provides desirable frictional resistance: texture, shape, size,
and resistance to wear and polish. Selecting the source of the aggregates
largely determines such factors as texture, mineralogy, and chemical com-
position. Factors like particle shape and size are affected by mineralogy but
are also significantly affected by the type of crushing process selected for
manufacture of the aggregate. Resistance to wear and polish is affected by
all the above plus the hardness of the various mineral constituents of the
rock, the size of the individual minerals in the rock, and various traffic fac-
tors.
While it is necessary for the designer to consider the texture charac-
teristics of the aggregates available for use in a mixture design, some agen-
cies do not permit the use of carbonate stone such as limestone which gen-
erally tends to polish under traffic. The quality of limestone aggregate
varies considerably; hence, some limestone aggregates do not tend to pol-
ish while others do.
Numerous states and the National Stone Association have shown that
it is possible through proper combinations of aggregates of different wear
rates to use most sources of limestone aggregates successfully in surface
mixes. Laboratory studies indicate that blending aggregates with different
polishing rates is effective because the frictional resistance of the blend is
proportional to the weighted average based on the percentage of the coarse
aggregate in the blend. However, blending requires careful testing and eval-
uations during preliminary studies leading to mix design. Blending is cer-
tainly worth considering because in the USA there is an enormous amount
of aggregates that tend to polish. Therefore, considerable effort should be
~-----x-pe-aa-©a-i-ri.-d~t~r-mining-h-ow-much-of-these--aggregates can be used in
combination with other aggregates to produce surfaces with adequate fric-
tional resistance. Consideration should also be made for tailoring the
blends of materials to meet the demands of traffic both in terms of volume
and speed of traffic.
254 Hot MixAsphalt Materials, Mixture Design and Construction
Laboratory devices which are. used to provide an accelerated estimate

of polishing and wear characteristics of aggregates or combinations of
aggregates are discussed in Chapters 3 and 8.
The primary use of these laboratory devices is to serve as a screening
test on coarse aggregates which are proposed for use in HMA surfaces. In
most states the principal aggregate sources have been used in surface mixes
and their. field performance has been observed. However, when a new
aggregate is introduced, these devices can be used to indicate whether the
aggregate will likely provide good frictional resistance for the life of the
pavement layer. If the polish value is low enough to be questionable, the
use of the aggregate can be limited to lower volume facilities until field
experience and measurements of frictional numbers can be performed and
its loss of frictional resistance with traffic observed.
Special aggregates can be used in the mix design to provide very high
frictional resistance mixes for special applications. Aggregates like slag,
expanded clay shale, and other manufactured products are sometimes used
in the surface course on very high volume roadways or for special situations
like a high speed racetrack where very high frictional resistance is required.
There is an additional discussion of frictional resistance in Chapter 8 on
the performance of HMA surfaces.
EXAMPLES OF MIXTURE DESIGN
This example problem includes Marshall, Hveem, and Superpave mix

design methods.
Aggregates were sampled from the contractor's stockpiles and sent to
the laboratory for preparation of mix designs. Normally, approximately
200 pounds (90 kg) of aggregate and 2-3 gallons (8-12 liters) of asphalt
cement are needed for each Marshall and Hveem mix design.
Approximately 500 pounds (225 kg) of aggregate and 3-5 gallons (12-20
liters) of asphalt cement are needed for a Superpave mix design. The sam-
ples are taken in accordance with standard sampling procedures discussed
in Chapters 2 and 3. Each aggregate sample is taken to be representative of
the average stockpile properties. After drying the aggregates, each aggregate
sample is split down to a size suitable for measuring the gradation.
The gradations of the four aggregates along with the specifications for
_ _the_gradatio.n..o.£the_to.talaggregate_are__sho:w.n.in-T.able.A-l.4..'-1'.he-next-Step,-----
is to determine the percent of each aggregate to be used so that the result-
ing blend meets the specified limits. The calculations are shown in Table
4-15. The aggregate blend is determined by trial and error method, and
after two trials the aggregate proportions are determined to be 25 percent
of aggregate No. 1, 15 percent of aggregate No. 2, 35 percent of aggregate

No. 3, and 25 percent of aggregate No. 4. The first trial using 25 percent
of each aggregate was outside the specification range on sieve sizes No. 8,
16, and 30 (2.36, 1.18, and 0.600 mm). The bulk specific gravity of the
combined aggregate (G,b) was determined to be 2:620. The specific gravi-
ty of the asphalt cement was determined to be 1.030.
Table 4-14. Gradation of Aggregates and Specification Limits
Sieve Aggregate Aggregate Aggregate Aggregate Specification

Size Number 1 Number 2 Number 3 Number 4 Limits
1/2 inch (12.5 mm) 100 100 100 100 100
3/8 inch (9.5 mm) 89 100 100 100 95 - 100
No. 4 (4.75 mm) 3 58 99 100 47 -77
No. 8 (2.36 mm) 1 10 81 99 52 - 70
No. 16 (1.18 mm) 1 8 71 94 46 - 63
No. 30 (0.600 mm) 1 5 46 86 37 - 57
No. 50 (0.300 mm) 1 4 26 68 24 - 39
No. 100 (0.150 mm) 1 3 15 18 8 - 25
No. 200 (0.075 mm) 0.1 2.2 10.0 2.5 2-8
Approximately 20 aggregate specimens were prepared using the deter-

mined proportions of each aggregate. The blended aggregate was heated to
the temperature specified in ASTM D1559 for the Marshall method and
D1560 for the Hveem method. The total weight of each specimen was suf-
ficient to prepare a sample of specified diameter and height, i.e., 4-inches
(101.6 mm) in diameter and 21/2-inches (63.5 mm) high.
Marshall Mix Design

The heated aggregate samples were mixed with various amounts of
asphalt cement so that some were above and some were below the expect-
ed optimum asphalt content. The samples were compacted with 75 blows
on each side with the standard Marshall hammer as specified in ASTM
D1559. Three samples were prepared at each asphalt content. A total of 18
_____s_am_ples were prepared (3 at each asphalt content for 6 different asphalt
contents). The test results and calculated properties are--shown in Table
4-16.
After compaction the samples were removed from the molds and
allowed to cool. The samples were then weighed dry in air (Wn), allowed
Table 4-15. Determination of Aggregate Proportions to Meet

Specification Requirements
Sieve Size
(1/2 Inch 3/8 Inch No.4 No.8 No.16 No. 30 No. 50 No:100 No.200
(12.5mm) (9.5mm) (4.75mm) (2.36mm) (1.18mm) (0.600mm) (0.300mm) (0.150mm) (0.075mm)
Trial No. 1
No.1
(25%) 25 22 1 0 0 0 0 . 0 0.0
No.2
(25%) 25 22 14 2 2 1 1 1 0.6
No.3
(25%) 25 25 25 20 18 12 6 4 2.5
No.4 .,,.,..
(25%) 25 25 25 25 24 22 17 4 0.6
Total
100 97 65 47 44 35 24 9 3.6
Specs 95·.
100 100 47 - 77 · 52- 70 46 - 63 37- 57 24 - 39 8 - 25 2-8
Trial No. 2
No.1
(25%) 25 22 1 0 0 0 0 0 0.0
No.2
(15%) 15 15 9 2 1 1 1 1 03
No.3
(35%) 35 35 35 28 25 16 9 5 3.5
No.4
(25%) 25 25 25 25 24 22 17 4 0.6
Total
100 97 70 55 50 39 27 10 4.4
Specs 95 -
100 100 47- 77 52- 70 46 - 63 37 - 57 24 - 39 8 - 25 2-8
to soak for 3 minutes in water and vveighed submerged in water (Wsub),

removed from the water, blotted dry and weighed in air (W550).
The bulk specific gravity of the sample Gmb is then determined by:
~
Gmb= - - - - -
WssD - Wsub
The volume of the sample in ml is equal to W 550-W,ub when weighed in
grams. The bulk density (y) of the sample in pcf is
Y= Gmb X Ow
~
()
::,'
-t
Welghtln lbeordl-
Sample A,phaO Weight la Water Welghtlll Bulk cal Mu. Vokk M...wed Coneded er .g
No. Content Ah-(D,Y) (SSD) Air(SSD) Volume Density Density AlrVoida VMA FIJlcd Slabllily st.,bUky flow
0 !:f
518.3 140.6
ta
1 5.0 1167.8 650.7 1169.0 2400 2400 11
2 1164.9 647.0 1166.2 519.2 140.0 2630 2630 .11 .....

3 1165.1 651.0 1167.0 516.0 140.9 2500 2500 12
?'
Average 1405 153.1 82 18.J 552 mo 11 :;i ~
)><
{I)
1 55 1166.4 652.4 1167.5 515.1 141.J 2520 f"lo ►
2520 11
'g.
2 ·1179.0 661.4 1180.6 5192 141.7 2690 2690 12 -~
{I) ~
3-
Average
1169.4 650.9 1171.0 520.1 140.3
141.1 1525 15 18.4 592

2650 2650
2620
13
12 t f
!ti
1 6.0 1170.4 656.7 1171.0 514.3 142.0 2620 2620 13 ~
,::i.. (l)
2 1181.l 664.7 1181.9 517.2 142.5 2710 2710 13
~ ~-
3
Average
1187.3 670.9 1189.0 518.l 143.0
142.5 151.3 5.8 18.1 68.0

2980 2980
2770
12
13
~- ::;
~
1 65 1174.2 661.6 1174.7 513.l 142.8 2BOO 2800 12 a° I
"0 0.
%
2
3
1185.3
1182.3
661.1
667.7
U86.0
1182.9
518.J
5152
142.7
1432
2730
2900
2730
2900
13
14 a.
0
0{I)
°"'<
Average 142.9 149.9 4.7 18.3 74.3 2810 13
1 7.0 nn.s 663.0 1177.9 514.9 142.7 :WO :WO 14

o'
...
2 1183.4 665.4 1183.6 518.2 142.5 2730 2730 14
a=
3
Avenge
1192.8 675.7 1193.3 517.6 143.8
143.0 1485 3.7 18.6 80.1

2790 2790
2780
15
14
5i:r
1 15 1111.9 663.3 1182.3 519.0 142.1 2650 2650 16 ~
2380 16
~
2 1173.0 6602 1173.5 513.3 142.6 2380
3 1182.2 666.1 1182.7 516.6 142.8 2590 2590 14
Average 142.5 147.1 3.1 19.4 84.0 2540 15 t::l

fl
Bulk ,pedJlc gravlly ot . _.... (O,b) • 2.620
1· ~
---l
The following calculations are macle for sample 1 at 5.0 percent asphalt
content.
1167.8
Gmb= - - - - - - = 2.253
1169.0 - 650.7
Specimen density (y) = 2.253 x (1) = 2.253 glml
The voids in total mix (VIM) are determined for each sample by com-
paring the average bulk density for each asphalt content to the theoretical
maximum density (TMD) for that asphalt content. The most common
method to determine the TMD is the Rice-method as specified in ASTM
D2041. The VTM is determined by the following equation.
or
nM = { 1 _ specimen;;: density) 100
For this problem, the average of the three samples at an asphalt content of
5.0 percent had a bulk density of 2.252 g/ml and the TMD for this same
mixture was determined to be 2.454 g/ml. The VTM is then
2.252 )
V.MA. = (1 - - - x 100
2.454
= 8.2%
The voids in mineral aggregare -(VMA) are determined by the follow-

ing equation for the mixture containing 5.0 percent asphalt content.
Y.MA = 100 [1 - Gmb (1 - Pi)]

Gsb
2.252
= 100 1 - _ (1 _
_....._ - 0.05)]
_ ~
[ 2.620
= 18.3
The voids filled with asphalt (VFA) are determined for the 5 percent
asphalt mixture as shown below.
= ( 18.3 - 8.2) 100

18.3
= 55.2 perce_nt
After the samples were weighed in air and water and all calculations
made, the samples were then tested for stability and flow. The load to fail-
ure (pounds) when tested in the standard Marshall equipment is the
Marshall stability. The flow is the deformation in 0.01 inches (0.25 mm)
from the start of loading to the point where the load begins to decrease.
The measured stability is the actual load measured, and the corrected sta-
bility includes a correction for sample volume. For this design all volumes
were in the required range (ASTM D1559); hence, no correction was
needed. The stability and flow results are shown in Table 4-16.
After all the data was collected, plots were developed to show the rela-
tionship between the various properties and asphalt content as shown in
Figure 4-24. Properties that are usually plotted include: unit weight,
Marshall stability, flow, voids in total mix, voids filled with asphalt cement,
and voids in mineral aggregate.
Now to select the optimum asphalt content the project specifications
must be known. For this mixture design the specifications are shown
below:
Test l:'roperty ho-j"ect Specifications

Marshall Stability lbs (N) 1500 (6667) min
Flow .01 inches (0.25 mm) 8 - 16
Voids in Total Mix (Percent) 3-5
Voids filled with Asphalt Cement 70 - 80
There are a number of methods used to select the optimum asphalt con-
tent once the data is developed. Each agency that is involved with pave-
ment construction has 1ts own method for selecting optimum.
Method 1. One commonly used procedure is that recommended by

NAPA (11), in which they suggest preparing the plots contained in Figure
4-24. Then the optimum asphalt content is determined by first finding the
asphalt content which corresponds to the median air void content (4 per-
cent) of the specifications. That asphalt content is then used to determine
the values for Marshall stability, VMA, flow and percent voids filled from
each of the plots in Figure 4-24. Each value is compared against the spec-
ification value for that property and if all are within the specification range,
the asphalt content at 4 percent air voids is the optimum. If some value is
outside the specification range, the mixture must be redesigned. Using the
example problem, the initial optimum asphalt content is 6.9 percent at 4
percent air voids. The following values were determined and compared
with the specifications.
Property Spec Result · Status

Marshall Stability 1500 lbs (6667 N) min 2775 lbs (12333 N) OK
Flow 8-16 14 OK
Voids filled with
asphalt cement 70~80 79 OK
VMA None Required 18.6 ---
Method 2. A second common procedure that is used requires that the
asphalt contents should be selected that provide the maximum stability,
maximum unit weight, and 4 percent voids in total mix (mid spec). Hence,
for the data developed the following asphalt contents are selected:
Property Selected Asphalt Cement Content

Peak of Stability Curve 6.5
Peak of Unit Weight Curve 6.8
4 percent VTM 6.9
Average 6.7
The three asphalt contents selected were averaged to determine the

tentative optimum asphalt content. The properties of the mixture must
now be evaluated to ensure that specification requirements are met at 6.7
percent asphalt content.
"1i'I
I
I
Ii
, I
Chapter 4-Hot Mix Asphalt Mixture Design Methodology 261 ! !
i !
'I
I
8 •
b
a..
146 (/'J
7
'" "
~ 145 96
0 l
:z: >
~144 a: 5
"1
3
1-143 . :.t4 I
~
I,' ~ r,,_
~-142 17 3
141 Ii' 2
5.0 5.5, 6.0 6.5 7.0 7.5 5.0 5.5 6.0 6.5 7.0 7.5
% AC BY WGT. OF MIX % AC BY WGT. OF MIX
~
..J
3050
90
>-
!::
2950 • II
82
~ 2850 ,_
..... ~74
~
(/'J
2750 , > I'
V
--.. ~66 (_;,

::J 2650 • I.,'
[7 58
j! 2550
(/'J
l '
a: 50
i 2450
5.0 5.5 6.0 6,5 7.0 7.5 5.0 5.5 6.0 6.5 7.0 7.5
% AC BY WGT. OF MIX % AC BY WGT. OF MIX
15 f]
I ""'
I,
~13 19.6
" <(
812 ~ 19.2
-
T II
3
--.;, >
~ 18.8 ,
I
310
LI..
18.4 ,....
i--..., v
9 18.0
5.0 5.5 6.0 6.5 7.0 7.5 5.0 5.5 6.0 6.5 7.0 7.5
% AC BY WGT. OF MIX % AC BY WGT. qF MIX
·O
Figure 4-24. Graphical illustration ofHMA Design Data by Marshall

Method
Property Spec Result Status

Marshall Stability 1500 lbs (6660 N) min 2800 lbs (12430 N) OK
Flow 8-16 14 OK
Voids in Total Mix 3-5 4.2 OK
Based on the above results the optimum asphalt content is selected to be

6. 7 percent. One difficulty with this second method for determining opti-
mum asphalt content is that not all mixtures exhibit the types of plots con-
tained in Figure 4-24. In fact, some mixes which contain from 75 to 100
percent crushed particles do not show a peak in the density curve, making
it impossible to apply the second method.
Hveem Mix Design

The Hveem mix design is performed in much the same way as
Marshall mix design. A kneading compactor is used to c~mpact samples
for the Hveem test {ASTM DI560). The calculation of properties such as
VTM, VMA, and VFA are the same regardless of which mix design
method is used. '
Since the same aggregates were used, assume that the same properties
were obtained for VTM, VMA, and VFA as obtained for the Marshall mix
design. These properties will not actually be the same, however, since dif-
ferent compaction methods are used with the two methods.
60
>-
I-
d
al
5o
<( ...........
I- V .......... ~
U) /
::::t / ....... r--..
w •
~ 40
:::c
0
'
w
I-
u
~ 30
a::
0
u
20
4.0 5.0 6.0 7.0
. ASPHALT CONTENT, PERCENT

Figure 4-25. Hveem Stability Test Results
'
An example of properties that may be evaluated to select optimum

asphalt content for the Hveem mix design procedures are Hveem stability
and VTM. For this mix design the VTM were required to be 3-5 percent
and the Hveem stability was required to be 40-55 percent. The 6.9 per-
cent asphalt content can again be selected for optimum; however, the
Hveem stability at 6.9 percent is 40 which is marginal (see Figure 4-25).
It would probably be better to decrease the optimum asphalt content to
approximately 6.3 percent which would result in a Hveem Stability equal
to 43 and VTM equal to 4.9. Past experience is always helpful in selecting
the optimum asphalt content since properties change once field produc-
tion begins and knowing the direction that changes usually occur is bene-
ficial.
Superpave Mix Design Example Problem
A hot mix asphalt surface course (38 mm thick) is to be designed to

carry 15,000,000 ESALs during the life of the pavement. A 9.5 mm nom-
inal size aggregate is required. The aggregates shown in Table 4-14 are
available for use. The control limits and restricted zone boundary are
shown in Table 4-10. Based on environmental information, a PG 64-22
is used as the binder.
Binder Evaluation J
Determine that the binder to be used meets the requirements for PG
64...:.22. For this example it is assumed that these tests meet the require-
ments discussed in Chapter 2.
Aggregate Evaluation
Tests are conducted to determine the appropriate aggregate properties

as specified in the specifications. These properties may include specific
gravity, absorption, L.A. abrasion, and sulphate soundness. There are typ-
ically no requirements for specific gravity nor absorption, but this data is
----~eeded-in-tl:te-mix-design-process-;-Test-results-indicate that the L.A. abra-
sion and sulphate soundness are within the specified requirements.
The aggregate consensus properties within 100 mm of surface is as

shown in Table 4-17.
I
i
j
i
!
Table 4--17. The aggregate consensus properties for 15,000,000

ESALs when used within 100 mm of surface
i !
',,,
Property Specification
1,,
Coarse Aggregate Angularity 95/90 minimum

Fine Aggregate Angularity 45 minimum
Flat & Elongated Coarse Aggregate Particles 10 maximum
Fine Aggregate Sand Equivalent 45 minimum
After the percentage of each aggregate used in the blend is determined

the blended aggregate can be evaluated to determine if the blend meets
specification requirements. Each of the four consensus tests is conducted
on each aggregate to determine its value. For three of the four properties,
the individual components can be mathematically blended to determine
the overall property. However for the fine aggregate angularity the test
must be conducted on the fine aggregate blend.
Tests on the four individual aggregates are provided in Table 4-18.
Table 4--18. Test Results for Individual Aggregates
Test Results
Aggregate Aggregate Aggregate Aggregate
Pro12e!];}'. 1 2 4
Coarse Aggregate Angularity 100/100 100/100
I. Fine Aggregate Angularity 48 47 42
Flat & Elongated Coarse Aggregate Particles 3 5
Fine Aggregate Sand Equivalen~ 52 58
After reviewing the test results for each aggregate shown in Table 4-18
and the requirements shown in Table 4-17, it is seen that all individual
aggregates meet the requirements for coarse aggregate angularity, flat and
elongated count, and sand equivalent test. The only question is whether or
... -~-- not the fine aggregate-meets-the-speGifiecl-requirement-for-fi.ne-aggregat·----
11 angularity. Aggregate 3 meets the minimum requirement of 45 while
I I: adggregathte 4 dhoesbnot meet the requirement. If aggregate 4 is used in thedmfcix
· !i esign en t e lend of aggregate 4 and aggregate 3 must be evaluate or
fine aggregate angularity.
Evaluation of Aggregate Structure
Three gradations should be evaluated to determine the best aggregate

structure. For this example try one gradation above the restricted zone and
two gradations below the restricted zone. Since aggregate 3 does not meet
the fine aggregate angularity requirement the amount used in the blend
will be minimized. The selected blends are shown in Table 4-19.
Table 4-19. Trial Blends for Evaluation of Aggregate Structure
Blend 9.5mm 4.75mm 2.36mm 1.18mm 0.6 mm 0.3mm 0.075 mm
100% of No. 1 89 03 1 1 1 1 0.1

100% of No. 2 100 58 10 8 5 4 2.2
100% of No. 3 100 99 81 71 46 26 10.0
100%ofNo.4 100 100 99 94 86 68 2.5
Trial Blend No. 1 99 80 58 51 37 24 5.9
Agg. No. 1-10% .
Agg. No. 2-25%
Agg. No. 3-50% '
Agg. No. 4-15%
Trial Blend No. 2 96 53 33 29 23 14 3.3
Agg. No. 1-35%
Agg. No. 2-30% (
Agg. No. 3-25%

Agg. No. 4-10%
Trial Blend No. 3 · 97 59 34 30 23 15 3.6
Agg. No. 1-25%
Agg. No. 2-40%
Agg. No. 3-25%
Agg. No. 4-10%
Trial blend No. 1 goes above the restricted zone while blends No. 2
and No. 3 go below the restricted zone. There is not a large difference
----between-the-gradati0n~0f-aggregate-fer-0lend-2--and blend 3. With the
aggregates available, this is about the limit of variation that can be done
and still stay within the control limits and below the restricted zone. Even
this small variation may result in a significantly different aggregate struc-
.ture. A plot of the three trial gradations is shown in Figure 4-26.
The fine aggregates are blended for each of the three trial gradations to
evaluate the fine aggregate angularity. The values are determined to be 46
for gradation 1 and 47 for gradations 2 and 3.
Trial Blends
After estimating the optimum asphalt content for these blends, speci-
mens are prepared to evaluate the aggregate structure. In this case the opti-
mum asphalt content was estimated from past experience with similar
materials. The optimum asphalt content for all three aggregate blends was
estimated to be 5.0 percent.
The number of gyrations required for this design is 128 as shown in
Table 4-11 (based on 15,000,000 ESALs and 41-42°C average design high
air temperature). N and Nm are shown to be 9 and 208 respectively. So, all
mixtures are compacted up to 208 gyrations. The densities at Nd and Ni
are determined by calculating the volume from height measurements dur-
ing compaction and making corrections based on a comparison of mea-
sured density and actual density at Nm.
100.00
~ Blend 1
-0- Blend2
80.00 --A- Blend 3
bO
.E 60.00
"'
"'
"'
i:i...
1:l
Cl)
~
Cl)
i:i... 40.00
,,]
I
20.00
0.00 ........---::~,...,..-,,.....,..-----------------
0.075 0.3 2.36 9.5 12.5
Sieve Size (mm) Raised to 0.45 Power
Figure 4-26. Plot of three trial blend gradations
Table 4-20. Volumetrics for Trial Blends
Sample
Sample No. of Weight Ht Volume Uncorrected Uncorrected Corrected
Number Gyrations (grams) {mm) ctn3 Density TMD Density Density
Trial Blend
No.1
Sample 1 9 4693.6 123.6 2184.3 2.149 2.453 87.6 88.4
128 114.3 2020.5 2.323 94.7 95.6
208 JJ2.2 1982.8 2.367 96.5 97.4
Sample2 9 4684.0 123.3 2179.7 2.149 2.453 87.6 88.5
128 114.5 2022.7 2.316 94.4 95.4
208 112.2 1982.8 2.362 96.3 97.3
Trial Blend
No.2
Sample 1 9 4750.5 123.3 2262.4 2.100 2.447 85.8 86.5
128 114.5 2036.9 2.322 95.3 96.1
208 112.2 2005.3 2.369 96.8 97.6
Samele2 9 4744.3 128.0 2254.7 2.104 2.447 86.0 86.9
128 115.3 2034.7 2.332 95.3 96.3
208 113.5 2005.2 2.366 96.7 97.7
Trial Blend
No.3
Samele 1 9 4729.8 127.8 2259.1 2.094 2.449 85.5 86.6
128 115.4 2039.6 2.319 94.7 95.9
208 JJ3.7 2009.9 2.353 96.J 97.3
Samele 2 9 4721.1 127.6 2254.5 2.094 2.449 85.5 86.5
128 115.3 2037.6 2.317 94.6 95.7
208 113.7 2010.0 2.349 95.9 97.0
The trial blend samples are compacted and the volumetric results are
shown in Table 4-20. The test results are collected ori intervals of 10 gyra-
tions up to 100 and then by intervals of 25 through Nd and Nm. The
results are presented here for 9, 128, and 208 gyrations only since these
gyrations are used to evaluate the quality of the HMA. The height of these
samples is determined as the samples are compacted. The volume is then
calculated from the diameter (150 mm) and the measured height. The
·uncorrected density is then calculated from the weight of the sample and
the measured volume. The actual density is determined for each specimen
compacted to Nm. A correction factor is determined from the actual den-
sity at Nm and the measured density at Nm. This factor is applied to the
uncorrected density at other gyration levels to obtain the corrected densi-
ty.
:A-summary-of-the-compaction-data-for-the-trial-blends is shown in
Table 4-21. This data needs to be converted to that expected for optimum
asphalt content (that asphalt content that provides 4 percent air voids) to
fully evaluate the expected mix quality at optimum asphalt content. First
correct the asphalt content used (5.0 percent) to the estimated optimum.
This is done using the following equation:
where,
Pbe ,; estimated. optimum asphalt content

Pbi = asphalt content used in trial blends
Va = percentage of air voids
Table 4-21. Trial Blends Compaction Data
Aggregate Voids in
Bulk Total Voids o/oGmm o/oGmm
Identification Gravity %AC Mix VMA Filled at Ni at Nm
Blend No.I 2.631 5.0 4.5 15.4 70.8 88.4 97.4
Blend No. 2 2.625 5.0 3.8 14.8 74.3 86.7 97.6
Blend No. 3 2.622 5.0 4.2 15.0 72.0 86.6 97.2
So the estimated optimum asphalt contents for trial blends 1, 2, and 3 are
5.2, 4.9, and 5.1 respectively. Now estimate VMA at the estimated opti-
mum asphalt content using the following equation:
VMAe =\!MA;+ C X (4-V.)
where:
VMA. = estimated VMA at optimum asphalt content
VMAi = initial VMA for trial blend
C = constant: 0.1 when air voids are less than 4.0%

0.2 when air voids are greater than 4.0%
Va = air voids
The estimated VMA at optimum asphalt content is 15.3, 14.8, and 15.0
;-- respectively-forolenas~1-;-2-;-ana~-:-Esfimate the percent ofvoiasfilTe-ar-w-It_,_h_ _ _ __
asphalt at optimum asphalt content using the equation below:
VFAestimated= 100 (½'11A,- 4 .0)

½'11A,
For the three trial blends then the VFA values for blends 1, 2, and 3 are
73.8%, 73.0%, and 73.3% respectively. Now estimate the percent of Gnim
at Ni for optimum asphalt content. The equation to be used is:
Estimated percent of Gmm @Ni= percent of Gmm for trial@Ni - (4.0- VJ
So the estimated percent of Gmm @ Ni for optimum asphalt content is

88.9, 86.5, and 86.8 respectively for blends 1, 2, and 3. Now estimate the
percent of, Gmm at Nm for optimum asphalt content. Use the following
equation:
Estimated percent of Gmm @Nm= percent of Gmm for trial @Nm - (4.0- VJ
Substituting.into the equation gives 97.9, 97.4, and 97.4 for blends 1, 2,
and 3 respectively.
Now summarize the estimated volumetric and compaction data for the
optimum asphalt content as shown in Table 4-22.
Table 4-22. Trial Blend Summary of Estimated Properties at

Optimum Asphalt Content
Estimated Estimate Estimated

. o/o Gmmat %Gm at Optimum Asphalt Estimated Estimated
Blend Ni~ Nm Content VMA Voids Filled
1 88.9 ,- 97.9 5.2 15.3 73.8
2 86.5 97.4 4.9 14.8 73.0
3 86.8 97.4 5.1 15.0 73.3
After reviewing the ·data, it appears that blend No. 3 is the best choice.
Blend 1 barely meets the Ni and Nm requirements for percentage of Gmm·
Blend 2 does not meet the VMA requirements for 9.5 mm nominal size
aggregate.
Determine Optimum Asphalt Content
Now compact two samples each at 0.5% below estimated optimum, at

estimated optimum, at 0.5% above estimated optimum, and at 1.0%
above estimated optimum for blend No. 3. The results for the eight com-
pacted samples are shown in Table 4-23.
Table 4-23. Compaction Data for Selection of Optimum Asphalt

Content
~-6 Percent Asphalt Content

Sample No. I Sample No. 2
Gyrations - Average
Uncorrected Corrected Uncorrected Corrected Corrected
Ht.mm Vol cm3. Densiry Densiry Ht.mm Vol cm3 Density Density Densitv
5 131.4 2322.0 81.2 82.2 127.3 2249.2 81.i 82.3 82.3
9 127.4 2260.9 83.3 84.4 124.1 2193.3 83.3 84.4 84.4
20 123.9 2189.2 86.2 87.3 120.0 2120.1 86.2 87.4 87.4
30 121.8 2152.5 87.7 88.8 117.8 2081.7 87.8 89.0 88.9
40 120.4 2128.0 88.7 89.8 116.5 2059.0 88.8 90.0 89.9
50 119.2 2107.0 89.6 90.7 115.5 2041.5 89.6 90.8 90.7
60 I 18.3 2091.3 90.3 91.4 114.6 2025.8 90.3 91.5 91.4
70 117.7 2079.0 90.7 91.9 113.9 2011.9 90.8 92.1 92.0
80 117.2 2070.2 91.1 92.3 113.4.._ 2003.1 91.2 92.5 92.4
90 116.7 2061.6 91.6 92.8 113.1 1997.9 91.6 92.9 92.9
100 116.3 2054.5 91.9 93.1 112.6 1989.2 92.0 93.3 93.2
128 115.3 2037.1 92.7 93.9 111.6 1971.7 92.8 94.l 94.0
208 113.5 2005.6 94.2 95.4 109.7 1938.6 94.4 95.7 95.5
5.1 P~rcent Asphalt Content
Gyrations Sample No. 1 Sample No. 2 Average
Ht.mm Volcm3 Densirv Densirv Ht.mm Vol cm3 Densitv Density Densitv
5 130.9 2313.3 82.9 84.1 133.3 2355.3 83.0 84.2 84.1
9 127.5 2253.1 85.0 86.2 130.1 2298.2 85.1 86.2 86.2
20 123.0 2174.2 87.9 89.1 125.7 2222.0 88.0 89.2 89.2
30 '120.8 2134.9 89.3 90.6 123.6 2184.0 89.6 90.8 90.7
40 119.3 2107.9 90.4 91.6 122.l .2158.6 90.6 91.8 91.7
50 118.1 2087.1 91.2 92.5 121.1 2139.5 91.4 92.6 92.5
60 117.2 2070.5, 91.9 93.2 120.2 2124.7 92.1 93.3 93.2
70 116.5 2058.1 92.4 93.7 119.5 2112.0 92.6 93.9 93.8
80 115.9 2047.8 92.8 94.l 118.9 2101.4 93.1 94.3 94.2
90 115.4 2039.4 93.3 94.6 118.4 2092.9 93.5 94.7 94.7
100 114.9 2031.l 93.6 94.9 118.0 2084.4 93.8 95.1 95.0
128 114.0 2014.5 94.3 95.7 117.0 2067.6 94.6 95.9 95.8
208 112.2 1983.4 95.8 97.2 115.1 2033.7 96.2 97.5 97.3
5.6 Percent Asphalt Content
Gyration Sample No. 1 Sample No. 2 Average
Ht.mm Vol cm3 Density Densitv Ht.mm Volcm3 Densiry Density Densirv
5 128.0 2261.9 83.6 84.6 129.3 2284.9 83.9 85.1 84.8
9 124.6 2201.9 85.9 86.9 126.0 2226.6 86.1 87.3 87.1
20 120.3 2125.9 89.0 90.0 121.8 2152.4 89.1 90.3 90.1
30 118.1 2087.0 90.6 91.7 119.7 2115.3 90.6 91.8 91.8
40 116.6 2060.5 91.8 92.8 118.2 2088.8 91.8 93.0 92.9
50 115.5 2041.1 92.7 93.8 117.1 2069.3 92.6 93.9 93.8
60 114.6 2025.1 93.4 94.5 116.3 2055.2 93.3 94.6 94.6
70 _113.9.- -2012.8- _9,Lo _9.5,.1 11-!'i.c)- 2041.1 93,9 --95.2 95.
80 113.3 2002.2 94.5 95.6 114.9 2030.4 94.4 95.7 95.6
90 112.8 1993.3 94.9 96.0 114.4 2021.6 94.8 96.0 96.0
100 112.3 1984.5 95.3 96.3 114.1 2016.3 95.1 96.4 96.4
128 111.5 1970.4 96.0 97.1 113.0 1996.9 96.0 97.3 97.2
208 109.7 1938.6 97.6 98.7 111.4 1968.6 97.4 98.7 98.7
Table 4-23. (Continued)

6.1 Percent Asphalt Content
Gyrations Sample No. 1 Sample No. 2 Average
Ht.mm Volcm3 Densitv Densitv Ht.mm Vol cm3 Densitv Densitv Densitv
5 134.4 2375.0 84:9 85.9 131.6 2325.6 84.5 85.5 85.7
9 130.9 2313.2 87.2 88.1 128.2 2265.5 86.7 87.8 88.0
20 126.4 2233.7 90.3 91.4 123.5 2182.4 90.0 91.0 91.2
30 124.1 2193.0 92.0 93.1 121.4 2145.3 91.6 92.7 92.9
40 122.6 2166.5 93.1 94:2 119.7 2115.3 92.9 94.0 94.1
50 121.3. 2143.5 94.1 95.l 118.5 2094.1 93.8 95.0 95.0
60 120.5 2129.4 94.7 95.8 117.7 2079.9 94.5 95.6 95.7
70 119.8 2117.0 95.3 96.4 116.9 2065.8 95.1 96.2 96.3
80 119.3 2108.2 95.7 96.8 116.3 2055.2 95.6 96.7 96.7
90 118.8 2099.4 96.1 97.2 115.8 2046.4 96.0 97.1 97.1
100 118.3 2090.5 96.5 97.6 115.3 2037.5 96.4 97.5 97.5
128 117.3 2072.9 97.3 98.4 114.5 2023.4 97.1 98.3 98.3
208 115.9 2048.l 98.5 99.6 112.9 1995.1 98.5 99.6 99.6
The optimum asphalt content is determined from the compaction and

volumetric data. The optimum is selected at the asphalt content that pro-
vides 4 percent air voids. Figure 4-27 shows the relationship between air
100.00
~ 4.6%AC
-B- 5.1 %AC

96.00 - -~ s·.6-%---'A:e
-&- 6.1 %AC
92.00
J
~
88.00
84.00
80.00 -+---T--i--T-r-rT"T-rr---r----.---.-.............- . . ' - ; . . . - ~ - . - ~ ~ ~

\
10 100 i 1000
Number of Gyrations !
Figure 4-27. Percent of Gmm versus number of '~yrations
voids, number of gyrations, and asphalt content. It can be seen from this
plot that the optimum asphalt content is 5.2 percent and that the require-
ments for Ni and Nm are met.
Now verify that the compaction and volumetric properties ai:e satis-
factory at 5.2 percent asphalt content. The properties include percent of
Gmm at Ni, percent ofGmm at Nm, VMA, and voids filled. This can be done
best by plotting each of these values versus asphalt content as shown in
Figure 4-28. The data is calculated as done earlier for the trial blends. The
data is summarized in Table 4-24.
6.00 88.00
5.00 z.; 87.00

~
{]. 4.00
~ ,,
j
-... 86.00
-=-< 3.00
0
ii
2.00 ~ 85.00
""
1.00 84.00
4.40 4.80 5.20 5.60 6.00 6.40 4.40 4.80 5.20 5.60 6.00 6.40
Asphalt Content, % Asphalt Content, %
15.80 100.00
15.60 i.; 99.00
~ E 98.00
i 15.40
....t.5
> 0
97.00
15.20
J 96.00
15.00 95.00
4.40 4.80 5.20 5.60 6.00 6.40 4.40 4.80 5.20 5.60 6.00 6.40
Asphalt Content, % Asphalt Content, %
90.00 \
80.00
~
t 70.00
-- - -----i;o:oo-
4.40 4.80 5.20 5.60 6.00 6.40
Asphalt Content, %
Figure 4-28. Mix Properties versus Asphalt Content

Table 4-24. Compaction and Volumetric Properties for Range of

Asphalt Contents
Asphalt Voids at Voids Percent Gmm Percent Gmm

Content Nd Gmm VMA Filled at Ni at Nm
I
4.6 6.0 2.465 15.7 61.8 84.4 95.5
5.1 4.2 2.445 15.2 72.4 86.2 97.3
5.6 2.8 2.426 15.1 81.4 87.1 98.7
6.1 1.7 2.406 15.3 88.9 88.0 99.6
An inspection of Table 4-24 and Figure 4-28 shows that the mix at
5.2 percent asphalt content should have 4.0 percent air voids and should
meet the requirements for V1v1.A, voids filled, percent Gmm at Ni and per-
cent Gmm at Nm. Interpolating between the values indicates that at opti-
mum asphalt content the VMA is 15.2, the voids filled is 74.2, the percent
Gmm at Ni is 86.4 and the percent Gmm at Nm is 97.6. All of these values
meet the specified requirements. Therefore, the optimum asphalt content
is selected to be 5.2 percent.
After the mix design is completed water susceptibility tests should be
conducted. If the mixture fails to meet the specified requirements for water
susceptibility the mix design will have to be adjusted.
REFERENCES
1. Kandhal, P.S., and W.S. Koehler. Marshall Mix Design Method:

Current Practices. Proceedings, Association of Asphalt Paving
2. Crawford, C. The Rocky Road of Mix Design. National Asphalt
Pavement Association, Hot Mix Asphalt Technology, Winter 1989.
· 3. Richardson, C. The Modern Asphalt Pavement. John Wiley & Sons,
New York, Second Edition, 1912.
4. Vallerga, B.A., and W.R. Lovering. Evolution of the Hveem Sta-
bilometer Method of Designing Asphalt Paving Mixtures. Proceedings,
Association of Asphalt Paving Technologists, Volume 54, 1985.
5. White, T.D. Marshall Procedures for Design and Quality Control of
------,/'\,--sphaJ.t-Mixtui:es.-P-roceedings,-Association of Asphalt Paving
6. Foster, C.R. Development of Marshall Procedures for Designing
Asphalt Paving Mixtures. National Asphalt Pavement Association,
NAPA IS84, 1982.
7. Background of Superpave Asphalt. Mixture Design and Analysis,

Federal Highway Administration, Report Number FHWA-SA-95-
003, 1995.
8. Goode, J.F., and L.A. Lufsey. A New Graphical Chart for Evaluating
Aggregate Gradations. Proceedings, Association of Asphalt Paving
9. Nijboer, L.W: Plasticiry as a Factor in the Design of Dense Bituminous
Road Carpets. Elsevier Publishing Co., Inc., New York, 1948.
10. Asphalt Concrete Mix Design and Field Control. Federal Highway
Administration, Technical Advisory T 5040.27, 1988.
11. Arena, P.J. Evaluation of the Gyratory Compactor for Use m
Designing Asphalt Concrete Mixtures. Louisiana Department of
Highways, Research Report No. 26, LA, December 1986.
12. Finn, EN., and].A. Epps:.Compaction of Hot Mix Asphalt Concrete.
Texas Transportation Institute, TTI Research Report 214-21, Texas
A & M University, August 1980.
13. Mix Design Methods for Asphalt Concrete and Other Hot Mix Types.
The Asphalt Institute, MS-2, May 1984.
14. Mix Design Techniques-Part I. NAPA TAS-14, National Asphalt
Pavement Association, Instructors Manual, April 1982.
15. Foster, C.R. The Effect of Voids in Mineral Aggregates on Pavement
Performance. National Asphalt Pavement Association, NAPA IS96,
1986. 1
.·
16. Superpave Asphalt Mixture Design: Illustrated Level I Lab Methods,
Federal Highway Administration, Report Number FHWA-SA-95-004,
1995.
17. Superpave Level I Mix Design, Asphalt Institute, Lexington, KY,
Report Number SP-2, 1995.
18. Superior Performing Asphalt Pavements (Superpave): The Product of
the SHRP Asphalt Research Program, Transportation Research Board,
Report Number SHRP-A-410, 1994.
19. Brown, E.. R. and Cross, Steve, A National Study of Rutting in Hot
Mix Asphalt (HMA) Pavements; Journal of the Association of Asphalt
Paving Technologists, Volume 61, 1992, p. 535-573.
20. Kandhal, P.S., Field and Laboratory Evaluation of Stripping in Asphalt
..Pavements: State of the Art Report. Tran.s.P-OJ:.tation_Res.ear.ch_Re.c.oJ:1,..,__ _ __
1454, 1994.
21. Kandhal, P.S. et. al., Performance of Recycled Mixtures in State of
Georgia, Report No. FHWA-GA-94-9209, January 1994.
CHAPTER 5. CHARACTERIZATION
OF ASPHALT MIXTURES
INTRODUCTION
Material characterization, as presented in this chapter, is the measurement

and analysis of the response of HMA mixtures to load, deformation, and/ or
the environment at various rates of loading and tempe~c1:t:iires-:-Tests current-
lyused are generally empiricalinnatu£5ui ongoing ~esearch is yielding tests
that measure fundamental engineering properties. Empirical tests (such as
Marshall stability and flow, and Hveem stability) are more commonly used
but are not reliable when conditions are outside those in which the tests were
developed. For example, these tests are not very useful for predicting perfor-
mance as axle loads and tire pressures continue to increase. The emphasis of
this chapter is on mixture properties and on the tests used to measure those
p~ies that are r~_¥$cito perfor~anct. Pe~ormance of the m~ure is pri-
marily dependent upon the propert1es of the rhlxture and not so much upon
the individual properties of t h ~ a t e . Unfortunately at this
time, testing of the mixture properties alone is not sufficient to ensure satis-
factory performance hence the individual materials must also be evalu~ed.
There.are occasions when the asphalt binder and aggregate are adequate but
the mix fails because of poor compaction, use of an incorrect binder cont~nt,
poor adhesion or some other problem as~ciate~ " \ ' \ T i t h ~
The driving public considers a HMA pavement to be satisfactory if it is
smooth and provides adequate frictional properties. However, pavement
engineers understand that rutting, cracking, raveling, etc., can lead to rough-
ness, loss in friction properties, and poor performance. Table 5-1 provides
----~gu-id.ili'lee-e0-!'1eer-a-ing-t:he-p-roblems-that-are-observed-in-asphalt pavements _gnd
the likely causes of these pro~s. The,...test procedures discussed in this
chapterare used to evaluate the mix for the pavement problems shown in
Table 5-1 and to identify the likely causes of these problems. A more detailed
discussion of pavement distresses is provided in Chapter 8.
Table 5-1. Asphalt Mixture Properties Affecting HMA

Serviceability
Serviceability Problems Performance Problems Likely Causes of

Performance Problems
Roughness Rutting High binder content

Use of excessive amounts of
rounded aggregate
Water sensitivity
Thermal cracking Stiff binder
Low binder content
A high dust to asphalt ratio
Fatigue cracking Low binder content
Asphalt cement too stiff
,_ Water sensitivity
Inadequate support
.Reflection cracking Not usually mix related
C
Ravelling Low compaction
Low binder content
High dust to asphalt ratio
,;- Water sensitivity
Friction Resistance Polishing Poor aggregate quality
Bleeding High binder content
Water sensitivity
MATERIAL CHARACTERIZATION TECHNIQUES

./
Although the need for materials characterization has been around for
many years, and a great deal of work has been done, there ~!!!e agree-
ment throughout the industry with regard to the types of tests and test pro-
cedures required to predict mixture performance Tne approa2;. to solving
the problem has, at best, '""5een piecemeal involving many different
researchers, from many different agencies working relatively independent-
ly yet striving for an optimal solution to a singular problem without con-
sideration for all of the ramifications. For the past 50 years or more, pave-
ment design agencies have always pressed for an immediate answer to their
needs and problems. Long term, well thought out, sequential efforts have
usually been rejected because of the time arid expense involved. The effect
... ---has been that structural-and-mixture-design--procedures-have-suffered-a:nA------
an integrated design and analysis system has not evolved.
The Strategic Highway Research Program (SHRP) was undertaken in
an attempt to remedy this situation. SHRP involved extensive research to
develop performance-based asphalt cement specifications, fundamental
Chapter 5-Characterization of Asphalt Mixtures 277
tests, and an improved structural design and analysis system. The SHRP
work was completed in 1993. Additional' work is being performed to
develop a model that will predict ~an.ce. This effort should be com-
pleted prior to tlieyear 2000.
CRITERIA FOR ASPHALT MIXTURE TESTS
Asphalt mixtures in pavements are subjected to a wide range of load

and environmental conditions. The ~sponse to these conditions is ~
plex and involves the elastic, viss~,and plast~ :.h.~~!-~?f the
material. Thus, satisfactory mixture and structural aesign for HMA pave-
merits requires an understanding of both the load-deformation and the
strength properties of the materials to be used for the range ofloading con-
ditions. Strength represents a limiting condition. Lo~d_:__d~formatio11 char-
acteristics, however, can relate to a single load application or to many load
applications.
There are a number of computer programs available that use a layer
elastic analysis to evalu~esses and straiQ~ill-~.P3Y,e..~~Elforgty~i::J.9ad-
ing conditions. · I nel5asic material properties required as input for layer
elastic analysis are the modulus of elasticity and Poisson's ratio of each
pavement layer, and the subgrade. Unfortunately, most, if not all, pave-
ment materials including the subgrade are n<?!~e!¾SJic (i.e., the modulus
varies with temperature, load, or stress state). For bituminous-bound lay-
ers, the modulus varies with t~erature and rate of loading. Thus it is
important to measure the cliaracteris~~tnea-;phai77n.7?tures as related
to thermaj ci;gckip,g,, fatigue c~~cking, and permanent defoE!Ilation, and to
C:Onsickr the adverse effects bfmo1sfore se,11~siti'::ii~~4~gin'g. The type of
test used to determine these mixture characteristics depends on the follow-
ing general criteria:
1. Material variability and project size;

2. Ability to estimate fundamental properties;
3. Ease of testing; and
4. Reproducibility of test results.
-Materfal-Variability-and-Projecr-Siz~~--
Although not directly related to characterization, the size and cost of

the project and the inherent variability of the materials must be considered
in setting up the materials testing program.
Variability. Material variability must be quantified for meaningful

quality control of a HMA project. It is obvious from observing pavement
distresses that this variability is one of the most significant factors to be
considered in pavement performance. For example, if 10 percent of a
pavement fails, then the entire pavement has probably failed in terms of
performance. All pavement failures begin with some small percentage of
the pavement showing signs of distress with this percentage increasing with
time until repairs are made.
The concept of variability and its relationship to.failure is illustrated in
Figure 5-1 which shows the variations of stress and strength for a hypo-
thetical pavement. The area of overlapping curves represents a failure con-
dition, i.e., the stress exceeds the mixture strength. If the variation in
material properties increases, then the probability of failure also increases.
Similar examples could be shown involving other properties or a combina-
tion of these properties.
Closely related to inherent material variability is the question of how many
tests should be conducted? An elaborate and extensive testing program on a
small amount of material taken from a population that is quite variable yields
a great deal of information, but it is meaningful to only a very small portion of
the pavement. Likewise, a very limited program for a highly variable material
provides limited information that may not accurately describe the quality of
the material. A realistic approach involves conducting sufficient tests to deter-
mine the average property value and variability of the material so that the per-
cent within any desired range can be determined.
When material properties are evaluated, the precision and bias of the test
method being used should be known. Knowing the precision and bias allows
the engineer to use the test data to determine if there is a significant difference
between the test results and the desired results and it also allows evaluation of
two differ~nt test results to determine if they are significantly different from
each other. This comparison can be made for two tests conducted in the same
laboratory by the same technician or between two laboratories. A precision
and bias statement is included in relevant ASTM test methods.
Project Size. Project size and cost are also important in establishing a
testing program to characterize a HMA mixture. As the size of a project
increases, the variability of material properties may also increase since
material properties may change with time and as equipment wears. When
·-·---ffro]ecf s1ze of cost-increases, the JUSt1fication for a more extensive testing
program increases. Thus, the extent and nature of the testing program ulti-
mately relates to the variability of the material and test method, the cost of
the project, and the consequences of failure. More test results allow the
engineer to more accurately measure the true quality of the mixture but
Chapter 5-Characterizatio.n o.f Asphalt Mixtures 279
excessive testing can be expensive and may not be justified when the test-
ing cost exceeds some established percentag~of the overall project cost.
en
'iii Q.
Q.
0
0 0
0 t\l
t\l II 'iii
11 "iij Q.
.c.
.c.
Q.
-o
-o
g'(.\J
g'I")
G)
G) _:II
.._ II 0
(X)
C/)C/) I 0
(X)
U) Cl) G) I
(\J C\J
O'I
G)
O'I
...
C
...
C
G)
I
I
G) >
> 0 <( I 0
<( I
,;;t" V
(\J I C\J
I
0 0
0 I 0
(\J ·u; I C\J en
a. a.
\
OU) \ OU)
(0 U) \ U)
U)
-w \ -w
er. \ er.
1- \ I-
0 U) OU)
(\J C\J
C C
_g _g
0 0 0
(X) :§ (X)
>
G)
>
G)
C C
0
...
-c
C 0 0
-c
~
V
-c m V
-0
C
n C
C II 0
en
en-
G) Cl)
'en Cl)
en- 'en
... Q. G> en n
cno
-ct\l
II
Cl)
... Q.
'en 0 Cl)
-ct\l
~n ~ II
'5CI) '5CI)
.5 .5
G)
G)
O'I
C O'I
0
G)
G)
~ ~
Figure 5-1. Graphical Presentation of Failure for Two Conditions with
the Same Mean Value But Different Variability (ll
,,
':1
Ability to Estimate Fundamental Properties
Another criterion relates to the ability of the test to measure the fun-
damental properties previously mentioned. Empirical test results are only
of value to those empirical design procedures from which they were devel-
oped. Attempts at using empirical tests to estimate the fundamental prop-
erties through correlations are poor at best and should be used with cau-
tion only when better information cannot be obtained. Most correlations
between actual performance and test properties have shown that the phys-
ical properties of the asphalt mixture have produced the best correlation
with performance. Substantial research is still needed to develop practical
fundamental tests that can be used in mix design and quality control test-
ing to evaluate mixture performance at a high level of reliability.
Ease of Testing
Ease of testing is also an important criterion to be applied to any pro-

posed test method. Often an "imperfect" test method is favored because
of its simplicity and the possibility of conducting the test without costly
equipment, excessive time, or extensive personnel training. A test which
can be readily conducted is highly desirable.
Simplicity and low cost should not, however, be the primary basis for
selecting a giv'ep. test or testing program. Compared to the total cost of
designing, constructing and maintaining a pavement, the cost of the test-
ing program is usually minimal. It is important that tests be utilized that
relate to performance so that the service life of a HMA mixture under var-
ious traffic and climactic conditions can be estimated.
Reproducibility of Test Results
Another criterion relates to testing error. A test method ideally should

be written such that two laboratories testing the same mixture obtain
approximately equal test results. One measure of this reproducibility is the
coefficient of variation obtained between a number of laboratories for
identical laboratory prepared and tested specimens of a given mixture. The
variation obtained using this technique represents inherent test method
+-----variationassociated-wirh-sample-preparation-;-tesceqi:ripment; ancl person--------
nel experience. For a test i:o be useful it must have minimum error caused
by test equipment variation and technician performance so that the true
variation in mixture properties is .measured by the test and not just the
variation in the test method. Tests with high testing error are not useful jn
measuring material properties, since the true properties cannot be deter-

mined with confidence.
CURRENT TESTS
A wide variety of test methods and procedures have been developed

over the years, many of which are still being used today. These test meth-
ods can be classified into different categories. These categories are in-pl~e
versus laboratory tests, empirical versus fundamental tests, and those tests
grouped accordi;;'g to the mode of test (e.g., direc:t tension, indirect te~-
sion, compression, shear, flexural, or torsion). Tests which yield'these"~-
damental properties directly are desirable. Variations in testing techniques
exist and there is a tendency based on usage to favor one test method for
one material and another test method for another material.
Examples of in-place tests used to indirectly measure engineering
properties of HMA mixtures are deflection measurlng_dqjces such as the
Benkl~ma2, beam, d ~ t , and f~J.~.~K.Y!'.~ight.~kfte.ctometer. Field tests
basically can be used only to evaluate an existing condition. Thus, they can
be used to evaluate the subgrade for a proposed pavement prior to design,
an existing pavement to determine its basic structural condition, or an
existing pavement to determine its support characteristics for use in reha-
bilitation design. Results generally must be considered in terms o£the con-
ditions which exist at the time that the tests are conducted, and involve
back calculation techniques to obtain elastic moduli for the various pave-
ment layers and composite subgrade. These nondestructive field tests,
while important, are not considered in this chapter.
Laboratory tests that are currently used to characterize asphalt mix-
tures include evaluation of the individual asphalt cement and aggregate
materials followed by evaluation of the HMA mixture. Asphalt cement
tests that are most often used for characterization of mixture performance
are discussed in Chapter 2. Thermal cracking in HMA mixtures is one dis-
tress that is primarily controlled by the properties of the asphalt cement.
Since tensile stresses in a HMA mixture are primarily resisted by the·
asphalt cement, the development of cracks is often related to the ability of
------'tne-asphalt-cernent-to-undergo--Stra~-=~j_thout.failure.
The effects of aggregates on mixture performance are discussed in
Chapter 3. The aggregate primarily provides the support for applied traf-
fic loads. The angularity, surface texture, gradation, and hardness are
important factors in resisting p'errna:nent deformati~n. ··----
-··-------~----- ., _____ .,
Tests on asphalt mixtures are conducted to evaluate the overall quality

of the HMA. These tests, to some limited extent, evaluate the adequacy of
the asphalt cement quality, aggregate quality, asphalt con!ent, and d~~i!Y.
Ideally, these inixture tests should prov1cl.e guidance on the overall perfor-
mance of the asphalt mixture. Unfortunately, mixture tests cannot be used
alone to evaluate overall performance of a mixture and therefore, tests must
be performed on the aggregate and the asphalt cement to help ensure that
the mixture performs satisfactorily.
The mixture to be evaluated is prepared in the laboratory or taken
from the HMA facility during production. Accurate characterization of
the mixture requires that it be compacted to a density equal to that which
will be achieved in the field after a few years of traffic. The density of a
HMA mixture in the field increases throughout the life of the pavement
with most of the increase occurring during the first two to three years,
assuming that the traffic does not change significantly. A properly
designed and constructed mixture normally has 6-8 percent air voids
immediately after construction and has approxim~ 4 percent air voids
after 2-3 years of traffic. ==-
The two general types of tests that are used to characterize asphalt mix-
tures are physical and mechanical. Physical tests include measurement of
density, air voids, voids in mineral aggregate, voids filled with asphalt
cement, and aging characteristics. These properties, except for aging char-/"
acteristics, are discussed in detail in Chapter 4. The mechanical tests
include measurement of Marshall stability and flow, Hveem stability andr
cohesion, gyratory shear index (GSI), dynamic modulus, resilient modu-\
lus, flexural stiffness modulus, indirect tension, creep, and moisture sus..U
ceptibility.
Characterization of the asphalt mixture, when the Marshall mix design
procedure is used, generally includes measurement of air voids, voids in
mineral aggregate, voids filled with asphalt cement, Marshall stability, flow,
and moisture susceptibility. Mixture characterization with the Hveem mix
design procedure includes measurement of air voids, voids in mineral
aggregate, Hveem stability, and moisture susceptibility. Jhe other tests
such as gyratory shear index, dynamic modulus, resilient modulus, flexur-
al stiffness modulus, indirect tensile stre.r'i.gth, compressive strength, fatigue
and creep are primarily used for research purposes. These tests are impor-
1-----cant, andsomewi1Uikely-be-used-r0uti-nely~i-n-the-fumre,bur-tl:rey-havenoct-------
generally been used for mix design or quality control to date. After com-
pletion of the SHRP program, some of these tests now used only for
research are likely to be used rou,tinely in mix design and construction
quality control of HMA mixtures.
Physical Properties
Air Voids. The physical property most often correlated .to per-
formance is air voids in the mixture (see Chapter 4 for calculation of air
voids). Re~shown that the amount of air voids in a HMA mix-
ture affects the ~ t y -and durability. When the in-place air voids
decrease to less than 3 percent, rutting of the asphalt mixture is likely to
occur (2, 1, ~ due to plastic fl.ow. However, when the air voids are above
approximately 8 percent, the mix is permeable to air and water, and the
rate of oxidation
__, ...__,.- . ~-
of the asphalt binder is significantly
-·-··---···-··~··incre;:i.sed
....
(i, g, 2)
resulting in premature cradang and/or ravefling. The rate of oxidation is
a function of the air voids as vidl as th~--aiph.tlt film thickness. The labo-
ratory compaction effort (50-75 blows/side for Marshall Method) is select-
ed to provide a density in the HMA that provides air voids approximately
equal to that expected in-place after 2-3 years of traffic. Air voids in lab-
oratory compacted samples then provide an estimate of the minimum air
voids expected after traffic.
Voids Filled with Asphalt Cement. The voids filled with asphalt
cement are inyersely related to the air voids. As the percentage of air voids
approaches _{ero:the-p~rceritage ofvoias filled with asphalt cement approach-
es I 00. The asphalt mixture is initially constructed to some percentage of
voids filled (usually 50-70 percent) with asphalt cement. This perce'ntage-of
voids filled increases as the asphalt mixture continues to densify under traf-
fic. When the voids filled exceed approximately 80%, to ~5%,. the asphalt
mbcture typically becomes unstable and rutting is fik:elyt~ occur.
,__ .,--···•,-.v•·•·=··- ,p•--..,. . --- • ·- -~ •
Voids in Mineral Aggregate. The voids in mineral aggregate (VMA)

is primarily a function of the aggregate gradation and of the shape and sur-
face texture of~aE.~~g~E,t When the VMAis too low, suffu:i~--~sphalt
cement cannot oe aaded to the mixture to provide a satisfactory film thick-
ness without overfilling the voids. Mixtures.with low VMA are also more
sensitive to a small chang;;:;;_sphalt content than a mixture that has high
VMA. Most agencies do not specify a maximum VMA; however, if the
VMAis too high the mix stability is likely to be reduced to an undesirable
-----i.evel-:-'Fhe-guid-an:ee-p-revided- in--Ghapter-4-for-VMA should be followed.
Aging. Aging of asphalt mixtures affects the stiffness of the asphalt

binder and may result in excessive stiffness. To .be able to adequately char-
acterize asphalt mixtures requires that samples being evaluated be aged to
simulate in-place properties of the HMA mixture. The SHRP program

developed aging tests for the asphalt cement and for HMA mixtures.
The asphalt cement is aged with the rolling thin film oven test to sim-
ulate aging during plant production. The asphalt cement is also aged in
the pressure aging vessel (PAV) to simulate aging for a few years ofin-place
service. Aging the asphalt cement alone is not a preferred way to predict
properties of the asphalt mixture during its life, although this is a good and
easy way to get a rough measure of the potential for the asphalt cement to
I
age.
Since the rate of aging of the asphalt cement in a mixture is affected by
the mixture properties such as in-elace air void§.,}_he entire mix~-~b.ould
be aged. The Superpave mix design requires ~hat the mixture be aged for
2-2Chours after mixing and prior to coE}l?acti~n. This process simulates
tl:ie";gi':ng of the mixnire'duriniprodu~tion a~d placement in the field. It
also allows the asphalt cement to be absorbed into the ~se_aggregat_e to
simulate what happens d_1::1ting_E<>g_sg~~n.------ ·
Superpave also aeveloped a method to age the compacted mixture to
simulate in-service aging in the field. The process involves placing the
compacted samples in an oven for an extended period of time to allow for
oxidation of the asphalt binder. After aging the samples in this way, they
can be tested in a condition similar to that on the roadway and thus should
provide a better measure of expected P<:Efo!mance.
Mechanical Tests
The two most common tests used to evaluate or design HMA mixtures
are the MarJ~-~!.~!~£ig t:y test and the Hveem ~!"~2H}]:'.;..,,,test. Both of these
0
tests and mix design procedures have served the industry well and will con-
tinue to be used until better test methods are developed and accepted._
Since both of these tests are empirical, and not highly correlated to perfor-
mance, improved test methods are needed.
Marshall Stability. The Marshall Test (ASTM D 1559) was developed

by the Corps -of Engineers in the 1940s and is based on methods and con-
cepts formulated by Bruce Marshall of the Mississippi State Highway
-~--~Department. -The-purpose-0Lthe-test-wasc..t0-m&as-1:1-r-e-1:he-st-r-ea=b:-0f-a:A--- - - - -
1' 5"-'"'
asphalt mixture that had been compacted to a standard laboratory com-
I pactive effort. This test is also used as part of the Marshall mix design pro-
cedure for selecting a design asphal_t content, and in the quality control of
asphalt mixtures.
The Marshall Test procedure involves applying a compressive load to a

cylindrical specimen (4 inches (101.6 mm) in diameter and 2.5 inches
(63.5 rrim) high) through semicircular testing heads, as discussed in
Chapter 4. The temperature of the specimen is 140°F (60°C); the load is
applied at a rate of 2 inches per minute. The temperature of l 40°F (60°C)
was selected since that approximates the maximum pavement temperature
in the summer, thereby providing the weakest condition for the HMA
mixture. Marshall stability is the maximum load which the specimen can
withstand or the load which must be applied to cause the specimen to fail.
The flow index, or flow value, is the total vertical deformation (in 0.01
inches (0.25 mm)) of the specimen at maximum load. The fl.ow is the total
deformation from the beginning of loading until the load begins to
decrease. One property that is sometimes used to characterize asphalt mix-
tures is the Marshall stiffness index which is the Marshall stability divided
by flow. This is an empirical stiffness value and is used by some engineers,
especially in Europe, to evaluate the quality of asphalt mixtures. A higher
value of stability divided by flow indicates a stiffer mixture and, hence,
indicates the mixture is likely more resistant to permanent deformation.
There is very little performance data to indicate that the Marshall stabili-
ty/fl.ow is related to performance. The Marshall stability test is discussed
in more detail in Chapter 4.
Hveem Stabilometer. The Hveem stabilometer was developed by

Francis Hveem while working with the California Division of Highways.
The stabilometer test was developed as an empirical measure of the inter-
nal friction within a mixture. However, the strength or stability of a HMA
mixture involves both cohesion and internal friction. Thus, a companion
test using the cohesiometer, was developed to measure the cohesion char-
acteristics. In summary, a vertical axial load is applied to a HMA mixture
specimen (4 inches (101.6 mm) in diameter and 2.5 inches (63.5 mm)
high) and the displacement is measured. The specimen temperature is
l 40°F (60°C) at the time of test to produce mixture properties that simu-
late the most critical field condition. Values of stability can vary between
0 and 100; the stability of water is O and the stability of steel is 100.
Different agencies have modified the procedure and equation slightly.
-----1'.he-i::ubber-membr-ane-in--the--st-abilometer-pro:v:ides the confining pressure
which the· asphalt cement would normally provide (cohesion). Thus, sta-
bilometer values are relatively insensitive to asphalt cement characteristics
but are indicative of aggregate characteristics. The Hveem test is discussed
in more detail in Chapter 4.
Cohesiometer. The cohesiometer was developed to measure cohesion

or the tensile characteristics of HMA mixtures and is a companion test to
the stabilometer. The specimen (4 inches (101.6 mm) in diameter and 2.5
inches (63.5 mm) high) is tested at 140°F (60°C). A constant rate ofload
is applied by means of a lever arm which causes the tensile loading to be
applied to the specimen from which an empirical cohesiometer value or an
estimate of the tensile strength is obtained. A measure of deformation can
also be obtained. The cohesiometer is seldom used now since the testing
error is quite high and the measured cohesion has little correlation with
performance. Also, most HMA mixtures pass the test. More recently, the
indirect tensile test has been used in most laboratories to estimate tensile
properties.
Modulus of Elasticity. The dynamic complex modulus and ~ t

modulus tests are used to m_s:.asJJ,.re modulus of elasticity. Papazian (Ji) pre-
serifecl the concept and definitions concerning the complex modulus and
Seed, et. al., (2) introduced the concept of resilient modulus. In. both
tests, repeated loads are applied to a specimen and the displacements are
measured. These values are measured on uniaxial compression or indirect
tensile specimens. The modulus value is calculated as the ratio of stress to
recoverable (resilient) strain under repeated loading conditions. A major
difference between the resilient modulus and complex modulus is that in
elastic as well as elastic deformations are measured in the complex modu-
lus test. Another· difference is the complex modulus test uses sine wave
loading whereas the resilient modulus test uses a square wave loading. The
calciilafions for stresses and strains in the resilient modulus test are dis-
cussed under the section on Indirect Tension.
A third stiffness test uses a beam subjected to repeated flexure to mea-
sure the modulus of elasticity. A flexural stiffness modulus is calculated
from the center point deflections under an applied load. The calculations
for stiffness using this test method are provided under flexural fatigue.
A majority of current stiffness measurements are obtainectusiiig a ver-
sion of the repeated load indirect tensile test (diametral test). This test is
usually performed over a range of temperatures, and the relationship
between resilient modulus and temperature is then shown on a plot.
Resilient-Modulus-(-Repeated-b0ad-Indirect-Tensile}fMR7:-B-e.:li:nse-o-f
its simplicity and applicability to test field cores, the MR test has become
popular with many laboratories. This is the most common method of
measuring stiffness modulus for HMA. A test setup somewhat similar to
that for the indirect tensile test is used to measure MR. The primary dif-
ference is that the test equipment must be able to apply repeated loads.
The procedures for conducting this test are described in ASTM D4123
and the data is typically plotted as shown in Figure 5-2. The equations for
computing the tensile str,ength and strain to failure for the indirect tensile
test can be used for computing stress and strain in the resilient modulus
test. The applied stress is calculated the same way as that for tensile
strength; however, the specimen is not loaded to failure when measuring
resilient modulus. For the resilient modulus test, the specimen is normal-
ly loaded to a stress level between 5 and 20 percent of indirect tensile
strength. Therefore, prior to conducting the resilient. modulus test, the
tensile strength must be measured or estimated. The load is typically
applied for 0.1 seconds along with a rest period (no load) of 0.9 seconds.
Hence, the HMA sample receives one load cycle per second. Other com-
binations of loading times and rest periods have also been used.
The equations that have been derived to estimate Poisson's ratio differ .
slightly for 4-inch and 6-inch samples. The two equations are shown
below:
µ =
H - 0.27
3.59 V for 4 inch (IO 1.6 mm) samples
µ = 4.09 {f - 0.27 for6 inch (152.4 mm) samples
C
a b
(\ '\ (', r
I I \ I '
I \ / l
Time
(al Load-Time Pulse
a = load duration
b = recovery time
c = cycle time
(bl Vertical Deformation Versus Time

Figure 5-2. Load and Deformation Versus Time Relationship
where,
µ = Poisson's ratio;
H = horizontal deformation, inches; and
V = vertical deformation, inches.
As stated earlier, Poisson's ratio is usually assumed to be 0.35 and, hence,

is not usually calculated from the above equation. However, Poisson's ratio
does vary with temperature. Typical values that have been used are 0.25
for 41 °F (5°C) and 0.40 for 104°F (40°C).
The equation derived to calculate the resilient modulus (MJ is the.
same for 4-inch (101.6 mm) and 6-inch (152.4 mm) samples. The equa-
tion is:
where,
MR = resilient modulus, psi;
P = applied load, pounds;
H = horizontal deformation, inches;
t = sample thickness, inches; and
µ = Poisson's ratio.
Assuming that Poisson's ratio is 0.35, the above equation can be further
reduced to:
M R-_ 0.62P
Ht
The resilient modulus test is becoming more routine within many labora-
tories because of the emphasis being placed on the measurement of funda-
mental engineering properties. Thus far, there is no good correlation
between MR and rutting, but MR at low temperatures is somewhat related
to cracking. It has been shown that stiffer mixtures .at low temperatures
tend to crack earlier than more flexible mixtures.
Example 5-1. Assume that a repeated load of 100 pounds (444 N) is

applied to a HMA specimen. The circular specimen is 2.5 inches (63.5
mm) tall. The horizontal deformation is 0.00015 inches (3.8 microns).
•<e-----Determine-the-resilient-modulus.--------------------
Remember that
1
MR-_ 0.62P
Ht
substituting,
M _ 0.62(100)
R - 0.00015(2.5)
solving,
MR= 165,000 psi (1138.5 MPa)
Stiffness Modulus by Indirect Means. One method of determining

the stiffness modulus of conventional asphalt cements and HMA mix-
tures was originally developed by Van der Poel ( 1 O) and presented in the
form of a nomograph. The procedure involved has been described in
numerous references and is discussed in Chapter 2. Reference 11 con-
tains a complete description, including the modificati~ns developed by
Heukelom and McLeod. It is rapid and simple to use, employing val-
ues of penetration and softening point to derive asphalt stiffness for the
desired time and temperature. Then, the asphalt stiffness is used to
derive the mixture stiffness. Since these correlations involve empirical
properties, the derived stiffness values of the mixture are at best only
approximations.
Dynamic Complex Modulus. The dynamic modulus, sometimes

called complex modulus, is determined by applying sinusoidal vertical
loads to cylindrical samples while measuring the deformation. The test
procedure for determination of the dynamic modulus is prov1ded in
ASTM D3497. The height to.diameter ratio of the samples is required
to be at least 2 to 1 to minimize the effect of frictio1:1 at the top and bot-
tom of the samples. A height to diameter ratio as low as 1.0 has also
been used when a friction reducing material is placed between the spec-
imen and loading platens to reduce the end effects. The diameter of the
sample must be at least twice the maximum aggregate size. The sample
sizes that are most commonly used are 4 x 8 inches (101.6 x 203.2 mm)
and 6 x 12 inches (152.4 x 304.8 mm).
The applied load usually varies up to 3 5 psi (241. 5 kPa) and is usu-
ally defined as a percentage of the mixture strength. Specimens are typ-
ically tested at three different temperatures and at three loading frequen-
cies. T~e common temperatures used are 41, 77, and 104°F (5, 25, and
-----0.!?.Q-a-l'l:cl-1;he-e0m-m0n-l0adi-ng-frequern:;ies-an!-l-,-4, and 16 hertz.
The stress applied to the sample is determined by dividing the applied
load by the cross-sectional area. The strain is measured by attaching strain
gauges on opposite sides df the sample and averaging the measured strain
for the two gauges, or it is calculated from vertical deformations measured
with linear variable displacement transducers (LVDTs). A typical test

set-up for determining dynamic modulus is shown in Figure 5-3. A typ-
ical plot of data collected during the test is shown in Figure 5--4. The data
shows that a repeated load (of constant duration and magnitude) is applied
and the resulting deformation is measured.
The dynamic modulus is calculated by dividing the repeated stress by
repeated strain.
where,
E = dynamic modulus;
cr = applied stress; and
- £ = measured vertical strain.
Load Cell
Loading Head
Specimen
Strain Gages
--- --- - - - - - - - - - - - - - - H - - - - - - - - - - - - - - - - -
Figure 5-3. Typical Test Set-Op for Dynamic Modulus~

Poisson's ratio can also be determined from the test set-up by measuring
the horizontal deformation during the test. Attempts to measure Poisson's
ratio of HMA mixtures have usually been unsuccessful.
The primary purpose for measuring the dynamic modulus is to deter-
mine the stress-strain relationships in. a pavement structure under an
--
appliedJoad. Because the dynamic modulus~ 1s not a measure of strength,
a~alt mixture with a higher dynamic modulus does not necessarily
have higher strength. A higher dynamic modulus indicates that a given
applied stress results in lower strai1:;jp. the mixture.
This test h~ limited use because of the time involved, complex equip-
ment required to conduct the test, and height requirement of test speci-
mens. This test cannot be used to test cores taken from pavements, since
the pavement layers are normally no more than two inches thick and the
cores must be at least 8 inches (203.2 mm) tall for a 4-inch diameter spec-
imen. In some cases, cores have been stacked on top of each other to pro-
vide the required 8 .inches (203.2 mm), but this practice is questionable.
LOAD
I\ ,,
I I I
I '
I
I I
---
J I \
I I I I I 1 I I I I I I I I I I I I I I I I I I I I I I I I I I I I
O,l sec.
STRAIN
,_ I"'-- -- -- __ -- _, _ _ ,_ -- - ---
,
i,.-i- 1,i-
· - - - - - - - - ·--
,.,i- 1 ......
"" , --i-.. ,i-

"'-
,.
""
I I I I 1 I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I
Figure 5-4. Recorded Trace from a Dynamic Complex Modulus Test (1)
Flexural Test. The repeated flexural test is conducted to evah.1ate the

fatigue properties of an asphalt mixture and, hence, to estimate pavement life
U2., 13., 11) for fracture. Repeated haversine loads are applied at the third
points of a beam specimen. The load rate is variable, but is normally 1 to 2
cycles per second. This produces a constant bending moment over the center
portion of the beam. A load is often applied in the opposite direction, forcing
the beam to return to its original position to maintain the zero position during
the rest period. The deflection caused by the load is measured at the center of
the beam. A schematic of the apparatus is shown in Figure 5-5.
KEY
I. Reaction Clamp
2. Restrainer
3, End Plates
4. Specimen
5. Loading Rod
6. Stop Nuts
7. Load Bar
8. Piston Rod
9. Thompson Ball Bushing
10. LVDT Holder
I I. LVDT
"REPEATED FLEXURE APPARATUS

Figure 5-5. Repeated Flexure Apparatus for Determining Stiffness and
Fatigue Characteristics (1)
The stress and strain at the outer fibers, and the stiffness modulus after
about 200 load applications, are calculated using basic relationships for
stresses and strains in beams. Simplifying this information provides the
following equations:
E = Pa(3l 2-4a2)
s (48Id)
Chapter 5-Characterization of Asphalt Mixtures. 293
where,
cr tensile stress in the outer fibers, psi;
=
tensile strain in the outer fibers, inches/inch;
£ =
E5 = flexural stiffness modulus, psi;
a = distance between support and first applied load,
inches; ·
P = total dynamic load with 1/2 P applied at third
points, lbs;
b = specimen width, inches;
h = specimen height, inches;
1= reaction span length, inches;
I = moment of inertia of specimen, in4; and
d = dynamic deflection of beam at the center, inches.
This test can be conducted in the constant stress or constant strain

mode ofloading. In the constant stress mode the applied load, Pis repeat-
ed until failure occurs. In this case, failure results when the sample actual-
ly breaks. For small stresses, the time to conduct this test for one sample
may be several hours, several days, or even several weeks. The log of the
number of cycles to failure is plotted versus the log of applied stress (Figure
5-6). In the constant stress mode of loading, materials with higher stiff
1/)
0..
1/)
1/)
-
....
Cl)
Cl)
"C
-~
0..
0..
<(
0
C"I
0
_j
Log of Cycles to Failure, Nt

Figure 5-6. Typical Plot of Fatigue Data Using Constant Stress Mode of
Loading
ness values generally perform better than materials with lower stiffness val-
ues. Another plot that is often made for a constant stress mode ofloading
is log of initial strain versus log of cycles to failure.
In the constant strai~ mode of loading, the deflection is maintained
constant and the applied load is allowed to decrease with increasing cycles.
Failure is more difficult to define in the constant strain test since the sam-
ple usually does not break. Failure is normally taken as the point where the
applied load reaches some preselected percent of the original applied load.
For instance, 25 and 50 percent of the original load are typical values that
are often selected as failure. Two plots are normally prepared from the data
obtained in this test. The first is the log of applied strain versus the log of
cycles to failure and the other is the log of initial stress versus the log of
cycles to failur~. Materials that are more flexible (lower stiffness) perform
better in the constant strain test when the log of applied strain is plotted
against log of cycles to failure.
Example 5-2. A beam is being tested for flexural fatigue. The reac-
tions are 12 inches (305 mm) apart and a total load of 100 pounds (444 N)
is being applied at third points (50 pounds (222 N) at each 1/3 point).
After 1000 stress applications, the measured deflection is 0.0011 inches
(0.28 mm) per load application. Determine the flexural stress, flexural
strain, and stiffness modulus. The cross sectional dimensions of the beam
are 3 inches by 3 inches (76.2 mm X 76.2 mm).
The flexural stress has been determined to be
3aP
a=-
bh2
hence,
= 3(4)(100)
0
- 3(3)2
Simplifying the equation, the flexural stress for the load conditions in
this example is
cr = 44.4 psi (306.4 kPa)

Chapter 5---Characterization of Asphalt Mixtures 295
The flexural strain can be determined from the following equation
12hd
e
312 - 4a 2
hence,
e= 12(3X0.0011)
3(12)2 - 4(4)2
e = 108 X 10-6 inches/inch (mm/mm)
Now the flexural stiffness can be determined by dividing stress by

strain or by substituting into the following equation:
E = Pa(3l2 - 4a 2)
s (481d)
Recall that the moment of inertia (I) for a rectangular cross section is:
bh 3
l=-
12
Substituting,
I= 3(3)3
12
solving,
I= 6.75 inch4 (281.0 crn4)
Now substituting into the equation for stiffness modulus
E = 100(4) [3(12)2 - 4(4)2)

s 48(6.75)(0.0011)
solving,
Es= 413,000
It can also be shown that the stiffness modulus is equal to stress divid-
ed by strain. For this example:
E = .!!_ = 44.4
• e 108 X 10-6
solving,
Es= 411,000 psi (2835.9 MPa)
The difference between the two calculated stiffness values is caused by

rounding errors.
The flexural fatigue test takes a significant amount of time to conduct

since approximately 10 samples have to be tested to produce enough data
to develop the required plots (13.). The time to conduct sufficient tests to
develop a plot is likely to exceed 2 weeks. In addition, the equipment
needed is complicated and expensive. For these reasons, this test is nor-
mally used for research and not for evaluating mixtures during mix design
or quality control.
Experience has shown that thick asphalt pavements (greater than 5
inches (BO mm)) generally perform close to the constant stress mode of
loading while thin asphalt pavements perform close to the constant strain
mode of loading. It follows then that thicker sections require stiffer HMA
mixtures and thinner sections require more flexible mixtures for best
fatigue performance. In practice, however, the mixture stiffness is not
modified for different pavement thicknesses.
In the constant stress mode of loading, the general equation for the
plot of log of applied stress versus log of cycles to failure is:
N = K (..!..)111
1 1 C1
where,
Nf = number of cycles to failure;
--flexural-stress1.n-extreme-fiber;-psi;-arr,1---------
K1,n2 = regression constants.
A similar equation is produced when comparing the log of cycles to failure

to log of strain.
where
Nf = number of cycles to failure;
£ = flexural strain in extreme fiber, inches/inch; and
K2,n2 = regression constants.
The same two equations can be used when analyzing initial stress or initial
strain. For example, the initial stress, determined in the constant strain
mode ofloading, can be plotted versus cycles to failure (1.1). The constants
KI> K2 , np and n 2 are determined from a regression analysis of the labora-
tory data.
Rallhby 8.
Sterling
n,-3.87
K~-3.65xl 0 11
T=77°
Pell et al
n2 -5.3
K~•l.lxl0 19
T=32°F
Pell et al
n2 -3.9
K~-3.0xl0 12
T=50°F
Kennedy et al
n2 -3.88
K~-3.0xl0 11
T=75°F
0
10 ,02
Stress, psi
Figure 5-7. Typical Stress-Fatigue Life Relationships (1.i)
This flexural fatigue test is used to estimate the fatigue life of a ,pave-
ment. For a given pavement and loading condition the maximum flexur-
al stresses and strains can be computed. By knowing the stresses or strains
that the pavement experiences and by knowing the relationship between
stress or strain and cycles to failure, the number of traffic passes to failure
can be estimated. Examples of test results are shown in Figure 5-7. Since
it is impossible to duplicate field conditions in the laboratory, the predict-
ed traffic applications to failure is not a good estimate of actual applications
to failure. Even in a highly controlled laboratory environment, the results
of flexural fatigue tests are highly variable. Thus, this test is normally used
to compare various mixtures and rank them for relative performance and
is not used to provide accurate estimates of the 'amount of traffic to failure
for specific projects.
Indirect Tension. The indirect tensile test is performed by loading a

cylindrical specimen with a single or repeated compressive load which acts
parallel to and along the vertical diametral plane (Figure 5-8) (ASTM
D41~3). This loading configuration develops a relatively uniform tensile
stress perpendicular to the direction of the applied load and along the vertical
diametral plane, which ultimately causes the specimen to fail by splitting
along the vertical diameter (Figure 5-9).
A 0.5-inch (12.7 mm) wide curved loading strip is used for 4-inch
(101.6 mm) diameter specimens to provide a uniform loading width which
produces a nearly uniform stress distribution. The equations for tensile
stress and tensile strain at failure have been developed and simplified ( 16,
1Z, 1..9). These equations assume the HMA is homogenous, isotropic, and
elastic. None of these assumptions are true, of course, but estimates of
(
properties based on these assumptions are standard procedure and are use-
ful in evaluating relative properties of HMA mixtures. The equations for
the indirect tensile stress and strain at failure are provided below:
(1
2P
·=-
z 1tdt
6P~-------------------
y
1tdt
where,
c;x = horizontal tensile stress at center of specimen, psi;
cry = vertical compressive stress at center of specimen,
psi;
Ef = tensile strain at failure, inches/inch;
P = applied load, lbs.;
d = diameter of specimen, inches;
t = thickness of specimen, inches; and
x, = horizontal deformation across specimen, inches.
The above equation applies for 4-inch (101.6 mm) diameter samples hav-
ing a 0.5 inch (12.2 mm) curved loading strip and for 6-inch (152.4 mm)
diameter samples having a 0.75 inch (19.0 mm) curved loading strip.
Figure 5-8. Indirect Tensile Figure 5-9. Indirect Tensile

Test During Test at Failure
Loading
The indirect tensile test provides two mixture properties that are use-
ful in characterizing HMA. The first property is tensile strength, which is
often used in evaluating water susceptibi_!Lt)'.:.J)Lmixtures. For water sus-
ceptibility, the tensile strength is measured before and after water condi-
tioning of samples to determine the retained tensile strength as a percent
of the original tensile strength. A high number typically indicates that
-----o-g,ood-perfor:rri:a:ne:e-is--e-:x:peetecl:---whi-1-e-a-lew-number-typieally ·indicates that
poor performance is expected. The tensile strength is also sometimes used
to help evaluate c~acking potential of an asphalt mixture. However, the
second property determined from the indirect tensile test, which is tensile
strain at failure, is more useful for predicting ~ t e n t i a l . Mixes
1
that can tolerate high strains prior to failure are more likely to resist crack-
ing than mixtures that cannot tolerate high strains.
The indirect tensile test is normally conducted at a deformation rate of
2 inches/minute (50.8 mm/minute) and ata temperature of77°F (25°C).
A deformation rate of 3 inches/minute (76.2 mm/minute) should be used
for 6-inch (152.4 mm) samples if a strain rate similar to that for 4-inch
(101.6 mm) diameter samples is desirable. In practice, the same machine
speed is normally used for 4-inch (101.6 mm) and 6-inch (152.4 mm)
samples; however, the measured tensile strength for 6-inch (152.4 mm)
samples is expected to be smaller since the strain rate is lower. Tensile tests
can also be conducted at other temperatures (especially low temperatures)
to evaluate mix properties closer to the anticipated performance tempera-
ture. When tested _at lower temperatures, the data can be used in thermal
crack prediction models. '-
Examples 5-3. A 4-inch (101.6 mm) diameter core is obtained from

an existing ·paveme;nt. The core thickness is determined to be 2 inches
(50.8 mm) and is tested in the indirect tensile mode of loading. The fail-
ure load was determined to be 835 pounds (3711 N) and the horizontal
defor~ation was determined to be 0.0023 inches (58 microns).
Determine the tensile strength and strain at failure.
Recall that the tensile strength for a 4-inch (101.6 mm) diameter
specimen is equal to
cr = 2P
rcdt
substituting into the equation,
cr = 2(835)
3.14(4)(2)
solving,
cr = 66 psi (455.4 kPa)
The tensile strain at failure for a 4-inch (25.4 mm) diameter sample
(assuming 0.5 inch loading strip is used) is
e = 0.52x,
1
The horizontal deformation was measured to be 0.0023, hence,
solving
e
1
= 0.0012 inches/inch
Chapter 5----'Characterization of Asphalt Mixtures 301
Creep and Permanent Deformation. The creep test was developed to

estimate the r~_.ming potential of asphalt mixtures. This test is conducted
by applying a static load to a HMA specimen and measuring the resulting
permanent deformation with time. A typical creep plot is shown in Figure
5-10.
Loading - - ---Unloading
Strain
Time
Figure 5-10. Typical Creep Relationship
There are 3 general ways that the creep test can be performed:
1. Uniaxial static unconfined; /'-

2. Uniaxial static confined; and / , ,
r ;;; c,,;i_.etv>~
t ~ ~•\,v'-
3 • Indl.rect tens1·1e. ~ ) . / , ~ 1t.
I
The indirect tensile creep is typically used for l9w temperature cracking
potential, whereas the uniaxial creep (confined or unconfined) is used for
rutting potential.
The permanent deformation test closely relates to actual field traffic.
This test requires that a confining pressure be exerted on the sample and a
repeated load be applied to represent repeated traffic loads. This test is
expensive to set up and difficult to perform. For these reasons the dynam-
ic test is not commonly used to estimate permanent deformation at pre-
sent.
The,most popular creep test is the static unconfined. This test involves
-----ap·pfrcation-of-a-staric-lnad-w-a-sample-fora-specified-time and tempera-
ture, and measurement of deformation as the sample deforms. This test is
inexpensive and easy to conduct, but the ability of the test to predict per-
formance is questionable. With the unconfined compression test, the
applied pressure cannot usually exceed 30 psi (206.9 kPa) (otherwise the
sample fails), and the test temperature cannot usually exceed 104°F
(40°C). In-place asphalt mixtures are typically exposed to truck tire pres-
sures of approximately 120 psi (828 kPa) and maximum temperatures of
140°F (60°C) or higher. Therefore, this test does not closely simulate
in-place conditions. ,
One type of creep test that has been used that is relatively simple and
easy to perform is the static confined test,. which more closely relates to
field conditions. By applying a confining pressure (usually approximately )
20 psi (138 kPa)), the sample can be tested at a vertical pressure up to 1:20
psi (828 kPa) (or higher) and at a temperiture up to 140°F (60°C). These
test conditions are more closely related to actual field conditions than those
for static unconfined. Research suggests that the static confined creep test
does a better job of predicting field performance than the static unconfined
creep test.
An integral part of the creep test that may be utilized is the rebound
portion of the curve obtained after the load is removed. Creep in pave-
a
ments is not necessarily problem if significant rebound occurs after the
load is removed. In the laboratory, a static load is typically applied for one
hour; then the load is removed and the sample is allowed to rebound for
another hour.. The permanent deformation is the maximum deformation
minus the amount of rebound. When comparing various mixtures i~ is this
permanent deformation that is important for estimating performance.
Creep compliance is often calculated from the creep test results. The
creep compliance, J,, is. calculated by dividing the strain by the applied
stress as follows:
ft_,,; ci ~t any test temperature, T, and time ofloading, t

where,
~\ = strain at time t, inches/inch;:, and
cr = applied stress, psi.
In addition to normal permanent strain characteristics, the permanent.
strain properties used in VESYS (12, 2flJ involve ~o basic parameters,
GNU and ALPHA, whicli Wused to describe the permanent deformation
characteristics of HMA mixturd and to predict rutting. VESYS is a pave-
ment design procedure utilizing viscoelastic. concepts that was developed
for the Federal Highway Administration.
The theory assumes that the logarithmic relationship between the
number~of·repeated-loads-and--permarrenntrain7.s esserfti:allylinear over a
range ofload applications (Figure 5-11) and can be described by the equa-
tion
Ea = IN5
-0
Cl)
c
::i C:
...
E"a
:::,
u-
0 U)
<-c:
-CD
0 C:
t:I
EE
..c: ...
-CD
15 a..
t:l'I
0
_J Arithmetic Value I
Logarithm of Number of Load Repetitions

Figure 5-11. Assumed Logarithmic. Relationship between Permanent
Strain and Number of Load Repetitions (12)
where,
Ea = accumulated permanent strain;
I = intercept with permanent strain axis (arithmetic ,
strain value, not log value) (Figure 5-11);
N = number of load applications; and
S = slope of the linear portion of the logarithmic
relationship.
GNU is defined as:
and ALPHA is defined as:
where,
£r = resilient strain which is considered to become
constant after a few load applications (Figure 5-12).
he-origina:l-corrcept:-crti:-lized,epeated-load7:estnoproduce accumu-
lated permanent strain relationships. Subsequently, sequential creep tests
were utilized. An evaluation of different tests indicates that the permanent
strain relationships for tests, which involve tensile stresses, are similar but
different from those for the triaxial compression tests.
t
w CNl = resilient
strain
V__-1'_ _.,__
€a = accumulated
permanent strain
Number of Load Repetitions

l
Figure 5-12. Typical Relationship betwee~ Strain and Number of Load
Repetitions (1.2.)
Gyratory Testing Machine (GTM). The GTM developed by the

Corps of Engineers has been shown to be an effective tool in the evalua-
tion of HMA mixture quality. This machine (Figure 5-13) has the capa-
bility to compact HMA mixtures using a kneading process that simulates
Figure 5-13. Gyratory Testing Machine

the action of rollers during construction. The GTM has the flexibility of
·varying the vertical pressure, gyration angle, and number of gyrations to
simulate field compaction equip_lll~f!~ and subsequent traffic.Typically, the
vertical pressure applied is 12(Ypsi (828 k:Pa), whichis approximately equal
to truck tire inflation pressures. The s;ttings for the gyration angle and the
number of revolutions vary between laboratories but typical values are 1
degree gyration angle and 300 revolutions. ·
During compaction of a specimen in the GTM, several mixture prop-
erties are determined. The gyratory shear index (GS!) is a measure of mix-
ture stability and is related to permanent deformation. The GSI is deter-
mined by dividing the intermediate gyration angle by the initial angle.
The gyration angle increases during compaction for unstable mixtures due
to plastic Bow of the asphalt mixture. The gyration angle does not increase
significantly for stable mixtures. GSI values close to 1.0 have been shown
to be typical for stable mixtures and values significantly above 1. 1 usually
indicate unstable mixtures.
The GTM also has the capability of measuring the shear strength of
the mixt_~r~ _during__compaction The pressure required to producetfie
desired gyration angle is measured and can be converted to shear strength.
Strength, which is measured during compaction at high temperatures, is
primarily a measure of aggregate properties, since the viscosity of the
asphalt is low and the mixture has little cohesion. It is these aggregate
properties that must provide the support to resist permanent deformation
caused by traffic.
The GTM can be used for mix design or quality control of HMA.
This equipment does a good job of achieving the ultimate density that is
obtained in the field. It also has the flexibility of being adjusted to simu-
late the tire pressures of any traffic type including cars, trucks, and aircraft,
and is the only device which can monitor the change in mixture response
with densification. This becomes critically important for traffic densifica-
tion studies or effects.
Laboratory Wheel Tracking Tests. Initial adoption of Superpave

includes performance graded asphalt cements and a volumetric mix design
procedure. The volumetric mix design procedure alone has no proven rela-
tion to performance. The Superpave performance tests had not been final-
ized at the time that Superpave volumetrics was adopted and it appeared
that a significant amount of time would elapse between the time when
-----Super-pa:v-e-v0hu1'le1=r-ies-was-acl:0ptea-ancl-the---per-formance based mixture
tests would be available.
Many specifiers looked for another test that could be used until the
Superpave mixture tests were finalized. The most promising test to evalu-
ate the potential of an HMA mixture to ru5 appeared to be some type of
Figure 5-14. Laboratory wheel tracking device.
laboratory wheel tracking test (Figure 5-14). Several manufacturers have

produced some type of laboratory wheel tracking tests. Some of these
-
devices conduct tests underwater and so~e. in a temperature controlled
~
chamber. At least one device allows testing to be done _gry_or under water.
These tests are often referred to as proof tests. They do seem to pro-
vide some laboratory measure of potential for rutting. However, there is lit-
tle published information correlating these tests with actual observed per-
formance.
Moisture Susceptibility Tests. Numerous test methods have been

developed and used in the past to predict the moisture susceptibility of
HMA mixes (.21, 22, 2.3.). However, no test has any wide acceptance. This
is due to their low reliability and lack of satisfactory relationship between
laboratory and field conditions. Only selected test methods which are
commonly used by some agencies will be discussed briefly.
Quantitative or Subjective Tests
1-----+.--Soili:ng-Wa:1:er-Test (ASTM D3625 or a variation}:-boose-HM:A-:rnix-i~----

added to boiling water. ASTM D3625 specifies a IO-minute boiling
period. The percentage of the total visible area of the aggregate that
retains i ~ i n g ~ boiling I; estimated as above or oelow
95 percent. This testcanl>eused for initial screening of HMA mixes.
Chapter 5-Characterization of A?phalt Mixtures 307
Some agencies use it for quality control during production to deter-

mine the presence of antistripping agent. This test method does not
involve any strength analysis. Also, determining the stripping of fine
aggregate is very difficult. This test method generally favors liquid anti-
stripping agents over lime.
2. Static-Immersion Test (MSHTO Tl82): A sample of HMA mix is
immersed in distilled water at 77°F (25°C) for 16 to 18 hours. The
sample is then observed through water to estimate the percentage of
total v_is.ible_ar_e.a.__oLthe....aggregate which remains c2_~t~9- as above or
below 95 percent. Again, this method does not involve any strength
test.
Quantitative Strength Tests
l. Lottman Test (NCHRP 246): This method was developed by Lottman

~ under the National Cooperative Highway Research Program 246.
Nine specimens 4 inches (102 mm) in diameter and 2½ inches (64
mm) high are compacted to expected field air void content. Specimens
are divided into 3 groups of 3 specimens each. Group 1 is treated as
control without any conditioning. Group 2 specimens are vacuum sat-
urated (26 inches or 660 mm Hg) with water for 30 minutes. Group 3
specimens are vacuum saturated like Group 2 and then subjected to a
freeze (0°F or -18°C for 15 hours) and a thaw (140°F or 60°C for 24
hour) cycle. All 9 specimens are tested for resilient modulus (MiJ
and/or indirect tensile strength (ITS) at 55°F (13°C) or 73°F (23°C).
A loading rate of 0.065 inch/minute (1.65 mm/minute) is used for the
ITS test. -
Group 2 reflects field performance up to 4 years. Group 3 reflects
field performance from 4 to 12 years. Retained tensile strength (TSR)
is calculated for Group 2 and Group 3 specimens as follows:
TSR = ITS of Conditioned specimens
ITS of Control specimens
A minimum TSR of 0.70 is recommended by Lottman and Maupin

(11, 25) who reported values between 0.70 and 0.75 differentiated
between stripping and nonstripping HMA mixtures. It has been argued
that the Lottman procedure is too severe because the warm water soak
-------of'-tlie vacuum saturated and frozen specimen c:iiiaevelop internal
water pressure. However, Stuart (2§) and Parker and Gharaybeh (.2Z)
generally found a good correlation between the laboratory and field
results. Oregon has successfully used this test with modulus ratio in lieu
of tensile strength ratio (TSR).
308 . Hot Mix Asphalt Materials, Mixture Design and Construction
2. Tunnicliff and Root Conditioning (NCHRP 274): This method was

proposed by Tunnicliff and Root under the NCHRP Project 274 (2.i'J.
They proposed six specimens to be compacted to 6-8 percent air void
content and divided into two groups of three specimens each. Group 1
is treated as control without any conditioning. Group 2 specimens are
vacuum saturated (20 inches or 508 mm Hg for about 5 minutes) with
water to attain a saturation level of 55 to 80 percent. Specimens satu-
rated more than 80 percent are discarded. The saturated specimens are
then soaked in water at 140°F (60°C) for 24 hours. All specimens are
tested for ITS at 77°F (25°C) using a loading rate of 2 inches/minute
(51 mm/minute). A minimum TSR of 0.7 to 0.8 is usually specified.
The use of a freeze-thaw cycle is not m~gated in ASTM D4867-88
which is based on this method. The freeze-thaw cycle is optional. The
primary emphasis is on saturation of the specimen which for a short
duration of about 24 hours has been reported to be insufficient to
induce moisture-related damage (22).
3. Modified Lottman Test (MSHTO T283): This method was proposed
by Kandhal and was adopted by MSHTO in 1985 (3.Q). It combines
the good features of Lottman test (NCHRP 246) and the Tunnicliff and
Root test (NCHRP 274). Six specimens are compacted to 6-8 percent
air void content. Group 1 of three specimens is used as a control.
Group 2 specimens are vacuum saturated (5 5 to 80 percent saturation)
with water, and then subjected to one freeze and one thaw cycle as pro-
posed by Lottman. All specimens are tested for ITS at 77°F (25°C)
using a loading rate of 2 inches/minute (51 mm/minute), and the TSR
is determined. A minimum TSR of 0. 7 is usually specified.This method
is gaining acceptance by the specifying agencies.
4. Immersion-Compression Test (MSHTO T165): Six specimens 4 inch-
es (102 mm) diameter x 4 inches (102 mm) high are compacted with a
double plunger with a pressure of 3,000 psi (20.7 MPa) for 2 minutes
to about 6 percent air void content. Group 1 of three specimens is
treated as control. Group 2 spocimens are placed in water at 120°F
(49°C) for 4 days or at 140°F (60°C) for 1 day. All specimens are test-
ed for unconfined compressive strength at 77°F (25°C) using 0.2
inch/minute (5.1 mm/minute) loading rate. The retained compressive
strength is determined. Many agencies specify at least 70 percent
-+:(_ _ _ _ _ retained strength. · This test has produced retained strengths near 100
percent even when stripping is evident. Stuart (3.1) has attributed this
to the internal pore water pressure and the insensitivity of the compres-
sion test to measure the moisture induced damage properly. Lack of sat-
isfactory precision has been a major problem with this test.
i
,:I
I
5. Other Tests: Moisture-vapor susceptibility, swell test, and a film strip-

ping test are used by California DOT. Retained Marshall stability is
used in Puerto Rico and some other states.
Evidently, a wide variety of test methods are being used by various
agencies. However, no test has proven to be "superior" and can correctly
identify a moisture-susceptible mix in all cases. This means that many
HlviA mixes which might otherwise perform satisfactorily in the field, are
likely to be rendered unacceptable if these tests and criteria are used. The
use of these tests has simply encouraged the increased use of antistripping
agents in many states.
There are still many concerns and requirements related to the test
methods which need to be addressed:
1. Proliferation of test procedures and criteria.

2. Reproducibility of most test methods is not satisfactory. For example,
small variations in air void content of the specimens can significantly
affect the TSR results.
3. Need to consider minimum wet strength (if the desired value can be
established) of the conditioned specimens rather than relying solely on
the TSR value. For example, some additives increase both dry and wet
strengths but might have a low TSR value.
4. Lack of satisfactory correlation between laboratory and field perfor-
mance.
However, based on a survey of states (22) it appears that the Modified
Lottman Test (AASHTO T283) is the most appropriate test method avail-
able at the present time to detect moisture damage in HlviA mixes. A min-
imum TSR of 0.70 is recommended when using this test method. This cri-
terion should be applied to the field-produced rather than laboratory-pro-
duced mixes. AASHTO T283 has been included in Superpave mix design
procedures.
The Strategic Highway Research Program (SHRP) had two research
contracts dealing with moisture susceptibility of HlviA mixes. SHRP pro-
ject A-003A "Performance Related Testing and Measuring of
Asphalt-Aggregate Interactions and Mixtures" attempted to develop an
improved test method to evaluate moisture susceptibility. SHRP project
A-003B "Fundamental Properties of Asphalt-Aggregate Interactions
Including Adhesion and Adsorption" studied the fundamental aspects of
-----a-s-pnalt-aggregate bond.
A Net Adsorption Test (NAT) was developed under SHRP A-003B
completed by the National Center for Asphalt Technology. It is a prelim-
inary screening test for matching mineral aggregates and asphalt cement
(32) and is based on the principles of adsorption and desorption. A solu-
tion of asphalt cement and toluene is introduced and circulated in a reac-

tion column containing the aggregate sample. Once the solution temper-
ature has been stabilized, 4 ml of solution is removed and the absorbance
is determined with a spectrophotometer. Fifty grams of minus No. 4 (4.75
mm) aggregate is then added to the column, and the solution is circulated
through the aggregate bed for 6.5 hours. A second 4-ml sample of the
solution thc;n is removed from the column and the absorbance is again
determined. The difference in the absorbance readings is used to deter-
mine the amount of asphalt that has been removed from the solution
(adsorption) because of the chemical attraction of the aggregate for the
molecular components of the asphalt cement. Immediately after the sec-
ond solution sample is taken, 575 µm of water is added to the column.
The solution is then circulated through the system for another 2 hr. A final
4ml of solution is taken from the column at the end of this time. The
increase in the absorptivity is a measure of the amount of asphalt cement
that is displaced by water molecules (desorption). Additional validation
data are needed for the NAT.
An Environmental Conditioning System (ECS) was developed in
SHRP A-003A (33) in which HMA samples are exposed to wetting and
accelerated hot-cold cycling to represent actual field exposure, including
repeated loading to simulate traffic. The modulus of the HMA specimen
and change in air and water permeability are monitored during the condi-
tioning after each cycle, and tensile strength and stripping are measured at
the conclusion of conditioning. Both warm- and cold-climate condition-
ing can be performed. Modulus ratio and water permeability ratio are cal-
culated after completing each conditioning cycle. A provisional MSHTO
standard, Designation TP34, "Standard Test Method for Determining
Moisture Sensitivity Characteristics of Compacted Bituminous Mixtures
Subjected to Hot and Cold Climate,Conditions," is available. The ECS
system is expensive but versatile; however, sufficient field-validation data
· are not available at this time (1995) to warrant its use in lieu of MSHTO
T283.
REFERENCES
1. Kennedy, T.W., R. Haas, and F.R. Meyer. Characterization of

-----~Pav:emen.L-Materials__fo.r_EundamentaL-..St1:uctur.al-Ana-l3/.ses.,-----
Inter-American Conference on Materials Technology, Caracas,
Venezuela, 1975.
2. Brown, E.R. and Stephen Cross. A Study of In-Place Rutting of
Asphalt Pavements. Proceedings, Association of Asphalt Paving
3. Ford, Miller. Pavement Densification Related to Asphalt Mix

Characteristics. Paper presented at 67th annual Transportation
Research Board, January 1988.
4. Huber, G.A. and G.H. Herman. Effect of Asphalt Concrete
Parameters on Rutting Performance: A Field Investigation. Proceed-
ings, Association of Asphalt Paving Technologists, Volume 56, 1987.
5. Zube, Ernest. Compaction Studies of Asphalt Concrete Pavement as
Related to the Water Permeability Test. Paper presented at the Annual
Meeting of the Highway Research Board, 1962.
6. Brown, E.R., R. Collins, and J.R. Brownfield. Investigation of
Segregation of Asphalt Mixtures in State of Georgia. Paper presented
at the Annual Meeting of the Transportation Research Board, 1989.
7. Santucci, L.E., D.D. Allen, and R.L. Coats. The Effects of Moisture
and Compaction on Quality of Asphalt Pavements. Proceedings,
8. Papzian. The Response of Linear Viscoelastic Materials in the
Frequency Domain with Emphasis on Asphaltic Concrete.
Proceedings, International Conference on the Structural Design of
Asphalt Pavements, University of Michigan, 1962.
9. Seed, H.C., C.K. Chan and C.E. Lee. Resilience Characteristics of
Subgrade Soils and Their Relation to Fatigue Failures in Asphalt
Pavements. Proceedings, International Conference on the Structural
Design of Asphalt Pavements, University of Michigan, 1962.
10. Van der Poel, C. A General System Describing the Viscoelastic
Properties of Bitumens and Its Relation to Routine Test Data. Journal
of Applied Chemistry, Volume 4, May 1954.
1 l. Haas, R.C.G. A Method for Designing Asphalt Pavements to
Minimize Low-Temperature Shrinkage Cracking. The Asphalt
Institute, Research Report 73-1, January 1973.
12. Epps, J.A. and C.L. Monismith. Influence of Mixture Variables on
the Flexural Fatigue Properties of Asphalt Concrete. Proceedings,
Association ofAsphalt Paving Technologists, Volume 38, February 1969.
13. Kallas, B.F. and V.P. Puzinauskas. Flexural Fatigue Tests on Asphalt
Paving Mixtures. Fatigue of Compacted Bituminous Aggregate Mixtures,
American Society for Testing and Materials, STP 508, 1972.
14. Brown, E.R. Evaluation of Fatigue Properties of Recycled Asphalt
Concrete. Proceedings, Sixth International Conference on Structural
Design of Asphalt Pavements, Volume I, 1987.
15. Kennedy, T.W. and J.N. Anagnos. Procedures for the Static and
Repeated Load Indirect Tensile Tests. University of Texas, Research
Report 183-14, August 1983.
-----1.6.-Hadle.y,_Williaro O~.,-W.Jlonal.d..Hudsl>n,~d Thomas W. Kennedy.
An Evaluation of Factors Affecting the Tensile Properties of
Asphalt-Treated Materials. University of Texas, Research Report I
98-2, August 1972. 11
17. Hadley, William 0., W. Ronald Hudson, and Thomas W. Kennedy.

A Method of Estimating Tensile Properties of Materials Tested in
Indirect Tension. Center for Highway Research, University of Texas,

Research Report 98-7, July 1970.
18. Anagnos, James N. and Thomas W Kennedy. Practical Method of
Conducting the Indirect Tensile Test. University of Texas, Research
Report 98-10, August 1972.
19. Moavenzadeh, F., J.E. Soussou, and H.K. Findakly. Synthesis for
Rational-Design. Volume II, Final Report for FHWA Contract 7776,
January 1974.
20. Moavenzadeh, F., J.E. Soussou, H.K. Findakly, and B. Bradeneyer.
Synthesis for Rational Design-Operating Instructions and Docu-
mentation. Final Report for FHWA Contract 7776, January 1974.
21. Kandhal, P.S. Field and Laboratory Investigation of Stripping in
Asphalt Pavements: State of the Art Report. Transportation Research
Record 1454, 1994. ~
22. Hicks, R.G. Moisture Damage in Asphalt Concrete. NCHRP
Synthesis of Highway Practice No. 175, October 1991.
23.· Kandhal, P.S. Moisture Susceptibility of HMA Mixes: Identification
of Problem and Recommended Solutions. National Asphalt
Pavement Association Quality Improvement Publication (QIP) 119,
December 1992.
24. Lottman, R.P. Predicting Moisture-Induced Damage to Asphaltic
Concrete-Field Evaluation. TRB, NCHRP Report 246, 1982.
25. Maupin, G.W The Use of Antistripping Additives in Virginia. Proc.
Assoc. of Asphalt Paving Technologists, Vol. 51, 1982.
26. Stuart, K.D. Evaluation of Procedures Used to Predict Moisture
DamageinAsphaltMixtures.FHWAReportFHWAJRD-86/091, 1986.
27. Parker, F. and F. Gharaybeh. Evaluation of Indirect Tensile Tests for
Assessing Stripping of Alabama Asphalt Concrete Mixtures.
28. Tunnicliff, D.G. and R.E. Root. Use of Antistripping Additives in
Asphaltic Concrete Mixtures, TRB, NCHRP Report 274, 1984.
29. Coplantz, J.S. and D.E. Newcomb. Water Sensitivity Test Methods
for Asphalt Concrete Mixtures-A Laboratory Comparison.
1, 30. Resistance of Compacted Bituminous Mixture to Moisture Induced
11
Damage, Test Method T283-85. AASHTO, Part II-Methods of
,I Sampling and Testing, August 1986.
'i'.i 31. Stuart. K.D. Moisture Damage in Asphalt Mixtures-A State-of-
the-Art Report. FHWA, Report FHWA-RD-90-019, August 1990.
·1 32. Curtis, C.W, K. Ensley, and J. Epps. Fundamental Properties of
·,."'·
_____
Asphalt-Aggregate Interactions Including Adhesion and Absorption:
FinalRep..ort....S~Q03.B.J9_9-_.___ _ _ _ _ _ _ _ _ _ _ _ _ __
1
f .••. •ii, 33. AAl-ShwalailCmi, S. anMd _R.L. TeUrr~l. thEvalEuati?n of WaalteCr Damd_ ~ge, of
sp t oncrete 1xtures smg e nvironment on monmg
' System (ECS). Journal of the Association of Asphalt Paving
I Technologists, Vol. 61, 1992.
CHAPTER 6. EQUIPMENT AND
CONSTRUCTION
INTRODUCTION
This chapter includes a discussion of equipment and procedures
involved in the construction of HMA. The equipment consists of HMA
facilities, trucks, placement equipment, and compaction equipment.
Construction specifications and procedures are also discussed along with
the most common methods of selecting contractors. This chapter provides
the guidance needed, after a HMA mixture is designed, to manufacture,
place, and compact the HMA.
HOT MIX ASPHALT FACILITIES

The basic purpose of a HMA facility is to properly proportion, blend,
and heat aggregates and asphalt to produce a HMA that meets the require-
ments of the job mix formula. Aggregates are typically a blend of 2 to 4
separate stockpiles of materials while the asphalt cement is typically a
paving grade asphalt cement with or without additives.
The two principal types ofHMA facilities in use today in the USA are
batch and drum mix facilities. A third type, continuous mix facility, is still
in service but is seldom encountered. No new continuous mix facilities
have been built in several years. Both the batch and drum facilities accom-
plish the samt: purpose, and the HMA material produced from either using
the same ingredients should be similar. Both types of facilities are dis-
cussed in detail in this chapter. A brief description of the layout of these
two types of facilities is made before the individual components are
described in detail.
Typical Layout
Drum Mix Facility. A schematic diagram of the components in a typ-
ical drum mix facility is shown in Figure 6-1. The :flow of materials is basi-
EB
..0
® -uB>"
c
>
C
0
@~
:iI
0
:c
'0
II>
:,1 II>
LL
11
'0
I 0
I, u
'·
@
II
Figure 6-1. Schematic Showing Major Components of a Typical

Drum Mix Facility
Chapter 6-Equipment and Construction 315
cally from left to right. The aggregates are typically moved from the stock-
pile area to the cold feed bins via a front end loader which charges the cold
feed bins. The minimum required number of bins depends on the num-
ber of different aggregates to be used in the HMA. Most facilities have 4-5
cold feed bins. The metal cold feed bins have slanted sides with a vibrator
attached (if gravity feed is inadequate) to keep a constant supply of mate-
rial fl.owing out of the bottom of the cold feed bin. Typically an adjustable
gate and a variable speed feeder are located at the bottom of each cold feed
bin to proportion the material from each bin to meet the job mix formula
QMF) gradation. A gathering conveyor brings the material to the cold
feed elevator which moves the material to the dryer. An automatic weigh-
ing system on this elevator continuously weighs the amount of aggregate
plus moisture going into the drum mixer. In the control room, a correc-
tion to the total weight is made for the ·moisture content of the aggregates
so the asphalt proportioning system can pump the proper amount of
asphalt cement into the drum mixer. The moisture correction is made by
physically determining the moisture content of the aggregate stockpile or
aggregate from the gathering conveyor and providing this information to
the computer.
A typical drum mixer is of a parallel fl.ow design in which the aggre-
gates move in the same direction as (parallel to) the exhaust gases, although
a number of counter fl.ow units are also being built. Therefore, the aggre-
gate is fed into the drum at the burner end and then it is dried and heated
as it moves down the drum toward the discharge end. The asphalt cement
is pumped from the asphalt cement storage tank and enters the drum at a
point about 1/3 the drum length from the discharge end. Some plants use
a coater which is similar to a pugmill. The asphalt cement is added to the
aggregate in the coater after the aggregate exits the drum. When the
asphalt cement is added into the drum, it is pumped into the bottom of
the drum at about the same location that the mineral filler and/ or bag-
house fines are reintroduced. Adding asphalt cement and dust in close
proximity allows the asphalt cement to trap the fines and to coat them
before they are picked up by the high velocity exhaust gas stream. The
exhaust gases are passed through a dust collection system where enough of
the dust is trapped and removed to meet emission requirements.
--------1-he-asph-alt-e:ement-e0ats-1:he-aggregates-as-the-materials move down
the lower 1/3 of the drum. The HMA exits the drum through a discharge
chute into a conveyor system which transports the HMA into a surge silo
(or other type of storage bin) from which it is discharged into the trucks
via an automatic scale system. It is then transported to the laydown site.
316 Hot Mix Asphalt Materiais, Mixture Design and Construction
All material movements from the cold feed to the surge silo are mon-
itored from the control room of the I-IMA facility. Sensors at various loca-
tions in the facility monitor conveyor speeds, aggregate weights, tempera-
ture, and other critical functions which affect efficient operation.
Batch Facility. A schematic diagram of the components in a typical

batch HMA facility is shown in Figure 6-2. Each of the major compo-
nents of the batch facility are numbered and identified in the legend. A
comparison of the first two or three components in the batch and drum
mix facilities shows that the cold feed proportioning systems for both types
are similar. The primary difference, up to the drying operation, is the
automatic weighing system on the cold feed conveyor of the drum facility.
The proportioned material is fed by the cold feed conveyor into the dryer.
The dryer in a batch facility is typically of counter flow design in
which the aggregates flow in the drum in opposition to (counter to) the
flow of exhaust gases. The dryer has a burner which provides the heat ener-
gy for evaporating the moisture in the aggregate and then heating the
aggregate up to a discharge temperature of typically 285 to 300°F. The
heated aggregate is discharged from the dryer and lifted to the top of the
tower plant via an enclosed hot elevator. The hot, dust laden exhaust gases
from the dryer are passed through a dust collection system to remove dust
particles so that emission standards are met. The collected dust is returned
to the hot material elevator or filler silo and is reintroduced into the HMA
as required.
The tower unit contains a series of components that screen, store, pro-
portion by weight, mix, and discharge the HMA into a truck or storage
silo. After the aggregates exit the hot elevator at the top of the tower, they
are discharged onto vibrating screens that are designed to separate the
material into a number of sizes. After screening the aggregates, they are
stored in hot material bins. The stored aggregates are later proportioned
to produce the HMA. The finest material is separated first into a bin by a
vibrating screen; oversize material is caught by a scalping screen and dis-
charged from the tower. Other material sizes are separated by vibrating
screens and stored in the remaining bins. The control system then pro-
portions the hot aggregates from the bins into the weigh box which is
mounted on a set of scales. These dry aggregates are discharged into .the
pugmill and the hot asphalt cement, which has been weighed and stored in
----a-w-eigh bucket, 1s sprayed mto the pugmill after a few seconds of'~d~ry_m_i~x--- - - - -
ing. The pugmill is typically a twin-shaft, counter-rotating mixer designed
to coat the aggregate quickly with asphalt cement. After mixing, the HMA
is discharged directly into a truck or is transferred into a storage silo where
trucks may be loaded for transport to the laydown site.
Figure 6-2. Layout of a Batch Facility Showing Major Components

In addition to the co-tr1ponents mentioned for both drum and batch

HMA facilities, there are also units for storing and dispensing mineral filler
or other dry additives, including lime and reclaimed asphalt pavement
(RAP). There is also a heated storage tank for the asphalt cement for both
types of facilities.
Drum Mixers
The purpose of the drum mixer is to dry and heat the aggregate eco-
nomically and to mix the aggregate with the asphalt cement, so that the
HMA produced at the discharge end is uniformly coated and is at the
desired temperature. In order to accomplish this purpose, the drum mixer
incorporates a variety of elements which are described below.
Aggregate Feed. The virgin aggregates from the cold feed bins are
delivered to the drum through a series of conveyors and introduced into
the drum either through an inclined chute or on a slinger conveyor. The
inclined chute feeds the aggregate by gravity into the drum at an angle that
directs the material toward the discharge end and toward the bottom of the
drum and away from the flame. The slinger conveyor, located below the
burner, is operated at a speed high enough to throw (sling) the aggregate
away from the burner toward the discharge end so that it does not inter-
.i
··' fere with the flame.
·•1:!
·'·i i
1
' Drum. The aggregate enters the drum and begins to move down the
1llii
'11i! inclined drum by gravity action. The time the aggregates spend in the
1
!t drum (dwell time) depends on the length of the drum, rotational speed of
l'!i! • the drum, angle of the drum, design and m.imber of flights, and the aggre-
!i j gate size. Typical dwell times vary from 3 to 4 minutes.
... Early drum mixers used the same drum length-to-diameter ratio of
\I 4: 1 as used in batch facility dryers, even though the drum design was par-
::.1.:.11;,1 allel fl.ow while the batch design was counter fl.ow. RAP materials are
: 11
I added about the midpoint of the drum in some systems, necessitating
:: .1 additional length to dry and heat this material properly. Therefore, drum
I' ! length to diameter ratios have increased to 5. I or 6: 1 for drums of more
lil recent design.
~~I---th_e_d-±;::!i;::::7!:::;~u;:t;;f~~-r:~::7~t::;:~;::l~~:~:~;~s-----
: jil of drum facilities operating at high efficiency. As the drum diameter
increases, the production rate increases nominally as the square of the
drum diameter for aggregate with l~w moisture content. As moisture con-
Table 6-1. Typical Production Rates of Parallel and Counter Flow

Drums of Different Diameters (1)
Drum Maximum Production Rate for Various Percentages

Drum Diameter Exhaust of Moisture*, Tons per Hour
Design (feet) Volume,
(cubic feet 3 4 s 6 8 10
per minute)
6 28,000 286 233 196 169 131 106
Parallel
Flow 7 38,500 390 318 267 230 178 145
8 50,000 511 416 350 301 234 190
9 63,500 648 528 444 382 297 240
10 78,500 802 654 550 473 367 298
6 28,000 322 262 221 190 148 119

Counter
Flow 7 38,500 439 358 301 259 200 163
8 50,000 575 468 394 339 263 214
9 63,500 730 595 500 430 334 270
10 78,500 903 736 619 533 413 336
* 310°F (154°C) Stack temperature/ 300°F (149°C) discharge temperature.

. (1 ton= 1364 kg, 1 foot= 0.305 meters)
tent increases, the production rates are increasingly affected because so

much additional water must be evaporated. Therefore, while drum diam-
eter is the equipment design factor that affects potential production rate
the most, aggregate moisture content is the factor that most affects the field
production rate.
Flights are the metal elements welded or bolted to the inside circum-
ference of the drum that are designed to control the movement of the
aggregates as they pass through the drum. While each manufacturer uses
0
different p atterns, shapes, numbers, and locatfons for the flights in their
drums, all have been designed in an attempt to get the aggregates heated
and dried, using the minimum amount of fuel, and to efficiently coat the
-----.a-g:greg.a.re.....w.ith..as.phalLc.eme=n-t.~---------
The first set of flights (kicker flights) are configured at a steep angle rel-
ative to the length of the drum in order to push the aggregates down the
drum and away from the flame. If the aggregates penetrate the flame near
the entry p9int, flame quenching can occur, causing inefficient combustion
and allowing unburned fud droplets to adsorb onto the aggregate surfaces.
Thin films of unburned fuel on aggregates can significantly .reduce the
adhesion of the thin film of asphalt cement to the aggregate, reducing
durability and strength of the HMA. Pushing the aggregate away from the
flame allows the flame to expand, burn efficiently, and radiate heat down
the drum. ·
The next set of flights (cup, notched, and tapered flights) lift part of ,
the aggregate from the bottom of the drum and, as the drum rotates, aggre-
gates fall from thf flights through the hot exhaust gases. More of the aggre-
gate is lifted and cascaded through the exhaust gas flow as the aggregates
proceed down the drum. As the.material moves down the drum, the flights
are designed to pickup larger and larger amounts of the drying aggregate.
The flight designs change from tapered to notched to cup flights, each of
which picks up more material than the previous one. Once past the cup
flights, the cascading aggregate produces a curtain across the drum, increas-
ing heat transfer efficiency and preventing the flame and very hot gases
from reaching the entry points of either the RAP or the asphalt cement.
Near the middle of the drum length, immediately before the RAP
material inlet, some manufacturers have installed devices to retard the vir-
gin aggregate flow. These aggregate retention rings, also called donuts or
dams, are attached to the circumference of the drum and act as an obsta-
cle to flow, causing aggregates to build up in front of the ring. When this
larger quantity of aggregate is picked up and cascaded across the drum, a
denser veil of aggregate occurs immediately in front of the RAP inlet,
which helps to improve drying and to prevent the flame from penetrating
to the RAP, which would produce blue smoke and associated environmen-
11;1, tal problems. Other manufacturers use kicker flights in this region of the
ii'' drum to kick the aggregate back upstream toward the front of the drum.
'ri1I Either arrangement produces a concentration of aggregate just before the
:·i
,,I RAP inlet.
,IiI; The mixing flights are located just past the middle of the drum. These
1
!. flights, often called J-flights because of their shape, are designed to pick up
;···:il.·1
the aggregate and asphalt cement, mix them by producing a rolling action,
1
11
l'i and then throw the combined material through the exhaust gases. As the
l!lf asphalt cement encounters the moisture laden exhaust gas, it begins to
J,]1--i\--~£=oam,.in_c.r_e_as_ing_its_v:.ohim.e_.__The_c.aating_o£.the,_ag.gi:egates-with-asphal-t-----
lI
1
1 cement occurs very quickly and the continued cascading of coated aggre-
1
'!.'I
'i'I gate through the exhaust gas stream completes the mixing phase of the
, :I drum mix process.
j i A set of discharge flights are located at the end of the drum. These
ti
1,,,1
,!
1111
flights occupy avery short segment and are designed to move the material
out of the drum at either the side or rear through a discharge chute.
Burners. It is important to understand how the combustion process

operates because it is the key to cost and efficiency. Fuel is burned to pro-
vide the heat energy for evaporating the water from the aggregates and for
heating the aggregate and HMA to the desired discharge temperature.
Burners are simply the devices used to efficiently burn the fuel and there-
by produce heat for the drying process. These burners are specifically
designed for heating and drying aggregates and use combinations of tech-
nology found in other combustion applications.
In forced air burners, 100 p~rcent of the combustion air is pulled into
the burner by the blowers. The air is forced through a constricted area to
a point where the fuel is introduced. The fuel is introduced, under pres-
sure, through a nozzle containing a small opening. It is then atomized into
small droplets, dispersed into the air, forced through the blowers, and then
ignited by a pilot light. For complete combustion to occur, sufficient air
must be supplied and the droplets of fuel must be small, approximately 80
microns in· diameter.
A second type of burner is the induced draft burner. This type has an
exhaust fan at the end opposite the burner which pulls 70 percent of the
air for combustion past the burner. The burner contains a nozzle to atom-
ize the fuel and a pilot for ignition. Automatic adjustments are made to
control the quantity of fuel fl.owing into the nozzle so as to maintain a con-
stant firing rate.
The .type of burner used in HMA facilities includes both forced air and
induced draft features to form a hybrid type of burner. In many facilities,
70 percent of the required air flow is induced by the exhaust fan (secondary
air) and 30 percent is forced air through the combustion air blower (pri-
mary air). Figure 6-3 shows a schematic of this hybrid type of burner. In
HMA facility applications, the purpose of the forced air blower is to pro-
vide air to atomize the fuel. The balance of the air required for complete
combustion is provided as induced air by the exhaust fan.
No matter what type of burner is used, a minimum amount of air
must be provided to ensure complete combustion. This minimum amount
of air for 100 percent combustion is defined as the stoichiometric condi-
----~tro~n:-T-ypica±ly-25-percent-excess-air-is-provided-to-ensure that extra air is
available during the fraction of a second during which combustion occurs.
However, the burner is basically a fuel atomizer and injector and only pro-
vides the fuel and part of the air required for combustion. The balance of
EXHAUST
FAN
30% FORCED AIR

70% INDUCED AIR
ASPHAL T1 PLANT BURNER

Figure 6-3. Typical HMA Facility Hybrid Burner Using Both Forced
and Induced Air for Combustion (.2)
"'
the air m.ust be provided by the exhaust fan. Therefore, sufficient air
flow must be provided by the exhaust fan to handle the exhaust gas com-
ponents made up of steam and water vapor, combustion products, excess
air, and leakage. Reference 3. provides an indication of the relationship
between these exhaust gas components for a drum mix facility operating
at 250 tons per hour (TPH) (227 Mg/Hr) production rate (see Table
6-2). The volume of steam and water vapor increases more significant-
::i:: ly than any other component of the exhaust gas stream as material mois-
:,iii
1
,','1':':
[!,1!
ture content increases.
'i:j1 i
1
To appreciate the effect of steam and water vapor on production refer

IJ.:'II to J:able 6-1. Tables 6-1 and 6-2 show that to maintain a 250 TPH (227
Mg/Hr) production rate, the size of the required drum increases substan-
1
1 :,il tially as moisture content increases: at 5 percent moisture a 7-foot drum is

I
lill! required, at 10 percent moisture a 9-foot (2.74-m) drum is required, and
1
', 1 at 15 percent moisture a IO-foot (3.05-m) drum does not have sufficient
l :, exhaust capacity to maintain a 250 TPH (227 Mg/Hr) production rate.
,1111Thus, to maintain the same production rate in various parts of the USA,
·1,]}I. where stockpile_moisture c~ntents vary signi_ficantly, requires different size
iif1\I·,• h.ardware an~ different ~ap1~al costs. In Flonda, a IO-foot (3.05-1:1) drum
h
~+----may-be· requ1red-to-mamtain--a-z-:50-'F-PH-c22-Y-MgiHr)-J>rodlict1on rat=e-,- - - -
while in New Mexico or Arizona, a 5- or 6-foot (1.52 or 1.83-m) drum
may be adequate because of the differences in stockpile moisture.
It seems apparent then that the single largest factor affecting the pro-
duction rate of a certain size drum facility is stockpile moisture content.
Table 6-2. Exhaust Gas Components and Their Volume for Different
Aggregate Moisture Contents (J)
Exhaust Gas Volume (CFM) at Various Percentages of Moisture

Component 1 5 10 15
Steam and Water Vapor 2,440 12,200 24,390 36,580
Combustion Products 10,140 16,900 26,190 35,480
Excess Air 1,800 3,000 4,640 6,280

Leakage 760 1,690 2,910 4,120
System Gas Volume Total 15,140 33,790 58,130 82,460
Data Based on 250 TPH (227 Mg/Hr) production, 6 percent asphalt, 310°F (154°C) exhaust tem-
perature, 5 percent leakage, 25 percent excess air, operation at sea level. 1 cubic foot = 28.3 liters.
Therefore, contractors should pay particular attention to those measures

which minimize the retention of moisture in the aggregate handling area
and should train the front-end loader operators to always work in the dri-
est portions of the stockpiles while charging the cold feed bins. In regions
where regular heavy rainfalls occur during the construction season, mea-
sures such as placing stockpiles on paved surfaces, covering stockpiles, or
even using a covered permanent building may be justified.
To minimize the cost of producing a ton (909 kg) of HMA. the facil-
ity operator should concentrate on those factors that affect energy. These
factors include moisture in the aggregate and leaks in the facility that allow
air to enter the exhaust stream anywhere between the baghouse and the
burner. Typical leakage locations include the fl.op gate where the mix is dis-
charged from the drum, seals between the various elements of the facility,
lids at the top of the baghouse, and the discharge point from the cold feed
into the drum. Leaks at any of these locations adversely affect production
because leaked air occupies a portion of the available exhaust gas volume
fl.ow and must be heated without producing any of the output product,
HMA.
For economical production of HMA, the burners must operate effi-
ciently. All burners have automatic controls which alter the fuel input to
maintain a constant mix discharge temperature. If changes are noted in
fuel consumption, the exhaust stream gases can be analyzed using a gas
-----,a:0:aly-z-€-r-a-n.tl.-appmp-r.iate..adj.ustments-0.r...rep.airs....made.._ ... _
Drying and Heating Process. At the tip of the burner flame the tem-
peratures exceed 2,500°F (1371 °C) and as the induced air begins to fl.ow
around the flame its temperature begins to rise rapidly to about 1400°F
O . ·:,.···:·,,.;-,•,,•:,,.,,. . Exhaust
Ag~i::~~te~==~1:60:0t:::·:· ·=· ·:· ·~---•~'-,:;,.,:_~i;k~rt~~,~~:~:\~;~i~;~·i~·~· .:i~~~~~s·~~T-i::;:::-=i~;~~~~~I

14001-=::----------------I
1200H-----'---------------I
•F 1000H-r-----="'C''""'----------1
S00t-tt------~...,_---------1
600 t-t+--"'=-=-===c:'---""-..::--------1/ 310°-330°
400 Gas Temp
200 Mix301!Temp
"
AdjIniusttimalen17/
Figure 6-4. Temperature ;t>roftles of Both Exhaust Gases and
Aggregate in a Typical Drum Mix Facility (1)
(,760°C) in a typical drum mix facility. It is this I 400°F (760°C) exhaust

gas that heats and dries the aggregates. Figure 6--4 shows the temperature
profile of both the exhaust gases and aggregate as they flow through a typ-
ical drum mix facility. The aggregate temperature profile has three dis-
tinct zones. In the first zone, initial adjustment, the aggregate is heated to
about 100°F (38°C) but at that point the hot exhaust gases begin to evap-
orate the water off the surface of the aggregates. The second zone, con-
stant drying, is the zone in which most of the water evaporation occurs
and in which the aggregate is heated 'to between 180 and 200°F
(82-93°C). The third zone, aggregate heat up, is the zone in which the
aggregate is heated up to the mix discharge temperature. Near the end of
the third zone the asphalt cement is pumped into the drum for mixing
·with the heated aggregate.
The temperature profile of the exhaust air is also shown in Figure 6-4.
The air temperature rises rapidly to about 1400°F (760°C) and then
decreases steadily as heat is transferred to heat the drum, evaporate the
water and then bring the aggregate up to the mix temperature. In the mix-
ing zone, the exhaust and aggregate temperatures are within 10 to 30°F
(6 to 17°C) of each other and in an efficiently operating HMA facility the
difference between the mix discharge~p_e.r.atur.e_ancl.the...exhaus.t.gas_t.\..Jerou.=-------
perature should be 25°F (14°C) or less.
Figure 6-5 shows a temperature profile for a drum that has been mod-
ified to include RAP material at about the midpoint of the drum. Notice
that the exhaust gas profile is substantially the same as in Figure 6-4.
/ • : ·,.,.........,....Recycle Asphalt
, .....~~-~ ···•.. t
Liquid
Asphalt
lnilial _J
Adjustment I Il£onstant I Superheated
Drying f"Virgin Aggregate
-+ Dry & Heat Rap
t- .. _JI
Mean Aqgreqate Temoeralure
Mixing
Figure 6-5. Temperature Profiles for a Drum Mix Facility with 50

Percent RAP (1)
However, the aggregate temperature profile is quite different with the con-
stant drying zone much shorter. The virgin aggregate is kept in the hot
zone of the drum and superheated to about 500°F (260°C). Kicker flights
or dams are included in the middle of the drum to increase the aggregate
dwell time in the first half of the drum. The superheated aggregate is used
to heat the 50 percent RAP material and develop the temperature profile
in Figure 6-5. Notice that the mean aggregate temperature drops as the
RAP material is heated and its moisture evaporated in the "dry and heat
RAP zone" of Figure 6-5. The mix is brought up to discharge temperature
in the last portion of the drum.
The heat transfer process in a drum occurs by three distinct mecha-
nisms: (1) by convection-heat from the exhaust gases, (2) by conduc-
tion-heat from hotter aggregate particles transfers to cooler ones as a
result of physical contact between the two surfaces, and (3) by radiation-
the hot me~al flights and drum shell radiate thermal energy to the cooler
aggregates in the immediate vicinity of these hotter metal parts which have
been heated by the burner gases.
Asphalt Cement Injection. As shown in Figures 6-4 and 6-5, the

a:spforlrcementis-introduced---into-the-dmm-through-a-pipe-entering from
the rear of the drum. The diameter of the pipe is sized to match the pro-
duction rate of the facility with typical sizes ranging from two to four inch-
es. The asphalt cement is usually dumped into the bottom of the drum
and not sprayed through a nozzle. The actual location of the asphalt
cement discharge varies widely but generally tends to be beyond the 2/3
point down the drum toward the discharge point. For most facilities, the
asphalt cement discharge point can be adjusted to match drying conditions
in the drum.
One manufacturer has eliminated the asphalt feed line from some
models and has added a shaft mixer, called a coater, at the discharge end of
the drum which is similar to the design used in a continuous mixer. A sub-
sequent variation by the same manufacturer is the double drum mixer , ·
which has concentric drums, with the mixing taking place in the annular
space between the inner and outer drum walls.
Fines Feeder System. Mineral filler is delivered to the HMA facility

by tank truck and conveyed pneumatically to a vertical storage silo. The
mineral filler flows out of the silo by gravity into a vane feeder at the bot-
tom of the silo. Some mineral filler silos are equipped with a weigh pod
located under the silo that reverse weighs the flow of material into the
drum delivery system to ensure the accuracy of flow. This feeder rotates in
response to an input signal from the HMA facility computer control. Such
feeders are_ normally equipped with an aeration device to keep the filler
moving uniformly into the vanes. The filler from the vane feeder enters
the delivery pipe for transport to the drum mixer and typically is either dis-
charged directly into the bottom of the drum or fed into a mixing box
where it is coated with asphalt cement before being dropped into the bot-
tom of the drum.
The mineral filler can be discharged either upstream or downstream of
the asphalt cement discharge point. Since the mineral filler is very fine and
dry, the material can become airborne in the high velocity exhaust
airstream if it enters the drum upstream of the asphalt cement discharge
point. If fed into the drum downstream of the asphalt cement discharge
point, most of the mineral filler is coated with asphalt cement before it can
become airborne and lost. Some of the filler is almost always lost unless it
is mixed with the asphalt cement prior to entering the bottom of the drum.
Some drum mix facilities are equipped with mixers which are designed to
coat the mineral filler with enough asphalt cement to keep it from becom-
ing a1rborne.
In addition to commercial mineral fillers, baghouse fines can be fed
- - - - 0 -ac-k-in1:0--1che-cl1°1:1:m-0n-th0se-faei-lities-equipped-with-b-a:glrouse-airp·o+l,-=--- - - - -
tion control systems. The baghouse is filled with rows of fabric filters
through which the exhaust, gases flow prior to exiting through the stack.
The fine material collected in the.baghouse is available for inclusion back
into the drum. Depending on the gradation of the material being dis-
charged from the drum, all or part of these baghouse fines are delivered to
a surge silo and are metered through a vane feeder into the supply pipe. If
mineral filler and baghouse fines are both fed back into the drum, a com-
mon supply pipe or separate pipes may be used. Some facilities use screw
conveyors rather than vane feeders to return baghouse fines to the drum.
Batch HMA Facilities

Batch facilities produce HMA in separate batches rather than contin-
uously as in a drum mixer. The size of the batch is controlled by the capac-
ity of the pugmill, the mixing chamber where heated aggregates and
asphalt cement are mixed. Typical batch quantities range from 3,000 to
10,000 pounds (1,364-4,545 kg).
Until the 1970s, the batch facility was the primary type used to pro-
duce HMA in the USA. The increased capacity and portability of the
drum mixers introduced in the 1970s made it the choice of most contrac-
tors. Currently approximately 95 percent of the facilities being manufac-
tured are drum mixers. However, about 70 percent of the operational facil-
ities in the USA are still batch type. It is apparent that as batch facilities
are retired, most are being replaced by drum mixers.
Several of the elements included in the discussion of drum mixers are
also found in batch facility layouts. For example, the aggregate cold feed
systems and emission control systems all serve the same function in both
types of facilities. Therefore the following discussion centers on those fea-
tures of batch facilities that are different from drum mixers.
Aggregate Dryer. Typical aggregate dryers in a batch facility are of the

counter flow design, i.e., the aggregate from the cold feed enters the dryer
at the opposite end from the burner and flows in opposition to (counter
to) the direction of the exhaust gases. Since the cold, wet aggregate enters
just before the exhaust gases exit the dryer, the exhaust gas temperature is
generally lower than that exiting a typical drum mixer. In fact, the con-
trolling factor for the exhaust gas temperature is the dew point tempera-
ture of the gas stream. The exhaust temperature must be greater than the
dew point if the pollution control system includes a baghouse.
Temperatures above the dew point prevent moisture vapor from condens-
-----+-ingon d:re-dustca:kes-which-accumu±ate-on-rhe-air-bag{abric. If conden-
sation occurs, it turns the dust cake to mud. To prevent this, most bag-
house manufacturers recommend that exhaust gas temperatures stay above
250°F (121 °C). The dew point seldom is as high as 200°F (93°C). The
exhaust temperature of a counter flow dryer of a batch facility can be 50 to
60°F (28-33°C) cooler than that of a parallel flow dryer of a drum facili-
ty. The increased thermal energy transfer efficiency of the counter flow
designs not only produces aggregate at the same temperature for lower cost
but lower exhaust temperatures are easier on the baghouse.
The dryer is also equipped with longitudinal flights which lift the
aggregates and cascade them through the hot burner gases. The tempera-
ture of the aggregates exiting the dryer is the primary factor that deter-
mines the temperature of the final mixture. If the aggregates are too hot,
mixing causes the asphalt cement to oxidize excessively, producing prema-
ture hardening and possibly premature cracking of the HMA mixture. If
the aggregates are not hot enough, coating the aggregates is difficult.
Once the aggregates leave the dryer, they are transported by bucket ele-
vator to the top of the tower unit containing tlie screening unit, hot bins,
and pugmill. Each of these units is described in subsequent sections.
Burners. Burners for a batch facility are the same as those used and
discussed for drum mix facilities.
Screens and Hot Bins. The hot screening unit is shown schematical-
ly in Figure 6-6. The hot aggregates are fed into the hot, vibrating screens
from the left. On the first screen, the oversized material is scalped and
moved to the right over the slanted vibrating screen. The oversize materi-
al is fed into a discharge chute and usually dropped to the ground below.
The coarse aggregate is retained between the top and second screen and
moves into Bin No. 4 on the far right. The material retained between
screen numbers 2 and 3 is the medium coarse aggregate which vibrates
through the screen into Bin No. 3. Bin No. 2 contains the intermediate
fine material that is retained between screen numbers 3 and 4. Bin No. 1
contains the finest aggregate particles. The material in these four bins must
be combined in proportions that will produce the job mix formula QMF)
developed in the mixture design. In addition to the aggregates in these
bins, two additional aggregate components that may be added to produce
the JMF are mineral filler and RAP material.
The surface area of the hot screens must be balanced with the output
from the dryer, the aggregate proportions from the aggregate cold feed sys-
tem, and the capacity of the pugmill. If the materials from the cold feed
----,1·-re-out-of-proportion-wi-rh-rhe-quanti--ry-ofm--a1:erial.1equired-fr"'""o=m~e~ac=1chc----=of-----
the bins, then some bins will be overcharged and others will be under-
charged. Such proportioning problems could result in the mix being
rejected because it does not meet the specifications or the output of the
facility being reduced because of delays in waiting for the undercharged
0 Rejected
oversize
aggregates
Figure 6-6. Schematic of Hot Screening Unit of Batch Facility (2)
bins to fill. Efficiency is adversely affected because of the lower production

rate and the additional handling of material that overflows from the bins
receiving excessive aggregate quantities. The screens may wear or tear,
allowing oversized material to enter a bin for finer material. This condi-
tion, if severe, produces a mix coarser than the JMF and a shortage of fine
material in the mix. When the screen is plugged (blinded), fine material
does not efficiently pass through the screen which results in some finer
aggregates going into coarser bins. This condition produces a mixture finer
than the JMF and a shortage of coarse material in the mixture which can
then cause a lack of coating on some aggregate particles. Increasing the rate
of feed of material into the facility reduces the screening efficiency and
results in material over-riding the screens and going into coarser bins.
This results in the production of a mixture with a finer gradation.
The hot bins should be large enough to store a sufficient amount of
material to permit the pugmill to operate at capacity. Since Bin No. 1 con-
tains the fine material, and a large percentage of it is used in typical mixes,
it is the largest bin. The partitions between the bins must be designed to
prevent intermingling of aggregates from adjacent bins. Overflow pipes are
included in each bin to help eliminate the accumulation of~excess materi-
al which could overflow into an adjacent bin or interfere with the screen ii
above the bin, resulting in carryover and screen damage. Each bin should
also have detectors to alert the operator when the aggregate level is running :I
below or above a preset level. "II'
:II
:1
ii
/i
,,!j
J
The amount of material to be used in a batch from each bin can be cal-
culated by multiplying the percentage of the batch to be taken from that
bin times the aggregate batch weight.
Weigh Hopper and Asphalt Bucket. The desired amount of mater-

ial from each bin is weighed into the weigh hopper. At the same time the
aggregate is weighed, asphalt cement is pumped into a weigh bucket prior
to being sprayed into the pugmill.
The aggregate weigh hopper and asphalt cement bucket are suspend-
ed from scales. Once the total charges are weighed out, the aggregate is
', II dumped first into the pugmill and is then followed by the asphalt cement.
Both the aggregate weigh hopper and the asphalt cement bucket should be
checked periodically for calibration using certified test weights.
Pugmill Mixer. The pugmill mixer is a twin shaft mixer designed to

mix the material sufficiently to coat the aggregate with asphalt cement.
Since the pugmill is the limiting unit in a batch HMA facility, the mixing
time should be as short as possible to obtain uniform asphalt coating on all
aggregate particles. Excess mixing time tends to degrade the aggregate and
oxidize the asphalt cement. After the aggregates are introduced into the
pugmill, a short period of dry mixing occurs, followed by wet mixing as the
asphalt cement is sprayed into the pugmill. The length of the mixing time
can be established using AASHTO T195 or ASTM D2489. In this pro-
cedure a sample of HMA is secured and the material retained on the
3/8-inch (9.5 mm) sieve is removed. Each of these larger particles is visu-
ally evaluated as being completely coated or partially coated. The mixing
time is established as the time required to provide at least 95 percent coat-
ed particles retained on the 3/8-inch (9.5 mm) sieve. If less than 95 per-
cent of the particles are coated, the mixing time is lengthened, additional
material is mixed, and another count is made.
The pugmill mixer consists of a mixing chamber and twin, counter-
rotating shafts with paddles at the ~nd of each of several paddle shanks
(Figure 6-7). The paddles are configured to produce maximum efficiency
in mixing and must be positioned so that the clearance between the tip of
the paddle and the walls of the mixing chamber is less than one half the
maximum aggregate size. Otherwise dead zones can develop where the
l>H----material is not mixed ancl: coateawiln aspnalt cement.
The batch size should be consistent with the batch rating of the pug-
mill to ensure that all material is within the live zone of the mixer. Batches
which are too light should be avoided since inadequate material movement
occurs when insufficient material is in the mixing chamber. After mixing,
~~/4~~~~~/4~
PPPPP~P~~--~..XEA
(a) Typical Paddle Arrangement
I
LIVE
ZONE
(h) Live Zone of Counter Rotating Shafts in Pugmill
Figure 6-7. Pugmill Mixer Elements (2)
the HMA is dropped into a truck or temporarily stored in a surge or stor-

age silo.
Air Quality Control System

Drying units on batch and drum mix facilities contain a small amount
of dust in the exhaust gas streams. The amount of airborne dust is a func-
tion ofthe size and weight of the material being dried, the velocity of the
air in the dryer and, in the case of some drum mixers, the presence of
asphalt cement to coat the smaller particles and keep them in the drum and
_____,ouLof..the__exhaus.t..gas_s_tx.e_am. Because some fine material becomes air-
borne, pollution control equipment must be present to capture this dust
before it is discharged into the atmosphere. These emissions must not
exceed the various state and federal air pollution limits.
The air fl.ow inside a drum is primarily a factor of the diameter of the
drum. Since the generally accepted industry standard for velocity of air
through a drum is about 1,000 feet (305 m) per minute, the exhaust fan
must be sized to ensure that adequat1:: air flow is provided. When drum
facilities were first introduced, exhaust air velocities were between 600 and
800 feet (183-244 m) per minute, but were found to be inadequate in
areas where high production rates were required for aggregates with mod-
erate to high moisture contents. The maximum exhaust gas volumes con-
tained in Tables 6-1 and 6-2 are based on an exhaust gas velocity of 1,000
feet (305 m) per minute.
The velocity of the exhaust gas is a primary factor affecting the poten-
tial amount of airborne material for an aggregate with a given size distrib-
ution. The amount of airborne dust increases with the square of the gas
velocity.
One other factor that affects the amount of dust entering the pollution
control system in a drum mixer is the location of the asphalt cement inlet.
The closer the asphalt cement inlet is to the flame, the larger the quantity
of airborne dust that is captured by the asphalt cement and prevented from
being picked up by the dust collector system. Moving the asphalt cement
inlet closer to the flame produces two potential problems. First, the
asphalt cement may be oxidized or stripped of its low volatile components
which produce blue smoke in the stack. Second, the aggregate is not dried
as well before being coated. The facility operator must then optimize the
location of the asphalt cement inlet to minimize the amount of dust and
smoke in the pollution control system while at the same time meeting the
HMA moisture content requirements of the specifying agency.
The air pollution control system i.q. most HMA facilities includes one
or more of the following types of dust collectors (1) primary dry collectors,
(2) wet collectors, or (3) fabric filters (baghouses). Each type of collector
is discussed in the following sections.
Primary Dry Collectors. There are two types of primary collectors

currently in use, the knockout box and the centrifugal .collector, predomi-
nantly on batch facilities. Dry collectors are placed immediately at the rear
of the dryer where the hot exhaust gases enter the ductwork between the
dryer and the stack.
The knockout box is simply an.expansion chamber where the cross-sec-
-----"-tioi;i;!! ;trea increases significantl:x: therebyJ.e.d.ucing_the_yelo.ci.:ty_o£the..exhaus.___ _ __
gases. With the reduced velocity, the larger dust particles fall out of the
airstream and accumulate at the bottom of the collector which is typically
sloped enough to allow the dust particles to concentrate at a central point.
This material can then either be wasted or fed back into the HMA.
The centrifugal collector has a tangential inlet which forces the heavier dust
to the outside wall where it slows down and falls to the bottom and is either
wasted or fed back into the HMA The cleaned exhaust air goes up through the
inner cylinder enroute to the remaining pollution control equipment. These
primary collectors are between 60 and 85 percent efficient and when used with
;i. wet scrubber, reduce the volume of sludge that must be handled.
Wet Collectors/Scrubbers. The wet collectors operate on the princi-

ple that airborne dust can be forced to drop out of the airstream by increas-
ing the particle weight by spraying it with small droplets of water. In order
A. Venturi throat showing water spray.
EXHAUST GAS
WITH WETTED OUST
PARTICLES
---
OUST LADEN
EXHAUST GAS
FROM THERMOORUM
TUBE ADJUSTABLE WATER

VENTURI SPRAY
THROAT NOZZLES
CLEANED EXHAUST GAS

SEPARATED FROM
OUST LADEN WATER
--------~-----------~---
' SLUDGE
' - TO SETTLING PONO
B. Circular air flow in wet collector.
Figure 6-8. Schematic of Ductwork of a Wet Collector

to minimize the amount of water to be handled and to increase the con-

centration of dust particles in the area to be sprayed, the cross-sectional
area of the ductwork is reduced as shown in Figure 6-8. The water spray
nozzles are placed immediately upstream of the venturi throat (Figure
6-8a) and aimed at the reduced cross-section. The exhaust gas with the
wetted dust particles is carried into a cylindrical separator. There the wet-
ted dust particles follow a helical path down the separator, encounter the
wall by centrifugal force, and flow down the wall where the particles are
intercepted by a skimmer at the bottom and carried to a settling pond
(Figure 6-86). A cutaway diagram of a typical horizontal wet collector is
shown in Figure 6-9.
Figure 6-9. Cutaway Diagram of a Typical Wet Collector
As the name implies, settling ponds are used to settle (separate) the
dust from the water and reduce the total volume of water required for the
operation by recirculation. Water, laden with dust, enters the settling pond
and the heaviest dust settles quickly to the bottom of the pond. Typically,
---+---_ _th_e_s_e_s_e_tt_li_n--o_f1onds are divided into sec_tions_with_the.._w.ate.r_b.ecoming----
cleaner as it moves from one section to another. Water in the last section
of the pond is drawn off the top and circulated back to the scrubber unit.
The collected fines must be removed from the ponds and disposed of since
excess buildup adversely affects the efficiency of the ponds.
When using a scrubber unit, the dust particles in the sludge are lost to
the HMA, thereby making the gradation of the produced materials differ-
ent from that of the incoming aggregate. If a large volume of dust is cap-
tured in the exhaust gas, a significant change can occur in the mixture gra-
dations particularly in the fine aggregate fraction. If the materials in a par-
ticular area of the country are so low in fines that mineral filler is required,
the use of a wet dust collection is not recommended since the dust cannot
be fed back to the plant.
Fabric Filter (Baghouse). Baghouses are very efficient and can remove
over 99 percent of the dust from the exhaust gases. The operational prin-
ciple for the fabric filters is very simple. The dust-laden exhaust gases are
pulled through a filter cloth that traps the dust on one side but allows the
clean air to pass through the filter cloth. The operational principle is sim-
ilar to that of a common home vacuum cleaner with the primary difference
between the two units being one of size. The size (total cloth area) of a
baghouse is determined by the volume of air it must filter. Experience has
shown that an air volume to cloth area ratio (volume in cubic feet per
minute/square feet) between 5: I and 6: I is very effective in cleaning the air
without reducing plant production due to a lack of dryer air volume.
Typically a baghouse contains between 300 and 800 separate bags.
These bags are sewn into the shape of a cylinder closed at one end and
mounted over a metal cage to prevent them from collapsing during the fil-
tering process. They are constructed from a felted, nylon fabric that can
withstand temperatures up to 450°F (232°C) and repeated flexing cycles
during the cleaning process.
The dust-laden exhaust gases are brought through the ductwork from
the primary collector to the front of the first set of filter bags (Figure 6-10).
As the dirty exhaust gases encounter the filter fabric, the dust particles are
trapped on the outside, and the clean air continues to the exhaust fan and
out the stack (Figure 6-1 I). The dust particles on the outside of the filter
bag build up with time and form a dust cake which is very important for
proper functioning of the baghouse. As a dust coating builds up on the
bag, the filtering efficiency of the bag increases and finer and finer dust par-
ticles are trapped. However, if the coating on the bags becomes too thick,
the bags are blinded and air is unable to penetrate the cake, causing the
baghouse to stop functioning. Because of this the dust cake must period-
ically be removed from the bag. This is accomplished by the cleaning
----m-ecnams-rn:--'TliI':-d:uscfa:tlroff-ene-outstcl.eohne-bags-aha-falls to the bot-
tom of the baghouse and is either fed back into the HMA or wasted. Since
there are so many bags in a collector, the cleaning is usually performed in
groups so that only a small percentage of bags will be off-:line at any given
time in order to maintain efficiency.
Figure 6-10. Cutaway View of a Typical Baghouse (2)
Cleaning is usually accomplished by flexing the bag with a pulse, or

jet, of compressed air. The cleaning cycle is activated by timers or pressure
sensors. In the case of timers, the frequency of the cycle and the duration
of the compressed air pulse is adjustable. If equipped with pressure sen-
sors, the frequency is automatically adjusted based on a selected pressure
drop through the collector.
· The efficiency of a baghouse collector is affected by many factors. If
it has been si:z;~d correctly and maintained properly, the single most impor-
tant operating parameter is the pressure drop across the bags (Figure 6-12).
Since pressure drop is a measure of resistance to flow, the lower the drop,
the higher the volume of air that can be handled. Some reasons for
increased pressure drops are:
Figure 6-11. Schematic Functioning of the Filter Fabric in Cleaning

the Dust from the Exhaust Gases
1. Production set higher than rated capacity-This liberates excess

quantities of dust.
2. Moisture-If the temperature of the gas stream falls below the dew
point of water, condensation deposits on the bags, and combines
with the collected dust, forming mud and blinding the bags.
3. Cleaning System-Either the timing of the cleaning system mal-
functions, or the pressure of the compressed air drops below 90 psi,
which is too low to permit the air to flow through the bags properly.
TRANSPORTATION AND LAYDOWN OPERATIONS

Two of the most critical operations on the job site are the transporta-
tion of the Hi\1A from the production facility to the construction site, and
the actual laydown operations. Transportation operations are critical
because of their effect on the project cost and the quality of the HJ\t1A.
_J __
Figure 6-12. Pressure Drop Across Filter Bags in a Baghouse

Laydown is critical becaus_e it significantly affects the quality of the HMA

surface. Each of these operations is discussed in the following sections.
Transportation Operations
The transportation portion of the project involves the trucks used to
haul the HMA from the HMA facility to the paver, and the equipment and
procedures used. These procedures include loading the HMA into the
truck, weighing and ticketing, protecting the HMA from cooling, trans-
porting to the job site, dumping into the paver hopper, and returning the
truck to the 'HMA facility. This whole transportation operation is often
referred to as the truck cycle and must be analyzed carefully when deter-
mining how many trucks of a specific size are required to balance the out-
put from the HMA facility wipi the compaction and laydown equipment
at the road. The number of trucks needed for a project is affected by the
length of waiting times for loading and unloading, the loading time, time
to ticket and cover the HMA, the distance between the production site and
the laydown site, the average truck speed, the output of the HMA facility,
and the availability of surge or storage facilities. In this section, the dis-
cussion begins with a description of the types of trucks used to haul HMA
from the production site to the laydown site. The primary types of vehi-
cles, categorized by the method of discharge ofHMA, are end dump, bot-
tom dump, and live bottom (conveyor) vehicles. Regardless of the dis-
charge method, all v.ehicles should have metal bodies which are clean,
smooth, and free of holes. Insulation extends the haul distance (time) that
is acceptable for a project.
End Dump Vehicles. These vehicles are typically one of three AASH-
TO classes: a single unit vehicle with either a tandem rear axle (3D) or
triple rear axle (4T); or a tractor-semitrailer with tandems on the rear of
both the tractor and the semitrailer (3-S2). Each of these vehicles has a
maximum gross weight which is controlled by the weight restrictions
imposed by either state or federal statutes. The federal bridge formula
restricts the gross vehicle weights (GVW) and payload of vehicles, primar-
ily as a function of axle spacing.
The end dump vehicle must be capable of lifting and rotating the
- - - - r r t ruck-bed-ro-discharge-aH--H-MA-wi:thour-rhe-rea:rof-rh-e---rraiterpressirr·o-----
down on the paver hopper. 'When discharging HMA from the truck to the
paver hopper, the truck bed should be elevated slightly before the tailgate
is opened to allow HMA to flood the hopper with a surge of asphalt mix-
ture thereby preventing a trickle of coarse material into the hopper which
produces segregated spots. However, filling the hopper should not cause
material to spill in front of the paver.
Bottom Dump Vehicles. The bottom dump vehicles are gondola type
3-S2 vehicles which unload by dumping directly onto the roadway or into
a spreader box beneath the vehicle which then discharges onto the road.
For vehicles that dump directly onto the road, the discharge rate is con-
trolled by the width of the opening of the clam shell jaws that are in the
bottom center of the trailer bed. The bed has slanted sides to facilitate
smooth flow of material from the bed into the discharge area. Control of
the width, and therefore the volume, of the material is critical since the
material is often supplied to the paver by a windrow pickup machine. This
machine has no significant storage capacity to store excess material or to
supply extra material if there is a deficiency of material in the windrow in
front of the paver. Therefore the quantity of material discharged by the
bottom dump vehicles must be accurately controlled if a windrow pickup
machine/paver unit is used to spread the HMA.
'!
Live Bottom Vehicles. There has been an increase in the use of live
bottom trailers which contain a conveyor at the bottom of the trailer.
These vehicles have been used successfully on some projects to help mini-
mize segregation of materials, especially at end of truck loads. The slat
conveyor pulls material from the truck bed and discharges it into the hop-
per of the paving machine without elevating the bed. Use of this type of
vehicle increases capital investment, because of the additional cost of the
trailer, and increases operating costs as a result of additional maintenance
on the conveyor system. Some advantages are increased payload, and lower
cost per ton mile compared with 3D and 4T vehicles, and increased safety
when compared with rear dump 3-S2 vehicles.
Truck Loading Procedures. No matter which type of vehicle is used

it is necessary that the truck bed be cleaned and coated with a lubricant to
prevent sticking. Petroleum based products, such as diesel fuel, should not
be used because of environmental problems and potential detrimental
effects to the asphalt mixture. Rather, lime water, soapy water, or suitable
commercial products should be sprayed on the bed and the excess lubri-
-----·ca:nrdrairred-priorto-loading-with-HMA-:---Bse-of-lubricant-prevents stick-
ing and buildup of cooled HMA, thereby lowering the unit cost of trans-
portation.
When loading the HMA into the truck, multiple dumps should be
used to minimize segregation caused by the larger aggregates rolling down
a) Single Dump Loading Which Produces Segregation Around All

Edges of the Truck Bed
b) Multiple Dump Loading Which Minimizes Segregated Areas

Within the Truck B~d
Figure 6-13. Truck Loading Procedures
the sides of the cone of dumped material. The typical pattern of segrega-
tion of these coarse materials in a truck bed filled with a single dump is
shown in Figure 6-I-3a. The segregation in the truck bed shown in Figure
6-13a shows up on the road behind the paving machine as a distinct pat-
tern since the segregated areas may be scattered across the entire paving
lane. The segregation is increased when the paver wings are lifted and the
material is dumped into the paving hopper. This segregation is highlight-
i ed at the end of truck loads.
~,'+--------1-he-slight-segregati0fl-i-n-the-tmek-bed-produced-by-multip-le-dt1mp-----
as shown in Figure 6-13b may not show up at all on the road, especially if
dumps 1 and 2 are made close to the ends of the truck bed thereby pre-
venting the segregation that occurs at the front and rear when single dumps
are used. With the segregation concentrated in the center of the truck bed,
the paver augers have an opportunity to slightly remix and move the large
aggregates around to minimize the segregation on the roadway.
The gate configuration on the silos or surge bins should be large and
elongated to drop a large mass of HMA material quickly and thereby pre-
vent significant coning of the drop which accentuates segregation. The
worst drop situation is a slow trickle of material into the center of the bed
which allows the coarse aggregate to run down the pile producing segrega-
tion at the edges. Studies have shown that one of the biggest causes of seg-
regation on the roadway is improper use of the storage silo.
Protection During Haul. When the weather is cool or the haul time
is long, the HMA should have a protective covering to prevent excessive
cooling and the formation of a crust on the surface. It is important, how-
ever, that protective covers be securely tied down to prevent air from get-
ting between the cover and the asphalt mixture. Therefore all covers
should lap over the sides of the truck bed and be snugly tied down all along
the sides and rear of the truck bed.
Factors Affecting Truck Cycle. As indicated earlier, the truck cycle

consists of the different segments of a typical trip from a HMA production
site to the job site where paving occurs, including the return trip. In con-
sidering techniques for reducing the cost of truckirig, one must either
increase the payload per trip or reduce the total trip time by reducing one
or more of the elements of the trip. The number of vehicles required to
haul the daily output from a HMA facility is a function of the total tons
produced, the total truck cycle time, and the payload per truck.
A surge bin or storage silo is often used to decrease the length of time
required to load a truck with HMA and also to reduce the time waiting in
line to be loaded. The surge bin is typically a circular silo that has a coni-
cal end (funnel shape) where the HMA is discharged through a set of gates
at the bottom of the cone into the truck bed. The gates are designed to
open and close quickly so that a large quantity of HMA can be discharged
as a mass into the truck bed. There are several elements of a surge bin or
storage silo that must be designed very carefully to ·prevent segregation
either in discharging the mix into the silo or in discharging the mix from
-----the-silo--i:0.-1:0---t--h-e-1:r-1:1-ek•.- - - - - - - - - - - - -
The basic purposes of surge bins and storage silos are:
1. To replace truck beds as a storage site for HMA, thereby minimiz-

ing the number of trucks required;
2. To allow the HMA facility to continue operation when trucks are

not available for discharge;
3. To allow every batch to be of uniform weight and of the optimal
size;
4. To provide a temporary storage location for mix needed but not yet
received for laydown, thereby allowing adjustments to equipment
or production of another mix without affecting field laydown oper-
ations;
5. To allow production to begin before either trucking or laydown
and compaction operations start, because of the short term storage
available in the surge bin or storage silo; and
6. To accurately discharge mix into the haul vehicle to maximize the
payload for each trip. This can be dbne especially if a batching
weigh hopper is used as the discharge device underneath the surge
bin or storage silo or if truck scales are underneath the discharge
device.
The primary operational difference between a surge bin and a storage

silo is that the surge bin must be emptied at the end of each day's work
because mix remaining in the surge bin cools and is unsuitable for use the
next day. Storage silos, however, may be insulated and can be heated to
ensure that the mix stays at elevated temperatures for periods ranging from
several hours to several days. However, long-term storage may cause prob-
lems in some cases. Two problems that may occur due to long-term stor-
age are drainage and oxidation of the asphalt cement. Figure 6-14 shows
a storage silo that is being fed by a bucket elevator. The HMA is received
by a batcher at the top which drops a full load of asphalt mixture into the
insulated silo. The silo discharges HMA into a weigh batcher at the bot-
tom which then discharges the HMA into the truck in multiple drops to
minimize segregation. A storage silo generally has a larger capacity than a
surge bin.
It is difficult to make small quantities of any mix design with a drum
mix facility. Hence, without silos, customers for small projects must wait
until the contractor is producing the required mixture. However, if mul-
tiple storage silos are available, several mixes can be prepared at off-peak
times, stored, and made available to these customers at all times. Changing
-------1rnm-0J1e-ra-i~-design.-t0-aFJ:0ther-i-n.-a-cl:-r-l:lffl-m-ix-f-aci-l-i:ty-prodttces-a-mixt1:J:re>------
during the changeover that may not meet the specifications for either
design. Maintenance costs and the total quantity of material wasted are
also lower for continuous operations than for start-and-stop operations
required when silos are not available.
Figure 6-14. Storage Silo Showing Bucket Elevator and Weigh Batcher
One manufacturer has recently introduced a temporary storage device

for the roadway called a material transfer vehicle (MTV) or shuttle buggy.
The device basically serves as a mobile surge bin of 35 ton (32 Mg) capac-
ity that can rapidly unload trucks and move back to the paver to supply
HMA. When the MTV is used in paving operations, the paver has a large
25 ton (23 Mg) hopper with vertical sides that replaces the standard paving
hopper with wings. This larger hopper permits the paver to operate inde-
pendently of the MTV and improves the paver mobility, especially in tight
geometric situations such as residential and sh:opping-centerpaving. The
MTV has a mixer which remixes all HMA to minimize segregation before
it is loaded into the paver. Because the paver operates independently of the
truck, it can pave continuously at a slower speed without affecting the
truck cycle or decreasing HMA facility production. This device offers
many practical advantages for special applications such as paving test tracks
and race tracks where high superelevation, stringent grade control, and
unusual construction requirements are included in the construction con-
tract. Experience has shown that the use of a MTV results in improved
surface smoothne_ss.
Asphalt Pavers
Principles of the Asphalt Paver. As seen in Figure 6-15, the paver is
the prime mover which receives material from a truck, stores it on board
and has a conveying system that takes the material from the hopper to the
rear of the machine and dumps it on the prepared surface. After being
dumped in front of the screed, a pair of laterJ augers distribute the mate-
rial transversely across the width of the screed. The paver is comprised of
two basic machines: the first unit is the tractor, which can be mounted on
rubber tires or tracks that propel the paver and which provides power for
running the mechanical systems; the second unit is the screed unit at the
rear of the paver with its pivot arms that are mounted on the tractor. The
screed unit is free to rise and fall since. no torque is exerted at the pinned
connection point on the tractor. The plan view of the paver shown in the
bottom of Figure 6-15 shows the two material handling systems that (i)
move the material from the hopper at the front of the paver to the rear of
the paver and (2) distribute the material transversely across the paving lane.
The most critical feature of the paver is the self-levelling screed unit
(Figure 6-16) which determines the profile of the material being placed.
As the screed is towed down the road there are a number of forces acting
on the unit. First, there is a towing force which is provided by the tractor
unit and which is exerted at the tow point. Second, there is a force by the
asphalt mixture resisting the towing force; this material is shown in Figure
6-17 immediately in front of the screed unit. Third, the weight of the
screed unit presses down vertically on the material unµerneath the screed.
Fourth, there is additional compactive force applied by the screed either
through tamping bars, vibrators, or both. Fifth, there is a frictional force
between the bottom of the screed and the material underneath it. Sixth,
there is a resistive vertical force against the bottom of the screed from the
material being compacted. These forces are seeking an equilibrium posi-
-4'11-----t-1o_n_a_t__,.l times, therefore a free body diagram can be drawn to demon-
strate the relationship between these forces as shown in Figure 6-18.
The HMA in front of the screed is loose but the material discharged
behind the screed is slightly compacted. The screed weight plus the vibrat-
ing and tamping energy it imparts produces the increase in density. If the
PUSH ROLLERS
AUGERS
''
)
. , , .. __ ;r ------ ' ___ , .
MATERIAL
F\.OW
MATERIAi.
FLOW
' ,
,~.----'--------'- -- ....,
( ~
'·------ ----- -----
<:::::E..RECTION of TRAEEL < FRESHLY PLACED MAT
Figure 6-15. Schematic of Paving Machine
LEVELLING ARM
~~:g;-~~. '
~~~~t~,I~N SCREED
Figure 6-16. Elements of the Self-levelling Screed Unit

t F2~f~~~
Fl (~!'.:~~i::;;;~~:.~_:_;,►;";:).•:. ;-,,·.,. :--.-· •:'t :- .. .•.• , ..--t-.:• .. ,-~~~ • ..!-~•.••·• :
Figure 6-17. Forces Acting on the Screed in Tow
F 1 = Towing Force
I
F2 = Weight of F = Closing force required by equilibrium which
1 5
Screed is due to both the resistive force of the HMA
~under the screed and the compaction effort
1
1
supplied to the screed.
F, = Force of Material in
front of screed
Figure 6-18. Free Body Diagram of Forces on Self-levelling Screed
screed is set up to handle a given volume of material at the front of the

screed and the density at the rear of the screed is higher than at the front,
the thickness of the material at the rear is less than at the front. This dif-
ference in thickness corresponds to an angle at which the screed slides or
skis on the HMA. Since the screed is free floating, it settles at an angle that
produces a balance of the forces, as shown by the dashed line in Figure
6-18.
In order to make adjustments in the thickness of the material, there is
a screed depth crank at the rear of the screed which is adjusted to lower or
raise the rear pivot point in Figure 6-17. Lowering this pivot point
increases the angle of attack of the bottom of the screed on the HMA
!-W1-------=w=1,-:,cl·1 causes the dosmg force shown by the dashed line inFigure 6--~1~8-.- - - -
As the screed rises, the material extruded behind the screed is thicker. The
response of the screed to this change in angle is not immediate. In fact, it
takes a distance that corresponds to at least four times the length of the
towing (levelling) arm to reach a flew equilibrium. Therefore, it is criti-
cally important that thickness adjustments be made only as a result of

unusual conditions at the paver and not as a normal procedure by the
screed operator. Figure 6-19 is a plot that shows the response of the paver
as a function of horizontal distance in terms of levelling arm length versus
percent of the distortion (change in thickness) that is introduced at the rear
of the screed. Notice that 63 percent of the distortion is recovered by the
first levelling arm length, 87 percent by the second, 95 percent by the
third, and 98 perc~nt by the fourth. After five levelling arm lengths, 99
percent of the distortion is adjusted for and equilibrium has again been
achieved. The practical result of this plot is that, for a 10-foot (3.1 m) lev-
elling arm length, the screed has not recovered from a step input change at
the screed until 50 feet (15.2 m) of mixture is placed. Therefore, an over-
active screed operator that constantly adjusts the thickness crank prevents
the screed from averaging out the high and low spots in the pavement.
A second source of input distortion into the screed is the vertical rise
or fall of the wheel base of the tractor unit. As the vehicle moves over the
undulations in the existing surface, the towing point of the screed moves
up and down.
When a paver stops, the force F4 (Figure 6-18) changes from the equi-
librium condition that existed when the paver was moving. The screed can
settle slightly, or the material in front of the screed can cool, depending on
the time interval between loads and the temperature conditions of the sur-
face and the air temperature. Therefore, when a new load of material
arrives, the force produced by the head of material in front of the screed
can be different, causing the screed to rise or fall to achieve a new equilib-
rium condition. These effects can be eliminated or minimized if the paver
moves continuously. Sometimes it is not practical to keep the paver mov-
STEP-FUNcT1ON DISTURBANCE
TRANSIENT RESPONSE OF SCREED ..
I DIRECTION OF PAVING ., SCREED PATH
-----~-----
I --'-----.!. ____ i _____,__ 'i,,._ _____ _
-
T
.....: .....: .....: ~
:
~~ ~ ~ ~ :
,, 633/o I ; : I ;
1•--------,. , ...... ;i.,, ....... ..... ...... .._..
Tl
STEP AMPLITUDE= 100%
I 7' I
TOW POINT PATH
..I.-
I
1L
I
2L
I
3L 4L SL 6L
L=LEVELLING ARM LENGTH
Figure 6-19. Response of Self-levelling Screed to a Manual Change in

Thickness Control
ing continuously; in these cases the paver should be stopped quickly but
smoothly, loaded as quickly as possible, and accelerated to paving speed
again quickly and smoothly.
A second job-related factor that affects the screed is a change in paving
speed. During the paving operation, the material passing under the screed
is subjected to certain weight and compactive forces that cause the density
of the HMA to increase as it Bows from the front to the back of the screed.
If the paver speed is increased, the mix Bowing under the screed is exposed
to the compactive forces for a shorter period of time, thereby providing a
lower density material. This lower density material does not provide as
high a resistive force as the higher density material produced at the lower
speed, with the result that force F4 (Figure 6-18) is lower. The resistive
force increases to achieve equilibrium and the- screed falls. Therefore, if a
smooth pavement is desired, it is important that the paver maintain a con-
stant speed to avoid the screed undulations caused by speed changes.
Control of the head of material in front of the screed is maintained by
the proper setting of the Bow gates and augers on the paver. An adjustable
Bow gate is located above each of the slat conveyors immediately in front
of the augers. The purpose of the Bow gates is to control the amount of
material that is supplied by the slat conveyor to the augers. It is important
that the Bow gates be set at a height that delivers sufficient material to the
auger chamber to have only about 50 percent of the auger sticking out of
the mix and to cause the augers to turn when the paver is moving forward.
The auger motion is controlled by sensor elements that activate the augers
when the depth of material in the auger chamber falls below a preset value.
Each of the augers operates independently and distributes the materi-
al transversely across the paving lane. At the junction of the two augers, in
the center of the paver, is the auger gear box. In order to force HMA under
the gear box, there is typically a reverse auger adjacent to the gear box, to
ensure that sufficient material is forced into that location.
Historically, screeds produced compaction by the action of a tamping
bar which was mounted in front of the screed plate (Figure 6-20). A typi-
cal tamping bar moved up arid down 5/32 inch (4.0 mm) at a rate of 1500
tamps per minute. The compaction actually occurred just in front of the
beveled face of the tamping bar, which confined the material when coupled
..------~-_,ci.th._the_fo..cw..ard.moYement-o£the_p.av.:er__The..m.ateriaLcame-in-contact-w.ith.__ _ __
the face of the bar and was forced down when the bar moved down. This
tamper bar compaction system had a large number of moving parts, required
substantial regular maintenance for efficient operation, and was very heavy.
The compaction produced by the bottom 1/4 inch (6.4 mm) of the
24"
Tamping bar
of Motion
Plate
60"
-i·lr"t-
5132 in. eccentricity
1500 lamps/minute
Figure 6-20. Schematic of a Screed with a Tamping Bar
tamper bar proved to be detrimental because slight indentations in the

HMA surface were produced by each stroke of the bar.
This problem with the tamping bar marking the pavement was over-
come with the introduction of the vibratory screed (Figure 6-21). The
vibratory screed consists of a curved, trough-shaped plate that has an elec-
tric vibrator mounted to it. The vibrator typically operates at 3600 vibra-
tions per minute at an amplitude of a few thousandths of an inch. This
screed compaction technique was quickly accepted into the market. It pro-
duced a very smooth surface because of the sand and asphalt that was
worked up to the surface by the low amplitude, high frequency vibratory
action of the screed. This smooth surface was significantly different from
the relatively open texture produced by the tamping bar compactor. In
some countries, tamping or combination screeds are still popular.
During the early days of the vibratory screed, some difficulties were
24"
Electric Vibrator ~ 3600 VPM
I I
Direction
of Motion
\A
t.h I -
~
Figure 6-21. Schematic of a Vibrating Screed
experienced with mixes that tended to build up on the curved portion of

,the face of the screed and suddenly lift the weight of the screed over the
built-up material, producing a slight hump in the road. This problem was
solved by adding a strike off plate to the front of the screed to prevent the
build up of material around the top part of the curved face of the screed
(Figure 6-22).
Screeds have other adjustments and controls to facilitate the proper
placement of the HMA mat. Screed heaters are provided to increase the
temperature of the screed plate before paving begins each day. Heating the ·
screed plate to approximately 300°F (149°C) ensures that the first materi-
al placed does not stick to the screed plate and tear, giving the mat a rough
texture. After paving a short distance, the HMA keeps the screed temper-
ature hot a~d the heater can be turned off. ·
Another adjustment on the screed is the crown control. Some pave-
ment lanes are crowned to assist in draining rainfall off the road surface.
The crown can be adjusted to meet the geometric requirements of the
roadway by adjusting a turnbuckle device to flex the bottom of the screed
to produce the desired crown.
Situations arise where the width of paving exceeds the basic width of
the paver, which is typically 8 feet (2.4 m) for small pavers and 10 feet (3.0
m) for larger pavers. When such situations arise, the paving width can be
extended using either rigid screed extensions or hydraulically-operated
screed extensions. The rigid extensions are bolted to the paver screed in typ-
ical increments of 0.5, 1, 2, or 3 feet (15.2, 30.5, 61.0, 91.4 cm). For
extended paving widths with paver screed extensions, auger extensions
should be used in order to provide an adequate head of material in front of
the screed extension. Care should be taken to ensure that the alignment of
the front and rear of the screed extension is matched to that of the paver
Strikeoff
Direction Plate
Figure 6-22. Vibratory Screed with Strike-Off Plate

screed. When variations in paving width occur often, it may be more eco-
nomical to have the flexibility of the hydraulically extendible screed as
compared to the down-time required to add and adjust the rigid screed
extensions. One problem with the hydraulically extendible screed is that
the surface texture is different and the HMA density is lower under the
screed extension than that under the basic screed unit.
Automatic Paver Screed Controls. If the paver is operated on a level

grade, the floating screed lays a level, uniform mat. Since the existing sur-
face being paved is seldom level, paver manufacturers have developed grade
controls to modify the elevation of the tow or pull point to adjust for vari-
ations in the grade or elevation over which the paver passes.
In an attempt to smooth out the vertical movements of the tow point
of the paver, all paver manufacturers have introduced automatic paver
screed control systems. The objective of these control systems is to keep
the elevation of the tow point constant, even though the tractor unit is
moving vertically in response to the changes in profile of the surface being
paved. Thus, by maintaining the constant elevation of the tow point, the
angle of attack of the screed is also maintained at a constant level. This
allows the screed to do a better job of reducing the amount of material
placed on high spots, increasing the amount of material placed in low
spots, and producing a smoother mat surface.
Types of Grade Reference Systems. Three basic types of grade refer-

ence systems that have been developed include the erected stringline, the
mobile reference, and the joint matching shoe. These reference systems
typically use a combination oflongitudinal grade sensors to. establish posi-
tion in the direction of traffic and may•,indude a slope control device to
control transverse cross-slope of the paving lane. It is possible to use either
erected stringlines on both sides of the paver or longitudinal grade sensors
on both sides of the paver. When dual grade sensors are used, the
cross-slope may not be constant but rather is a function of the elevations
on both sides of the paver.
Theoretically an erected stringline provides the smoothest possible
HMA mat behind the paver. In practice, however, this is often not the
case. One of the most significant difficulties is maintaining the correct ele-
vation of the stringline. High tensile strength line isneecl.ed to keep the
line taut without appreciable sags, and supports are typically required every
25 feet (7.6 m) on tangents and every 10 feet (3.0 m) on horizontal curves.
Since the elevations must be set by a survey crew, the process is time-con-
suming and expensive. In addition, any errors in the stringline grade are
reproduced by the paver.
Another disadvantage of the erected stringline is that all equipment
and personnel must stay away from the line so that the line is not dis-
turbed. Therefore, truck drivers must be very careful when entering and
leaving the area in front of the paver. Unless grade is an important crite-
ria on a paving project, the added cost and care associated with the use of
an erected stringline is seldom warranted. However, the use of stringlines
is essential for many airfield paving projects since this method controls the
actual grade. On airfields, for instance, where the final pavement must tie
in with taxiways, concrete slabs, and other existing structures, and where
close controls are specified for grade changes, the stringline method is a
necessity. On these large paved surfaces, such '2.S airfields and parking areas,
the use of stringlines also prevents bird baths.
Laser technology has recently been applied to the paving operation,
offering an alternative to the stringline grade control system. A heli-
um-neon laser transmitter is used to establish a horizontal reference plane
over the job site. Current laser technology permits the establishment of a
horizontal plane, a single grade plane or a dual grade plane for both grade
and slope control. These systems can be coupled with existing paver con-
trols for automatic or manual operation and have been demonstrated to
work in both dust-laden and in high-electronic and light-noise areas like
airports. Advantages claimed for these systems include more uniform
depth control and smoother riding quality than for conventional automat-
ic control systems. The laser method is difficult to use when the pavement
surface is designed to have frequent grade changes.
The purpose of the mobile reference is to average the effect of devi-
ations in the existing pavement surface over a distance greater than the
wheelbase of the tractor unit. Therefore, the mobile reference system
expands the base length of the paver in an attempt to smooth out the
profile of the surface on which the paver is moving (Figure 6-23).
Manufacturers have used several types of mobile reference systems
including:
I. Tubular skis ranging in length from 20 to 40 feet (6.1-12.2 m);

2. Tubular skis with wheels at each end; and
__,______3. Floating beams witluhoes.,____________________
This third type of reference system includes a short shoe, or ski, which
slides on an adjacent pavement or curb to which the mat being placed is to
be matched. This type of mobile' reference is only used when the grade
Figure 6-23. Mobile Reference System
being sensed is smooth. Because of its short length, this reference system
does not remove any variations occurring in the adjacent pavement surface
but rather duplicates that grade in the mat being placed.
Effect of Specifications on Paver Operations. The portion of the

specifications dealing with the quantity of HMA to be placed on a project
is most often written in one of three forms: (1) those requiring a nominal
thickness of compacted mat, (2) those requiring placement of an average
amount of mixture per square yard, or (3) those requiring a specified ele-
vation for the finished pavement surface.
To meet a nominal thickness requirement, it is necessary for the con-
tractor to place a mat that has an average thickness approximately equal to
the nominal thickness. The variability in thickness of the mat depends on
the roughness of the existing pavement. The more uneven the existing
pavement, the more difficult it is to meet the nominal thickness require-
ments.
When the specifications require that an average amount of mix be
----~.Pla_q:d_p~r_ s_quare yard,_tli~ specification is really an average yield type of
specification. The average amount of mixture being placed can be quick-
ly calculated for the first truck load of mix by measuring the length of pull
and multiplying this length by the pavement width and dividing that prod-
uct into the weight of the HMA in the first truckload. Appropriate con-
versions are necessary, since the average amount of mixture is typically

specified in units of pounds per square yard. For example, if a 20 ton (18.2
Mg) load of HMA paves 200 linear feet (61 m), 12-feet wide (3.7 m), the
spread rate is 150 pounds per square yard (57.1 kg/m2).
When the specifications require that a HMA pavement be built to
some predetermined elevation, it is essential that stringline or some other
method be used to control elevation. It is also possible to use the string-
line in conjunction with slope control devices.
COMPACTION
Compaction is the process by which the volume of air in a HMA mix-
ture is reduced through the application of external forces. The expulsion
of air enables the mix to occupy a smaller space thereby increasing the unit
weight or density of the mass.
The compacted mixture should have sufficient voids to allow the
asphalt cement to expand and contract as temperature changes without :fill-
ing the voids resulting in flushing. The voids should be high enough to
allow for some subsequent traf:fic~induced densi:fication during the first
few years of service without the void content falling below about 3-4 per-
cent for dense-graded mixtures. If the void content in a dense-graded
asphalt mixture drops below about 3 percent, significant permanent defor-
mation can occur. However, for dense-graded mixtures, the pavement life
is reduced about 10 percent for each percent increase in voids above 7 per-
cent, according to Linden, et al. (i).
The compaction process is affected by the condition of confinement of
the HMA being compacted. In the laboratory, confinement is provided by
the sides and bottom of the mold and the hammer. Compaction energy in
the laboratory process can be calculated. However, in the field, confine-
ment is provided by the surrounding HMA material, the underlying layer
and the compactor contact area in its zone of influence. The amount of
compaction energy put into field material cannot be determined.
However, its effectiveness can be determined by measuring the increase in
density produced by the compaction energy.
;[i
1.:1.
Mixture Factors Affecting Compaction
For effective compaction t<;> occur, the compactive forces exerted by the
roller must exceed the forces resisting compaction within the mixture. The
mixture resistance is a result of the _combined effect of the aggregate phase
and the asphalt cement phase which fills the voids in the aggregates (.g). An
analog for the aggregate-asphalt cement system is the soil-water system of

soil mechanics where the resistance to a normal force is described as:
S=c+crtan<p
where,
S = shearing resistance of the soil;
c = cohesion;
cr = applied vertical stress;
<p = angle of internal friction due to the interlocking
resistance of the particles.
In comparing the soil-water and aggregate-asphalt cement systems,

one must immediately recognize that the cohesive forces for the asphalt
mixture are much greater than that for the soil-water system. This is due
to the fact that the viscous resistance of the asphalt cement is much greater
than the apparent cohesion between soil and water.
The angle of internal friction increases with the angularity of the par-
ticles in the aggregate mass, i.e., a mix made from rounded river gravel and
sand has a lower angle of internal friction than a mix made from crushed
aggregate particles. Generally, the higher the percentage of crushed parti-
cles in an aggregate mass, the higher the angle of internal friction and the
more difficult it is to compact.
"When liquid asphalt cement is added to an aggregate mass being com-
pacted, the lubricating effect of the liquid initially reduces the internal fric-
tion between aggregate particles allowing them to slide against each other
into a denser configuration. However, at high asphalt contents, the asphalt
cement forces the aggregates apart and does not allow the aggregates to be
moved into a denser condition.
The properties of the asphalt cement do not significantly affect the
mass viscosity of the mixture. Since mass viscosity is the primary force
resisting compaction, the aggregate factors which affect it are much more
significant than the asphalt cement properties (Q). For compaction, it
appears that:
1. The voids factor should have a minor effect, since most mixtures
are designed for approximately 4 percent voids;
2:-Tne fitler-15itumen factor is quite-imp-ortant~ as~ir-affects the mass
viscosity of the matrix that surrounds the coarse aggregate particles;
3. The effectiv;e particle size of the filler is important since generally
the smaller the effective diameter of the filler, the greater its mass
viscosity; and
4. The volume concentration of the coarse aggregate probably has the

greatest potential effect on mass viscosity; especially as the maxi-
mum size of the particle increases, since there is an attendant
increase in the volume concentration of aggregate.
The shear rate is also very important and it is one factor over which
the contractor has absolute control. Therefore, if the mass viscosity for a
particular mixture is higher than normal, the contractor can reduce the
speed of the rollers in order to decrease the shear rate to produce shearing
stresses high enough to move the aggregate into a denser position. Of
course, the contractor might be able to increase the weight of the roller to
produce the same effect.
Compaction Equipment
Self-propelled compactors are used to provide the compaction energy
required to reduce the volume of freshly laid HMA sufficiently to produce
the specified density. This compaction train usually consists of two or
more rollers whose objectives are:
1. To achieve the required density to meet the specifications; and

2. To provide a smooth surface.
i, !
These two objectives are accomplished through the use of breakdown

and intermediate rollers to achieve the desired density and a finish roller to
remove roller marks. The breakdown and intermediate rollers may be one
or more rollers, depending on the speed of the paver and the productivity
of the breakdown roller. If the breakdown roller can achieve the required
density, an intermediate roller may not be necessary. A finish roller is usu-
ally required.
The breakdown roller is the first compactor to roll the freshly laid
HMA and is usually a steel-wheeled roller but can be a pneumatic-tired
, roller (PTR). The steel-wheeled roller can be static or vibratory but in
,1 1
either case the first compaction drum to encounter the mix should be the
1
: :.'i!: one providing the drive ·energy to propel the roller. This arrangement
allows the loose HMA to be tucked under the drive drum. If a non-dri-
en-drnm·encounters--freshlyfaid-HMA-;-drere-f·satena.eficyfofl:ne Mrn-----
1 :,I! to shove horizontally and produce a ridge of material immediately in front
'11.11 ! of the drum which remains after the roller reverses directions and produces
,I a bump. Hairline cracks (checking) also develop when an asphalt mixture
Ji is rolled in this way. A PTR can also be used as a breakdown roller but it
l!I
i II
II!
I
is sometimes difficult to remove roller depressions with the finish rollers.
The hot asphalt mixture also tends to stick to the rubber tires causing sur-
face deficiencies. Therefore, a PTR is most often used as an intermediate
roller. Finish rollers are steel-wheeled rollers.
Rolle;s operate in a rolling zone behind the paver whose length is
determined primarily by the number of compaction passes required to
achieve density. The number of passes are in turn determined by the effec-
tiveness of the roller in producing the required density, the compactability
of the HMA being placed, the width of the roller, the width of the paving
lane, the speed of the roller, and the cooling rate of the mat. Compaction
must be accomplished before the mat cools to a temperature below approx-
imately 175°F (80°C), called the cessation temperature. The cooling rate
of the mat is affected by the mat thickness, the temperature of the mat
when placed, the ambient temperature, the temperature of the base on
which the HMA is placed, and the wind conditions. A sample of the types
of data available showing cooling time versus factors affecting cooling rate
is shown in Figure 6-24.
Compaction always results from the application of pressure over a con-
tact area (SJ. The number of contact points in the aggregate mass increas-
es as the particles are moved into more intimate contact. Thus compaction
occurs only when the HMA is confined as a result of the roller, the base
support, and the surrounding mix particles. If sufficient confinement pres-
sure is not present, the HMA will move laterally producing decompaction
rather than moving vertically producing compaction. This possibility for
a decrease in density must be considered when developing a roller pattern
(Figure 6-25). As the temperature of the mix drops and the number of
contact points increases, there is an increased resistance to further com-
paction. This process continues until no additional densification occurs.
At this refusal density point the internal resistance of the mix equals or
exceeds the compaction contact pressure. If the refusal density does not
equal or exceed the specification density, the operator must increase the
compaction pressure. Modification of the mixture design to improve
in-place density is a common mistake made on many paving projects. As
a general rule, the mix design should not be modified for the purpose of
increasing in-place density.
The ability of the roller operator to change the contact pressure
improves the operator's chances of achieving the specification density. The
options available to the operator for increasing the contact pressure
include:
1. Increasing the ballast load (steel-wheeled or rubber-tired rollers);

2. Increasing the tire pressure (rubber-tired rollers); and
01 lO
02
0£ Cl)
Ov w
m; IC) 5
09 z
(./)
Cl)
01
~~
~
u
09 J:
I-
t() I-
<(
~
C\J
0
-----,---.....-------....-----+-
0 0 0 0
0
IC) I'() C\J
S3lnNIVi ',:j. <;LI 01 700:) 01 1 'tri ~0,:j 3Vill

Figure 6-24. Time Allowed for Compaction, Based on Temperature
and Thickness of Mat and Temperature of Underlying
Base (Z)
___ Desired _______ _

>- Density
f,--
C/)
z
w Too Many Passes
0
LOW 0 2 3 4 5 6 PASSES
COMPACTION CURVE
Figure 6-25. Establishing a Roller Pattern Using a Test Strip
3. Increasing the magnitude of the dynamic force component (vibra-

tory rollers).
Steel-wheeled rollers were the earliest rollers used in HMA paving

operations and were used exclusively until tractor-drawn pneumatic-tired
rollers were introduced in the late 1930s (2). In the 1950s, a self-propelled
pneumatic-tired roller was introduced and by the mid-1950s was widely
accepted in many areas of the United States. Rollers with hydrostatic
drives were introduced in the early 1960s, and by the end of that decade
nearly all mahufactured rollers had hydrostatic drive, which provided bet-
ter control of rolling speed and smoother starts and stops when reversing
directions. By the late 1960s, the vibratory-steel-wheeled roller began to
gain popularity and the use of the pneumatic-tired rollers began to decline.
In the late 1960s, the first vibratory rollers specifically designed for HMA
compaction were introduced in the United States. These machines had
single vibrating drums with pneumatic drive wheels. Larger, heavier, vibra-
tory rollers were introduced in the 1970s having single and dual drums
with sizes up to 25 tons (23 Mg).
Steel-Wheeled Rollers. These rollers typically are of the tandem and

three-wheel types and wme in a variety of weights and configurations.
The three-wheeled tricycle rollers typically weigh between 8 and 14 tons
(7 Mg and 13 Mg) (1JJ) . The two rear wheels are typically 55-69 inches
(140-175 cm) in diameter and 18-20 inches (46-51 cm) wide. This roller
is typically used as a breakdown roller during the initial part of the rolling
operation which occurs immediately behind the paver. These rollers are
not widely used anymore. Their biggest advantage is their ability to apply
large pressure, due to the narrow back wheels, an:d thus improve density in
areas such as longitudinal joints.
Tandem steel-wheeled rollers are two axle rollers with both drums
having the same width. They typically weigh from 5 to 14 tons (4.5 to 13
Mg) (12). These rollers can be ballasted with water or sand for extra
weight. Typically, the,drums vary from 48 to 60 inches (122-152 cm) in
diameter and are 42 to 54 inches (107-137 cm) wide. Many of these
rollers have a single drive drum to propel the machine. The drive drum is·
often larger in diameter than the steering (tiller) drum.
Since hot asphalt tends to stick to the steel drums, all rollers have a
series of mats, mounted on a transverse bar, which are kept wet and in con-
tact with the drum surface The mats are kept wet to help minimize stick-
ing, especially when the drum is cold. The roller also has a scraper across
the steel wheels to remove any asphalt mixture that sticks to the wheels.
The water system for wetting the mat is used during breakdown rolling but
is not used for finish rolling.
The actual compactive effort supplied by steel-wheeled rollers is deter-
mined by the contact pressure between the steel drum and the HMA mix-
ture being compacted. The vertical load is determined by the gross weight
of the roller and ballast. The contact area is determined by the diameter of
the drum and the depth of penetration of the drum into the HMA mate-
rial, which is a function of the compaction characteristics of the HMA
(Figure 6-26). Table 6-3 contains contact pressure information for sever-
al rollers of different drum diameters and static linear loads for different
depths of penetration of the drum into the freshly laid mat (h 1-h2 from
Figure 6-26). & the penetration depth increases for a given roller, the
contact pressure decreases since the contact area is larger. For a given con-
tact pressure, larger drums have lower angles of contact than smaller drums
since the angle is a function of the arc length, AIA2, divided by the radius
of the drum (Figure 6-26). Since larger drums have smaller contact angles,
they also have a lower component of horizontal force that tends to push
against the mat being compacted.
Pneumatic-Tired Rollers. These rollers are usually called either rub-

ber..:...tired rollers or pneumatic-tired rollers (PTR). Their name is derived
from the pneumatic (air filled) rubber tires used to compact the HMA.
The inflation pressure in the tires can be varied to produce the desired con-
R = radius of steel wheeled roller
h, = thickness of mat before compaction pass
h1 = thickness of mat after compaction pass
8= angle between points of contact of mat and drum

before compaction A., and alter compaction A1 •
Figure 6-26. Contact Angle of a Steel Drum with a HMA Mat Being
Compacted (11)
tact pressure on the mat. The design of the pneumatic-tired rollers provides
a kneading action by the tires, which are individually mounted to the roller
frame. Typical configurations of pneumatic-tired rollers provide 4, 5, 6 or
7 tires on the front of the roller and 3, 4, 5 or 6 tires on the rear with the
rear tires aligned to provide compaction for the material passing between
the front tires. These rollers typically vary in wheel weights from 1,500 to
10,000 pounds (0.7-4.5 Mg) as reported by Geller (JQ). The tires are
smooth, i.e. they have no tread, and must be able to withstand high infla-
tion pressures for long periods of time. The size and ply rating of the tires
can vary significantly. Inflation pressure is controlled by tire ply rating but
the combination of contact pressure (controlled by tire inflation pressure)
and area (controlled by weight of roller) are important in producing densi-
fication of the HMA. High contact pressures are only attainable with high
tire ply ratings, whether for the 7.50 or 20-inch (19 or 51 cm) rim sizes. If
----~•-P_'.f.R-i-s-t0-be-1:1sed-fe.r-G0-mpaGti0-B.-pur:poses,-then-the-tire.inflation pres-
sures should be as high as the behavior of the HMA will permit without
severe rutting which may be difficult to remove by the finish roller.
PTRs can be used in the breakdown or intermediate roller positions.
In the breakdown mode, PTRs usually cause deformation that may or may
.,,
Si ~
i .
flass Drum Static Lineal Penetration Depth, in.
-
c:r
n,
0\
Diameter Load ~
12-15\Ton Static 69" Conditions 1/16 1/8 3/16 1/4 3/8 1/2 5/8 3/4 f(l n0
3°Wheel
Contact Pressure
n,
t!,. a
Without
Ballast
341 pli psf 174 112 92 83 , 66 58 53 48 ~a
n, ~
t!,. ..,
n, n,
With 429 pli psf 220 142 116 104 83 73 66 60 ~~
I Ballast ~·=
@: a
::i::
$a
~
53 45 38 ~-
rem
8-12 /ron Static 60" Without 180 pli psf 96 64 33 29 26
A
Ballast
"' ~ ►
With
Ballast
250 pli psf 132 88 74 63 53 46 41 36
~~... 1
~
i • 0 ~
Over 14 fon Vibratory
T~ndem
60" 190 pli psf 101 68 56 47
'
1
40 35 31 27
=
"'
;:!.
e.
Y'
9-111/2 Ton 48" Low Range 130 pli psf 78 51 43 38 31 26 24 22 ~ ~

i
Vibrat ry Tandem = i
High
Range
175 pli psf 105 68 58 51 42 35 32 30
I...
0
t:1
~
<§"
6-9 Top Vibratory
Trdem
41" Low Range 115 pli psf 74 50 41
48
36 29 25 23 21
= [
High 135 pli psf 87 58 42 35 30 27 24 t, n
Range .g ~
e- ao.
"' 0
I g
not be removed with finish rolling. Therefore, the mix behavior deter-
mines whether a PTR can be effectively used as a breakdown roller (12).
Geller (1Q) points out that the effectiveness of a PTR as a breakdown roller
can only be determined after the next roller levels the surface, so that a
nuclear gauge can sit fl.at on the surface and density measurements can be
performed.
Traditionally, PTRs have been used in the intermediate roller position.
Geller (1Q) recommends ground contact pressures between 75 and 90 psi
(517 and 621 kPa) which are most easily met by the 20-inch rim with 16
or 18 ply tires. In addition to producing densification, the PTR in the
intermediate position can help to remove roller checking produced by a
steel-wheeled breakdown roller. Checking is the term used to describe
fine, hairline, transverse cracks which occur in the surface of a HMA mat
typically under a steel-wheeled roller. Checking is normally caused by
horizontal movement of the mixture during compaction with the
steel-wheeled rollers. This can be caused by mixture problems (tender
mixture) and/or improper rolling techniques.
Since asphalt cement sticks to cold rubber tires, wetting mats, like
those described for steel-wheeled rollers, are used with individual mats
contacting each tire. Once the tires become warm, the asphalt cement does
not stick to the tires and the water is no longer needed. Rollers should not
be parked on the hot mat since this causes depressions that cannot be
removed with additional rolling. If the tires cool, they have to be brought
back to operating temperature in the same way as for initial cold starts.
Pneumatic-tired rollers are reported to offer several advantages when
compacting dense-graded HMA (12):
I. They provide a more uniform degree of compaction than

steel-wheeled rollers;
2. They provide a tighter, more dense surface, thus decreasing the per-
meability of the layer;
3. They provide increased density that many times cannot be
obtained with steel-wheeled rollers; and
4. They compact the mixture without causing checking and they help
to remove any checking that is caused with steel-wheeled rollers.
------¥ibrat-o-ry-Steel-"Wh.eeled-R-0llen.-T'.h@-m0st-recent-development in
rollers for compacting HMA is the vibratory steel-wheeled roller or vibra-
tory roller. Vibratory rollers require more operator discipline than nonvi-
bratory rollers in carrying out the roller pattern. Selecting the wrong force
level, rolling too fast, and making too many passes with the vibrator
engaged can cause probleml!. However, current roller designs have numer-
ous features to assist the operator in maintaining roller pattern discipline
and the features being added are improving UJ!).
Five types of vibratory rollers which may be seen. in use for HMA com-
paction include:
1. Single drum-rigid frame;

2. Single drum-articulated frame;
3, Double drum-rigid frame;
4. Double drum-single-articulated frame; and
5. Double drum-double-articulated frame.
Schematics of these different configurations 'are shown in Figure 6-27.

The most popular of these rollers are the single drum-articulated frame
and the double drum-single-articulated frame.
Single Drum Vibratory Rollers
Rigid Frame Articulated Frame
Double Drum Vibratory Rollers

(one or both drums may vibrate)
Rigid Frame Single Articulated Frame Double Articulated Framt0
Figure 6-27. Types of Vibrat(?ry Rollers Commonly Used for

Compaction of HMA (.LJ)
Vibratory rollers are the only type of HMA compactors that have a
dynamic load component and are typically lighter than non-vibratory
steel-wheeled rollers because the total, compactive effort is the sum of the
static and dynamic loads. Typical drum widths for vibratory rollers are 58 (
to 84 inches (147 to 213 cm) with typical drum diameters of 40 to 60
inches (102 to 152 cm). The dynamic load applied by the vibratory roller
is the major force component producing densification of the HMA.
The dynamic load is produced by an eccentric weight attached to a
rotating shaft in the center of the drum. When the eccentric weight is
rotating toward the surface being compacted, the total applied load is the
sum of the static and centrifugal loads. When the eccentric weight is rotat-
ing away from the surface being compacted, the total applied load is the
static load minus the centrifugal load. Geller (lff) reports that vibration
reduces the mechanical friction during compaction but results in an
increased mechanical interlock afterwards. The vibration is especially
effective on harsh mixtures which have high volume concentrations of
coarse aggregate.
In order to compact a particular HMA layer effectively, the frequency
and amplitude must be selected to satisfy the conditions on the job so that
the total force applied is adequate to produce the density desired but not
so large that decompaction occurs because of the large vibrating force.
Generally thicker lifts require a higher amplitude and .a lower frequency
compared to thinner lifts. Monitoring changes in field density with each
roller _pass using a nuclear density gauge is probably the simplest way to
establish the proper settings for amplitude and vibration for a particular
project.
Several organizations have evaluated the relationship between impact
spacing and marking of the pavement surface (13.). For surface mixes, a
smooth ride is one of the primary criteria for acceptance; therefore, the
impact spacing should be short enough so that no vibration marks are
observed in the finished surface. To accomplish this, it is recommended
that the impact spacing be 1.5 inches (38 mm) or less (8 or more impacts
per foot).
Field Conditions which Affect Compaction

-------aGt-0i:s-that-affect-the-max-im:um-density-which-can-be--achieved by a
particular set of rollers in the field are HMA characteristics, HMA tem-
perature at laydown, layer thickness, temperature and firmness of the layer
on which the HMA is being placed, air temperature, wind velocity and
humidity, and solar radiation levels. Most of these factors directly affect
Table 6-4. Recommen<!ed Minimum Laydown Temperature, °F ~

Mat Thickness, inches
Base
Temp, °F 1/2 3/4 l I½ 2 u
20-32 2851
32-40 305 295 280
40-50 310 300 285 275
50-60 310 300 295 280 270
60-70 310 300 290 285 275 265
70--80 300 290 285 280 270 265
80-90 290 280 275 270 265 260
> 90 280- 275 270 265 260 255
~-
Rolling
Time2, min. 4 6 8 12 15 15
1
Inc;rease by 15°F (8°C) when placement is on base or subbase containing
frozen moisture.
2
Time available between recommended laydown temperature and cessation
temperature (175°F) (80°C) when attempts to compact the mat should cease.
the cooling rate of the HMA mat and the length of time the material is hot
enough to be compacted. Information on cooling rates and time available
for compaction is typically presented in either tabular or figure form. Table
6-4 contains one set of recommended minimum laydown temperatures for
a range of base temperatures and mat thicknesses. These compaction tem-
peratures are estimates and will vary with different asphalt cements and
aggregates. This table was developed for typical asphalt mixes used in the
United States (ld). Notice that for thin mats, the time available for rolling
is short. For example, a 3/4-inch (19 mm) mat placed at the recom-
mended minimum laydown temperature has only 6 minutes to be com-
pacted to achieve the target density. The roller speeds cannot be increased
significantly without adversely affecting density; hence, additional rollers
may be required when paving at low temperatures.
Rolling Patterns
---+------tiffer.ent.rolling_patterns..ace-used-on-mat.s-with-G0-n-sta1H-c-cF0ss-sl0-p@--atr------
com pared to those with crowns or super elevation. In rolling unsupport-
ed mats with constant cross slope, the first pass of the roller should be on
the outside (low) edge of the mat? with each successive forward pass mov-
ing toward the high edge of the mat as shown in Figure 6-28. The next
CD
Figure 6-28. Recommended Rolling Pattern for Thin Unsupported

Mats(< 2 inches) with Constant Cross Slope (15)
forward pass should overlap the first forward pass by at least 6 inches (150
mm). The roller passes should end at different points to prevent the devel-
opment of a bump. The last forward pass of the roller; at the high edge,
should overlap the upper edge by at least 6 inches (150 mm).
Rolling of transverse joints is important to ensure that a smooth dense
joint is obtained. If possible, transverse joints should be rolled in a trans-
verse direction. This requires that runoff boards be used adjacent to the
HMA mixture to prevent damage to the edge. The runoff boards should
be the same thickness as the asphalt mixture being placed. Many times,
due to traffic considerations and working room, the transverse joint must
be rolled in a longitudinal direction and, thus, runoff boards are not need-
ed. When rolling in a longitudinal direction, the hot mat tends to push
away from the cold asphalt mixture resulting in a depression.
Longitudinal joints are important because of their extent on most
HMA paving projects. It is important that longitudinal joints be com-
pacted properly to ensure a smooth joint that is barely noticeable when
_ _ _ __.,_rossed.--Iflongitudinal-jointsare_noLcompacted properly, the joint may
be rough and noisy when crossed, it may be a future location for water
damage or raveling, and it may be a potential driving hazard. Longitudinal
joints are compacted using the rollers available for compacting other por-
tions of the HMA mixture.
There are several approaches to compacting longitudinal joints. When

placing a nyo lane pavement, the longitudinal joint can be effectively delet-
ed by paving in echelon (a second paver working closely behind the first
paver). The length of some projects is short enough that the second lane
can be placed with one paver before the first lane cools. When both sides
of the longitudinal joint are hot, compaction of the joint is not a problem.
A common method of compacting longitudinal joints involves placing
a second hot mat adjacent (2-3 inch (51-76 mm) overlap) to the first mat
after it has cooled. It is difficult to obtain high density in the joint using
this method. Another method is to heat the cold mat adjacent to the joint
with a joint heater as a new mat is being placed. Experience with this
method has been mixed-some good and some bad. A final method,
sometimes used on airfields and other wide pavements, is to cut back 2-3
inches (51-76 mm) of the cold mat adjacent to the edge with a cutting
wheel to expose a dean, dense, vertical face. The adjacent lane is placed
overlapping the cut face by 2-3 inches (51-76 mm). This technique has
been shown to provide good density but may not be practical on many
roadways where traffic uses the adjacent lane.
Rolling the longitudinal joints is performed using many different tech-
niques. However, almost all methods involve rolling the joint first, then
moving to the opposite side of the paving lane and working back toward
the joint. When vibratory rollers are used to roll longitudinal joints, care
must be used to ensure that the vibration does not damage the material on
the cold side of the joint. Damage to material on the cold side of the joint
depends on the type of aggregate in the HMA and the temperature of the
cold side during compaction. It is important that the vibratory roller not
crack the aggregates or asphalt mixture.
When rollers reverse directions, they must stop prior to moving in the
opposite direction. How that stop occurs is important. Stops must be
made gradually and the roller should be turned slightly so that the wheels
are not at right angles to the direction of travel during the stop. This may
be difficult to do with pneumatic-tired rollers but should be possible if tire
penetration into the mat is minimal. The roller operator must take care to
minimize any scuffing or shoving of the HMA material as the roller
changes directions. If the roller stops with the wheels at right angles to the
direction of travel and a roller mark occurs, it is very difficult to roll out.
---~B~oW-e..V-er,-a-.ma.rk-which...is-obliq.ue.-t0-the-d-i-r-ec;t-i-0n--0f-tr-a:v.d-Ga.fl-De-mme:-----
easily rolled out and is more difficult to feel in an automobile. Vibratory
rollers of modern design are equipped with adjustable on-off travel speed
settings at which vibration is automatically disengaged and re-engaged
when the roller goes through a reversal of travel direction. Parking a roller
on a fresh mat should be avoided because it may cause. a depression. It is

preferable that the roller be kept busy but if a long delay is encountered,
the rollers should he parked on a cold mat or as far away from the fresh
mat as is practical.
When rolling is complete a visual survey should be conducted to deter-
mine the condition of the freshly compacted surface. There should be no
roller marks that have not been removed by the finish roller. There should
be no undulation in the pavement surface either longitudinally or trans-
versely. The surface texture should be uniform with no scuff marks.
Most state DOTs require that the surface profile be checked with a
straightedge of some length, typically 10 feet (3.0 m). · The amount of
deviation from the straightedge is noted and compared to the specifica-
tions. Typically a maximum of 1/s-inch (3.2 mm) variation is allowed for
a 10-foot (3.0 m) length. The straightedge is used to check all transverse
joints and other visible surface irregularities. It is placed on the surface,
parallel to the centerline, and the maximum deviation between the road
surface and the bottom of the straightedge is measured.
More states are checking final quality of the surface by measuring the
smoothness of the road while driving over the surface at a prescribed speed.
In some states the smoothness rating is a subjective rating given by the
inspector or engineer. In other states smoothness is rated using some type
of roughness measuring device like a profilometer or roadmeter (Chapter
8). If a road is too rough, the smoothness is sometimes improved by grind-
ing the high spots.
Types of Density Specifications Used for Construction of HMA

The amount of voids in a HMA mixture is probably the single most
important factor that affects performance of the mixture throughout the
life of the pavement. For a given aggregate gradation the voids are pri-
marily controlled by asphalt content, compactive effort during construc-
tion, and additional compaction under traffic. The density requirements
and the methods of measuring density vary considerably from one user
agency to another. Some agencies construct a control test strip, measure
the density on the strip, and use that density as the target density for the
-----v-,roject.-O..th.er__agencies_com.paCLsamples.in....rhe__fieldJ.ahoratory during
construction and use that density or some percentage of it as the target
density. Still, other agencies measure the theoretical maximum density
(ASTM D2041) and use some percentage of that density as the target den-
sity. All of these techniques have been used successfully to build pavements.
that perform satisfactorily; however, all these techniques have also been
misuse1d, producing pavements that do not perform satisfactorily.
One difficulty in preparing density specifications is describing the
method of measurement. The two primary methods for measuring densi-.
ty are (1) taking cores from the in-place pavement and then measuring
bulk density in the laboratory, and (2) using nuclear gauges to measure the
in-place density. Most engineers agree that nuclear gauges do not measure
density as accurately as the core density procedures. Many agencies use the
nuclear gauge to develop roller patterns but specify that cores be taken and
tested for acceptance or rejection of the in-place mixture. However, some
agencies believe that a properly calibrated and. properly operated nuclear
gauge provides results accurate enough for acceptance testing. Some cores
should routinely be taken throughout the paving project to verify that the
1
nuclear gauge is in calibration.
The voids in an asphalt mixture are directly related to density; thus,
density must be closely controlled to ensure that the initial in-place voids
are within an acceptable range. There is considerable evidence to show
that the initial in-place voids for dense-graded mixtures should be no
higher than approximately 8 percent and never fall below approximately 3
percent during the life of the pavement Cl§,_ 12. l1J.,. 12,_ 2Q,_ 21). High
voids allow water and air to penetrate into the structure resulting in water
damage, oxidation, raveling, and cracking. Low voids lead to rutting and
shoving of asphalt mixtures. This recommended range of in-place voids
applies only to dense-graded HMA mixtures. Other mixtures, which are
specifically designed to contain higher or lower air voids, can perform
properly when designed using proper techniques.
Ford ( 16) showed in a study for the state of Arkansas that HMA mix-
tures should be designed and constructed so that the in-place air voids stay
above 2.5 percent. At void levels above 2.5 percent, Ford's data showed
that the expected rut depth would be less than 10/32 inch (7.9 mm)
(Figure 6-29) based on air voids determined from cores and rut depth
measurements of in-s_ervice pavements.
Brown and Cross (12) showed that significant rutting would likely
occur in mixes that had laboratory compacted voids of 3 percent or less.
When a suitable aggregate was used and the voids stayed above 3 percent,
little rutting was observed. Huber (11f), in a study of HMA pavements in
----\Gafl.-a@,e0n-si-derecl-a-ft1:l:l"l'l:0er-0-fea1:tSes-0f-rt1:tting--and-determined-that·on·-----
of the primary causes was air void content below 3 percent.
Zube (12) showed that dense-graded HMA pavements become high-
ly permeable to water at approximately 8 percent air voids. Figure 6-30
shows that as long as the voids are below 8 percent permeability should not
RUT DEPTH vs. AIR VOIDS

40
..c.
0
C:
C\I 30
-
I'()
Best Filled Equation
Log RUT = 1.188 - 0.6947 Log AIR VOIDS
R2 = 0.456, RMSE = 0.186
:C: 20
I--
CL
w
Cl
I- 10
:::)
0::
Figure 6-29. Relationship Between Air Voids and Rut Depth in

Arkansas (1.i)
be a problem, but the permeability increases quickly as the void level

exceeds 8 percent. Brown, Collins, and Brownfield (2Q), in a study of seg-
regated mixes in Georgia, also showed that HMA mixtures were imperme-
able to water as long as the air void content was below approximately 8 per-
cent. Figure 6-31 shows that the permeability increases rapidly at void
contents above 8 percent.
Santucci, et al. (2.1) showed that the retained penetration of asphalt
cement is affected by the air voids in the asphalt pavement. Figure 6-32
shows that the loss in asphalt cement penetration increases significantly as
the air voids exceed 8 percent. Therefore, if dense-graded asphalt mixtures
are constructed with air void contents below about 8 percent, rapid oxida-
tion which leads to cracking and raveling is avoided and adequate perfor-
mance results.
From these previous studies, one concludes that dense-graded HMA
mixtures should be constructed with an initial air void content below
approximately 8 percent and that the final air void content after traffic
should be above approximately 3 percent. Good performance should
------o-cc_u_r_if_ these limits are observed and smtaoleaggfegates and asphalt
cement are used.
Density Specified as a Percent of Laboratory Density. One method

for specifying field density requires that the in-place material be compact-
15-----,----,----,----r----ir-----,
----- --
--·
V
0
0 i---
0 ... i.--
101----c~>+-I/__.~-L'~_YO"'l------+-=0---+----+-~->----1
V O I) 0
~
Cb ..,.~ > o
I
CJ)
o,/o 0
a
0
>
o--_.......___....____._______._____._____,
0 I 00 200 300 400 500 600
FIELD PERMEABILITY ML/MIN
Figure 6-30. Relationship Between Air Voids and Permeability for 10
Projects in California (12)
ed ·to some percentage of the laboratory density. The Marshall method

specifies that laboratory specimens be compacted using either 50 or 75
blows with the Marshall hammer. In recent years, most states have
required 75 blows for high volume roads. Typically density specifications
require that field density be at least 95 percent of laboratory density and,
in some cases, at least 98 percent. Some specifications prohibit the
in-place density being greater than 100 percent of the laboratory density
because--premature-ruttrng-fs-likelTro-o-ccuras a resulrof towair voia"'s-.- - - - - -
In order for this type of specification to work effectively several items
are required. First, the material produced at the HMA facility must be
compacted in the field laboratory to establish a reference density and to
evaluate the amount of air voids in the mixture at the reference density. If
0.7
0.6
-.E
Q)
: :,
C:
-en
Q)
.c:
o.5
(.)
C:
-;_0.4
.J:l
C
Q)
E
a5 0.3
Cl.
Q)
-
.::::
C
a5 0.2
a:::
_0.1
0.0L.(")...lC>---O:;.....i...._ ___,,_ _ _-1-_...;..._ _ _ _ __

15 20
Total Mix, percent
Figure 6-31. Relationship Between Air Voids and Permeability for
Pavements in Georgia ~
the air void content of the laboratory compacted material is not satisfacto-
ry, the mix must be adjusted so that an acceptable air void level is obtained.
Often the only adjustment required is a slight modification of the asphalt
------ccomem-:--T--:he-clensi-ty-p·1:00:1:1:eea-d1¾r-ing-mixtui:e-design-should not be used
as the reference density since the laboratory produced material is usually
somewhat different from the material produced in the HMA facility.
Aggregates tend to break down during mix production, creating an
increase in dust, thus altering the properties of the compacted asphalt mix-
1,I ·100
90
:i:I sci
i ,70 10
I
60
i :I ~ 50
11'
-0 9
] 40
0 0
i ~-
C 30
15
13
I
.2
"§
ai
~ 20
CL
104'------'6e-----!:a,------+.1o:c----:'::12,---,---=1'-=-4---t,5·
Pavement Air Vold Content, %
Figure 6-32. Relationship Between Retained Penetration and Air

Voids (21)
ture. The material produced in the field may be different from that pro-
vided for mix design due to normal sampling and testing errors prior to
and during mix design.
In the development of the Marshall method, the density produced
with the Marshall manual hammer was correlated with density in the field
after traffic (22). Hence any other method of compaction (mechanical or
otherwise) must be calibrated to produce a density equal to that obtained
with the hand hammer or, better yet, should be calibrated to produce a
density equal to that obtained in the field after several years of traffic. The
procedures specified in ASTM Dl559 and MSHTO T245 for the
Marshall test require that the manual hammer be used or that the chosen
compaction method be calibrated to produce the same density as that
obtained with the manual hammer.
Suppose a mix is designed to have 4 percent voids and the field com-
paction is required to be at least 95 percent of laboratory density. This
specification results in up to 9 percent voids (4 percent at 100 percent lab-
oratory-density-plus-s-pel'tenn:noreac95 percent of laboratory density)
immediately after compaction and should result in approximately 4 per-
cent voids after several years of traffic. The initial voids (9 percent) may be
a little high, but the final voids (4 percent) should be acceptable. The high
initial voids may result in rapid oxidation resulting in an increase in crack-
ing and raveling.· If this mix is subjected to a high volume of traffic, then
a small rut (5 percent oflayer thickness, 0.10 inch (2.5 mm) for 2-inch
(51 mm) layer) results in the wheel paths. This 0.10-inch (2.5 mm) rut,
which is insignificant, is caused by densification under traffic.
·If a mixture is designed to have four percent air voids and is compact-
ed to a density greater than 100 percent, rutting may develop early in the
life of the pavement since the initial in-place voids are less than four per-
cent. If the laboratory compactive effort is sufficient to produce a density
similar to that obtained in the field after a few years of traffic, then past
experience has shown that the contractor cannot reasonably compact the
mix to a density higher than the laboratory density. Thus, if the density
on any project continually approaches or exceeds 100 percent of the labo-
ratory density, it is likely that the reference laboratory density is low for
various reasons and not likely that the contractor is applying excessive
compactive effort in the field.
This technique of evaluating field compaction results in good perfor-
mance for properly designed mixes if 1) laboratory samples are compacted
during construction to establish a reference density, 2) correct laboratory
compaction techniques are used for compacting laboratory samples, and 3)
a minimum compaction requirement is set to insure that in-place air voids
after field compaction are approximately 8 percent or less.
Density Specified as Percent of Theoretical Maximum Density. A

second method of compaction control for density is to require that the
contractor achieve some minimum percentage of the theoretical maximum
density (TMD). This compaction requirement provides a direct method
of specifying in-place air voids and an indirect method for controlling
compaction. This method involves taking a sample of the asphalt mixture
during ~onstruction and conducting tests to measure the TMD (ASTM
D2041). The bulk density of the asphalt mixture is measured after com-
paction of the in-place mixture and compared to the TMD.
This type of compaction specification requires that the TMD, which
is the reference density, be measured routinely during construction. The
TMD measured during mix design should be a starting point for con-
struction but should not be used as a reference density for the field pro-
duced material.
S-onre-states-do--not-compacr-samples-0f-the-field-produced materials
in the laboratory during construction. Some engineers believe that labo-
ratory compaction of samples is not necessary since the reference density is
based on the TMD which is determined without compacting samples.
HMA samples must be taken during construction and specimens com-
.pacted in the laboratory to control the construction process and to verify a

satisfactory,void content in the mixture. The voids in the laboratory com-
pacted samples must be measured and evaluated to estimate the in-place
voids after traffic. It is undesirable to build a pavement with an initial air
void content of 7-8 percent if the voids reduce to 1-2 percent after one
summer of traffic. The only way to estimate the final in-place voids
(which is probably the most critical property of an asphalt mixture) is to
compact samples in the laboratory using the specified laboratory com-
paction technique-(manual or equivalent) and then to measure the voids.
If those voids are not at an acceptable level (typically approximately four
percent), then the mix must be modified to produce acceptable voids.
This type density specification has often been misused because of over-
concern with achieving the required percentage of the TMD. On many
projects, so much emphasis has been placed on the initial density after
compaction that the asphalt content has been arbitrarily increased to
increase the field density as a percent ofTMD. The increased asphalt con-
tent fills the voids making it appear that incrl:!ased compaction is being
obtained. Lower in-place voids are obtained but not as a result of
increased compaction. The voids will continue to decrease under traffic
until the voids fall below approximately 3 percent at which time rutting
likely develops due to plastic flow of the asphalt mixture. The asphalt con-
tent is often increased when paving in cold weather or at other times when
density is difficult to achieve. If v~ids are high during construction in cold
weather or at other times, additional compactive effort and improved roller
patterns should be tried before modifying the mix design.
Therefore, this method of specifying density does encourage adding
extra asphalt cement and filler to make field density easier to achieve if den-
sity is the primary acceptance criterion (23). Laboratory compaction tests
must be conducted during construction to ensure that the voids in the lab-
oratory compacted samples are maintained within an acceptable range.
The TMD must be routinely measured on the material being placed to
obtain the correct TMD for comparison to in-place density. Additional
asphalt cement should not be added for the sole purpose of reducing the
in-place voids to satisfy the density requirements. The asphalt content for
the mixture should only be changed to improve' the properties (primarily
voids) of the laboratory compacted mixture.
Density Specified as Percent of Control Strip. A third method that

is used to specify density is to compare the bulk density of the in-place
asphalt mixture to the bulk density of a control strip that is constructed
using selected rolling procedures. Most specifications that use the control
strip method require that the control strip be compacted to some mini-
mum percentage of the design laboratory density or to some minimum
percentage ofTMD. If the specifications do not require some minimum
density for the control strip, then the inspector must closely evaluate the
compaction equipment and rolling procedures to ensure that reasonable
compactive effort is applied to the asphalt mixture. Any significant
changes in the mixture during construction require that a new control strip
be constructed and evaluated.
This method of density control allows the compactibility of a mixture
to be evaluated and included as part of the specification requirements.
Monitoring the density with a nuclear gauge provides the information nec-
essary to know when maximum density for the rollers being used has been
obtained. If this maximum density is lower than desired, the size of rollers
should probably be increased or the rolling techniques improved. A large
number of factors affect density and a change in any one of these factors
can make the test strip invalid. Common factors affecting density include
aggregate gradation (especially percent passing the Number 200 sieve),
asphalt content, moisture content, mixture temperature, air temperature,
layer thickness, firmness of the underlying layer, roller weight, roller pat-
tern, roller speed, etc.
A typical control strip specification requires some minimum density
for the control strip. The remaining construction must achieve some min-
imum percent of the control strip density. This specification can be made
simpler by requiring that the compacted mix meet some percentage of lab-
oratory density or TMD. A control strip could be constructed to ensure
that the specified density is achievable and an appropriate rolling pattern is
developed. For example, assume that a specification requires that a control
strip have a density of at least 94 percent ofTMD and that all HMA placed
after the control strip have a density of at least 98 percent of the control
strip. This specification could be made simpler by requiring that the
in-place mixture be compacted to a minimum density of 92 percent of
TMD since both of these examples result in similar compaction require-
ments (0.94 X 0.98 = 0.92).
The control strip method of specifying density can be used to obtain
satisfactory results. However, the specifications should be written so that
------1c.-he-initiaLin=place-v:oids...in._th.e..HMA...do__no_t_exceed_approximately 8 per-
cent, and the in-place voids after traffic do not fall below approximately 3
percent. To ensure that the voids never fall below 3 percent, laboratory
specimens must be compacted during construction to estimate the
in-place air voids after traffic.
Measurement of Density
Core Method. The core method of measuring density is the referee
procedure, which is the standard to which the nuclear gauge is compared.
Requiring that density be r measured with cores does impose a delay
between the completion of compaction and a determination of the densi-
ty results, since the pavement has to cool before cores can be taken and
then the cores must be dried to obtain dry weight. In most cases the den-
sity results using the core method are obtained the day after construction
is completed.
After cutting the core from the pavement layer, the freshly placed
material must be separated from the underlying material attached to the
core. In some cases paper or other bond breakip.g material is placed on the
existing surface prior to placing the HMA. When this is done, the core can
be easily separated; however, the location of the paper must be carefully
marked so that the core can be taken in the proper place. There are some
problems in using paper. or other material as a bond breaker, as there is
some concern that a lower density in the asphalt mixture may be obtained
over the bond breaker.
The method most often used to obtain core samples is to identify ran-
dom sample locations, to cut a full depth core through the asphalt pave-
ment, and to saw or otherwise separate the layer being tested from the
underlying material. This method is believed to be the most suitable
method of evaluating the overall density of the pavement.
Burati and Elzoghbi (2.1) showed that the variability of density test
results was lower with core measurements than with nuclear density mea-
surements. Their study included nuclear gauges made by three different
manufacturers which were used on two constrµ_ction projects. They found
that there was a statistically significant difference in the average density
measured from cores and with the nuclear gauges.
Nuclear Gauge Method. Nuclear gauges have been used for a num-
ber of years to measure the bulk density of HMA mixtures. This technique
has the advantage that results can be determined rapidly and nondestruc-
tively.
Most nuclear density gauges measure density of HMA in the back-
scatter mode. A gauge is placed on the selected spot on top of the pave-
---ment and a reading is made:The density determmeclrrom the readmg rep-
resents the average density for the top several inches of the pavement. For
instance, the average density may be representative of the top 6 inches
(150 mm) of the pavement, howeve::r, the thickness of the layer of interest
may only be 2 inches (51 mm). One way to eliminate the effect of the
underlying layers is to calibrate the nuclear gauge to produce the same den-
sity as that obtained from cores cut at the same spot. This calibration cor-
rects for the density of the underlying material. Even with this calibration,
errors still exist due to variations in layer thickness and variations in densi-
ty of the underlying layers along the length of the project.
In recent years, a nuclear. gauge has been developed that measures the
density of thin lifts. This new gauge should provide greater accuracy in
measuring density as compared to the previous gauges; however, the over-
all accuracy of these gauges is not known.
The best use of nuclear gauges is in the development of rolling patterns
by quickly monitoring changes in densification of HMA materials. Some
nuclear gauges have been attached to rollers to continually measure densi-
ty during the rolling operation. Used regularly, the nuclear gauge can be a
valuable quality control tool. Because of the possibility of errors with
nuclear gauges, it is not recommended that these gauges be used alone for
acceptance testing, even with daily calibration of the gauge. Some cores
should routinely be taken to verify the accuracy of the nuclear gauge and
to ensure that an acceptable density is obtained.
MIXTURE SEGREGATION
Segregation is another significant mat deficiency problem which can
cause performance problems. Segregation occurs when the HMA materi-
al behind the laydown machine does not have a consistent gradation across
the mat, usually as a result of the coarse aggregate separating from the fine
aggregate. The surface texture of a segregated mixture is very coarse when
compared to a normal mat surface. Segregation can be caused at any point
in the construction operation between the stockpiles and the laydown
operation (Figure 6-33). Because there are so many potential sources of
segregation, it is often very difficult to identify the cause or causes of a par-
ticular case of segregation. Based on the type and location of the segrega-
tion, the potential causes can be identified and checked to determine how
to correct the problem. An excellent segregation troubleshooting guide
was prepared by Brock (25). Segregation is a significant construction prob-
lem in much of the United States. The types of segregation discussed by
Brock are:
-----------+1.----1,Random----segregati-0n;-----------
2. Each side segregation which is evenly spaced between trucks;
3. Center line segregation;
4. Each side continuous segregation; and
5. One side segregation evenly spaced between trucks.
Figure 6-33. Typical Segregation in a HMA Paving Machine
When segregation occurs on a project, the cause must be determined

and the problem corrected. Otherwise, poor performance is likely to
result. Segregated areas are rough, tend to ravel, and are permeable to
water.
CONTRACT AND MATERIAL SPECIFICATIONS
General
Plans and specifications are used to describe to the contractor specifiy
cally how a project is to be built. The plans provide the dimensions and
layout of all aspects of the project. The specifications describe the materi-
als, workmanship, and other general requirements for the project. It is
- - - - · es~ential _that the specificatiJ::ms_ are_clear to .both the contractor..and..th<e-------
:
1
owner in describing what is required of the contractor. Failure to prepare

clear specifications often results in increased cost to the contractor result-
ing in claims that have to be evaluated by the owner and which, in many
cases, end up in court.
--
Bidding Methods
There are several approaches that have been used by owners to select a
contractor for a particular project. The three most common approaches
are low bid, best bid, and negotiated contract. Each of these methods have
some advantages and disadvantages that are discussed below.
Low Bid. The low bid method is the most common approach used by
public agencies for selecting a contractor. This method involves supplying
all the information about a project to a number of contractors and allow-
ing each contractor to submit a bid providing a detailed cost breakdown
for performing the work. It is assumed that each contractor will perform
satisfactory work since the plans and specifications describe the minimum
requirements necessary for a satisfactory project. The contractor with the
lowest bid is awarded the contract. This method encourages competition
among the contractors and results in a contract with lower initial cost. A
major problem with this method, however, is that competition is often so
keen among contractors that, in order to secure the contract, the winning
contractor may cut corners in his bid, making it difficult to perform the
construction for the contract costs. When this happens, construction qual-
ity may suffer.
Best Bid. The best bid method of selecting a contractor is similar to

the low bid method but involves a detailed analysis of the bidding docu-
ment before selecting a contractor. With this method the owner evaluates
each contractor's documents in detail to determine how the contractor
plans to perform the work and thus makes the selection based on both cost
and work plan. This method should produce a project that is more satis-
factory to the owner but also usually results in higher initial cost.
Negotiated Bid. A negotiated contract is often used on small projects

or projects that must be awarded in a very short time frame. This method
of awarding a contract involves identifying one or more contractors and
negotiating with each to determine the best construction approach as well
as the contract cost for the project. This method requires a close working
relationship between the owner and contractor to determine the desired
method for performing the contract and in establishing the cost. This
------m-,ethod-does-al-l-<:>w-1=he-0wner-1:0-p-re-seleet-t:he-Gontractor and, hence, helps
ensure construction of a satisfactory project in a timely manner, but it is
also likely to be more expensive than a low bid system. This method dis-
courages competition among contractors and is not widely used on large
projects.
Specification Development
The first step in ensuring satisfactory construction of a HMA is prepa-
ration of adequate specifications for the project. The second step is to
ensure the specification limits are met during mix design and construction.
Generally the two types of specifications used to describe asphalt pavement
construction are method and performance specifications.
Method Specification. The method specification describes in detail

the equipment and procedures used to obtain the desired quality of asphalt
mixture. This type specification requires sufficient detail to describe all the
variables required to obtain a satisfactory asphalt mixture. Items that affect
density, for instance, include air temperature, mix temperature, roller type,
roller weight, roller passes, roller tire pressure (if rubber tires) and roller
speed. All of these items have to be clearly described in a method specifi-
cation.
A typical compaction method specification may include the following
statements: "The air temperature at the time of placement shall be at least
40°F (4.4°C) and rising or shall be above 45°.F (7.2°C) if falling. Initial
rolling shall include at least 2 coverages with a 10-ton vibratory roller.
Intermediate rolling shall be performed with 6 coverages of a rubber-tired
roller. The minimum roller weight shall be 15 tons (13.6 Mg) and the tire
pressure shall be at least 90 psi (621 kPa). Intermediate rolling shall be
completed before the mixture cools below 175°F (79.4°C). The final
rolling shall be accomplished with a tandem steel-wheeled roller.
Sufficient passes shall be made with the final roller to remove all roller
marks and other pavement irregularities. All rollers shall be operated at a
speed not to exceed 3 miles per hour (4.8 km/hour). All rollers shall stay
as close as practical behind the paver or roller in front of it." This specifi-
cation becomes long and wordy because of all the detail required to speci-
fy adequate construction procedures.
· A method specification requires that an inspector be on the job site at
all times. If an inspector is not available during construction, it is impos-
sible later to determine whether the contractor complied with the specifi-
cation requirements. The method specification requires that the inspector
work closely with the contractor to point out any deficiencies as the work
is being performed. This type of specification is based on past experience
of rolling procedures and a knowledge of the requirements that produce
----,.atisfactory-d-errsity.
End-Result Specification. The end-result specification is normally

shorter and easier to write than a method specification. The biggest diffi-
culty in preparing an end-result specification is determining what proper-
ties to specify and what limits to set for these properties. For this type of
specification to be acceptable, these specified properties should be directly
or indirectly related to performance.
An end-result specification for density may state for example; "The
asphalt mixture will be divided into lot sizes of 1,000 tons (909 Mg) each
for evaluation. Each lot will be divided into 4 sublots and one random
sample will be taken from each sublot. The asphalt mixture will be cored
after compaction and the bulk density will be determined in accordance
with ASTM D2726. The average density of the four samples shall be at
least 98% of laboratory density. If the average density is less than 98%
then the mix will be accepted at a reduced price." A table for this reduced
. price is provided in the specifications. This specification does not require
that the inspector be on the project at all times; the tests can be conduct-
ed sometime after construction.
An end-result specification should clearly identify several items to be
effective (2.6). These items include:
l. Lot size;
2. Mixture properties to be evaluated;
3. Number of tests per lot;
4. The point of random sampling;
5. The sampling method;
6. Number of measurements per test result;
7. Test or measurement method;
8. Target value of measured characteristic;
9. Realistic tolerance; and
10. Action to be taken in case of noncompliance with specifications.
A lot is that amount of material which is evaluated for pay purposes in

quality control or quality acceptance testing. Each lot is evaluated as a sep-
arate project based on test results from a specified number of random sam-
ples taken from i:hat lot. The lot size is often set to be equal to the amount
of material that is placed in one day. The lot size can also be set equal to
an amount of production, such as 1,000 tons.
The number of tests for each lot varies between specifications but is
usually set at 4 to 5. When evaluating quality, a larger number of tests pro-
vides a better estimate of the true population. Even so, the number of sam-
-----ples-0-0ta.1:fl:ed-aoo-t-@st-e--d-mt1st-h€-l-i-mited-due-to-testing-cost and time.
Typically, the lot is divided into sublots with one random sample taken
from each sublot. These stratified random samples help to spread the test-
ing out so that more than one test is not required within a short period of
time.
The point of material sampling can be identified as in-place, truck,

stockpile, conveyor belt, or other locations in the production process.
Specifying the point of sampling reduces the sampling error between sam-
ples and minimizes testing variability. '[he only place where truly
end-result samples can be obtained is from the roadway. For convenience
however, HMA samples, except for density testing, are often· taken from
the loaded trucks.
In some cases, the test method can make a difference in the measured
test results. For example, the density of HMA measured from a core is
likely different than that measured with a nuclear gauge. Because of this,
the test method should be specified to ensure that the contractor and
owner use the same methods so data may be compared. The two general
types of samples that are obtained are random and representative. From a
statistical standpoint, random samples are more desirable. A sample is con-
sidered random if any potential sample of mix, in the lot or sublot being
tested, has an equal chance of being selected as any other potential sample.
A representative sample, which may be significantly different from a ran-
dom sample, is obtained so that it represents the average of the lot or sublot
being tested. If the average gradation of a stockpile is desired, then a rep-
resentative sample of the stockpile is needed. Typically a representative
sample is obtained by taking a number of small random samples and com-
bining them to produce an average sample.
The specification requirements must be established as a function of the
number of tests. It is desirable that the test results be near the target value
and have low variability. An average test property, calculated from a large
number of tests on random samples, provides a good estimate of the true
average. A small number of samples may not provide a good estimate of
the overall average because of high variability. Smaller sample sizes increase
the chance of inferior material being accepted or good material being
rejected. The number of samples specified to be tested to determine the
average test property then is a trade off between cost for sampling and test-
ing and desired accuracy of test results.
The test method used to measure a test property must be specified
because various test methods have different accuracies. For instance,
asphalt content can be measured by the nuclear gauge and by several
asphalt extraction methods. The results obtained from these tests are
----tt·epead~Flt-0H-fh@-t-@St-me1;h<:>a-usea-1:0-measttre-t:he-aspha:lt-content;-henc•p,-,-----
the test method should be specified or the tolerances should be adjusted to
fit the test method.
A target value for any test must be detailed in the specifications. This
target value is the value which the contractor should try to achieve in the
field. It is impossible to produce a mixture that exactly meets the target

value so some tolerance around the target value must be allowed. The tol-
erance should be wide enough to allow for normal variability of materials
but tight enough to ensure that low quality materials are not accepted.
Sometimes the tolerance in specifications is too tight and is therefore
restrictive, resulting-in rejection of high quality materials. Sometimes the
tolerance is too wide allowing low quality material to be accepted. The tol-
erance should be a function of normal material variability, test variability
(test method), and the effect of the material property on mixture perfor-
mance. In cases where the test variability is considerably higher than mate-
rial variability, it is often difficult to evaluate the acceptability of the mix-
ture. A reasonable tolerance must be used for the specifications to be legal-
ly enforceable.
A major problem with many specifications is caused by the lack of a
clear description of the action to be taken when the specifications are not·
met. Many specifications reduce the contractor's pay when marginally out-
side the specification requirements. As the test results deviate more from
the specification requirements, the amount of pay is further reduced.
Another method for handling material that does not meet the specification
requirements is a go - no go approach. With this method the construc-
tion operation is stopped when test results do not meet specification
requirements. The mixture problem is identified and corrected prior to
restart of production. This method may allow for the material not meet-
ing specifications to remain in-place or it may require removal and replace-
ment.
In a specification, the word shall usually denotes what the contractor
must do and the word will usually denotes what the owner will do. Being
consistent in the use of such words m_inimizes problems during inspection
and evaluation of the asphalt mixture. In some specifications, many words
and phrases, which are difficult to interpret and understand, are used. The
use of such ambiguous words should be minimized or else be very clearly
defined. Some of these ambiguous words and phrases are: to the satisfac-
tion of the engineer or as approved by the engineer, approximately, reject, rea-
sonably, clean, sound, durable, and dry.
-----.Sampling-of-Aggi:eg.ate.-and-Calibration-of_Eeeders.
Taking satisfactory samples at the HMA facility during construction is

an essential first step in proper evaluation of the aggregate, asphalt cement,
and HMA. The sampling program may include sampling at stockpiles,
individual aggregate feeders, combined aggregates on conveyor belt, hot

bins, asphalt cement storage tanks, loaded trucks, or in-place.
A number of random samples are normally taken during the produc-
tion process when it is desired to determine the average and variability of
the properties being evaluated. Representative samples are taken in such a
way that the material secured has properties equal to the average of a spec-
ified amount of material. In the mix design process, a representative sam-
ple of each stockpile is taken since it is the average properties that are desir-
able when only one test (mix design) is conducted.
Aggregate Sampling. Once a sample (whether random or represen-

tative) has been taken it must be reduced to the proper test size prior to
testing. For aggregates this can be done by the quartering method or by
using a sample splitter (ASTM C702). Either method is satisfactory, but
the sample splitter is easier to use and is preferred by most laboratory tech-
me1ans.
Obtaining a satisfactory sample from an aggregate stockpile is not easy
because larger particles tend to roll down the outer surface of the stockpile
and thus accumulate near the bottom. Because of this, a sample of the
aggregate taken near the bottom of a stockpile is coarser than one taken at
a higher point on the stockpile. This problem is somewhat minimized
when aggregates are properly stockpiled by sizes. For instance, typical
aggregate stockpiles include coarse aggregate(½ inch- No. 4) (12.5-4.76
mm), fine crushed aggregate (minus No. 4) and natural sand (fine
uncrushed aggregate). When stockpiling in this manner the range of
aggregate sizes in a stockpile is minimized, making it easier to obtain an
accurate sample and to produce a quality HMA material. It is not good
. practice to stockpile aggregate consisting of all sizes (for example ½ inch to
No. 200 material) (12.5-0.075 mm) together in one stockpile since segre-
gation is almost always a problem. Producing a consistent HMA mixture
when all sizes of aggregates are stockpiled together is very difficult, and it
is also difficult to obtain a representative sample from such a stockpile.
The HMA aggregate can also be sampled during production after it
flows from the cold feeders onto the gathering conveyor belt. Each feeder
is set to provide the desired amount of material into the HMA facility. The
>---~e._e_ders_ar.e_calib.rated_s_o_that__the__plant--oper.ato.r-ean-qu.idcl:y--.set-the-fo1::d-1=at.~-- - - - -
to meet any desired gradation of material. The most accurate way to sam-
ple this material is to stop the belt for a short period of time and remove
all material within a given length pf belt (usually 5 feet) (1.5 m). This
material can then be tested to determine aggregate gradation. Sampling
can also be accomplished during production by using a diverter chute, but

this may be difficult because of the high rate of feed which provides a large
sample in just a matter of 1-2 seconds of diversion. For instance, if the
output of a plant is 300 tons/hour (273 Mg/hour), the amount of materi-
al moving past a point in one second is approximately 170 pounds (77 kg).
After the feeders are calibrated and material is fed into the dryer, addi-
tional aggregate samples can be taken at batch facilities. However, no
aggregate samples can be taken after this poirit in a drum mix facility. In
a batch facility, the aggregates are separated into sizes in hot bins. There
are locations directly underneath the hot bins where aggregate samples can
be taken as the aggregate in each bin is dropped into the scales. The gra-
dation of the aggregate in each bin, along with the proportion of material
desired from each bin, can be used to determine the total gradation of the
aggregate going into the mixture.
Aggregate Feeder Calibration. Aggregate feeders are calibrated one

at a time using the same material that is fed through any particular feeder.
For Instance, if cold feeder No. 1 is used to feed coarse aggregate, then it
must be calibrated with that coarse aggregate. There are typically two ways
to vary the fl.ow of material through an aggregate feeder: changing the gate
opening and varying the feeder belt speed (if feeder belts are used). Each
bin has a gate that can be opened to various heights to change the rate of
flow of material. During the calibration process, the gate is usually set at
some desired opening based on the size of the aggregate and the approxi-
mate rate of material to be fed through the feeder. After a sample .of aggre-
gate is obtained, the speed of the feeder belt is changed to vary the amount
of material on the belt. For a given setting, the material is fed through the
feeder and onto the gathering conveyor and then onto the conveyor that
feeds into the dryer. To secure a sample, the belt is stopped and the aggre-
gate on a selected length of belt is removed and weighed. The process is
repeated at various belt speeds until several rates of feed have been mea-
sured. The amount of material supplied by the feeder in tons/hour is then
determined by the following formula:
R= 1.8 WS
B
where,
= rate otfeea, tons1hour;
W = weight of sample from B feet of belt, lbs;
B = length of belt sampled, feet; and
S = speed of belt, feet/second.
The speed of the belt on which aggregate samples are taken can be easily
determined, by measuring the belt length and determining the time it takes
to make one revolution.
After the rate of feed is determined for several feeder belt speeds, the
data is plotted on a graph similar to that shown in Figure 6-34 as X, Y, Z
or Q. After calibration of the feeders, the belt speed for any aggregate
being used can be selected to feed any desired amount.
Some HMA facilities (usually drum mix) have weight sensors on the
belt between the cold feeders and the dryer. These weight sensors allow the
operator to calibrate the cold feeders from the control room. In this case,
belt samples do not have to be taken so the entire process is faster. Care
must be taken, however, to ensure that the weight sensors are correctly cal-
ibrated. This can be verified by feeding the- aggregate into a truck for a
given amount of time and determining the weight of the material collected.
Feeder Number I
"-
::,
0
-
I
( /)
C:
0 Feeder Number 2
I-
"O
..
(I)
(I)
-
LL
0 Feeder Number 3
-(I)
0
c:::
Feeder Number 4
Belt Speed, ft/minute

Figure 6-34. Calibration Plot for an Aggregate Cold Feed at Various
Belt Speeds
Example 6-3. An aggregate feeder at a HMA facility is being cali-

brated. The feeder gate is set to an opening of 6 inches. The rate of flow
of material is controlled by a feeder belt underneath the aggregate feeder.
At a setting of 20 percent of maximum belt speed the aggregate is fed onto
the conveyor belt. A sample of aggregate on a five foot section of this con-
veyor belt was obtained and the material weighed. The weight of materi-
al was 35 pounds (15.9 kg). The feeder belt speed was then increased to
40 percent of maximum and the five foot section of aggregate weighed 80
pounds (36.4 kg). The belt speed was increased to 75 percent of maximum
and the weight of material removed from the five foot section was 150
pounds (68.2 kg). The conveyor belt was 100 feet (30.5 m) long and it
took 20 seconds to make one revolution. Determine the tons/hour of
aggregate fed to the plant for each belt speed setting.
Remember that
R= 1.8 WS
B
The rate of feed for the 20% belt speed is,
R = 1.8 (35) (100/20)

5
R = 63 tons/hour
The rate of feed for the 40% belt speed is,
R = 1.8 (80) (100/20)

5
R = 144 tons/hour
The rate of feed for the 75% belt setting is,
R = 1.8 (150) (100/20)

5
R = 270 tons/hour
These three measured tonnage rates should be plotted along with the
corresponding belt speed. The relationship between belt speed and tons/
hour are approximately straight lines. After being plotted, these graphs can
be used to select the belt speed for any desired tons/hour of aggregate.
Asphalt Cement Sampling

Tests on asphalt cement are conducted during the mix design process
and are not performed daily as part of quality control for HMA construc-
tion. Samples may be needed from time to time in order to determine var-
ious properties of the asphalt cement. A sample is usually collected in
one-gallon containers. These samples may be obtained from the asphalt
cement delivery truck or from the asphalt cement storage tanks. On many
occasions, samples are taken and saved to be tested later if early perfor-
mance problems are observed in the HMA mixture.
Sampling HMA
Samples of HMA are normally taken either from the loaded truck or
the roadway behind the paver. Because some segregation of the asphalt
mixture occurs when the truck is loaded, care must be exercised when col-
lecting samples. The coarse aggregate coated with asphalt cement tends to
• roll down the side of the mound of HMA mixture and congregate next to
the sides and the ends of the truck bed. Though this problem always
occurs to some extent, as long as it is minimized, problems during laydown
are not noticeable. This type of segregation affects test results if the sam-
ple is not properly taken. A segregated sample containing a high percent-
age of coarse aggregate has a low percentage of asphalt cement due to the
low surface area. Normally several small samples are taken at various
points along a truck load of HMA. These samples are usually taken just
beneath the surface at the top and on the sides of the truck load of HMA.
These small samples are combined to produced one large sample used for
testing.
When HMA is sampled on the roadway, care must be taken to ensure
that contamination with tack coat material is minimized. The roadway
sample can be representative (meaning several small samples) or random
(sample from one location) depending on what is being evaluated by the
test result. In most cases, random samples are required so that variability
of the construction process can be evaluated.
STATISTICAL .CONCEPTS
Variations in test results on paving materials and mixtures depend on
variability of materials, equipment, and testing procedures. Well con-
trolled projects provide mixture properties with low variability while poor-
ly controlled projects result in higher variability in mixture properties.
Understanding the variability of materials requires some knowledge of sta-

tistics.
Most properties of asphalt mixtures, when plotted as a histogram, pro-
duce a graph similar to the one shown in Figure 6-35. A curve drawn
through the upper points of this frequency diagram has the typical bell
shape of the normal distribution curve. Asphalt mixture properties that
have been shown to be normally distributed include aggregate gradation,
asphalt content, density, voids in total mix, stability, flow, and most other
properties. Knowing that these mixture properties are normally distrib-
uted allows a small number of tests to be used to predict the population
distribution. Some terms that should be understood when evaluating nor-
mal distributions are:
•1
E (X; - 'i)2
Variance= s2 = -'=-1 - - - -
•1
E <xi - x)2
i=1
Standard Deviation = a = {s2 =
Coefficient of Variation = CV= ~ x 100

X
Standard Error of Mean -

= a = -
a
.;n;
where,
result for the ith sample;
x 1 = test
x = average of measured test results;
n1 = number of test results;
n,_= number of samples that are averaged to y1eld-a~single test
result.
Example ~ : Ten in-place density test determinations were made on

one lot of material. The in-place density is expressed as a percent of the
>..
u
C:
Q)
:::,
0-
Q)
:....
LL
Asphalt Content
Figure 6-35. Histogram of Distribution of Asphalt Content
theoretical maximum density. Determine the estimated variance (s 2), stan-

dard deviation (cr), and coefficient of variation (CV) for the population
from which the samples were taken.
Test
No. Density ~d~ ~i::1U2
1 93.0 0.0 0.00
2 92.1 -0.9 0.81
3 94.2 1.2 1.44
4 93.3 0.3 0.09
5 91.5 -1.5 2.25
6 95.0 2.0 4.00
7 93.5 0.5 0.25
8 92.2 -0.8 0.64
9 91.6 -1.4 1.96
10 ~ 0.2 0.04
x = 93.0 I:= 11.48
Many calculators automatically make the required statistical calculations so

there is seldom a need to perform the tabulation shown above. However,
the calculations are shown here to illustrate the required steps in deter-
mining the statistical properties.
S2 = l: (X; - i)2 = 11.48 = 1.28
n1 -1 10-1
the estimated standard deviation is,
a = .[s2 = 1.13
and the estimated coefficient of variation is,
a 1.13
CV = -= x 100 = - - x 100 = 1.21 percent
X 93.0
Now, what if the specification requires that all density test results exceed 91
percent ofTMD? Using the estimated average density of the lot as 93.0
and the estimated standard deviation as 1.13, what is the probability that
a test result from a random sample is outside the specification require-
ments?
To analyze the problem, the data is reduced to a standard normal devi-
ate in order to use the standard probability distribution tables for a normal
distribution:
where,
Ka = standard normal deviate;
y = random variable;
x = estimate of the population average; and
cr = estimate of the population standard deviation.
So for the problem being evaluated, x = 93.0 and cr = 1.13. For this prob-
lem the lower specification limit of 91 is the random variable y. Then sub-
stituting x and cr into the equation for Ka,
Ka= (y- x)/cr = (91-93)/1.13 = -2/1.13 = -1.77
Ka and x are identified for the standard distribution shown in Figure 6-36.
From the statistical Table 6-5 it can be seen that for Ka= I.77 the nor-
mal distribution function has .9616 (96.2 percent) of the area below x +
1.77 and 3.8 percent of the area above x + 1.77 since the total area under
the normal distribution function is 1.00. Due to symmetry then x - 1.77
p.as 3.8 percent of the data below that point. Given the estimated average
and standard deviation of the population, it would be expected that
approximately 4 percent of the individual test results would fall below 91
percent or fall outside the lower specification limit.
x= 93 percent of TMD
Ka = -1.77 Ka = 1.77
91 percent of TMD 95 percent of TMD
Figure 6-36. Probability Distribution Function for Density

Now, if a given number of the individual sample values are averaged to
determine the density, the variability is reduced, the overall average
remains the same, and the percent below 91 percent is decreased. The
standard error of the mean is defined as:
CT=cr/Tu
cr = standard error of the mean;
cr = population standard deviation; and
n = number of individual samples which are included in
each average.
Assume for instance that four test results are averaged to produce one
reported mean test result. Then
cr = cr /Tu= ·1.13 / "'1/4 = 0.565

Table 6-5. Probability Distribution Function for a Normal

Distribution
Ka
Ka .00 .01 .02 .03 .04 .05 .06 .07 .08 .09
o.o .5000 .4960 .4920 .4880 .4840 .4801 A761 .4721 .4681 A641
0.1 .4602 .4562 .4522 .4483 .4443 .4404 .4364 .4325 .4286 A247
0.2 .4207 .4168 .4129 .4090 .4052 .4013 .3974 .3936 .3897 .3859
0.3 .3821 .3783 .3745 .3707 .3669 .3632 .3594 .3557 .3520. .3483
0.4 .3446 .3409 .3372 .3336 .3300 .3264 .3228 .3192 .3156 .3121
0.5 .3085 .3050 .3015 .2981 .2946 .2912 ..2an .2843 .2810 .2776
0.6 .2743 .2709 .2676 .2643 .2611 .2578 .2546 .2514 .2483 .2451
0.7 .2420 .2389 .2358 .2327 .2296 .2266 .2236 .2266 .2177 .2148
0.8 .2119 ;2090 .2061 .2033 .2005 .19n .1949 .1922 .1894 .1867
0.9 .1841 .1814 .1788 .1762 .1736 .1711 .1685 .1660 .1635 .1611
1.0 .1587 .1562 .1539 .1515 .1492 .1469 .1446 .1423 .1401 .1379
1.1 .1357 .1335 .1314 .1292 .1271 .1251 .1230 .1210 .1190 .1170
1.2 .1151 .1131 .1112 .1093 .1075 .1056 .1038 .1020 .1003 .0985.
1.3 .0968 .0951 .0934 .0918 .0901 .0885 .0869 .0853 .0838 .0823
1.4 .0808 .0793 .ona .0764 .0749 .0735 .0721 .0708 .0694 ~1
1.5 .0668 .0655 .0643 .0630 .0618 .0606 .0594 .0582 .0571 .0559
1.6 .0548 .0537 .0526 .0516 .0505 .0495 .0485 .0475 .0465 ~
1.7 .0446 .0436 .0427 .0418 .0409 .0401 .0392 ~ .0375 ~7
1.8 .0359 .0351 .0344 .0336 .0329 .0322 ~14 .0307 .0301 .0294
1.9 .0287 .0281 .0274 .0268 .0262 .0256 .0250 .0244 m39 .0233
2.0 .0228 .0222 .0217 .0212 .0207 .0202 .0197 .0192

.0188 .0183
2.1 .0179 .0174 .0170 .0166 .0162 .0158 .0154 .0150
.0146 .0143
2.2 .0139 .0136 .0132 .0129 .0125 .0122 .0119 .0116
.0113 .0110
2.3 .0107 .0104 .0102 .00990 .00964 .00939 .00914 ~ .00866 .00842
2A .00820 .00798 .00776 ,00755 .00734 .00714 .00695 ~76 .00657 .00639
2.5 .00621 .00604 .00587 .00570 ~ .00539 .00523 ~ .00494 ~

2.6 .00466 .00453 .00440 .00427 .00415 .00402 .00391 .00379 .00368 .00357
2.7 .00347 .00336 .00326 .00317 ~7 .00298 .00289 .00280 .00272 m264
2.8 .00256 .00248 .00240 .00233 .00226 .00219 .00212 .00205 .00199 .00193
2.9 .00187 .00181 .00175 .00169 .00164 .00159 .00154 .00149 .00144 .00139
Ka .o .1 .2 .3 .4 .5 .6 .7 .8 .9
3 .00135 .03968 .03483

.03687 .03337 .0•233 .03159 .0'108 .04723 .04481
4 .04317 .04207 .04133
.05854 .05541 .05340 .05211 .05130 .0'793 .0"479
5 .0"287 .0"170 .07996
.07579 .0'333 .07 190 .07 107 .0"599 .0"332 .0"182
s- -:0"98T -:0•530- --:09282- c-;0't49- -.o•"777- -;010402- -,0'°206- -.011104- .011 523 .011 260
· ♦ From Toblu tfArtlM ilt Two Toil.r/J/ld ilt OM Toil tftM Nonnol.C,,,.,., by Ftoderict B. CroxlOD. O>pyrighl 1949 by Prmlice-Hall,
Inc.
and the standard normal deviate, Ka, then becomes
Ka= (y- x)/cr = (91 - 93)/0.565 = ---:3.52
From the statistical table it can be seen that for Ka= 3.52 the area of
the curve below x + Ka is 0.9998, this means that 99.98% of the data falls
below x + 3.52 and only 0.02% is above x + 3.52 or below x - 3.52. By
averaging 4 samples for each data point, the percent of material expected
to be .below the specification requirements is reduced from approximately
3.8 percent to approximately 0.02 percent.
Past experience often provides a good indication of the expected value of
cr for various mixture properties for a typical project. However, cr is unique
for each project and cannot be accurately assumed for any project. After suf-
ficient data is generated for a project, cr can be computed, the variability of
test results can be determined, and a judgement can be made as to whether
good control occurs (cr is low) or poor control occurs (cr is high).
A large amount of data has been collected to evaluate the typical vari-
ability of pavement properties (26). Typical values for asphalt content and
density are shown below:
Property Standard Deviation

Asphalt content, percent by weight 0.20
Density of HMA, percent of laboratory density 1.02
Other work has shown that the standard deviation for asphalt content can
be as high as 0.30.
Standard deviation values for aggregate gradation have been shown to
be related to percent passing a particular sieve size. Typical results are
shown in Figure 6-37. Now consider a practical problem encountered on
many paving projects. HMA is being produced and tested by both the
owner and the contractor. The asphalt content is being measured using the
solvent extraction test. .Because of high aggregate absorption, some of the
asphalt cement is trapped in the pores of the aggregates and cannot be dis-
solved by the solvent, therefore it appears that the asphalt content is low.
The contractor's tests show more asphalt cement removed by the solvent
than does the owner's tests; hence, t:F-o_p_o_p_ulations....oLdata...exisL(.Eigun,___ _ _ __
6-38). Neither the owner's nor the contractor's data represents the actual
asphalt content since both sets of tests leave some amount of asphalt
cement in the aggregate. The bias in the test results is defined as the dif-
ference between the true value of the property and the actual test result
5
z
0
i= 4
<(
>
W3
---r--..--
0
~- -- -7-
EXTRACTION-
_--:;:
~2
<(
0
z1
<(
~~
.,,,-r- HOT BINS-
-
r- ...._
....
I-
C/) 0 1/
0 10 20 30 40 50 60 70 80 90 100
PERCENT AGGREGATE PASSING
Figure 6-37. Typical Standard Deviation Values for Aggregate
Gradation (26)
>,,
(..)
C:
Q)
::,
er
Q)
:,...
LL
% Asphalt
Figure 6-38. Comparison of Owner's Test Results, Contractors Test
Results, and True Value
observed. The bias from the owner's tests is different from the bias from
the contractor's tests. The primary objective in testing is to measure the
true property by reducing the bias to approximately zero and to minimize
the variability associated with the testing process. If both the owner and
contractor average a number of samples for each data point, the variability
of the data points is reduced; however, the bias is not affected. The bias
can only be reduced by removing more of the asphalt cement from the
aggregate during· the extraction test.
Example 6-5. Specifications must have a realistic tolerance as dis-

cussed earlier. Suppose that a specification for HMA requires that the den-
sity be between 93.0 and 96.0 percent o(tpeoretical maximum density. If
the standard deviation for percent density is 1.0, is this a realistic tolerance?
If the contractor builds the HMA pavement to an average density of 94.5
percent, how much of the test data can be expected to be outside the spec-
ification requirements?
Since Ka= (96.0 - 94.5)/1.0 = 1.5, it can be seen from Table 6-5 that
93.3 percent of the data lies below 96.0 percent ofTMD and that 6.7 per-
cent of the data lies above 96.0 percent ofTMD. If the contractor aver-
ages 94.5 (midrange) approximately 13.4 percent of the tests (6.7 percent
at each end of the curve) will be outside the specification range. This is
probably not a realistic specification if individual test values are to be used
to determine compliance..
What happens if each reported result is averaged for each set of four
tests? The standard error of the mean of four test results is calculated as
shown below: ·
Now if the average is 94.5, Ka becomes (96 - 94.5)/0.5 = 3.0. Using

Table 6-5 for a Ka of 3.0 shows that 94.5 + Ka would have 99.87 percent
of the data below that value, while 94.5 ± Ka would contain 1.00 -
2(.0013) = 99.74 percent of the data. Hence, only 2 tests out of 1000
would be expected to be outside the specifications if the average is 94.5 and
the standard deviation is 1.0. This specification is much more realistic,,__,i=n,____ __
terms of being achievable by the contractor, when the standard error of the
mean is used. This problem shows that the tolerance for a given mixture
property can be reduced when the standard error of the mean is used
instead of the standard deviation.
QUALITY CONTROL/QUALITY ASSURANCE (QC/QA)
Good QC/QA practices are essential to obtain a satisfactory product.

QC normally refers to those tests necessary to control a product and to
determine the quality of the product being produced. These QC tests are
usually performed by the. contractor. QA refers to those tests necessary to
make a decision on acceptance of a project and hence to ensure that the
product being evaluated is indeed what the owner specified. These QA
tests are normally performed by the owner.
Testing of the asphalt mixture during production is essential to ensure
that a satisfactory product is obtained. The tests that should be performed
during manufacture and placement of asphalt mixtures may include aggre-
gate gradation, asphalt content, temperature, mixture properties of labora-
tory samples, theoretical maximum density, in-place density, smoothness,
elevations (if applicable), and visual inspections.
Aggregate Gradation. There are several aggregate properties that are

important but routine testing during construction is usually limited to gra-
dation only. Visual inspections must be made to ensure the aggregate
being produced is similar to the aggregate tested during development of
the mix design. For QC/Qj\. testing, aggregate samples are typically taken
from the stockpile, cold feeder belt, hot bins (if applicable), and extracted
asphalt mixture. The gradation of the aggregate from the asphalt mixture
is of most importance since this is the end product; however, the aggregate
gradation must be controlled at the other points to ensure that the grada-
tion of the final product is satisfactory.
The stockpile should be sampled and tested during_ the mix design
process and approved for use. Once production begins, it is only necessary
to sample new material that is added to the stockpile since the overall
stockpile gradation has already been determined. The new material added
to the stockpile must have the same gradation as the original stockpile,
within reasonable tolerances, otherwise the gradation of the final mixture
is affected. Causes in gradation variations at the stockpile include changes
at source, segregation during hauling or stockpiling, and sampling and
testing errors.
The second typical location for taking aggregate samples is the cold
----~e.e.der..helt.__Ihis_b_elt_c.0.11.tains the combined agg~gate being.fed into the
HMA facility. Variability of gradation that results at this point is caused
by variations in stockpile gradations, segregation of aggregate, improper
loading of cold feed bins, improper setting of individual cold feed bins,
and sampling and testing errors.
The third location for sampling aggregate is in the HMA batch facili-
ty hot bins, (drum mix facilities do not have aggregate hot bins). Causes
for variability here include improper gradation fed from cold feeder, errat-
ic feed from dust collector system, changing production rate (screening
efficiency changes with production rate), blinding screens, holes in screens.
or bin walls, and sampling and testing errors. The hot bins, if operated
correctly, will partially correct for gradation fluctuations coming into the
plant.
The fourth location for determining gradation is from the asphalt mix-
ture. The sample is normally taken from loaded· trucks but can be taken
' of the asphalt paver. This test, which is performed on the
from the back
finished product, must be controlled because it is the one on which accep-
tance of the mixture is normally based. Variability of gradation at this
point (for a batch plant) could involve incorrect hot bin gradations, incor-
rect percentage of material from each hot bin, segregation of mixture (in
storage silo or in truck), and sampling and testing errors. For a drum mix
facility, the causes of variability at this point include improper cold feed
gradation, erratic feed from the dust collector, segregation of aggregate
travelling through the plant or in the storage silo, and sampling and test-
ing errors.
Evaluation of the gradation at several locations allows the engineer to
troubleshoot the gradation problem and quickly identify the location
where it is occurring. For instance, if the stockpile gradation is satisfacto-
ry but the cold feed gradation changes, then the problem areas are likely to
be segregation of mixture, improper loading of cold feed bins, or sampling
and testing errors. These items can be quickly checked and modifications
made to correct the problem.
Asphalt Content. Another mixture property that must be evaluated is

asphalt content. The asphalt content of a mixture is very important to
ensure satisfactory performance. A mixture with low asphalt content is not
durable, and one witp. high asphalt content is not stable. The actual
asphalt content directly affects mixture properties, such as asphalt film
thickness, voids, stability (Hveem or Marshall), and Marshall flow.
Therefore, it is important to monitor asphalt content, but it is really these
mixture properties that need to be controlled.
The asphalt content of a mixture is measured by an extraction test
(ASTM D2172) or with a nuclear gauge (ASTM D4125). The nuclear
gauge is used to measure asphalt content much the same way as it is used
to measure moisture in soils. The extraction test involves adding a solvent
to the asphalt mixture to dissolve the asphalt cement. The asphalt cement
and solvent are then passed through a piece of filter paper, but the aggre-
gate is not allowed to pass. This is not a highly accurate test but it is wide-
ly used for measuring asphalt content. One advantage of the extraction
test is that it allows determination of the aggregate gradation of the mix-
ture. A disadvantage of the extraction test is that the solvent used is haz-
ardous and is difficult to dispose. Biodegradable solvents are being used by
some states but additional work is needed prior to their widespread accep-
tance. The current asphalt cement metering devices, such as variable speed
pumps, give accurate control of asphalt content. These devices should be
calibrated by the HMA producer on a regular basis.
The National Center for Asphalt Technology (NCAT) has developed
a test method to determine the asphalt content of the HMA mixtures by
ignition. The test method is based on research started in 1990 at NCAT
(22). In the NCAT ignition method, a sample of HMA mixture is sub-
jected to an elevated temperature of 538°C (1000°F) in a furnace to ignite
and burn i:he asphalt content from the aggregate. NCAT's work has result-
ed in a test procedure and equipment that automatically measures the
asphalt content in 30-40 minutes. The grading of the aggregate can then
be determined using standard sieve analysis. Based on round robin stud-
ies conducted by NCAT in which 12 laboratories participated, the accura-
cy and precision of the NCAT ignition test was found to be better than
those of the solvent extraction method. Therefore, this test method is
increasingly replacing solvent extraction methods, which are being elimi-
nated due to growing health and environmental concerns associated with
the use of chlorinated solvents.
Improper asphalt content can be caused by several factors in a batch
HMA facility. These causes include inaccurate aggregate scales, inaccurate
asphalt cement scales, leaking valve in asphalt cement pot, segregation, and
sampling and testing errors. Causes of incorrect asphalt content in a drum
mix facility include inaccurate aggregate belt scales, improperly calibrated
asphalt cement meter, incorrect moisture content correction for aggregates,
segregation, and sampling and testing errors. Properly sampling the
asphalt mixture from a truck load of material is difficult. When a sample
------of'mixture is obtained wiililiigh coarse aggregate content aue to segrega-
tion after mixing, the measured asphalt content will be lower than speci-
fied since the coarse aggregate has a lower surface area than the fine aggre-
gate.
Field Management of Volumetric Properties

The Federal Highway Administration (FHWA) Demonstration
Project No. 74 has clearly shown that significant differences exist between
the volumetric properties of the laboratory designed and plant produced
HMA mixtures. The volumetric properties include voids in the mineral
aggregate (VMA) and the voids in the total mix (VTM). The FHWA pro-
ject concluded that a field mix verification of the material produced at the
HMA facility should be included as a second phase in the mix design
process. Mix verification is defined as the validation of a mix design with-
in the first several hundred tons of HMA production. Field management
of HMA provides a viable tool to identify the differences between plant
produced and laboratory designed HMA ~ixes and effectively reconcile
these differences (2Jl). The National Center for Asphalt Technology
(NCAT) conducted a statistical analysis of field data from 24 FHWA
demonstration projects to develop practical guidelines for the HMA con-
tractors to reconcile these differences, thereby assisting them to consistent-
ly produce high quality HMA mixes (22). The field data was analyzed first
to identify and, if possible, quantify the independent variables (such as
asphalt content and the percentages of material passing No. 200
(0.075 mm) and other sieves) significantly affecting the dependent vari-
ables such as VMA and VTM. The following conclusions were drawn
from the NCAT study:
1. Significant differences exist .between the volumetric properties of

the laboratory designed and plant produced hot mix asphalt.
2. VMA is most affected by the amount of material passing No. 200
(0.075 mm) sieve (P200) and the relative proportions of coarse and
fine aggregates.
3. VMA can be increased by reducing the amount of P200
(0.075 mm) material or natural sand in the HMA mixes. VMA can
also be increased by moving the aggregate gradation away from the
maximum density line (MDL) especially for HMA mixes with no
natural sand.
4.VT.M is most affected by asphalt content, P200 (0.075 mm) mate-
rial, and the relative proportions of coarse and fine aggregates.
5. VTM can be increased by reducing asphalt content and P200
(0.075 mm) material.
Flow charts were developed as general guidelines for reconciling the

VMA and VTM difference benveen the laboratory designed and plant pro-
duced HMA mixes.
Construction. The temperature of the mixture during production

should be closely monitored. The temperature must be just high enough
to provide good coating on the aggregate and to allow for satisfactory com-
paction. When the mixture is heated more than needed, additional oxida-
tion and loss of volatiles occur to the asphalt cement reducing mixture
durability. Compaction may also have to be delayed to allow the mix to
cool to the proper compaction temperature.
The TMD (theoretical maximum density) should be measured on a
daily basis, since this is needed to calculate voids in the mixture and is used
in some cases to specify density. A change in TMD indicates a change in
asphalt content, aggregate gradation, specific gravity or absorption of the
aggregate, or sampling and testing errors.
Compacting the asphalt mixture to a satisfactory in-place density is
required for satisfactory performance. Two methods that are used for check-
ing in-place density are nuclear gauges and cores. The nuclear gauge is gen-
erally preferred for QC testing while cores are desirable for QA testing.
When the nuclear gauge is used for QA testing, some cores must be taken
routinely for verifying the accuracy of the nuclear gauge. Low density can be
caused by inaccurate reference density, improper gradation or asphalt con-
tent, a mixture temperature which is too high or too low, a layer thickness
which is too thin (cools quickly and bridges under rollers) or too thick,
improper rolling techniques, inadequate rollers, sampling and testing errors.
During construction, smoothness is checked with a straightedge or
some other device to ensure a satisfactory ride. Many items cause rough-
ness in a completed asphalt mixture. Some of these items include: under-
lying material that is too rough, excessive stops with pavers or rollers dur-
ing placement, too many screed adjustments, variation in mixture temper-
ature, hand work, and segregation of materials.
Some projects, especially on large areas such as parking lots and airfields,
require the finished pavement to be built to some designed elevation. This
is necessary to ensure that bird baths do not exist and to ensure smoothness.
This is normally controlled with stringlines during construction and verified
after construction by measuring elevations across fne surface. ·
Visual inspections are always necessary during QC/QA evaluations of
asphalt mixtures. Test results can never identify all the problems that may
occur during HMA construction. Visual inspection can identify potential
trouble areas and these problems can be corrected, in some cases before a
mixture deficiency occurs. In other cases, samples may be obtained from

mixtures that visually appear to be deficient and evaluated to determine if
the mixture meets specification requirements.
Testing and Quality Control Charts

Testing Frequency. It is impossible to establish one desired testing
frequency for all projects. Factors that may affect testing frequency
include the size of project, importance of project, and variability of mate-
rials. However, the following suggested number of tests is provided as a
rough guide to the actual minimum number that should be conducted on
a project.
Test Frequency
Stockpile gradation 1 per day
Cold feed gradation 1 per day
Hot bin gradation (if applicable) 1 per day
Extracted asphalt content and gradation 2 per day
Laboratory compacted samples
voids, stability, flow 2 sets per day
Theoretical maximum density 2 per day
Temperature Regularly throughout day
Visual inspection Regularly throughout day
In-place density 6-10 per day
Smoothness Regularly throughout day
Elevations After completion of lot
Technician Qualificatioits. Sampling and testing errors can be a high

percentage of the overall variations in test results. For this reason it is
essential that testing technicians be highly qualified to ensure that sam-
pling and testing errors (for some tests this is over 50 percent of variabili-
ty) are minimized. Many states have begun to require certification of tech-
nicians (state and contractor) that work on state or federally funded pro-
jects. More emphasis needs to be placed on this technician certification
program to ensure that all projects have qualified personnel performing
--QC-and-QA-tests.------------------------
Quality Control Charts. Analysis and evaluation of test results must

be performed during the progress of work for adequate control of the pro-
ject. The best way to monitor the quality of work during construction is
with control charts. Control charts are simple methods of graphically dis-
playing the QC data as it is developed. The two most common types of
control charts are plots of individual test results and plots of running aver-
ages of the test results.
Example 6-5. Assume that the results in Table 6-6 are obtained for
the density of a HMA mixture. The specifications indicate that the densi-
ty must be at least 92 percent and no more than 96 percent of theoretical
maximum density for the running average of four samples. Plot control
charts showing the data and specification limits as shown in Figure 6-39.
(I) Upper Specification Limit

C'I
~
(I)
961-------------~-----------
~ 5 95
0
g'LL.94
·2~ 000°
c: en 0 0 0 0
::, 0 93
a:: ...J 0
00
~o 921----------------.::-------------
.iii
C: 91 Lower Specification Limit
(I)
Cl
I 2 3 4 5 6 7 8 9 10 II 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Test No.
Figure 6-39. Density Control Charts Showing Data and Specification
Limits
These control charts show no obvious trends nor data outside the spec-
ification limits on the running average of four plots. For proper evaluation
of the process being used to produce the mixture, the engineer must look
for trends moving upward or downward and for test results outside the
specification limits. Identification of these trends during construction
allows the problem to be solved before it becomes significant. Figure 6-39
does show some slight trends which are likely due to normal material vari-
ation. Test numbers 10 and 19 show that the running average is approach-
ing the lower limit. When this happens the contractor should probably
-----t·ak.e some action to increase: dre density.
It is helpful to note on the control charts any significant changes that

are made in the process. This note should indicate what change was made
and when it was made. Identifying the change and the test number at
Table 6-6. Density Data for a HMA Project
Density,
Density, Percent of 4-point
Test No. Theoretical Maximum Running
Density Average
1 93.0
2 94.5
3 92.1
4 95.1 93.J
5 93.3 93.8
~-
6 94.2 93.7
7 93.5 94.0
8 91.6 93.2
9 92.5 93.0
10 93.7 92.8
11 94.2 93.0
12 92.2 93.2
13 91.5 92.9
14 95.6 93.4
15 93.6 93.2
16 94.1 93.7
17 92.3 93.9
18 91.7 92.8
19 92.1 92.4
20 94.3 92.5
21 96.0 93.4
22 92.2 93.6
"
23, 91.3 93.4
24 93.5 93.2
~
...
0
0:::
...
Q)
">C
Q)
-( /)
C
(l)
I
I
Test No.
Figure 6-40. Effect of Process Changes on Density Test Results
which that change occurred allows the engineer to determine if the change
affects process control as shown on Figure 6-40. The increase in roller
weight appeared to increase the density of the mixture. Changing roller
operators did not appear to affect the average density, but the variability
between test results appears to be higher for the new operator.
Another plot sometimes used to analyze variability of test results is the
range plot shown in Figure 6-41. As the range values increase or decrease,
the variability changes in a similar way.
-
( /)
C
--
..J
Q)
O'\ Cl)
(/)
C :::,
Cc
0::>
~¢
(/)
C
Cl)
-------,O--t:======::::::::::::::::::::::::::::::::::::::::::::::::::::::;;;;;;;;:;;;;;;;;:~~-
T est No.
Figure 6-41. Density Range Plot Sh~wing Range of Last Four Data
Points
The properties that are typically plotted on control charts for HMA
construction are: asphalt content, aggregate gradation (for selected sieve
sizes such as 3/s-inch, No. 8, No. 50, and No. 200), in-place density, voids
in laboratory compacted samples, theoretical maximum density, laborato-
ry density, stability, and Bow Other properties not mentioned above may
be plotted as needed.
REFERENCES
1. Brock, J .D. Dryer Drum Mixer. Astec Industries, Chattanooga, TN,
undated.
2. Scherocman, J.A., T.W _Kennedy, M. Tahmorelsi, R. Holmgreen.
Construction of Asphalt Concrete,-- Pavement. Center for
Transportation Research, The University of Texas at Austin,
September 1986.
3. Brock, J.D. and E.G. Mize. The Drum Mix Process. Astec Industries,
Technical Bulletin NumberT-110, Chattanooga, TN, 1984.
4. Fugitive Dust Control and the Hot Mix Plant. National Asphalt
Pavement Association, NAPA Report IS-73, 1980.
5. Linden, et al. Effect of Compaction on Asphalt Concrete Perfor-
mance. Transportation Research Board, Transportation Research
Record 1217, 1989.
6. Nijboer, L.W Plastici-ty as a Factor in the Design ofDense Bituminous
Road Carpets. Elsevier Publishing Co., Inc., New York, N.Y., 1948.
7. Principles of Construction of Hot Mix Asphalt Pavements. The
Asphalt Institute, Manual Series No. 22 (MS-22), January 1983.
8. Compaction Handbook. 6th Edition, Hyster Company,
Construction Equipment Division, Kewanee, IL, October 1986.
9. Tunnicliff, D.G., R. W Beatty, and E.H. Holt. A History of Plants,
Equipment and Methods in Bituminous Paving. Proceedings,
Association of Asphalt Paving Technologists, Volume 43A, 1974.
10. Geller, Myron. Compaction Equipment for Asphalt Mixes. Placement
and Compaction ofAsphalt Mixtures, American Society for Testing and
1
1
Materials, Special Technical Publication 829, ET. Wagner, Ed., 1984.
:1,li, 11. McLaughlin, A.L. Field Compaction of Bituminous Mixes for Airport
Pav~ments. Federal Aviation Administration, Report No .
..i,1.----..L.o.=IOUJ.=7-7-=42.,..ApriLL9-7-t-.------------------
1
'
12. Brown, E.R. Experiences of Corps of Engineers in Compaction of

Hot Asphalt Mixtures. Placement and Compaction ofAsphalt Mixtures,
American Society for Testing and Materials, Special Technical
Publication X29, FT Wagner, Ed., 1984.
13. State of the Art: Vibratory Compaction of Asphalt Pavements.

Transportation Research Board, Transportation Research Circular
Number 242, April 1982.
14. Quality Control for Hot Mix Asphalt Manufacturing Facilities and
Paving Operations. National Asphalt Pavement Association, Quality
Improvement Series 97/87, 1987.
15. Hot Mix Asphalt Paving Inspection-Training Manual. Texas State
Department of Highways and Public Transportation, 1983.
16. Ford, M.C. Pavement Densification Related to Asphalt Mix
Characteristics. Paper presented at the 1988 Annual Meeting of the
Transportation Research Board, Washington, DC, 1988.
17. Brown, E.R. and S. Cross. A Study of In-Place Rutting of Asphalt
Pavements. Proceedings, Association of Asphalt Paving Technologists,
Volume 58, 1989.
18. Huber, G.A. and G.H. Heiman. Effect of Asphalt Concrete
Parameters on Rutting Performance: A Field Investigation.
Proceedings, Association of Asphalt Paving Technologists, Volume 56,
1987.
19. Zube, E. Compaction Studies of Asphalt Concrete Pavement as
Related to the Water Permeability Test. Paper presented at 41st
Annual Meeting of the Highway Research Board, Washington, DC,
1962.
20. Brown, E.R., R Collins, and J.R. Brownfield. Investigation of
Segregation of Asphalt Mixtures in State of Georgia. Paper presented
at 68th Annual Meeting of the Transportation Research Board,
Washington, DC, 1989.
21. Santucci, L.E., D.D. Allen, and R.L. Coats. The Effects of Moisture
and Compaction on the Quality of Asphalt Pavements. Proceedings,
22. White, T.D. Marshall Procedures for Design and Quality Control of
Asphalt Mixtures Proceedings, Association of Asphalt Paving
23. Brown, E.R. Density of Asphalt Concrete-How Much is Needed?
Paper presented at 69th Annual Meeting of the Transportation
Research Board, Washington, DC, 1990.
24. Burati, J.L., Jr. and G.B. Elzoghbi. Correlation of Nuclear Density
Results with Core Densities. Transportation Research Board,
25. Brock, J.D. Hot Mix. Asphalt Segregation: Causes and Cures.
National Asphalt Pavement Association, Quality Improvement Series
110/86, 1986.
26. Brown, E.R. Statistical Quality Control Procedures for Airport

Pavement,Materials. Federal Aviation Administration, Report No.
FAA-RD-73-199, October 1985.
27. Brown, E.R., N.E. Murphy, and S. Mager. Historical Development
of Asphalt Content Determination by the Ignition Method. Asphalt
Paving Technology, Vol. 64, 1995.
28. D'Angelo, J.A. and T. Ferragut. Summary of Simulation Studies from
Demonstration Project No. 74: Field Management of Asphalt Mixes.
Asphalt Paving Technology, Vol. 60, 1991.
29. Kandhal, P.S., K.Y. Foo, and J.A. D'Angelo. Field Management of
Hot Mix" Asphalt Volumetric Properties. ASTM Special Technical
CHAPTER 7. SPECIAL MIXTURES,
RECYCLING, AND ADDITIVES
There are numerous special HMA mixtures for various applications.

However, the most commonly used special mixtures are discussed in this
chapter. Pavement recycling includes several types involving different pro-
cedures; however, only central plant hot asphalt recycling has been includ-
ed in this chapter. Additives and modifiers are also discussed.
OPEN-GRADED HMA MIXES

Although dense-graded HMA mixes are commonly used in the
United States, open-graded mixes have also been used for special applica-
tions. Two primary types of open-graded mixes that are discussed in this
chapter include those used for base courses and those used for surface
courses.
Open-graded Base Mixes

Depending on the application, these mixes have been called "Big Rock
Mixes" (J) or Asphalt Treated Permeable Material (ATPM) (2). Generally,
these mixes contain very little or no fine aggregate as shown in Table 7-1.
Coarse aggregate should be angular to provide good interlock between the
particles. Typically, asphalt content for these mixtures ranges from 1.5 to
2.5 percent. However, the asphalt film around the aggregate particles is
---~---.,.ignifica ody _rhicker_than._that_fo_Ld~ns_e=gr.ad~_d mixture~_Jherefore, the
asphalt cement oxidizes (hardens) at a much slower rate. Low mix tem-
perature is generally used to prevent runoff of the thick films during trans-
port to the paving site. A short mixing time is also required since the sur-
face area per pound of aggregate is small. These mixes are highly perme-
Table 7-1. Typical Gradation of Open-Graded Base Mixes (Percent

Passing)
Sieve Size Tennessee Gradation Pennsylvania Gradation
(after Re£ 1) (after Re£ 4)
2 inch (51 mm) ~ 100 ---

1¼ inch (38 mm) 75-90 100
/'
1 inch (25 mm) --- 90-100
3
/4 inch (19 mm) 50-70 ~- ---
¼ inch (12.5 mm) --- 25-60
No. 4 (4.75 mm) 8-20 0-5
No. 8 (2.36 mm) --- 0-2
No. 100 (0.15 mm) 0-5 ---
able and provide rapid drainage for subsurface water. However, to be effec-
tive, the layer must be either daylighted or connected to edge or tile drains.
Their cost is significantly lower than dense-graded HMA mixes primarily
because of the lower asphalt content. Two or three roller passes are enough
to seat the aggregate and compact the mix. Excessive rolling tends to break
the aggregate. "When this type of mix is placed between a concrete pave-
ment and a dense-graded HMA overlay, it minimizes reflection cracking.
The use of "big rock mixes" has been reported in Tennessee and
Indiana (1). These mixes have supported heavy truck traffic without dis-
tortion. The resistance to deformation comes from relatively large aggre-
gate size and good interlock of aggregate particles.
ATPM has been used by several states to improve the subsurface
drainage of both flexible and concrete pavements. Cedergren and Lovering
(3) recommended a two-layer drainage blanket for flexible pavement sys-
tems consisting of subbase and ATPM. In the case of a two-layer system,
the water being drained can pass upward through the underlying'-'p_e_r_m_e_-_ _ _ _ __
---a-,ble subbase layer mto the ATPM ana be quickly drained from any spot
underneath the surface of the roadway. However, in a single layer system
consisting of sub base material only, any water entering the sub base must be
carried all the way to the drainage outlet in that material. Thus, the effec-
tive drainage path for the two-layer ·system is the thickness of the subbase,
Chapter 7-Special Mixtures, Recycling, and Additives 413
and the drainage path for the one-layer system is half the width of the
roadway, assuming the water is removed from the roadway at the two
shoulders by daylighting or edge drains. The problem is compounded
when the roadway has multiple lanes and is very wide.
Permeability of ATPM generally exceeds 10,000 feet (3,000 m) per
day (.1). It has been used successfully to drain water from badly pumping
concrete pavements when used as a drainage layer between the concrete
pavement and dense-graded HMA overlay. However, the concrete pave-
ment needs to be cracked and seated prior to placing the ATPM if water is
to be drained from underneath the concrete.
Open-graded Friction Course (OGFC)

An open-graded friction course (OGFC) consists of a HMA mixture
that is designed to have a large number of voids so that water can drain
through and over the surface of this mixture. The large air void content is
created by using a larger percentage of coarse aggregate (usually less than
20 percent of the material passes a No. 8 sieve). Removal of water from
the pavement surface reduces the threat of hydroplaning and improves the
overall skid resistance. Halstead (i, Q) has authored two syntheses on
open-graded friction courses.
An OGFC mixture is produced at an asphalt mix facility in the same
way as dense-graded HMA. Again, the primary difference between the
two mix types is the gradatioll'of the aggregate. Rubberized asphalt is often
used in production of OGFC mixtures to improve the ability of the asphalt
binder to hold the aggregate in place. Even though the void content is
high, the asphalt film thickness is typically greater than that for
dense-graded HMA. The total asphalt content for an OGFC mixture is
about the same or slightly higher than that for dense-graded mixtures with
the same maximum size of aggregate.
The mix design procedure most often used for OGFC mixtures is to
select the asphalt content necessary to provide some specified film thickness.
One property that is usually evaluated in the mix design process is the drain-
ability of the asphalt binder through the aggregate matrix. Some states use the
mix design procedure developed by the Federal Highway Administration (2).
The FHWA recommended gradation for OGFC is as follows:

-----------+/-z-i-nch-E-H7"5-mm1----------l-OO--------
3/8 inch (9.5 mm) 95-100
No. 4 (4.75 mm) 30-50
No. 8 (2.36 mm) 5-15
No. 200 (0.075 mm) 2-5
Asphalt content is estimated from the surface capacity CK) -of the
aggregate retained on a No. 4 (4.75 mm) sieve. Kc is determined from the
percent of SAE No. IO oil retained by the aggregate which represents the
total effect of superficial area, the aggregate's absorptive properties, and sur-
face roughness. Void capacity of the coarse aggregate is determined by
using a vibratory compaction apparatus. The optimum amount of fine
aggregate is that which fills the voids between the coarse aggregate particles
without bulking the coarse aggregate. This aggregate mixture must provide
enough room for the required asphalt content and interconnected voids
needed for drainage. The FHWA procedure establishes optimum mixing
temperature by running asphalt drainage tests at different temperatures.
The use of rubberized asphalt as the binder results in a mixing tem-
perature that is generally higher than that for conventional asphalt mix-
tures. A mixing temperature that is too high allows the asphalt binder to
drain through the aggregate, resulting in a high variation of asphalt con-
tent throughout the mix and bleeding spots in the pavement. If the mix-
ing temperature is too low, the aggregate will not be completely coated.
The asphalt drainage problem limits the amount of storage time of the
OGFC in a storage silo. A long storage time results in excessive drainage
of asphalt through the mixture. When the OGFC is stored in a silo even
for short periods of time, it is essential that the mix be observed to ensure
that noticeable drainage does not occur.
The asphalt may tend to drain through the aggregate while being
hauled from the mixing facility to the laydown site. If this occurs, asphalt
may leak around the tailgate while the truck is waiting or may form a pud-
dle in the bottom of the truck and hence cause a rich spot on the pavement
underneath the truck when unloading. This asphalt on the surface of the
OGFC results in a slick spot.
When the OGFC arrives at the laydown site, it tends to look rich due
to the greater film thickness and the lack of fine materials. This rich look
is normal. Typically the OGFC is placed appro~imately 3/4 inch (19 mm)
thick using an aggregate of 1/2 inch (12.5 mm) maximum size.
An OGFC should not be used over an existing pavement surface that
is uneven. An existing pavement that is uneven prior to overlay with an
OGFC should be levelled with a dense-graded binder course prior to over-
laying with the OGFC. Another problem that occurs when overlaying
___uneven-pa.v.em~-nts-is-puU-i-ng-aflcl-t:ea:ri-ng-wherrdre-a-sph:alt paver places ffi'1:-"""e- - - - - -
OGFC thinner than approximately 3/4 inch (1,9 mm).
The OGFC should not be laid in cold weather. Several failures have
been attributed to loss of bond between the OGFC and an existing surface
because of cold-weather laydown. Most states specify a minimum air tern-
Chapter ?-Special Mixtures, Recycling, and Additives 415
perature in the range of 60 to 70°F (16 to 21 °C) when OGFC mixes are
placed.
There is no density specification for OGFC mixtures. The normal
procedure for rolling is to make 2-3 passes over the surface with a steel
wheel roller. Vibratory and rubber-tired rollers should not be used. The
vibratory roller degrades the aggregate and the rubber-tired roller tends to
pick up the asphalt, especially when a rubberized asphalt is used. Another
problem that occurs with the rubber-tired roller is excessive closing of the
voids due to the kneading action of the rubber tires. Low voids, if present,
do not allow good drainage of water .across and through the surface.
The quality control procedures for an OGFC pavement are essential-
ly the same as that for dense-graded HMA. The primary difference is that
there are no requirements for laboratory or field compaction of the mix-
ture. The required tests generally include asphalt content, aggregate gra-
dation, thickness, and smoothness. These tests are conducted routinely to
ensure that a satisfactory mix is produced.
Proper placement of an OGFC is an art and hence is difficult ro spec-
ify. The quality control process must include close inspection of the mix-
ture by a qualified inspector during production and placement of the mix-
ture. Deficiencies to look for include drainage of the asphalt, pulling and
tearing behind the paver, and improper rolling procedures.
An OGFC surface has been shown to improve skid resistance, espe-
cially in wet weather. However, there has been one problem with perfor-
mance that has reduced its use in some states. A number of states have
reported that stripping occurs in the layer directly underneath the OGFC.
Many states believe that they have solved this problem by using antistrip-
ping agents in the underlying layers, and they continue to use OGFC sur-
faces on a regular basis. Some states (.2) have reported that the use of sand
or cinder during snow-and-ice-control operations tends to fill the surface
voids of the OGFC, preventing proper water drainage.
STONE MATRIX ASPHALT (SMA)

Introduction
Stone Matrix Asphalt (SMA) has been used in the United States since
1991 (.@. Five states constructed SMA. projects that year and its use has
------=c=o=n,..,tinuecl: to grow sin-ce7:he-firsrprojects-were-constructed;-~Prior to its use
in the United States, SMA. had been used in Europe for more than 20
years.
When first developed, SMA. was used in Europe to minimize the
effects of studded tires. Studies had shown that this mix was more resis-
tant to studded tire wear than any other type of hot mix asphalt. Since that
time most European countries have disallowed the use of studded tires but
SMA has continued to be used because of other desired properties. The
primary reasons for using SMA mixtures in Europe today are its resistance
to rutting and its increased durability.
Just as in Europe, the primary reasons for using SMA in the United
States have been its improved resistance to rutting and its improved dura-
bility. The macrotexture of the mixture has also been shown to improve
surface drainage over that of dense-graded mixtures and to increase fric-
tion during wet weather.
Today SMA is a standard mixture type in several states. As its use has
increased, the price per ton of mix has decreased substantially to a point
slightly above (typically 20-25%) that for conventional HMA.
Description
Stone Matrix Asphalt is a mixture of crushed coarse aggregate, crushed
fine aggregate, mineral filler, asphalt cement, and stabilizing agent. The
stabilizing agent is used to prevent draindown of the asphalt cement and
typically consists of fibers and/ or polymers.
The SMA mixtures are designed to have a high coarse aggregate con-
tent (typically 70-80%) a high asphalt content (typically over 6%) and
high filler content (approximately 10% by weight). Using high coarse
aggregate content results in stone on stone contact that produces a mixture
that is highly resistant to rutting. A view of a typical SMA mixture and for
comparison a typical dense-graded mix is shown in Figure 7-1. Notice the
STONE-ON-STONE CONTACT
COARSE AGGREGATE FLOATING IN FINE AGGREGATE MIXTURE

Figure 7-1. Cross-sectional View of a Typical SMA and a Dense-
Graded HMA Mixture
high stone content for the SMA mixture. Also notice that the coarse aggre-
gate in the dense-graded mixture appears to be floating in the fine aggre-
gate matrix. Typically the coarse aggregate in the SMA mixture carries the
load while the fine aggregate in a dense-graded mixture must carry the
load. The SMA mixture is more resistant to rutting since coarse aggregate
can develop more shear strength than fine aggregate.
The SMA mixture is designed, produced, and placed using methods
similar to those for dense mixtures. There are some differences in SMA
and dense-graded mixtures that are discussed in the following paragraphs.
Materials
As stated earlier, the materials used to produce SMA include aggregate,
filler, asphalt cement, and stabilizer. Most SMA projects that have been
constructed have used all crushed aggregate to provide the maximum resis-
tance to rutting. Typical specifications (2) require 100% of the coarse
aggregate to have at least one fractured face and at least 90% with 2 or
more fractured faces. This is a premium asphalt mixture and therefore
high quality materials must be used for best performance.
The Los Angeles abrasion value required by some states for SMA is
lower than that required for dense-graded mixtures. Since the SMA mix-
ture has stone on stone contact there is more stress under traffic loads on
the coarse aggregate particles and many 'engineers believe that because of
this higher stress the L.A. Abrasion requirements should be lower. A typi-
cal L.A. abrasion requirement that has been used for aggregate for SMA is
30. However, due to a lack of unlimited quantities of high quality aggre-
gates, many states allow the L. A. Abrasion requirement to be higher.
The typical gradation requirement for SMA mixture is shown in Table
7...:2 (2). Notice that this mixture has a relative low percentage passing the
4.75 mm sieves (20-28%) and a relative high percentage passing the 0.075
mm sieve (8-10%). The low percentage passing the 4.75 mm sieve is nec-
essary to provide a mixture with stone on stone contact and to 'Provide a
mixture meeting the minimum VMA requirements (typically 17%). The
high percentage passing the 0.075 mm sieve is necessary to adequately stiff-
en the binder so that the mixture is rut resistant and to ensure that drain-
down of the asphalt cement does not occur during construction.
- - - - - ~ r h : e h'i:gh-p-ercentagro·f-rrrare·fra:l-p·assirrg-the-Et0'15-mrn:·sieve requires
that a commercial filler be added to the mixture. In dense-graded mix-
tures it is common to have up to 6 percent material passing the 0.075 mm
sieve coming from the fine aggregate source. The SMA mixtures however
have a lower percentage of fine aggregate resulting in a lower percentage of
'
Table 7-2. Typical Gradation Requi,rements for SMA Mixture (.2)
Sieve Designation Percent Passing
19.0 mm 100
12.5 mm 85-95
9.5mm 75 maximum
4:75 mm 20-28
2.36mm 16-24
0.60 mm ""·-
12-16
0.30 mm 12-15
0.075 mm 8-10
0.020 mm less than 3
0.075 mm material available fr9m this source. Consequently, it is normal

to need as much as 3-5 percent commercial filler in SMA mixtures.
The grade of asphalt cement used in SMA is typically the same or a
slightly stiffer grade than that used for dense-graded mixtures.
One potential construction problem with SMA is draindown of the
asphalt binder from the aggregate matrix. Using a stiffer binder will help
to minimize any potential draindown problem. The stiffer asphalt binder
will also help to resist rutting at higher temperatures. One potential prob-
lem with using a stiffer binder is thermal cracking. SMA mixtures have a
relatively high optimum asphalt content and increased film thickness
which tends to reduce the susceptibility to thermal cracking. So it is
believed that the negatives of using a slightly stiffer asphalt ceinent are
more than offset by the increased film thickness. Itis also believed that
thermal cracking will be less of a problem with SMA mixtures than with
dense-graded mixtures.
Polymer modified asphalts are often used to produce SMA mixtures.
----a1The-m0d-ifiC:ld-asphalts-t~B:d-t0-.r:001:1ee-1:he-cl-r-ail'l:d-ewn-e,f-the-a-spha:l:t-cemen-r------
and therefore help in the construction process. The modified asphalt
cements may also provide improved high and low temperature perfor-
mance when compared to asphalt <;::ements that have not been modified.
Modification can potentially be used to increase the asphalt binder stiffness
Chapter 7-Special l\'.Iixtures, Recycling, and Additives 419
at high temperatures and actually decrease the stiffness at lower tempera-

tures. As discussed in Chapter 2, the Superpave binder tests can be used
to evaluate these high and low temperature properties.
Some type of stabilizer must be used to prevent draindown of the
asphalt cement during construction. The types of stabilizers that have been
successfully used to do this are fibers and asphalt modifiers. On many
SMA projects both polymers and fibers have been used. In this case, the
purpose of the polymers should be to improve the quality of the asphalt
cement at high and low temperatures and to provide some resistance to
draindown. The primary purpose of the fibers is to help ensure that no
draindown occurs.
Two general types of fibers have been used in SMA mixtures: cellulose
and mineral fibers. There are a number of fiber properties that are impor-
tant including length and thickness but it is most important that the fibers
prevent draindown of the asphalt cement. As long as the fibers can be eas-
ily handled prior to and during construction, and prevent draindown, then
they have adequately served their purpose. The mineral fiber is typically
heavier than the cellulose fiber and therefore is used at a higher content by
weight in the mixture so that sufficient volume of fibers is provided.
Mix Design
The mix design procedure for SMA mixtures is similar to that for
dense-graded mixtures but there are some significant differences (1.Q). The
Marshall and Superpave design methods have been used to design SMA
mixtures. The criteria for both methods have been set to ensure coarse
stone on coarse stone contact and to ensure that high asphalt contents are
selected.
One step in the mix design process is to ensure that stone on stone
contact exists in the SMA mixture. One way to measure this is to place
only the coarse aggregate fraction into a mold and rod it to determine the
dry rodded unit weight. This represents the density of the coarse aggregate
for stone on stone contact. Certainly, if the density of the coarse aggregate
portion of the SMA is compacted to a higher density than the dry rodded
density, stone on stone contact would be expected to exist in the SMA.
This 1s evaluated by measuring the voicl.sinthecoarse aggregate (VCA) in
the SMA mixture by the following formula: ·
VCA = (100 (1
where
Gmb = Bulk specific gravity of coarse aggregate
Pca = Percentage of coarse aggregate by weight in total
mix
G,6 = Bulk specific gravity of total aggregate
Since VCA is inversely proportional to density, a lower VCA is need-
ed in the S~ mixture than in the compacted coarse aggregate only for
stone on stone contact.
When the VCA in the mixture is greater than that in the coarse aggre-
gate only compacted sample, the VCA can be decreased by reducing the
amount of fine aggregate. As the amount of fine aggregate is decreased, the
VMA will normally increase for SMA mixtures and the VCA will decrease
(Figures 7-2 and 7-3).
20
19
18
17
- 16
~
< 15
::.
> 14
13
12
11
IO
15 20 25 30 40 so
Percent Passing 4.75 mm Sieve(%)
Figure 7-2. Effect of the Material Passing 4.75 mm Sieve on VMA
60
55
50
45
- 40
~
~ 35
u
> 30
25
20
15
IO
15 20 25 30 40 50
Percent Passing 4.7S mm Sieve (o/o)
Figure 7-3. Effect of the Material Passing 4.75 mm Sieve on VCA

r
, Chapter 7-Special Mixtures, Recycling, and Additives 421
For the SMA mix design, the fine aggregate content is varied to deter-
mine which aggregate structure will provide adequate VMA and stone. on
stone contact. In other words, the percent passing the 4.75 mm sieve will
be reduced (beginning on the high side of the specification band) within
the allowable specification range until the minimum VMA and maximum
VCA requirements are met.
Work with SMA mixtures has.shown that for most aggregates stone on
stone contact begins to occur at approximately 30 percent passing the 4.75
mm sieve. This is the primary reason that the high side of the specifica-
tions for aggregate gradation for SMA mixtures are usually set at approxi-
mately 30 percent passing the 4. 75 mm sieve.
To ensure high optimum asphalt content in an SMA mixture the min-
imum VMA is set relatively high. A typical minimum requirement for
VMA is 17. The VMA typically begins to increase rapidly as the percent-
age passing the 4.75 mm sieve falls below approximately 30 percent. If a
VMA problem exists, it can generally be solved by reducing the percentage
passing the 4.75 mm sieve a few percentage points. Some aggregates tend
to breakdown excessively resulting in low VMA values.
Aftenhe gradation has been selected to provide stone on stone contact
and minimum VMA requirements, the asphalt content is adjusted to pro-
vide the desired air void level. SMA mixtures have been designed to have
air voids somewhere between 3 and 4 percent with most mixtures designed
closer to 4 percent air voids. Close control must be exercised when pro-
ducing SMA mixtures at lower air void content to ensure that bleeding
does not occur after the mix is in-place. The air void content of laborato-
ry compacted samples must be controlled during construction to ensure
that the air voids stay within the desired range.
SMA mixtures have been designed using the Marshall and Superpave
compaction procedures. When the Marshall compaction is used, 50-blow
compactive effort has been utilized. Additional compaction does not
increase density significantly but does tend to break down the aggregate
excessively. Some SMA mixtures have been designed with the Superpave
gyratory compactor. Although the number of gyratory revolutions have
not been established for SMA, mixes have been designed at 100 revolutions
for Ndesign·
After the mixture has been designed a draindown test should be con-
------.-1---U€t-ed--1:0-ens1:1Je-~ha-t-El:Fa:i-a-a-0-wn-is-n0-t-a-prebl~m-(-H).~A-simple test has
been developed to measure the draindown potential. The test requires that
a sample of SMA mixture be placed into a wire basket and placed into an
oven at a specified temperature (typically the mixing temperature) for a
specified amount of time. The asphalt binder that drains through the bas-
ket is caught and weighed. The amount of drainage from the loose SMA
mixture should not exceed some specified requirement typically 0.3 per-
cent for I hour. A photo of the basket is shown in Figure 7-4. If exces-
sive draindown occurs, the mix should be modified, usually by increasing
the amount of polymer or fiber.
Figure 7-4. Wire Basket Used in the Draindown Test.
Construction
The use of SMA mixtures requires some changes at the HMA plant
(12). A high amount of commercial filler, usually 3-5 percent, is fed into
the plant. Most plants are not set up to handle this high amount of filler,
and this can be a bottleneck in production capacity. The preferred method
of addition is with a silo where the filler is accurately metered into the mix-
ture. Some projects have fed the filler though one of the aggregate feeders,
but this approach is undesirable.
When a polymer is used, it is normally blended with the asphalt
cement prior to delivery to the HMA plant. However, in some cases the
polymer has been added at the plant. In this case, the polymers may be
blended with the asphalt cement and aggregate at the same time. One
potential problem with pre-blending a polymer is the potential for separa-
tion during handling and storage. The primary problems often encoun-
tered when blending at the plant are lack of proper facilities, inadequate
1 - - - - ~ lenciiiig;-aii.a-ina6ificy-to test mod1fielasphafr prior to iii:troduction into
_ _ _ __
the mixture.
When fibers are used, their introduction into the mixture must be
closely controlled. Even though the fiber content is small, it has a signifi-
cant effect on the mixture properties. The fibers can be weighed and added
in batches in a batch plant. The fibers are normally blown into the drum
at the desired rate in a drum mix plant.
The SMA mixtures are typically stiffer than conventional dense-grad-
ed mixtures and may require higher temperatures and more mixing time at
the plant. The primary reasons for the increased stiffness are the high filler
content, modified asphalt, and/or fibers.
Placement and compaction procedures are similar to that for
dense-graded mixtures. SMA mixtures are typically stiff and more difficult
to work. Because of this stiffness and large coarse aggregate content, con-
struction of good longitudinal joints is more difficult. However, experience
has shown that a good joint can be built. Draindown of the asphalt cement
may occur if the mixture is not properly designed or if the mixture is not
properly controlled, resulting in fat spots in the compacted pavement.
Rollers must be kept immediately behind the paver for good com-
paction. As the SMA mixture cools, it tends to set up quickly and experi-
ence has shown that compaction becomes very difficult once the mix
begins to cool. Rubber tire rollers are not used on SMA primarily due to
the mix sticking to the tires. Static and vibratory steel wheel rollers have
been successfully used. Under some situations, the vibratory rollers may
tend to excessively break down the aggregate. If this is a problem, only sta-
tic rolling should be used. SMA mixtures are typically compacted in-place
to an air void content of 5-6 percent.
Performance
Since the first SMA projects were constructed in the U.S. in 1991, the
performance history is insufficient to make final conclusions. However,
the Europeans have used SMA mixtures since the 1960's and have found
them to provide excellent performance.
SMA mixtures have been shown to provide excellent resistance to rut-
ting on high traffic volume roads. SMA mixture has also been shown to
be less susceptible to cracking. It appears that SMA mixtures have the
potential to improve stability and at the same time increase the durability
of HMA mixtures.
-----------------------~------
DENSE-GRADED LARGE STONE MIXES
Premature rutting of heavy duty asphalt pavements has been a signifi-
cant problem in recent years. This rutting problem has primarily resulted
from higher pressure truck tires and increased wheel loads. The design of
HMA mixtures, which served reasonably well in the past, needs to be re-
examined to withstand the increased stresses. Various asphalt additives are
being promoted to. i~trease the stability of HMA pavements at high tem-
peratures. However, most asphalt technologists believe that fundamental
properties of the aggregate component of the HMA (such as size, shape,
texture, and gradation) ~e inost important to ensure that a rut resistant
mixture is obtained. There is a general agreement that the use of large size
stone in the binder and base courses will minimize or eliminate the rutting
of heavy duty asphalt pavements.
The use of large stone mixes is not new. As reported by Davis (13),
Warren Brothers Company had a patent issued in 1903 which specified a
top size aggregate of three inches (76 mm). Most paving companies start-
ed to use smaller stone mixes at that time 'i:o avoid infringement of the
patent. Although the patent expired, such use is still prevalent today.
Marshall mix design procedures are used by 76 percent of the states in
the United States according to a survey conducted in 1984 (1§). The
equipment specified in the Marshall procedure (ASTM Dl559) consists of
a 4-inch (100 mm) diameter compaction mold, which is intended for mix-
tures containing aggregate up to I-inch (25 mm) maximum size only.
This has also inhibited the use of HMA containing aggregate larger than
one inch because this mixture cannot be tested by the standard Marshall
mix design procedures. There are other test procedures, such as gyratory
compaction, TRRL refusal test, and the Minnesota DOT vibrating ham-
mer, which use 6-inch (150 mm) diameter molds accommodating I½ to
2-inch (38 to 51 mm) maximum aggregate size (15). However, most
,agencies are reluctant to buy new equipment because of cost and/or com-
plexity. They prefer to utilize the existing equipment and/ or methodology
(such as the Marshall test) with some modifications. NCHRP's AAMAS
(Asphalt-Aggregate Mix Analysis System) research study showed that a
laboratory gyratory compactor better simulates the aggregate particle ori-
entation obtained in the field compared to that obtained with an impact
compactor used in the Marshall procedure (lfi). However, it was realized
in 1989 that it would be a few years before many agencies start to imple-
ment the AAMAS study's recommendations and use gyratory compactors
or other proposed compactors. In the meantime there was an urgent need
to start designing large stone HMA using modified Marshall design proce-
----dures-6-a:s·ed-cm-catrenrktrnwhxtge and experience.
The term "large stone" is a relative one. For the purpose of this sec-
tion, large stone is defined as an aggregate with a maximum size of more
than one inch (25 mm) which cannot be used in preparing standard
4-inch (100 mm) diameter Marshall specimens.
Development of Modifications to Marshall Procedures

A study was undertaken by Pennsylvania DOT in 1969 to develop the
equipment and procedures for testing the ability of 6-inch (150 mm)
diameter specimens to accommodate binder and base course HMA mixes
containing aggregate of 11/z-2 inches (38-51 mm) maximum size (.1.Z).
The nominal height of the 6-inch (150 mm) diameter specimen was
increased to 3¼ inches (95 mm) to provide the same diameter/height ratio
that is used for a 4-inch diameter x 21/z inch high (100 mm diameter by
63.5 mm high) specimen. The weight of the compaction hammer was
increased from 10 lbs. to 22.5 lbs. (4.5 kg to 10.2 kg) in proportion to the
increased face area of the specimen. The hammer drop was maintained at 18
inches (457 mm) so that the same mechanical compactor used for 4-inch
(100 mm) diameter specimens could be used to compact 6-inch (150 mm)
diameter specimens. The number of blows/face needed to compact 6-inch
(150 mm) diameter specimens is 50 percent more than that used to compact
4-inch (100 mm) diameter specimens. The increased weight of the hammer
and the increased number of blows used provide the same amount of com-
paction energy per unit volume to a 6-inch (150 mm) diameter specimen as
is obtained in a conventionally compacted 4-inch (100 mm) diameter spec-
imen. This results in about equal density for 6-inch and 4-inch (150 mm
and 100 mm) diameter specimens. The modified Marshall procedure was
included in the Pennsylvania Test Method 705 developed by Kandhal and
Wenger (1.!J). The details of test procedures and equipment, such as mold,
hammer and breaking head, are given in Reference (12). The proposed stan-
dard (now adopted by ASTM as D5581-96) essentially follows ASTM
D1559-82 (2{!), which is intended for 4-inch (100 mm) diameter speci-
mens, except for the following significant differences:
1. There is a mold assembly and breaking head to accommodate
6-inch (150 mm) diameter specimens.
2. Since the hammer weighs 22.5 pounds (10.2 kg), only a mechani-
cally operated hammer is specified.
3. About 4,050 grams of mix is required to prepare one 6-inch (150
mm) Marshall specimen compared to about 1,200 grams for a
4-inch (100 mm) specimen.
_______4~I~b"'~~ro~i~x~is plac.e.d.in....the_m_old..iru:::wo_ap_proximatel~~equal incre-
ments; spading is specified after each increment. Experience has
indicated that this is necessary to avoid honey-combing on the
outside surface of the specimen and to obtain the desired density.
5. The number of blows needed for 6-inch (150 mm) diameter spec-
imens that are 3¼ inches (95 mm) high is 11/z times the number
of blows needed for 4-,---inch (100 mm) diameter specimens that are
21/z'inches (63,5 mm) high in order to obtain an equivalent com-
paction level fof a given volume of material.
Since the hammer 'Yeighs 22.5 pounds (10.2 kg) and the number of
blows on each side is 75 or 112 depending on the anticipated traffic, some
crushing of the aggregate at the surface has been observed. However, it is
believed that the effect of this minor amount of crushing on Marshall
properties is minimal.
Analysis of Comparative Data

Comparative data of 4-inch (100 mm) v~rsus 6-inch (150 mm) spec-
imens obtained by various highway agencies and producers indicate that
the compaction levels obtained in 4-inch (100 mm) and 6-inch (150 mm)
molds (using the appropriate hammer and number of blows) are reason-
ably_ close (21). As expected, the repeatability of stability and flow tests is
significantly better when 6-inch (150 mm) diameter specimens are used
for large stone mixes. Therefore, 6-inch (150 mm) diameter specimens
should be used for designing large stone mixes.
Since higher stability and flow values are expected from the larger spec-
imens, there was a need to determine the stability ratio (stability of 6-inch
(150 mm) specimen/stability of 4-inch (100 mm) specimen) and the flow
ratio (flow of 6-inch (150 mm) specimen/flow of 4-inch (100 mm) spec-
imen). This was necessary so that minimum stability values and maximum
flow values for 6-inch (I 50 mm) specimens could be derived from the val-
ues specified for 4-inch (100 mm) specimens.
Table 7-3 summarizes the stability and flow ratio values obtained by
various agencies and producers on large stone base or binder mixes (maxi-
mum aggregate size 11/z-2 inches or 38-51 mm). The average of 11 sta-
bility ratios is 2.18, and the average of 11 flow ratios is 1.44. These mea-
sured values are very close to derived values as discussed in the following
paragraphs.
From a theoretical viewpoint, an external load applied to the circum-
ference of a cylinder may be considered as acting directly on the diametri-
cal cross section of the cylinder. This permits calculation of the stress in
pourrds-penqua:re-irrch:-'Fh.-e-sta:rrda-rd-6-i:nch ( 150 mm)specimeii. is-3 3/ , - - - - - -
inches high, which gives a diametrical cross section of 22.5 square inches.
The standard 4-inch (100 mm) specimen is 21/z inches high and has a dia-
metrical cross section ofl0.0 square inches (6450 mm2). Therefore, on the
basis of unit stress, the total load on a 6-inch (150 mm) specimen
~
("}
-
::,-
O"' .g
n Ef
~
--i
Agency (Year data obtained) No. of Blows Ratio I
~ Cl)
~ra
'Cl
4 inch 6 inch Stability Flow
I
"
D.
e.
PennDO1f (1969) 50 75 2.12 · 1.62
~ ~
X
PennDojr (1970)
I
PennDO1f (1988)
50
50
75
75
2.81
1.95
1.15
1.39
.s
0
...8
"
Y'
::0
PennDojr (1988) 50 75 2.17 1.58
l-+-)
ti)
"
C)
'<
I
PennDO1f (1989) 50 75 1.68 1.40 g. [-
<J<>
Jamesto'tn Macadam Inc. (1989) 50 75 1.89 1.24 I?. §

4 0..
Kentuckj, DOH (1988) * 75 112 2.08 1.34

~
e;►
America* Asphalt Paving Co. (1989)* 75 112 2.37 1.63 i:i,. ~-
""'
America* Asphalt Paving Co. (1989) * 75 112 2.5~ 1.52 :!a
America* Asphalt Paving Co. (1989) * 75 112 1.98 1.68
~
America* Asphalt Paving Co. (1989) ·• 75 112 2.40 1.27
~
0(I)
I No. of Mixes (N) 11 11
Mean 2.18 1.44
a-,
I '"I
I Std. Dev. 0.33 0.18 I:""'
*Note: The average stability and flow ratio for these five mixes compacted with 75/112 blows are 2.28 and 1.49, respectively.
~
...
Vl
0 .I>,
1:1 N
n --l
should be 2.25 times the load applied to a 4-inch (100 mm) specimen of
the same mix. This means the stability ratio should be 2.25.
Flow units measi,ired by the testing machine are the values for the
total movement of the breaking heads to the point of maximum stabili-
ty. When flow is ~nsidered on a unit basis (inches per inch of diame-
ter), the flow value for a 6-inch (150 mm) diameter specimen is 1.5
times that of a 4-incp. (100 mm) diameter specimen. This means the
flow ratio should be 1.5.
The average stability and flow ratio of specimens compacted with 75
and 112 blows (4-inch (100 mm) and 6-inch (150 mm) mold, respec-
tively) are 2 . 28 and 1.49 which are very close to the theoretically derived
values of 2.25 and 1.50, respectively.
It is recommended that the minimum Marshall stability requirement
for 6-inch (150 mm) diameter specimens be 2.25 times the requirement
for 4-in.ch (100 mm) diameter specimens. For example, if 1,000 pounds
(4,500 N) minimum stability is currently being specified using ASTM
D1559 ·(4-inch (100 mm) specimen), then 2,250 pounds (10,125 N)
minimum stability should be specified for large stone mixes using the
6-inch (150 mm) Marshall testing equipment. Similarly, the range of flow
values for 6-inch (150 mm) specimens should be adjusted to 11/z times the
values required for 4-inch (100 mm) specimens. For example, if the spec-
ified range for 4-inch (100 mm) specimens is 8-18, it should be adjusted
to 12-27 for 6-inch (150 mm) specimens.
Typical Mix Designs and Field Trials

The Kentucky Department of Highways (DOH) has completed a sub-
stantial number of large stone mix designs using the 6-inch (150 mm)
Marshall testing equipment. Other states are beginning to use the 6-inch
(150 mm) compaction equipment. Kentucky DOH has used large stone
mixes on coal haul roads carrying very heavy trucks (gross loads varying
from 90,000 to 150,000 pounds (41 to 68 Mg) or more). Tire pressures are
also higher than generally encountered, ranging from 100 to 130 psi (689
to 896 kPa) (22). Typical mix designs used are given in Reference 2.3_.
All mixes used by the Kentucky DOH have limestone aggregates and
a maximum aggregate size of 2 inches (50 mm) with a substantial amount
of material retained on I-inch (25 mm) sieve. This res_ults_ii::u1--signific.an_c._ _ _ __
amount of I-inch to 3/4-inch (25 mm to 19 mm) material in the mix. All
mix designs were developed using the 6-inch (150 mm) mold and 112
blows on each side. Asphalt content was varied generally from 3.0 to 4.5
percent in 0.5 percent increments. Either MSHTO Gradation #467 (11/z
inch to No. 4) or #4 (1 ½ inch to 3/4 inch) was used for coarse aggregate
to incorporate + 1 inch (25 mm) material in the mix. The data from field
trials in Kentucky indicate that achieving the desired density (compaction)
in the field does not appear to be a problem if the compaction process is
optimized (12.). The average in-place void content of three reported pro-
jects was approximately 6.5 percent.
Careful attention to details is needed to assure uniform delivery and
laydown of large stone mix without any significant segregation. The fol-
lowing factors are considered important (22):
I. Uniform component aggregate gradations and good stockpiling

practices are required.
2. Increased sampling and testing is desirable to assure good quality
control. Usual extraction tests for control of gradation and asphalt
content may be a problem due to difficulty in obtaining a repre-
sentative sample for testing. Bin samples, recombined at the prop-
er percentages, are better measurements of gradation for large stone
mixes. Printout data may be relied upon for asphalt content con-
trol.
3. Segregation in the surge bin is more difficult to control. This ten-
dency to segregate is extended to truck loading. However, segrega-
t.ion due to loading may be overcome by using a front, back, cen-
ter loading scheme for single unit trucks. A five drop loading
sequence (front, rear, center for the first three drops with the last
two drops between the front/center and the rear/center) may be
used for semi-trailer trucks.
4. Coarse particles accumulate in the receiving hopper wings of the
paver. This effect is reduced by not clearing coarse material from
the hopper until the end of each day's paving. The accumulated
coarse particles may be wasted.
5. Mixture in the receiving hopper should be maintained at a mini-
mum depth of \8 to 24 inches (457 to 610 mm) over the slat con-
veyor to prevent coarse particles, collected in the wings, from reen-
tering the mix and producing concentrations of coarse particles.
6. Receiving hopper gates should be set to provide nearly continuous
operation of the slat conveyor. Furthermore, to supply mix to the
--------~s=c=re~e=d~a=t'--'t=h=e--=r-=-esiuired rate, continuous operation of the distribution
augers is desirable. .
7. Depth of mixture in front of the screed must be maintained at a
constant level for the full screed width to assure a uniform spread.
Auger extensions, as needed, supply material uniformly to the end
plates. If extensions are not used, coarse particles tend to roll to the
out~r edge of the spread, creating a porous area with low density.
8. Paver speed is very important. The lowest rate of travel that accom-
modates production should be used. Slower rate of movement per-
m.its 111:rre uniform feeding of mixture under the screed and sup-
plies more vibrating compaction by the screed. Both permit better
positioning of coarse particles. Avoiding "stop and go" paving
reduces segregation, improves the texture of the spread, and elimi-
nates any tendency for screed settlement.
ASPHALT PATCHING MIXTURES

Asphalt patching mixtures are comh>inations of aggregates, asphalt
binders and, if needed, additives. These mixtures are used to repair pot-
holes, wide and badly spalled cracks, severely raveled areas, and small util-
ity cuts. There are three general types of asphalt patching mixtures (2d';.
I. Hot-mixed, hot-laid patching mixtures
These are essentially HMA mixtures consisting of well-graded
aggregates and asphalt cement. HMA mixtures must be used
while hot. These mixtures are the highest quality of all asphalt
patching mixtures; however, there are some drawbacks:
(a) Availability. Most HMA facilities, especially in the northern
United.States, close during winter months or during periods
of no construction. Hence, HMA is not available for patch-
ing in the winter months.
(b) Quantity. Unless there are numerous potholes, one needs less
than a truckload (usual amount obtained from a HMA plant)
of the mix. Even if a full truckload is obtained, the patching
operation is very slow and time consuming, and the mix usu-
ally cools before it can be completely used. The use of mix
that has cooled results in inadequate compaction. However,
this problem can be overcome by using hot boxes or small
portable recycling units.
If at all possible, most agencies prefer to use hot-mixed, hot laid
HMA mixtures because they constitute a permanent and durable
-+-------.e~at~c~h~----------------------------
2. Hot-mixed, cold-laid patching mixtures

These mixtures are produced in a HMA facility; therefore, mate-
rials are properly proportioned and thoroughly mixed. Aggregates
are dried and heated to a temperature substantially lower than that

used to produce a HMA mixture Liquid asphalt binders, such as
medium curing cutback asphalts and emulsified asphalts, are used.
These mixtures are stockpiled for use at a later date and they are
workable in all weather conditions. High mixing temperatures
should not be used; otherwise, volatiles will be driven off from the
bituminous binder, reducing the storageability of the patching
mix.
3. Cold-mixed, cold-laid patching mixtures

These mixtures are produced from aggregate and liquid asphalt
binders which are not heated prior to mixing. Mixing can be done
in a HMA facility (dryer is not used), a stabilization plant, or any
other means of mixing the aggregate and liquid asphalt. The mix-
tures are stockpiled and used whenever needed. These mixtures
have the lowest quality of all the asphalt patching mixtures.
Only the asphalt patching materials which can be stockpiled (Types 2 and
3) will be discussed in more detail in this section.
Desirable Properties
In a majority of cases the failures of cold mixes may be traced to
improper construction practices (such as "dump and run" where the mate-
rial is dumped into a pothole on the run with little or no compaction).
Failures are also often related to the lack of desirable properties in a stock-
pile patching mixture. The following properties are desirable in such mix-
tures (.21):
I.Stability
A patching mixture should be stable after placement and com-
paction to resist vertical and horizontal displacement due to
imposed traffic loads. Lack of stability causes dishing and shoving
of the mixture.
2. Stickiness
Stickiness causes the mixture to adhere to the underlying pave-
rrreurarrd-srde·s-ofth:-qmth:-o-I-e~an-d-is-esp·ecially-desirable when the
mixture must be feathered to thin edges. Quite often mainte-
nance personnel do not take time to clean and dry the hole thor-
oughly so that proper tack coat can be applied. In such cases a
stickier patching mix is especially helpful.
3. Resistance to Water Action

Stripping or separation of asphalt binder from the aggregate
occurs, caus_ing the mix to ;avel unless the mix is resistant to water
acti~n. This phenomenon is more prevalent when the mix is used
to patch potholes due to poor water drainage. Stripping also
results when the mix is not compacted adequately.
4. Durabilizy
Durability is the resistance of the patching mixture to deteriora-
tion and disintegration due to traffic and weathering forces (21).
_Usually, durability problems appear in the form of "raveling."
Hot-mixed, cold-laid stockpile patching mixtures, which are
manufactured in HMA facilities under controlled conditions, can
be considered "high performance" cold patching materials which
are generally more durable than col~-mixed, cold-laid mixtures.
The latter materials are used for temporary patching during win-
ter months and are usually replaced with a HMA mix during
warmer weather.
5. Skid Resistance
Adequate skid resistance is desirable for long, large patches.
Slipperiness can be caused if the mixture contains aggregates
which are easily polished and/or excessive asphalt binder which
causes flushing and bleeding.
6. Workability
The patching mixture must be sufficiently soft and pliable for easy
shoveling and raking, especially in cool weather. The mixture
should be free of large lumps which need to be broken. Poor
workability of the mixture generally produces poor patches. Such
mixtures are also hard to compact and thus do not stay in the pot-
hole for a long period of time.
7. Storageability
A stockpile patching mixture should remain pliable (workable)
when stored for the desired period of time (usually six to twelve
months). If the mixture does not contain the right type ofliquid
asphalt binder, it can lose the v.:datileuapidl)U.Ud..hecom.e-hard...__ _ _ __
and harder with time. Covering the stockpile with tarps or poly-
ethylene extends the storage time. The mixture should not con-
tain excessive binder which drains and puddles at the bottom of
the stockpile.
Challenges Of Mix ·Design

It is difficult to design stockpile patching mixtures because they
require opposing demands on the material for stockpiling and handling
and for performance after placement in the pothole (25). Some of these
opposing demands are as follows:
1. Aggregate Gradation
For good mixture workability, it is desirable to have an open gra-
dation, but after the mix is placed, the gradation needs to be
denser to improve stability and durability.
2. Aggregate Shape
To obtain good workability, angular aggregate should be avoided.
However, once the mix is in place, high angularity is desirable for
better stability.
3. Binder Viscosicy:
Lower binder viscosity is desired for storageability and workabili-
ty, but after placement and compaction, higher viscosity is desir-
able for better cohesion and stability of the mixture.
4. Binder Content
It is desirable to have higher residual asphalt content in the mix-
ture to obtain thicker films on the aggregate for stickiness and
durability, but there are limits due to the binder drainage problem
in the stockpile just after stockpiling while the mix is hot. A high
binder content may also produce an unstable mixture.
Use of highly absorptive aggregates can also pose problems. High

moisture content in such aggregates often causes stripping and/ or drainage
problems in the stockpile. Selective absorption of the lighter fractions of
the asphalt binder by aggregates leaves an asphalt film with undesirable
characteristics and significantly different properties than the original
asphalt binder.
It is not possible to use conventional methods of mix design generally
used for HMA, such as Marshall, Hveem, etc. Not only is the specimen
preparation and testing difficult, but the desired design criteria are
unknown for the stockpile patching mixtures.
New Concepts
In the past, use oflarger aggregate size (1/2-3/4 inch or 12-19 mm)
in the stockpile mixture has been promoted to obtain higher stability.
Such a mixture can be successful if the patching technique is ideal (such as
making edges vertical, cleaning, applying tack coat, and compacting ade-
quately). However, ideal patching techniques an: not always used, and
mixtures with larger aggregates tend to ravel prematurely under traffic,
resulting in failure of the patch. Another concept is to disregard the sta-
bility and make the mixture finer and more pliable so that it is more toler-
ant of abuse during the placement and performance under traffic. This
finer mix, if placed less than 3 inches (76 mm) deep in one lift in a con-
fined area, shoufd cbe stable. For deeper and/ or larger holes, the mixture
has to be compacted in layers.
The cohesive and adhesive qualities of a mix are mainly dependent
upon the composition of the mortar (asphalt binder plus fines). The pres-
ence of excessive fines or dust (material passing No. 200 (0.075 mm) sieve)
in the mixture results in a mortar that is lean, less tacky, and friable. It is
no coincidence that most of the expensive commercial. patching products
are made. from clean stone. Several extraction tests performed on such
products have revealed that the fines (minus 200 fraction) are usually less
than one percent. The absence of excessive fines causes mixtures to be very
tacky and sticky; therefore, tack-coating of the pothole is not required.
Many conventional stockpile patching mixtures do not perform satisfacto-
rily because of excessive fines. Such mixes are dull and friable, and lack the
cohesive and adhesive qualities.
Improved Formulation
In view of the challenges of mix design and the new concepts, the fol-
lowing characteristics appear desirable for a satisfactory and economical
stockpile patching mixture (25.):
1. Finer and Predominantly One-Size Gradation
Atypical gradation used very successfully by a state DOT is given
in Table 7-4. Ideally, two sizes of patching mixtures should be
Table 7-4. Recommended Composition of Stockpile Patching Mixture

3/8 inch (9.5 mm) 100
No. 4 (4.75 mm) 85-100
No. 8 (2.36 mm) 10-40
No. 16 (1.18 mm) 0-10
No. 200 (0.075 mm) 0-2
Residual asphalt content, percent 4.5*
*For aggregates which do not absorb more than one percent water by weight
available: a base-size (about 3/4 inch or 19 mm) aggregate mix-

ture to fill the deeper holes and a small-size (about 1/4 inch or 6
mm) aggregate mixture to fill shallow holes or to do skin patching.
However, two sizes of patching material are not normally utilized
because the maintenance forces want to handle only one patching
mixture. Therefore, the mixture normally used is a compromise
between these two sizes. A uniform gradation, consisting of 100
percent passing the 3/8 inch (9.5 mm) sieve, has the following
advantages:
a. The mix is pliable and workable.
6. Due to increased VMA, more asphalt binder can be incorpo-
rated into the mix for improving the durability.
c. The mix remains pliable for a prolonged period of time, con-
tinues to densify easily under traffic, and adapts to the chang-
ing geometry of the pothole. This characteristic enhances its
chances of survival.
2. Clean Aggregate
As discussed earlier, it is very important to keep the dust content
(minus 200 or 0.075 mm fraction) in the mixture low (less than
two percent) to impart tackiness. This significantly improves the
adhesive and cohesive properties of the mixture.
3. Angular Aggregate Shape
Angular aggregate shape is desirable for better stability. Since a
finer and predominantly one-size gradation is used, the effect of
aggregate angularity on the workability of the mix is minimal.
Angular crushed stone aggregate is an ideal material to use in
patching mixes.
4. Use of Least Absorptive Aggregate
Highly absorptive aggregates should be avoided. The aggregate
water absorption should be limited to approximately one percent.
5. Proper Binder Type
The liquid asphalt binder selected for stockpile patching material
must satisfy two conditions. First, it should have relatively low
viscosity at low temperatures (around or below freezing tempera-
tures in the northern United States) so that the mix is workable.
Second, it should not lose volatiles at a fast rate which would cause
the mix to become unworkaoleauringthedesired storage period.
Binders for stockpile patching mixtures consist of both cutbacks
and emulsions. Cutbacks include MC-250, MC-400, MC-800,
SC-250, and SC-800 grades. The low viscosity grade (250) is
preferred for longer stockpile life and when the mixture is likely to
be used during winter. The high viscosity grade (800) is usually

used during the fall and spring. Both medium-setting and
slow-setting emulsions are used. Some agencies prefer high-float
emulsions which are more workable (due to softer residues) at low
temperatures.
6. Adequate Binder Content
It has been determined that at least 4.5 percent residual asphalt
binder is required in a stockpile patching mixture made from an
aggregate whose iwater absorption is less than one percent. If the
aggregate absorbs water in excess of one percent, the residual
asphalt content should be increased an amount similar to the addi-
tional absorption. For example, an ::i.,ggregate absorbing 1.5 per-
cent water should have 5.0 percent minimum residual asphalt
binder. The factor limiting the maximum amount of the asphalt
binder used in the mixture is drainage in the stockpile just after
manufacture. The drainage can be minimized or eliminated by
using lower mix temperatures and limiting the stockpile height to
4 feet during the first 48 hours.
7. Proper Type and Amount of Antistripping Agent
The use of antistripping agents is a very important part of the
stockpile mixture formulation. A mixture should retain its coat-
ing in the stockpile under adverse weather conditions, during
handling, and in the pothole after placement. A stockpile patch-
ing mixture is more pervious than a dense-graded HMA and thus
has to withstand more severe weather and traffic effects. It has to
survive in poor conditions (such as poor base, inadequate
drainage, deteriorated adjacent pavement, etc.) which led to the
creation of the pothole in the first place. Rain or melting snow
provide water, and the pneumatic tires of vehicles provide high
pressures. This combination can emulsify the asphalt binder and
displace it from the aggregate. If sufficient stripping occurs by this
action, the traffic will dislodge the aggregate particles.
There are many commercially available antistripping agents on
the market for use with the medium-curing (MC) cutback
asphalts. Extensive testing has shown there is no single additive
----~-which-works-wi:th-a:lt-aggrega-rnype] (;;?j)~Therefore, it i'~s~e-ss,..,,e=n-------
tial that the type of antistripping agent and its amount be select-
ed after testing with the aggregate actually being used in the mix.
Some states require the asphalt supplier to conduct the wet coating
test, the static immersion test, and the stripping test using the job
aggregate.
Stockpile patching mix made from the improved formulation gave
excellent performance in the study conducted by the Strategic Highway
Research Program.
Production, Inspection, and Storage

Production. The production procedures for stockpile patching mix-
tures are essentially the same as for other asphalt mixtures. As mentioned
earlier, HMA facilities produce accurately proportioned, quality patching
mixtures. Most hot-mixed patching mixtures for winter use are produced
at the end of the construction season before the HMA facilities shutdown.
Liquid asphalt binders used in the preparation of hot-mixed, stockpile
patching mixtures usually are heated before mixing with the hot aggregate.
Care is taken to prevent extra loss of volatiles at high temperatures.
Suggested temperatures of the mixtures are given in Table 7-5 (21). Since
the cutbacks have low flash points, safety precautions (around open flames
and near the pugmill) are necessary.
Cold-mixed patching mixtures do not require that the aggregate be
heated. However, the liquid asphalt binders may be warmed slightly to
obtain good coating.
Inspection and Testing. Agencies should inspect the production of

hot-mixed patching mixtures and obtain samples for testing. The follow-
ing tests are generally performed for acceptance:
I. Mix Composition
Extraction tests are performed or an automated HMA facility's
printed tickets are examined to verify the conformance of the mix
composition to the job-mix formula.
Table 7-5. Discharge Temperatures for Hot-Mixed Stockpile Patching

Mixtures
Mix Temperature After Discharge

Asphalt Material from Pugmill
MC & SC-250 135;190--- -
MC & SC- 800 165-220

Emulsified Asphalt 50-170
438 H~t Mix Asphalt Materials, Mixture Design and Construction
2. Stripping Test
Different forms of stripping tests are employed by different agen-
cies. Some use. the ASTM Standard Designation D 1664
"Coating and Stripping of Bitumen Aggregate Mixtures." Some
states use a boiling water test in which the mixture is placed in
boiling water for 3 minutes and the retained coating is evaluated
(at least 90 percent coating is often required). There is no satis-
factory test which has been related to performance.
3. Workability Test
A scientific test for quantifying the workability is not available at
the present time. Most specifications require that the "mix shall
remain workable and capable of being handled and placed easily."
Some states use a subjective test. "-For example, in Pennsylvania
approximately 5 lbs (2.3 kg) of the mixture is cooled to 20°F
(-7°C) in the laboratory. The mixture should be capable of being
broken up "readily'' with a spatula having a blade length of
approximately 8 inches.
Storage. The evaporation of volatiles from the stockpile causes the

mixture to become stiffer, and its workability is reduced. Most agencies
specify that the patching mixtures "shall be capable of being stockpiled
over a period of 6 months without hardening and shall remain workable."
Some specify a period of 12 months.
Usually when the volatiles evaporate, a crust is formed on the stockpile
(~. The thickness of the crust can vary considerably, but usually it is
about 1 inch thick after a few months. The crust has two advantages.
First, it tends to seal the surface and helps shed water. Water that enters a
stockpile can freeze and harden the stockpile. It is very difficult to work a
frozen stockpile. Second, the crust helps retard further evaporation of
volatiles and keeps the material softer inside the envelope.
Stockpiles should be placed on level surfaces which can be reached by
trucks. The base should consist of a hard surface such as HMA pavement
so that there is no contamination due to dirt being scooped up with a
loader when filling trucks.
Stockpiles are increasingly being covered with a tarp or plastic sheet to
keep out water and snow especially before they form a protective crust.
~ - - - - 1 his-a:I-so-exterrdnlre-sro-ra:ge-rwe:-S-ome agencies keep the stockpile ms1de
a heated building for improved workability besides protection from water

and snow.
Very high stockpiles can consolidate the material at the bottom and
therefore should be avoided. The hot-mixed patching mixture should not
be stockpiled higher than 4 feet for the first 48 hours to prevent drainage
of the asphalt binder and consolidation. This allows the stockpile to cool
before it is stacked too high. Some cases of heat build-up and burning of
high stockpiles prompted this suggested requirement.
HMA MIXTURES FOR HYDRAULIC STRUCTURES

Dense-graded HMA mixes have been used in the construction of
watertight and erosion resistant linings for hydraulic structures such as
dams and reservoirs. Since HMA mixes do not contaminate water, these
linings have been used for storage of potable water. Such hydraulic mixes
are generally designed to have very low voids to ensure impermeability to
water. If the Marshall method is used, specimens are compacted with 35
blows each face. Higher asphalt content and minus 200 (0.075 mm) mate-
rial, compared to HMA mixes for highway pavements, are used to fill the
voids. These mixes, generally placed on slopes, do not get any additional
densification in service, unlike highway pavements which are subjected to
traffic. Higher asphalt contents also make these mixes durable by increas-
ing the asphalt film thickness around the aggregate particles.
Hydraulic mixes do not need high stability. However, the mix should
have adequate stability so that it can be placed and compacted satisfactori-
ly on slopes while hot. Normally, AC-20, AR-8000, or 60-70 penetra-
tion grade asphalt cements are used. The lining should be at least 2 inch-
es (50 mm) thick and should not contain more than 4 percent air voids
when placed. This ensures reasonable impermeability to water.
Conventional pavers and rollers are used to construct HMA linings on
mild slopes. However, this equipment is attached to cables and winched
up and down the slope. Steep slopes usually require modifications to the
paving equipment.
HOT MIX ASPHALT RECYCLING

Introduction
Asphalt recycling has become popular since the mid to late 1970s.
Prior to this time very little recycling had been performed since there was
-------v-er_y_Ii.ttle incentive to recycle. References 26 ana_ll_ contain detailed
guidelines for recycling pavement materials.
A number of developments in the 1970s led to this increased aware-
ness of recycling. Probably the largest single factor was the oil embargo
and subsequent increase in the price of asphalt. The increase in the price .
440 Hot Mi:X Asphalt Materials, Mixture Design and Construction
of asphalt made any reclaimed asphalt material a valuable asset. Before the
oil embargo, the price of asphalt cement was so low that the cost of remov-
ing, stockpiling, and recycling old asphalt pavements was more than that
for purchasing, mixing, and placing new material.
A second item that has had a great impact on recycling is the milling
machine. Prior to the development of the milling machine, old asphalt
pavement had to be ripped from the roadway, then crushed prior to using.
This process was awkward, required full depth removal of the HMA, and
usually required major rework of the roadway prior to overlaying. This
process also often required that the roadway be shut down for extended
periods of time. Another process that was used prior to the milling
machine was the heater planer. This process required a large amount of
fuel to heat th~ pavement, resulted in pollution, and damaged the removed
HMA ·by overheating. The milling machine has solved most all of these
problems observed with heater planers. If can remove any amount of
material desired and produce any desired grade. It does not prod,uce appre-
ciable pollution since heat is not required. The material removed with a
milling machine does not have to be crushed since it is fine enough, imme-
diately after being removed, to recycle. A milled surface can be opened to
traffic temporarily until the overlay has been completed.
A third reason that recycling has become popular is the development
of the drum mix facility. Before the development of the drum mix facili-
ty, engineers did not know how to recycle in a batch plant. Reclaimed
material could not be fed through the dryer since this would overheat the
asphalt and cause a major pollution problem. Since the development of
the drum mix facility, the Minnesota process has been developed for the
batch facility which superheats the new aggregate in the dryer. The super-
heated aggregate mixes with the reclaimed asphalt pavement (RAP) in the
weigh box and pugmill, thus transmitting the heat of the aggregate to the
RAP. In one of the drum mix processes the RAP is added about halfway
down the drum where it is mixed with the new aggregate and new asphalt
cement. There are other processes which are described in Chapter 6.
Candidates for Recycling

Any project that requires an overlay or major reconstruction is a can-
idate for recycling~ne in-place mixture does have to be evaluated to
ensure its acceptability for recycling. For example, some pavements may
be repaired more economically with an overlay. If the pavement is rela-
tively thin, the milling machine may break the old pavement into chunks
which will likely require additional crushing when used in a recycled mix-
ture. The removal of chunks makes it difficult to maintain grade and may
require a major rework of the base course. In this case it may be simpler
and cheaper to apply an overlay. A major item that must be considered
prior to repair is the overall condition of the existing pavement. A pave-
ment .exhibiting alligator cracks has a structural problem; hence, enough
HMA must be placed to provide adequate structural properties. The need
to add 2-3 layers of asphalt mixture may indicate that overlays would be
quicker and cheaper. If it is determined that only one layer is needed to
provide adequate support, then recycling should be considered so that the
damaged layer is removed and replaced during recycling. The cracks are
likely to reflect through the overlay causing poor performance if the exist-
ing mix is not removed.
Recycling may not be desirable if the aggregate in• the existing mixture
does not meet the specified requirements. For example, if the existing
aggregate is uncrushed, then recycling produces another mixture with a sig-
nificant amount of uncrushed material. The amount of RAP used in the
recycled mixture in this case has to be minimized. A high percentage of
material passing the No. 200 (0.075 mm) sieve in the existing mix may be
difficult to recycle. The milling operation creates additional material pass-
ing the No. 200 (0.075 mm) sieve, and it may be difficult to keep the per-
cent of aggregate passing this critical sieve size within the specification
requirements. If the aggregate in the existing mix has a tendency to pol-
ish, then care must be used in trying to recycle this mixture if it is used for
the surface course. Most of these problems can be overcome if small
amounts of RAP are used.
Many additives are being used to improve the performance of asphalt
mixtures. Some of these additives when used in an asphalt mixture may be
difficult or impossible to recycle. Some additives that have been used
include latex rubber, crumb rubber, sulfur, polypropylene, and carbon
black. .No additive should be used in an asphalt mixture if this additive
prevents the mixture from being recycled at a later date. Reheating mixes
that contain sulfur can be dangerous since sulfur combined with asphalt
gives off toxic fumes when heated to high temperatures. There has been
some indication that milling mixes that contain latex rubber may be diffi-
cult. On some projects, the .teeth on the milling machine have been
gummed up with this rubberized asphalt. There is no known way to deter-
------m1·nnh-e-aspfra:l:rcon-rent-rrora:sphait-binder-proper-ties-of old asphalt rub-
ber mixes. The rubber may also result in more pollution during produc-
tion of recycled mixtures. Several years ago asbestos was used as an addi-
tive in some asphalt mixtures to improve the durability. Obviously, the
asbestos has made.it difficult or impossible to recycle those mixtures.
There are many candidate pavements for recycling, but each project
has to be evaluated to determine if recycling is desirable. The recycling
process can be used to reduce reflective cracking, to match curbs and gut-
ters, to reduce construction cost, to maintain better traffic flow during con-
struction, to maintain overhead clearance, and to expose an underlying
deteriorated layer that should be removed or modified._ The existing pave-
ments that obviously should be considered for recycling are those with sig-
nificant rutting, cracking, or on city streets or other places where existing
structures such as curbs and gutters must be matched.
Design Considerations for Recycled Mixtures

A recycled mixture should be designed to produce an asphalt mixture
having the same properties as that in a new mixture. The procedure is basi-
cally the same as for new mixtures with the additional requirement for
asphalt binder analysis. One asphalt cement is used for new mixtures,
whereas two or maybe three could be used· for recycled mixtures (new
asphalt cement, old asphalt cement, and recycling agent).
The first step in designing a recycled asphalt mixture is to obtain sam-
ples of materials to be used in the recycled mixture. The new aggregate
should be sampled in the same way as that for new mixtures. The asphalt
cement or recycling agent should be selected based on past experience for
recycled mixtures within a given area or using some selected specifications
so that once the mix is designed to have the correct asphalt content, it will
also have the correct asphalt cement properties. An existing pavement that
has a very stiff binder material will require a softer asphalt cement or recy-
cling agent to produce the desired properties. Higher percentages of RAP
in a recycled mixture require that a softer asphalt cement or recycling agent
be used. It may be desirable to evaluate 2 to 3 different grades of asphalt
cement so that the one providing desired properties can be selected.
The preferred location for sampling RAP is from an existing stockpile,
but this is not always possible. On many large projects, the RAP obtained
from the project is used to produce the recycled mixture for the same pro-
ject. The material is milled at the same time as the recycled mix is being
placed. The best way to sample the asphalt mixture to be recycled is to mill
the surface for some short length to the same depth that it will be milled
during-corrsmrcti:o~n~Tiri·s-p·ro~cetlare-wi1lproduceRAP--having approxi--- - - - -
mately the same properties as that produced during consJruction. When
the material cannot be milled, one alternative is to cut several cores, remove
all material below the designed depth for milling, and use a small labora-
tory crusher to reduce the size of RAP sufficiently for preparing mixtures.
The cores in this case should be at least 6 inches in diameter to minimize

the number of cores that need to be taken ..
The approximate percentage of RAP to be used in the recycled mix-
ture should be determined. From a contractor's point of view, higher RAP
content generally results in lower mix cost. However, high RAP content
may cause pollution during production or may result in a low quality mix-
ture. Hence, an upper limit on the amount of RAP to be used in a mix-
ture is usually provided in the specifications. After the approximate per-
centage is determined, the percent of RAP and percent of each new aggre-
gate are established such that the specification requirements for gradation
are met. The gradation of the RAP aggregate is determined after the
asphalt cement has been extracted from the RAP mixture. The aged
asphalt cement must also be recovered from the RAP and tested for vis-
cosity at 140°F (60°C). Figure 7-5 is then used to determine the required
grade of new asphalt cement or recycling agent to provide the overall
desired binder properties.
I " IO 7 ........,.Aged asphalt viscosity = 46,600 P=

~ - 4.66 " I0 6 cP
I cP=I0"3 Pa • s
l c = {I.F - 32) /1,8
0
1.8% / 5.8% = .31

.__....10_ . 20
0
. . _.....,_~ ___.._....._ _,__.._. . . .~,;io 2
30 40 50 60 70 80 90 100
WEIGHT PERCENT RECYCLING AGENT IN BLEND
Figure 7-5. Nomograph for Viscosity for Design of Recycled Mixtures
(ASTM D4887)
The next step in the design process is to vary the amount of new
asphalt cement added to the selected combination of RAP and aggregate
blend. The amount of asphalt added is less than that for new (virgin) mix-
tures since the RAP contains a significant amount of asphalt cement.
The prepared samples of recycled mixture are compacted and tested in
the same way as that required for a new mixture as discussed in Chapter 4.
Most agencies require that the recycled mixtures have the same properties
as that required for new mixtures.
The final test that must be conducted on recycled mixtures, but which
is not performed on new mixtures, is a test on the recovered asphalt binder.
The asphalt cement is extracted from the mix by the extraction test (ASTM
D2172) and recovered from solvent by the Abson method (ASTM
D1856). The properties of the asphalt binder from this recycled mixture
should be approximately the same as that for asphalt cement recovered
from new mixtures. This test should be conducted on laboratory prepared
samples as well as samples during production to get an estimate of the
asphalt cement properties. To be accurate this test should be conducted on
asphalt mixtures actually produced in a HMA facility.
The properties of the recovered binder such as penetration and viscosity
must be within the specification limits. The grade of new asphalt cement added
to the recycled mixture must be changed or the percent of RAP used in a mix
should be changed to modify the resulting binder properties.
The mix design on a recycled mix is a preliminary design that must be
finalized during production of recycled mixtures at the HMA facility (211).
Modifications to the mix design ·will likely be required once the plant oper-
ation begins.
Production of Recycled Asphalt Mixtures

Production of recycled asphalt pavement is much the same as that for
virgin asphalt mixtures. The RAP is basically treated as an aggregate with
a slight modification to the way it is fed through the HMA facility.
A batch mix facility has to be modified, as shown in the schematic in
Figure 7-6, to add the RAP directly to the weigh hopper. This modifica-
tion requires the addition of a feeder bin, conveyor belt from feeder bin to
weigh hopper, and the appropriate instrumentation to control the convey-
----u-F-0~1t.-111e-new-ag-grega:1:e-is-s1:1:per-hea:1:ecl-i-n--tlre-dryer-in-rhi:s--pro-cess-prior------
to mixing with the RAP. A higher percentage of RAP will require a high-
er temperature for the new aggregate at the time of mixing since the RAP
is not heated. Since there are practical limits to which the new aggregate
can be heated, the maximum percent of RAP used in a batch plant is
PPER
NEW
BITUMEN
RECYCLING AGENT I
~
OF NEEDED>
\U0;)
Figure 7-6. RAP Addition to the Pugmill in a Batch Facility (after

Col")?s of Engineers)
446
figure 7-7. RAP Addition at Mid-length of the Drum in a Drum Mix

Facility (after Corps of Engineers)
approximately 50 percent of the total mixture. Most recycled mixtures

contain significantly less than 50 percent RAP.
In one type of drum mixer the RAP is fed into the mix at approxi-
mately the drum mid-point (Figure 7-7). The new aggregate must shield
the RAP from the direct flame in this type of drum mixer. If the percent
of RAP in a recycled mixture is too high, there will not be enough new
aggregate to shield the RAP from the open flame; as a result, excessive pol-
lution occurs and the asphalt cement in the RAP is damaged. This is not
a problem if a double drum (drum within a drum) facility is used. The
aggregate is heated in the inner drum, and the RAP and asphalt cement are
added to the heated aggregate in the outer drum. The amount of RAP
material that has been recycled in a drum mix plant is as high as 60-70 per-
cent; however, most agencies restrict the amount of RAP to 50 percent or
less. Again, most projects use significantly less than 50 percent RAP. There
are plants that use microwave heating that can use up to 100 percent RAP.
There are several items that affect plant production that need to be
understood when working with recycled mixtures. Excessive moisture in a
RAP stockpile is more difficult and expensive to remove than that in new
aggregates; In some drum mixers, the RAP is added about halfway down-
the drum, which is further away from the flame and reduces the time
allowed for drying. This excessive moisture requires more fuel for drying,
likely reduces the rate of production, and can lead to stripping of asphalt
from the aggregate if not removed. When a batch facility is used, the heat-
ed new aggregate must transfer its heat to the RAP. This mixing of RAP
and new aggregate takes place primarily in the pugmill and hence does not
allow much time for the moisture to be removed. It is a good practice to
cover the RAP stockpiles to reduce the amount of moisture in the RAP.
The properties of the RAP to be recycled vary depending on depth of
cut with a milling machine. The RAP properties are usually reasonably
consistent if the depth of cut is consistent, but if the depth varies, the RAP
properties also likely vary. When the asphalt mix is removed in two pass-
es with a milling machine, the material obtained from each pass should be
stockpiled separately. If these two layers of material are mixed, the varia-
tion between mix tests will be higher.
Batch and drum mix facilities should be set up to remove any oversize
material. When RAP particles are oversized, they will likely not break down
-------E-,0-m-pl.gt~l:y--d.t1-f-i-Rg-th~m-~i-ag-an.cl.-p-la~t:ment-and-hence-cause-construction
problems as well as performance problems. Most specifications require that
the maximum chunk size not exceed 1½-2 inches (38 to 51 mm).
Existing pavements to be recycled must be evaluated closely to deter-
mine potential problems if they are recycled. Additives that may cause
problems when recycled include sulfur, rubber, and carbon black. Some of
these create problems when milling, some cause more pollution during mix
production, and some have mix and material properties that are difficult to
measure and control.
Recycled mixtures can be designed and constructed to perform as well
as new mixtures. The necessary tests ahd evaluations to ensure quality con-
trol must be performed on a regular b~sis throughout the construction pro-
ject; otherwise, poor performance1is likely the outcome.
ADDITIVES AND MODIFIERS IN HMA

Although asphalt modifiers have been used over 50 years there is a
renewed interest over the past ten years. Tliis resurgence in interest can
primarily be attributed to the following factors:
1. The increased demand on HMA pavements. Traffic volume, traf-
fic loads, and tire pressures have increased significantly in recent
years causing premature rutting of HMA pavements.
2. The recently developed Superpave asphalt binder specifications
based on Strategic Highway Research Program (SHRP). This new
binder specification recommended in March 1993 requires the
asphalt binder to meet the stiffness requirements at high as well as
low pavement service temperatures. Most neat asphalt cements do
not meet these requirements in regions with extreme climatic con-
ditions and, therefore, need modification.
3. The environmental and economic pressure to dispose of some
waste materials and industrial byproducts (such as tires, roofing
shingles, glass, sulfur and ash) as additives in HMA.
4. Public agencies .willingness to pay a higher first cost for pavements
with a longer service life or which will reduce the risk of premature
distress (failure).
All asphalt binders and mixtures do not require modification. The fol-
lowing are some specific technical reasons for using additives and modifiers
inHMA:
------=•_Qb_tain._s_tiffe.unixtur.e.s_at_bigh_ser:v.ice-temper.atur,ewo-minimiz,c------
rutting.
2. Obtain softer mixtures at low service temperatures to minimize
non-load associated thermal cracking.
3. Improve fatigue resistance of HMA mixtures.
4. Improve asphalt-aggregate bonding to reduce stripping or mois-

ture susceptibility.
5. Improve abrasion resistance of mixture to reduce raveling.
6. Minimize tender mix problems during construction.
7. Rejuvenate aged asphalt binders.
8. Replace asphalt cement as an extender.
9. Permit thicker asphalt films on aggregate for increased mix dura-
bility.
10. Reduce Bushing or bleeding.
11. Improve resistance to aging or oxidation.
12. Reduce structural thickness of pavement layers.
13. Reduce life cycle costs of HMA pavements.
14. Improve overall performance of HMA pavements.
Asphalt modifiers are more commonly used in Europe compared to

the United States at the present time (1996). Some European countries
require contractor guarantees for HMA performance for several years. This
has prompted the use of asphalt modifiers to obtain performance and
reduce life cycle costs. Low initial cost is the governing factor in the
United States in awarding the contracts, which discourages the use of
modifiers (higher initial or first costs) unless specified (22). However, some
state highway agencies have now begun to specify modified asphalt
binders.
Ideal Pavement Binder

An ideal HMA pavement binder should be modified to achieve the
following improvements as illustrated in Figure 7-8 (3.Ql:
1. Lower stiffness (or viscosity) at high construction temperatures to

facilitate pumping of the liquid asphalt binder, and mixing and
compaction of HMA.
2. Higher stiffness at high service temperatures (summer) to reduce
rutting and shoving.
3. Lower stiffness and faster relaxation properties at low service tem-
peratures (winter) to reduce thermal cracking.
4. Increased .adhesion between asphalt binder and aggregate in the
presence of moisture to reduce stripping.
High service Spraying

temperatures pumping and
Low service
" mixing
temperatures
•--~ temperatures 1-----,.---1►
en
en Ideal modified Compaction
-
Cl)
C Asphalt-aggregate · temperatures
mix
!-
-8
>,
"in
en
>... Conventional_/
Cl)
"C
C
asphalt-aggregate
mix
ixi •••
, Practical •
limits of
modification
-50 0 100 200 300

Temperature (F)
Figure 7-8. Stiffness Characteristics of Conventional Binder and Ideal

Modified Binder (312)
In the past, some improvements in HMA characteristics have been

obtained through mix design and by changing grades of asphalt cement.
The use of modifiers and additives has now opened up new possibilities
(22). However, several key questions about modified asphalts must be con-
sidered (31).
(a) What improvement is needed? It must be decided which specific
characteristic needs to be enhanced. (b) What modifier will work best?
Not every modifier is suited for every situation. Sometimes, one property
of the HMA is improved but another is compromised. For example, some
liquid anti-stripping agents reduce moisture damage but have a tendency
_ _ __..,o....s.often_rhe_asphal:t...binder....thuS-corn-p-wm-is-i-Rg-th~m-t-r~sista0.€e-0-:f:..th@-----
HMA. (c) How should the modifier be incorporated into the binder
and/ or mixtures? This is likely to be a key issue as more modified systems
become available. (d) Should compatibility be considered? Everyone will
answer yes to this question but unfortunately, the technology is not fully
developed at this time to check the compatibility in all instances. (e) How
should the binder be stored to maintain its properties over a length of time?
(f) How should the binder be specified? (g) Does the modifier affect rou-
tine test results? For example, an accurate determination of asphalt con-
tent in HMA may not be possible when certain modifiers are incorporat-
ed in HMA. (h) Can the HMA be recycled? For example, this question
has been raised about the HMA containing crumb rubber modifiers. (i)
What is the effect on life cycle cost? Are the initial increased costs justified
based on life cycle cost analysis? (j) What are the effects of the modifier on
worker health and safety?
Classification
It is possible to classify additives and modifiers in different ways.
However, an excellent generic dassification system was first suggested by
Terrel and Walter (32). A modified version of this system is shown in Table
7-6. Several additives exist un.der each category; however, this is a dynam-
ic list that is rapidly changing. A discussion of each additive or modifier
type follows.
Fillersj
Mineral fillers include mineral dust from the crushing and screening of
aggregates (including baghouse fines), lime, portland cement, and fly ash.
Other materials such as carbon black and sulfur have been used primarily
to modify asphalt binder properties but they do have a role of filler. Fillers
may be used to (1) fill voids and hence reduce optimum asphalt content,
(2) meet specifications for aggregate gradation, (3) increase stability, and
(4) improve bond between asphalt cement and aggregate. Typically, an
increase in filler lowers the optimum asphalt content, increases the densi-
ty, and increases the stability (33).
Mineral fillers have been used primarily to fill in the voids between the
aggregate particles and meet specified gradations for HMA. However, it has
been documented extensively in the literature that mineral fillers such as bag-
------,h01:1Se-fu:i©s-<ie-affe-Gt--the-h~ha¥i.or-0£asphalt-binders-and-HMA-mixtures C31.,.
.35). Mineral fillers cause a general stiffening or reinforcing effect in HMA mix-
tures. However, the extent of stiffening effect is generally dependent upon
Rigden void content (voids between the fine particles in dry compacted state)
which in turn is affected by mineralogic type, size distribution, particle shape,
Table 7-6. Generic Classification of Asphalt Additives and

Modifiers
TYPE GENERIC EXAMPLES ~
I. Filler • Mineral filler: _crusher fines

lime
portland cement
tly ash
• Carbon black
2. Extender • Sulfur
• Lignin
POLYMERS
3. Rubber • Natural rubber

a. Natural latex • Styrene-butadiene or SBR
b. Synthetic latex • Polyduoroprene Latex ~-
c. Block copolymer • Styrene-butadiene-styrene (SBS), Styrene-isoprene-sytrene (SIS)
d. Reclaimed rubber • Crumb rubber modifier
4. Plastic • Polyethylene/Polypropylene
• Ethylene Acrylate Copolymer
• Ethylene Acrylate Copolymer
• Ethyl-vinyl-acetate (EVA)
• Polyvinyl cluoride (PVC)
• Ethylene propylene or EPDM
• Polyolefms
5. Combination • Blends of polymers in 3 & 4
6. Fiber • Natural: Asbestos

Rock Wool
• Man-made: Polypropylene
Polyester
Fiberglass
Mineral
Cellulose
7. Oxidant • Manganese salts
8. Antioxidant • Lead compounds

• Carbon
• Calcium salts
9. Hydrocarbon • Recycling and rejuvem1ting oils

• Hard and natural asphalts
IO. Antistripping Agents • Amines

• Lime
11. Waste Materials • Roofing shingles

• Recycled tires
• Glass
12. Miscellaneous • Silicones
.-o·ekilrgi;atcium-ch:roride gfanuies
and surface texture of the mineral fillers (31:). Some baghouse fines, primarily
those containing clay, can increase the potential for stripping in Hl\1A mixtures.
All fillers must be fed to the asphalt mixture consistently and in cor-
rect proportions; otherwise, the mix properties are adversely affected.
Excessive amounts of filler usually reduce the VMA to a point that suffi-
cient asphalt content for a durable mix cannot be added. High filler con-
tent also increases the aggregate surface area and thus greatly reduces the
asphalt film thickness. Some specifications place limits on the dust to
asphalt cement ratio. A typical specification requires the dust to asphalt
cement ratio to be benveen 0.6 and 1.2 by weight. An ideal specification
should require such a ratio based on volume and should take Rigden voids
(which generally control the stiffening effect) into account.
Although a mineral filler, lime is primarily added as an antistripping
agent to Hl\1A to minimize stripping problems. Lime will be discussed
later as an antistripping agent.
Carbon black can be considered as a "microfiller" because. it is very
fine. Different grades and types are available because carbon black is wide-
ly used in other industries, such as reinforcement for rubber tires. The
grade which has been used in Hl\1A consists of sub-micron size particles
(about 10 14 particles per gram). Since this material is very fine it could read-
ily become airborne and cause air quality problems.· Therefore, carbon
black is usually combined with a high boiling point maltene oil to form
pellets; which are easier to handle in Hl\1A facilities. Recommended con-
centrations of carbon black pellets range from 10 to 15 percent by weight
of asphalt cement. To obtain desirable low temperature properties in cold
climates while maintaining desirable high temperature properties, some
highway agencies have used a softer grade of asphalt cement compared to
the common grade generally used. The carbon black can be supplied in
polyethylene bags or in a bulk container. It can be preblended with asphalt
cement either at the refinery or at the Hl\1A facility. A 10 to 20°F (6° to
11 °C) increase in mixing temperature may be needed to compensate for
the stiffening effect of the carbon black. It has been claimed that the car-
- - - - - - v..on..black=.eeinfo.rced...as.phalLcemen.Lb.indecimp.r.oyesJfMA.Jesistance to
rutting, improves Hl\1A resistance to wear and abrasion, and reduces
oxidative hardening and embrittlement of binder with exposure to weath-
ering. The use of .carbon black in Hl\1A has been evaluated by various
researchers (3..6, 32).
454 Hot. Mix Asphalt Materials, Mixture Design and Construction
Extenders
The concept of extending asphalt binder became prominent after the
1973 oil embargo. It appeared at that time that the price of asphalt cement
would continue to increase and that asphalt cement might become scarce.
The Federal Highway Administration (FHWA) promoted the investiga-
tion of extending asphalt binders by partially substituting with sulfur and
lignin. Complete substitution of asphalt cement with sulfur (sulphlex) was
also evaluated (18). Both sulfur and lignin are largely byproducts of other
industries: sulfur from natural gas desulfurization and lignin from pulp
and paper production. The supply and price of sulfur have fluctuated
widely since the 1973 oil embargo, so the use of sulfur has been sporadic.
At the present time (1996), because of lowensphalt cement prices and ris-
ing sulfur prices, sulfur can not compete with asphalt cement.
Commercial use of lignin in HMA has also been negligible. Much of the
lignin is burned in-plant for fuel.
Sulfur is an abundant, non-metallic chemical element. It is a yellow,
crystalline solid at room temperature. It melts into a free-flowing,
straw-colored liquid at about 240°F (116°C). Above 310°F (154°C), liq-
uid sulfur becomes highly viscous. Within the range of 250°F to 290°F
(121 °C to 143°C), usually employed for HMA production and construc-
tion, the viscosity of sulfur extended asphalt (SEA) binder is lower than
that of asphalt cement alone. Sulfur has been used to modify HMA base,
binder, and surface courses.
Since the specific weight of sulfur (specific gravity is about 2.07 at
68°F or 20°C) is about twice that of asphalt cement, the specific weight of
SEA binder increases as the sulfur content increases. Sulfur contents have
ranged between 10 and 50 percent by weight of the binder; however, the
most frequent substitution rate has been between 20 and 35 percent by
weight (approximately equal to 10 and 17.5 percent by volume) of the
binder. Sulfur may be delivered in either solid or liquid form, usually in
bulk by rail car or truck. Sulfur has to be added to the asphalt cement just
prior to the HMA mixing operation. For a"liquid sulfur system, it is nec-
essary to have a separate liquid sulfur storage tank, sulfur pump, metering
system, and in some cases a preblending unit. For a solid sulfur system, a
storage pad area, a conveyor system, a small crusher, and an auger or a van
1-----eecl.erancl. 15lower need to be arranged. Vanous methods and eqmpments
have been developed for combining sulfur and asphalt cement prior to
their addition to the aggregate at the HMA facility (22). Good tempera-
ture control combined with adequate ventilation are very important to pre-
vent any build-up of sulfur dioxide (SO 2) or hydrogen sulfide (H 2 S) gases.
Guidelines for design, quality control, and construction of SEA pavements

are given in reference (32). Evaluation of SEA binders and mixtures has
C'" been reported by various researchers (1Q 1.L 12). Generally, the perfor-
mance of SEA pavements has been found to be equal to conventional
HMA pavements. It is believed that only a portion of sulfur modifies the
binder, the remaining portion crystallizes out and acts like a filler.
Hot molten sulfur poses the same dangers as hot asphalt binder or any
other hot liquid. Sulfur will burn if ignited. All sources of ignition such
as smoking, open flames, and sparks should not be permitted near liquid
sulfur. It is very important to keep the temperature of SEA mixtures below
300°F (I 49°C) at all times. Above this point, emissions of hydrogen sul-
fide, sulfur dioxide, and sulfur vapor may increase rapidly and pose a health
hazard.
Lignin ras not been used as an asphalt extender in commercial HMA
production. However, it has been evaluated in the laboratory as a poten-
tial substitute or extender of asphalt cement (11).
Polymers
It can be seen in Table 7-6 that both rubber and plastic materials are
referred to as polymers. Simply stated, a "polymer" is a very large molecule
made by chemically reacting many (poly) smaller molecules (monomers)
to one another in long chains or clusters. The sequence and chemical
structure of the monomers determine the physical properties of the result-
ing polymer. Copolymers consist of the combination of two different
monomers which can be either in a· random or block arrangement. For
example, polystyrene is a hard, brittle plastic whereas polybutadiene is soft
and rubbery. If these two distinctly different monomers are randomly
mixed and reacted together, a new polymer called copolymer with very dif-
ferent properties is created. Polymers can be engineered to obtain a broad
range of physical properties. However, they can be divided into two gen-
eral categories: elastomers (rubber) and plastomers (plastic).
Elastomers or rubbers used as asphalt modifiers include natural rubber,
styrene-butadiene rubber (SBR) latexes or SBR, styrene-butadiene-
styrene (SBS) block copolymers, styrene-isoprene-styrene (SIS) block
----·copolymers,polychforoprene-1-a:texes,and-crumb-rubber-modifiers (from
ground tires). Elastomers resist deformation from applied stress by stretch-
ing and recovering their shape quickly when stress is removed. These poly-
mers add very little strength to the asphalt cement until they are stretched.
However, their tensile strength increases with elongations (~.
Plastomers or plastics used as asphalt modifiers include polyethylene,

polypropylene, ethyl vinyl acetate (EVA), polyvinyl chloride (PVC), ethylene
propylene (EPDM), and polyolefins. Plastomers have a tough, rigid, three
dimensional network which is resistant to deformation. These polymers
exhibit quick early strength on loading but may fracture under strain (14).
More flexible and resilient HMA pavements usually result when elas-
tomers are used for modifying asphalt cements. On the other hand, plas-
tomers generally increase the stiffness moduli of HMA pavements.
However, results of modifications in both cases are highly dependent upon
the concentration, molecular weight, chemical composition, and molecu-
lar structure of a specific polymer. The crude source, refining process, and
grade of the neat asphalt cement binder are also equally important.
Many combinations of elastomers (rubber) and plastomers (plastic) are
possible to provide special properties to the modified asphajt binder to
meet a desired situation. '
Rubber. As shown in"Table 7-6, numerous rubber materials (or elas-

tomers) are available for modifying asphalt cement binder. Many of the
materials are available commercially under different trade names which
may be more familiar. Specific details of these products are available in
Reference 2_2. Rubber materials are very complex in nature and may not
impart their properties as a neat polymer when blended with asphalt
cement. Moreover, their properties may be diluted or even changed when
the modifier is used in HMA, especially in thin films. Each modifier may
react differently with different asphalt cements. Whether a particular com-
bination will yield improvements in the desired property is very difficult to
predict. It becomes necessary to test the modified asphalt binder.
The primary purpose of using rubber modifiers in HMA is to obtain
a stiffer HMA at high service temperatures, a more elastic HMA to resist
fatigue cracking at intermediate service temperatures, and a lower or
unchanged stiffness at low service temperatures to resist thermal cracking.
It is hoped that•Superpave mix design and evaluation procedures can be
used in the future to verify if the intended purpose can be achieved
through the use of rubber modifiers.
SBR latex is an off-white liquid which is supplied in 55-gallon
(210-L) drums or bulk. The dosage ranges from 2 to 3 percent of dry
1-----m·bfrers-otrd:-s~by-wetglrr-ofasp"hatrcement. Latex can be add:ed1-=to=-1=h-=o=-t-----
asphalt cement in a blending unit at the refinery and the modified binder
is supplied to the HMA facility. Water is vaporized in this process and the
resulting steam is removed through a cyclone separator. Latex can also be
added at the HMA facility. The sequence of addition is critical to obtain
desired results. The aggregate is fully coated with neat asphalt cement
prior to introducing the latex. For batch facilities, latex is normally intro-
duced 8-10 seconds after addition of neat asphalt cement. The total wet
mix time is about 10 seconds longer than the conventional wet mix time.
In case of drum mix facilities, latex should be introduced into the drum at
a point between the introduction of neat asphalt cement and the discharge
end of the drum. No significant changes in HMA laydown and com-
paction are normally required when SBR latexes are used.
SBS block copolymers are available in solid form as pellets, crumb, or
ground material in bags or bulk. The usual dosage rate is about 5 percent
by weight of asphalt cement. High shear mixing equipment is used for
blending this modifier with hot asphalt cement maintained at 350-380°F
(177-193°C). The modified asphalt binder can be prepared at the refin-
ery and delivered to the HMA facility like a conventional asphalt cement.
No agitation is required during storage. Since most mix design procedures
require the asphalt binder to be within specified viscosity ranges during
mixing and compaction, it will probably be necessary to increase the mix-
ing and compaction temperatures while conducting laboratory work.
However, field experience has shown no unusual difficulties at conven-
tional mix temperatures when processing and compacting the HMA pro-
duced with this high viscosity modified binder. It is believed that the ease
of handling is due to the relatively higher shear forces experienced during
manufacturing in the field compared to the lower shear forces used in lab-
oratory viscosity measurements. Therefore, supplier's recommendations
for mixing and compaction temperatures should be followed in the labo-
ratory and field.
Reclaimed rubber is primarily obtained from tires. About285 million
tires are discarded every year in the United States. Of these, about 55 mil-
lion are retreaded or reused (resold), and about 42 million are diverted to
various alternative uses such as combustion for generating power and addi-
tives for HMA mixes. The remaining 188 million tires are added to stock-
piles, landfills or illegal dumps. According to EPA estimates, 2 to 3 billion
discarded tires were available in 1992 (di'). Several states have enacted leg-
islation to regulate the scrap tire problem.
Crumb rubber obtained from tires can either be ambient ground
(grinding at room temperature or above) or cryogenically ground (grind-
i.--ng-b-efow-embrittlemenrremp-era:mre-;-liqui_--d-nitrogen--is -often - used).
Ambient ground crumb rubber has a sponge-like surface. Due to very
high surface area this rubber reacts with asphalt cement reasonably fast.
Some cryogenically ground rubbers may have undesirable particle mor-
phology (structure). This process produces clean flat surfaces which, in
turn, may reduce the reaction rate with hot asphalt cement. In one study,
.cryogenically ground rubber gave lower elastic recovery compared to the
ambient ground rubber (1§;.
The use of crumb rubber to modify asphalt cement has been devel-
oped over the past 25 years. Crumb rubber is primarily used in HMA
mixes by two processes; generally referred to as the wet process and the dry
process:
l. Wet Process (Asphalt-Rubber)

The wet process blends the crumb rubber with the asphalt
cement prior to incorporating the binder into the project. The
modified binder is commonly called "asphalt-rubber". Generally,
18-26 percent crumb rubber (16 mesh or smaller) by weight of
"asphalt cement is reacted with asphalt cement at 375 to 425°F (191
to 218°C) for 1 to 2 hours. The blend is formulated at elevated
temperatures to promote potential chemical and physical bonding
of the two constituents. The first technology which applied the
"wet process" is called the "McDonald Process". This process is
also used for constructing stress absorbing membrane (SAM) and
stress absorbing membrane interlayer (SAMI), and manufacturing
crack sealers. SAM is a seal coat which uses asphalt-rubber as a
binder. When SAM is placed as an interlayer between HMA cours-
es it is called SAMI.
2. Dry Process (Rubber-Modified Mix)

This process mixes the crumb rubber with aggregate before
incorporating the asphalt cement. About 3-5 percent of coarse
rubber particles (1/16-1/4 inch or 1.6-6.4 mm maximum size) by
weight of aggregate are generally used. The natural aggregate is
usually gap-graded to accommodate the rubber particles as aggre-
gate. The amount of crumb rubber used in the dry process can be
2-4 times that used in the wet process. The first application of the
"dry process" in the United States was called the "PlusRide
Process". It was claimed that ice debonded easily from the pave-
ment surface because of higher than usual resiliency of the rub-
ber-modified mixture.
-o-produce :m-a:cceptableaspl1alt=rul5ber 15inder using the wet
process, it is necessary to establish the digestion temperature and
time for a specific combination of asphalt cement and crumb rub-
ber. Viscosity of the blend is checked at different time intervals
during the blending and digestion process. Achieving a reasonably
constant viscosity indicates that the initial reaction is nearly com-

plete and the binder is ready to use. This initial reaction is not well
understood, but appears to be due to a chemical and physical
exchange between the asphalt cement and rubber particles in which
the rubber swells, causing an increase in viscosity. Continued mix-
ing of asphalt cement and rubber after initial reaction can begin to
reduce the viscosity of the blend as the rubber particles break down
during mixing. However, breakdown of rubber particles is not
rapid, and may require several hours at high temperatures.
If the asphalt-rubber blend is too viscous, an extender oil is usu-
ally added. However, if less than 10 percent crumb rubber (by
weight of asphalt cement) is used, extender oil is not needed.
The asphalt-rubber blend should be used as soon as possible
after the initial reaction. Until used it should be recirculated con-
tinuously. Blends have been allowed to cool in storage tanks and
reheated prior to use without difficulty.
When asphalt-rubber binder is used in HMA mixes the follow-
ing mixing temperature ranges have been used:
Dense-graded HMA mixes: 325-375°F (163-191 °C)
Open-graded HMA mixes: 275-325°F (135-163°C)
Stack emissions can be higher because of the elevated mixing temper-
atures. HMA mixes containing asphalt-rubber binder should not be sub-
jected to prolonged storage in a silo. Processing the mix through a surge
silo is acceptable. Diesel fuel should not be used on truck beds as release
agent because, besides contaminating the HMA mix, it makes the mix stick
more to the bed. Use of lime water, soap solution or silicone emulsion is
recommended instead.
It is necessary to maintain the mix temperature within the desired
range during placement of the mix. Static steel wheel rollers are generally
used for compaction. Rubber-tired rollers tend to pick up the mat.
Vibratory rollers tend to tear and shove the mat. If the roadway has to be
opened to the traffic right after compaction and the traffic has a tendency
to pick-up the mat, it is recommended that 1-2 pounds of concrete sand
per square yard of the mat surface be applied to the surface.
Recyclability of the HMA containing crumb rubber is a very impor-
-----t·a:r1:t-i-sSt1-e-wh:ieh-m1:1.-s-t-be-a-<id-r-es-sed-befere-1±si:ng-serap-tire-crumb rubber
on a large scale. There is only limited experience at the present time
(1996) in recycling such mixes. It is quite likely that recycling these mixes
in the future could pose air quality problems during production. What
will happen to the asphalt-rubber binder during recycling? How will the
recycled mix be designed? If these concerns are not addressed it may not
be possible to recycle the HMA mixes containing crumb rubber. In that
case, the disposal of RAP (reclaimed asphalt pavement) may become a
much greater problem than disposing the tires.
The "McDonald process" (wet process) and "PlusRide" (dry process)
as discussed earlier are patented processes. However, the original patent on
the "McDonald process" expired in 1992. Recently some newer concepts
which include the generic dry process, chunk rubber asphalt concrete and
continuously blended asphalt rubber, have been introduced (15).
The generic dry process allows incorporation of crumb rubber into
available "generic" aggregate gradations rather than gap-graded aggregates
required in the PlusRide system. Experimental field applications of the
generic dry process have been made in NewYork, Florida, Iowa, Kansas,
Oregon, and Illinois (15).
The continuous blending asphalt rubber concept has been used by
Florida DOT. It is a wet process which uses 5 to 15 percent of fine crumb
rubber (180 µ or No. 80 mesh) to avoid reaction tanks, and to facilitate
continuous blending. The mixing of crumb rubber and the asphalt cement
is accomplished in a self-contained portable blending/metering unit.
Plastic. Numerous plastic materials (plastomers) are available for

modifying asphalt cement binder as shown in Table 7-6. Specific details
of brand name products are given in Reference 22_. The handling, storage,
and use of plastic polymers is very similar to that for rubber polymers.
Low density polyethylene (LDPE) has been used as a modifier. Either
virgin or recycled polyethylene can be used. The modified binder is a
bi-phase binder system containing polyethylene and neat asphalt cement.
The polyethylene content is typically 4-6 percent by weight of the modi-
fied binder. A patented high shear blending process using special manu-
facturing equipment (mounted on a mobile blender) has been used at the
HMA facility to prepare the modified binder. The modified binder must
be stored with agitation to prevent stratification and migration of the poly-
ethylene upwards within the liquid asphalt cement. Since the modified
binder has a high viscosity, the mixing and laydown temperature should be
about 35°F higher than normally used for unmodified HMA mixes.
Compaction of modified HMA mixes is largely ineffective below 270°F
-----<·-Hl.£Q-beea1;1se-1:he-p0lye1:hyl:eae-aticl-i:1:i:ve-erysta±l-izes-bel-ow·this-temperatur.-.------
Ethyl vinyl acetate (EVA) is available as an asphalt modifier in the form
of translucent to white, free-Bowing pellets in bags or bulk. The dosage
rate varies from 2 to 5 percent by weight of asphalt cement. The EVA is
added to hot asphalt cement maintained in the 300-340°F (149-171°C)
range. Only moderate agitation (such as low shear, stirred pot) or circula-
tion is necessary for proper blending. Blends can be stored for several
weeks without any separation. The EVA must be compatible with the neat
asphalt cement to obtain the desired properties.
Combinations. The combination of two or more polymer types (such

as rubber and plastic) may cause improvements in the performance of
HMA pavements otherwise not possible with only one modifier. For
example, a plastic polymer may improve the rutting resistance in summer
but make no significant difference in cold weather thermal cracking.
Adding a rubber polymer to the plastic polymer may enhance the cold
weather properties. However, two different polymer types may not be
chemically compatible with each other and the combination could even be
detrimental. Therefore, it is necessary to investigate the combinations
thoroughly for adverse side effects. The cost of using two or more poly-
mers may also be prohibitive (22).
Fibers
The use of natural fibers such as asbestos has been well known for
many years. However, since asbestos was identified as a health hazard, it is
no longer in use in the U.S. Manufactured or synthetic fibers have been
promoted to replace asbestos. Fibers are most commonly used as rein-
forcement or stiffener in the manufacture of roofing and industrial water-
proofing membrane systems. They have the potential of reinforcing and
improving the tensile strength and cohesion ofHMA mixtures. Fibers also
permit higher asphalt contents than conventional HMA mixtures without
any significant draindown problems. This is especially important for
open-graded friction courses and stone matrix asphalt mixtures. The
resulting thick binder films from high asphalt contents increase the dura-
bility of paving mixtures. Fibers are generally added to the mixture at the
HMA facility. Most manufacturers supply fibers in easily melted plastic
bags because, being small and finely divided, they cannot easily be handled
in bulk. For drum mixers, pneumatic systems are available for air blowing
the fibers into the drum. The fiber manufacturers usually supply or oth-
erwise assist the HMA contractors with special equipments to handle their
------peeifie-fi.ber-s.-'J'.B.e-H-lvlA-m-Bf:fl:l-Fe-Gen-1,ainiag-filiC:lrs-may-be-difficult to lute
and place by hand.
Polypropylene fibers about 10 mm in length have been used as an
asphalt reinforcement. Typically, 0.3 percent fibers by total weight of
HMA are added. The specific gravity of polypropylene fibers is about
0.91. The HMA mix temperature should not exceed 300°F (149°C) and
the aggregate temperature should not exceed 290°C (143°C) to prevent
polypropylene fibers from softening or melting. In a batch facility, good
dispersion of the fibers is usually obtained by a 10-second dry mixing cycle
followed by the usual 35-second wet mixing cycle. The finished HMA
mixture should be examined for presence of clumps of fiber. If clumps are
present, the wet mixing cycle should be increased in 5-second increments
until a uniform .niix is obtained. For drum mix facilities, feed systems for
delivering fiber at a metered rate are available from the fiber suppliers.
Polyester fibers of different lengths have been used to reinforce HMA
mixtures. The length varies from 1/2 inch (13 mm) for base course mix-
tures to 1/50 inch (1/2 mm) for fine-graded mixtures. The specific grav-
ity of these fibers varies from 1.32 to 1.40.-The minimum melt tempera-
ture and the minimum ignition temperatures are usually 480°F (249°C)
and 1000°F (538°C), respectively. Typical dosage rate varies from 3 to 7.5
pounds per ton (1.5 to 3.8 kg per Mg) of HMA mixture (about 0.15 to
0.40 percent by weight of HMA mix) depending upon the application.
Additional asphalt cement binder ranging from 0.2 to 0.3 percent by
weight of the HMA mix needs to be incorporated in fiber-modified mix.
Polyester fibers are handled and mixed similar to polypropylene fibers.
Mineral fibers are used in Europe in dense-graded, open-graded, and
stone matrix asphalt (SMA) mixtures. Typically, mineral fibers are manu-
factured at high temperatures from diabase as a raw material. The usual
dosage rate of mineral fibers in SMA is 0.4 percent by weight of total mix.
A maximum fiber length of 6 mm is generally used in SMA.
Recommended specifications for mineral fibers for SMA and how they are
introduced in HMA facilities, are given in Reference£ Mineral fibers are
supplied in prepackaged condition or in bulk. Methodology and equip-
ment for metering bulk (loose) fibers into batch and drum mix facilities
have been developed and are available.
Cellulose fibers are extensively used in SMA in Europe with the pri-
mary purpose of increasing asphalt content without any draindown prob-
lem. These fibers have been used in many SMA field trials in the United
States. Specifications for cellulose fibers and guidelines on handling are
given in Reference ...iZ Typically a maximum fiber length of 6 mm is used.
---~A-bout..20....p_ercenr_o£fihers_p.ass__thr.ough-N.o.-I-O.O-(0..-l.;;-m-m....si<W.e)-.-Th.ec..------
dosage rate for cellulose fibers in SMA is usually 0.3 percent by weight of
the total mix. Cellulose fibers are supplied in prepackaged condition or in
bulk. The methodology and equipment for metering bulk (loose) fibers
into batch and drum mix facilities have been developed and are available.
Oxidants
Sometimes it is desirable to increase the stiffness ofHMA mixtures for
improved strength. This may be appropriate for low stability mixes such
as sand-asphalt. One approach is to use a harder asphalt cement at the
begin~ing. However, this approach may reduce the long term perfor-
mance. The other approach is to use oxidation catalysts which stiffen the
asphalt cement binder after the HMA is placed on the pavement. An
oil-like liquid containing a manganese compound was used in field trials
throughout the U.S. in the 1980s. Typical concentrations range from 1 to
4 percent by weight of binder. Some agencies chose to use a softer grade
asphalt cement than normally required (for example, AC-10 in lieu of
AC-20). Initially, this modifier lowers the viscosity of asphalt cement at
77°, 140°, and 275°F (25°, 60°, and 135°C). In some cases, this softening
may prevent the early placement of heavy traffic on the pavement. The
viscosity of the modified binder continues to increase on exposure to air for
some weeks. The HMA becomes significantly stiffer after this "cure" peri-
od. Early field trials in the U.S. showed excessive stiffening that resulted
in premature thermal cracking. Later formulations and use of soft asphalt
cement reduced this problem. The use of this modifier is not prevalent at
the present time (1996). It appears that this modifier may be more appro-
priate for use in base courses rather than surface courses.
Antioxidants
Asphalt cements undergo oxidative hardening during HMA produc-
tion and construction, and subsequently during service. Antioxidant addi-
tives consisting of lead or other compounds are available to minimize
oxidative hardening of asphalt cements, thereby increasing-the durability of
HMA mixtures. The experience with the antioxidant additives in HMA is
rather limited, therefore, no guidelines are available. Most of the research
work has been done in the laboratory and there is a lack of adequate field
experience and validation. Some field work has been donein ~11stralia to
retard oxidative hardening of binders used in chip seals. Because the
short-term aging of the asphalt cement during HMA production is mini-
mized due to the presence of antioxidant, there is a potential for tender mix
problems during laydown and compaction operations.
I
Ii,
Hydrocarbons
It may not always be poss~ble to obtain an asphalt cement of desired
properties (such as viscosity, penetration, and ductility) for a particular
project. However, modification of the available asphalt cement is possible
by adding either softer or harder hydrocarbon materials. For example,
aged asphalt binder in the recycled asphalt pavement (RAP) needs to be
softened or rejuvenated by the use of recycling agents. On the other hand,
some asphalt cements need to be hardened with hardening agents when
used in HMA for high stress areas such as intersections and access to toll
booths. A discussion of recycling agents and hardening agents consisting
of hydrocarbon materials follows.
Recycling Agents. These hydrocarbon materials are also referred to as

rejuvenating agents, softening agents, and recycling modifiers. Their pri-
mary purpose is to restore aged asphalt cement to requirements of current
asphalt cement specifications. In most states "soft asphalt cements" with
relatively lower viscosities are used exclusively as recycling agents.
However, some states especially in the western U.S. permit the use of soft-
ening agents or rejuvenating agents. Softening agents lower the viscosity
of the aged asphalt cement. Rejuvenating agents attempt to restore the
physical and chemical properties of the aged asphalt cement to that of a
new asphalt cement. Some examples of softening agents and rejuvenating
agents follow (22).
1. Softening agents:
(a) Asphalt flux oils (generally blended with asphalt cement to
reduce the viscosity)
(b) Lube stock (a fraction of crude oil that has viscosity similar to
lube oils)
(c) Lubricating or crankcase oil (usually highly aliphatic)
(d) Slurry oil. Use of these bottoms from the catalytic cracking
process is not recommended, due to potential health hazards.
2. Rejuvenating agent:
(a) Lube extracts (highly naphthenic or aromatic fractions
removed from lube stock by solvent extraction)
(b) Extender oils (aromatic oils from lube stock, mostly used for
--------<e•Xtmding-asphal.t=rubber-blendSci-----------------·--,
ASTM Designation D4552 Standard Practice for Hot-Mix Recycling

Agents (48) can be used to specify the desired grade of the recycling agent
based on viscosity at 140°F (60°C). This standard specifies a minimum
flash point of 425°F (218°C) for all grades to ensure safety. Maximum vis-
cosity ratio of 3 and maximum weight change of 3 or 4 percent, after the
neat asphalt cement is subjected to thin film oven test or rolling thin film
oven test, are intended to preclude light materials that would evaporate
quickly in the HMA facility or from the HMA pavement.
Most recycling agents are viscous liquids at room temperature. For hot
mix recycling, a storage tank, pump, and accurate metering system is
required at the HMA facility. A separate injection pipe is needed in the
event only a recycling agent is used without asphalt cement. Careful con-
trol of the quantity of the recycling agent is essential because small devia-
tions in dosage rates can cause large deviations in HMA behavior and per-
formance (22). The structural properties of recycled HMA are generally
satisfactory when the proper recycling agent is selected (1.2,_ .5.Q).
Hardening Agents. As mentioned earlier, hydrocarbon hardening agents

are sometimes used to modify the physical properties of neat asphalt
cements for certain applications such as use in high stress areas. Examples
of hardening agents are as follows (22):
1. Natural asphalts:
(a) Resin (Gilsonite)
(b) Trinidad Lake Asphalt
2. Refinery produced:
(a) Solvent de-asphalted bottoms
(i) Propane extracted pitch
(ii) "Rose unit" asphaltenes (butane extraction)
(b) Industrial grade asphalt cements
Gilsonite is a black, brittle, hard, granular asphalt material mined in
Utah. It has a penetration of 0 at 77°F (25°C), its softening point is
300-315°F (149-157°C), and its asphaltene content ranges from 60 to 75
percent. It has been used in HMA in high stress areas such as intersections,
bridge deck overlays, and access to toll booths. Gilsonite is substituted for
10-15 percent of the asphalt cement. The mix designs and optimum
asphalt content are normally not changed with the use of Gilsonite.
Gilsonite is supplied in plastic bags or in bulk. It can be preblended with
asphalt cement. The blend does not require constant agitation. For batch
facilities, Gilsonite can also be dry mixed with hot aggregates 325-340°F
-----c·163.:..-r71-o-C) oeforeintroductron oraspna:lrcement:-However, the dry
mix cycle is increased by 15 seconds to ensure uniform distribution. For
drum mix facilities, Gilsonite can either be preblended with asphalt
cement or introduced by RAP feed conveyor. Conventional methods and
equipment are used for Gilsonite-modified HMA mixes.
Trinidad lake asphalt (TLA) is one of the natural asphalts which were
discussed in Chapter 2. It contains 50 to 57 percent asphalt. The remain-
der is colloidal clay of volcanic origin, plus some inert organic matter.
Trinidad lake asphalt has a specific gravity of about 1.41, a softening
point in 200-207°F (93-97°C) range, and a penetration in 3-10 range.
It has been used in HMA for high stress areas such as intersections and
access to toll booths. It has also been used in Gussasphalt (51) which is a
pourable mastic asphalt concrete. A blend of 25% TLA and 75% AC-20
asphalt cement is recommended for heavy duty pavements. TLA is usu-
ally supplied in 520-pound (235-kg) drums for crushing and blending
' in suitable tanks. TLA can also be supplied in powdered form, which can
be added through filler lines. It can be preblended with asphalt cement;
however, the blend should be stored in suitaBle stirred tanks. Pumps with
bearings exposed to the blend, which contains fine silica based filler,
should not be used.
Antistripping Agents
Antistripping (AS) agents are used to minimize or eliminate stripping
of asphalt cement from aggregate in HMA mixtures. Stripping is discussed
in Chapters 4 and 8. Both liquid AS additives and lime additives are used
to resist stripping.
Liquid Antistripping A4ditives. Most of the liquid AS agents are

surface active agents which, when mixed w:th asphalt cement, reduce sur-
face tension and, therefore, promote increased adhesion to aggregate. The
chemical composition of most commercially produced AS agents is pro-
prietary. However, the majority of AS agents'currently in use are chemical
compounds that contain amines (.52. .53.). These AS agents must be"heat
stable", that is, they should not lose their effectiveness when the modified
J
asphalt cement is stored at high temperatures for a prolonged period of
time.
The simplest and most economical way is to mix the AS agent with the
asphalt cement in a liquid state prior to mixing the asphalt cement with the
aggregate. Although this method is most commonly used, it is inefficient
because only a portion of the AS agent reaches the aggregate-asphalt
---fl·ement-i-nt-erf-aee:-Bi-reet--a:p-p-lieati-on-of-the-A-S agent to-rhe-aggregate-sur-~-----
face is undoubtedly the most efficient and possibly the most effective.
However, a uniform dispersion is not possible because very small amounts
of AS agents (for example 0.5 percent by weight of asphalt cement) are
normally used, and the HMA mix contains a substantial amount of fines.
The amount of AS agent to be used is important. Too little may not

be effective and too much may be detrimental to the Hl\1A mix. The long
term effectiveness of liquid AS agents during the service life of the Hl\1A
pavements has not been fully established.
Some agencies maintain an approved list of AS agents and require the
contractors to use an AS agent in all Hl\1A mixes without conducting any
moisture-susceptibility test (.51). This practice has many serious disad-
vantages. Some Hl\1A mixes do not need any AS agent and, therefore, it
is uneconomical (and sometimes detrimental) to use these agents. Some
AS agents are asphalt cement and aggregate specific and, thus, are not effec-
tive in all Hl\1A mixes unless verified by tests. It should be left to the con-
tractor to select a suitable AS agent and its dosage to meet the test criteria
of the specified moisture-susceptibility test. Such criteria should be con-
stantly modified to reflect technological advancements and product devel-
opments from the suppliers of AS agents.
Lime Additives. Unlike liquid AS agents which are usually added to

the asphalt cement, lime is added to the aggregate prior to mixing with
asphalt cement. Many studies indicate that lime is a very effective anti-
stripping agent. However, its antistripping mechanism is not well
understood. Various mechanisms have been postulated: (a) lime inter-
acts with acids in the asphalt cement that are readily absorbed on the
aggregate surface, (b) lime provides calcium ions which can replace
hydrogen, sodium, potassium and other cations on the aggregate sur-
face, and (c) lime reacts with most silicate aggregates to form a calcium
silicate crust which has a strong bond to the aggregate and has sufficient
porosity to allow penetration of the asphalt cement to form another
strong bond (ii).
Both hydrated lime Ca(OH) 2 and quick lime CaO (in slurry form) are
effective, although the former is most commonly used. Dolomitic limes
(both Type S and N) have also been used as antistripping additives. i
I
However, as a carbonate CaCO3 lime is not as effective. Generally, 1 to 1½
percent of lime by weight of dry aggregate is used. Finer aggregates may
require higher percentages because of increased aggregate surface area.
Aggregates have been treated with lime by the following four methods
(ii):
E>ry hydrated li:me:Themainproblemin using dry lime is to main-
tain its coating on the aggregate surface until it is coated with
asphalt cement. It is more critical in drum mixers which tend to
pick up some of the lime in the exhaust gas flow. However,
Georgia DOT has successfully used dry hydrated lime in drum
mixers by injecting lime into the drum just ahead of the asphalt
cement. The pick up of lime by the gas stream is prevented by
modifications of the flights and providing suitable baffles inside the
drum. (5§). Some asphalt paving technologists believe that the use
of dry lime is not consistently effective, although many agencies
including Georgia DOT have had very satisfactory results with dry
lime.
2. Hydrated lime slurry: This method requires additional water to be
added to the aggregates which results in increased fuel costs and
reduced HMA production rates.
3. Dry hydrated lime to wet aggregate: In this method dry hydrated
lime is added to wet aggregate, us~ally containing 3-5 percent
water, and then mixed in a pugmill or tumble mixer to obtain a
homogeneous mix.
4. Hot (Quicklime) slurry: The use of quicklime (CaO) slurry has at
least two advantages: (a) its cost is equal to that of hydrated lime
but when slaked the hydrated lime yield is 25 percent greater, and
(b) the heat from slaking results in an elevated temperature which
helps in the evaporation of the added moisture. It should be han-
dled with caution because it can cause skin burns.
The relative effectiveness of the preceding four treatments based on

comparative laboratory and field studies have been generally inconclusive
and, therefore, increased fuel and equipment costs and decreased HMA
production rates associated with the wet process may not be justified at the
present time.
Waste Materials
Numerous waste materials result from manufacturing operations, ser-
j
vice industries, sewage treatment plants, households, and mining CiZ).
Legislation has been enacted by several states in recent years to either man-
date the use of some waste materials or to examine the feasibility of s u c h -
usage. The hot mix asphalt (HMA) industry has been pressured in recent
years to incorporate a wide. variety of waste materials into HMA pave-
___m.ents_.Jhis_has_raisecLthe...follo:.w.ing.legitlmate-GGn.c;er,1=1s:-~a}-~agine~i:ing
concerns such as effect on the engineering properties (for example, strength
and durability of HMA), impact on its production, and its future recycla-
bility: (b) environmental concerns such as emissions, fumes, odor, leach-
ing, and handling and processing procedures; and (c) economic concerns
such as first cost, life cycle costs, salvage value, and lack of monetary incen-
tives.
The waste materials can broadly be categorized as follows: (a) indus-
trial wastes such as cellulose wastes, wood lignins, bottom ash, and fly ash;
(b) municipal/domestic wastes such as incinerator residue, scrap rubber,
waste glass, and roofing shingles; and (c) mining wastes such as coal mine
refuse. A discussion of some waste materials which have been used in
HMA on an experimental basis follows.
Roofing Shingles. Approximately 90 million roofing shingle squares

are produced per year by 77 plants in the United States. About 1/3 of the
shingles are used on new houses and the remaining 2/3 of the shingles are
used for re-roofing houses. "When a house is re-roofed, often an equiva-
lent amount of old shingles is removed and discarded. Moreover, each
roofing plant generates scrap materials and seconds that can range from 5
to 10 percent of the production capacity. An estimated 10 million tons (9
million Mg) of waste roofing material are generated in the U.S. per year.
This is significantly more than the tonnage of discarded tires (about 3 mil-
lion tons or 2. 7 million Mg) per year. The disposal of old shingles and the
scrap material has created a difficult disposal problem (.5..8).
It is estimated that roofing waste contains about 36 percent asphalt
content, 22 percent hard rock granules (minus No. 10), 8 percent filler,
and smaller amounts of miscellaneous materials (52).
Roofing shingles have been used successfully in the HMA paving of
the parking lots at Disney World, Florida. Shingles need to be shredded to
at least ½ inch (12 mm) or smaller prior to introduction in the mix to
ensure meltdown and uniform dispersion in the HMA mixture. According
to cost estimates (.22), the HMA can be reduced by $3.40 per ton by intro-
ducing only 5 percent organic shingles.
Scrap Rubber. from Tires. This has been discussed already as

reclaimed rubber in the classification of polymers.
Waste Glass. "When crushed waste glass is incorporated in HMA mix,

the resulting mixture is sometimes referred to as "glasphalt".
-------severanaboratory ancl-neld evaluations of-glasJ>lfafrwere conducted in
the early 1970s in the U.S. and Canada. After no significant interest for a
decade, the potential use of glasphalt is now being reassessed. Three
states-Connecticut, Virginia and Florida~have conducted feasibility
studies (§Q_,_ fi.L @.
The Connecticut report (§fJ) gives an excellent review· of literature on

both laboratory and field evaluations of glasphalt since 1969. The follow-
ing general observations have been made in various studies:
1. Glasphalt was successfully mixed and placed in at least 45 locations
in the U.S. and Canada between 1969 and 1988. However, most glasphalt
has been placed on city streets, driveways, and parking lots, and not on
high volume, high speed pavements.
2. Potential problems with glasphalt include: loss of adhesion between
asphalt and glass; maintenance of an adequate level of frictional resistance,
especially with coarse particles; breakage of glass and subsequent raveling
under studded tires; lack of adequate and consistent supply of glass; and
increased production costs (estimated at $5/ton more than the conven-
tional HMA mix in Connecticut). ·-
3. Glasphalt should be used only as a base course to alleviate poten-
tial frictional resistance and surface raveling problems.
4. The maximum size of glass should not exceed 3/8 inch (9.5 mm)
in glasphalt, with hydrated lime added to prevent stripping. A maximum
of 15 percent crushed recycled glass should be allowed in HMA mixes.
5. The use of glass tends to reduce the VMA and air voids in Marshall
specimens; therefore, optimum asphalt content will also be reduced. (§J)
6. Marshall stability values decreased by 15-20 percent and dry indi-
rect tensile strength decreased by 20 percent when 15 percent of the screen-
ings were replaced with either coarse or fine glass particles. (@
7. Moisture conditioning of Marshall specimens caused a 15 percent
and 50 percent decrease in tensile strength when coarse and fine glass,
respectively, were incorporated into the HMA mixture. Retained tensile
strength ratio (TSR) values indicated that the: liquid antistripping additive
used was ineffective in reducing the moisture damage (@.
Miscellaneous
There are some miscellaneous materials that are not classified but have
been used as additives and modifiers in HMA mixes. Two materials: sili-
cone and deicing calcium chloride granules, are discussed.
Silicone. Silicones are semi-inorganic polymers consisting of silicon,

---0-:eygen,ancl-mga:nic-molecules:-Silicon·es-can-lYe-fl.ums, gels, or ruooer-It.lc·,_e_ _ _ _ _ _-!
solids. Dimethyl silicone fluid available from Dow Corning as DC 200
and from General Electric as SF 96 is the most commonly used silicone in
the HMA industry. Both silicone fluids are available in a wide range of vis-
cosities, however, the fluid with a viscosity of 1,000 centistokes at 77°F
(25°C) is most widely used. The silicone fluid is added to asphalt cement
binder in very small amounts, usually two to three parts per million (ppm)
to realize the following benefits:
1. Smooth production and laydown of HMA without any foaming

problems.
2. Oxidative hardening of the asphalt cement is retarded when the
HMA is held in storage bins.
Laboratory studies and field experience have shown that silicones do

not have any significant effect on the physical properties of the asphalt
binder nor any detrimental effect on HMA as long as the concentration of
silicones did not exceed 5 ppm. The cost of treating asphalt cement with
silicone is almost negligible.
Deicing Calcium Chloride Granules. The commercially available

product is called Verglimit which consists of calcium chloride with 5 per-
cent caustic soda and is impregnated with linseed oil. Verglimit is added
at a rate of 5.5 to 6.5 percent by weight of total mix to a wearing course
mix. It was developed to (a) retard the general formation of ice on HMA
pavements, and (b) facilitate ice/snow removal from the road surface. It
does not melt ice or snow nor does it replace regular winter road mainte-
nance. Verglimit expands when exposed to humidity and works up to the
pavement surface where it absorbs the moisture until it dissolves.
Since Verglimit is lighter than aggregate, the o_p~timum asphalt content
of the HMA mix needs to be increased by 0.1 to 0.3 percent. Addition of
Verglimit about 15 seconds before the end of wet mix cycle has been found
adequate for obtaining satisfactory coating, and minimizing the danger of
cracking it. After the Verglimit in the upper 1/16 inch (1.6 mm) of the
HMA pavement has dissolved, most HMA surfaces give a raveled appear-
ance. Laboratory studies (§3) have indicated that addition of Verglimit
may increase the susceptibility of HMA to moisture induced damage.
The purpose of using additives and modifiers in HMA mixtures is to

enhance the performance of HMA pavements. If this is not accomplished
in a cost-effective manner, then their use is not justified.
REFERENCES
1. Fehsenfeld, Fred M. Performance of Open-Graded Big Rock Mixes
in Tennessee and Indiana. National Asphalt Pavement Association,
Information Series 104, 1988.
2. Cedergren, H.R., J.A. Arman and K.H. O'Brien. Development of
Guidelines for the Design of Subsurface Drainage Systems. Federal
Highway Administration Report RD-73-14, February 1973.
3. Cedergren, H.R. and WR. Lovering. The Economics and
Practicability of Layered Drains for Roadbeds. Highway Research
Board, Highway Research Record 215, 1968.
4. Kandhal, P.S. and M.E. Wenger. Experimental Asphalt Stabilized
Base under Portland Cement Concrete. Pennsylvania Department of
Transportation Research Project 71-5, Final Report, February 1974.
5. Halstead, WJ. Open-Graded Friction Courses for Highways
Transportation Research Board, NCHRP Synthesis 49, 1978.
6. Halstead, WJ. Criteria for Use of Asphalt Friction Surfaces.
Tr~nsportation Research Board, NCHRP Synthesis 104, 1983.
7. Kandhal, P.S., R.J. Brunner and T.H. Nichols. Design, Construction,
and Performance of Asphalt Friction Courses in Pennsylvania.
Transportation Research Board, Transportation Research Record 659,
1977.
8. Bukowski,] .R., SMA in America. Paper presented at the 1993 Annual
Meeting of the Transportation Research Board, Washington, DC,
1993.
9. Guidelines for Materials, Production, and Placement of Stone Matrix
Asphalt (SMA). National Asphalt Pavement Association, IS 118,
August 1994.
10. Brown, E.R. et al. Designing Stone Matrix Asphalt Mixtures.
National Academy of Sciences, Transportation Research Board,
National Cooperative Highway Research Program, Report No. 9-8/1,
1995.
11. Brown, E.R. and RB. Mallick, A Laboratory Study on Draindown of
Asphalt Cement in Stone Matrix Asphalt (SMA). Transportation
12. Scherocman, J.A., "The Design, Construction and Performance of
----£1:0-FJ:e-Mas-t-ie-Pa:v-e·mef.l·t-ba-yers:-1:he-Gontinuing-Sto~Proree-ding·,--,- - - - - ~
Canadian Technical Asphalt Association, Vol. XXXVII, November
1992.
13. Davis, Richard L. Large Stone Mixes: A Historical Insight. National
Asphalt Pavement Association: Report IS 103/88, 1988.
14. Kandhal, P.S. Marshall Mix Design Methods: Current Practices.

Proceedings, Association of Asphalt Paving Technologists, Vol. 54,
1985.
15. Acott, Mike. The Design of Hot Mix Asphalt for Heavy Duty
Pavements. National Asphalt Pavement Association, QIP 111/86,
October 1987.
16. Von-Quintus, Harold. AAMAS Mix Properties Related to Pavement
Performance. Proceedings, Association of Asphalt Paving Tech-
nologists, Volume 58, 1989.
17. Comparison of 4 and 6-Inch Diameter Molded Specimens.
Pennsylvania Department of Transportation, Bureau of Materials,
Testing and Research, Status Report, February 21, 1969.
18. Marshall Criteria for Compacted Bituminous Specimens.
Pennsylvania Test Method 705, Pennsylvania Department of
Transportation, Field Test Manual, March 1983.
19. Kandhal, P.S. Testing and Evaluation of Large Stone Mixes Using
Marshall Mix Design Procedures. National Asphalt Pavement
Association Information Series 108, 1990.
20. Resistance to Plastic Flow of Bituminous Mixtures Using Marshall
Apparatus. ASTM D1559-82, American Society for Testing and
Materials, Vol. 04.03, 1988.
21. Kandhal, P.S. Design of Large Stone Asphalt Mixes to Minimize
Rutting. Transportation Research Board, Transportation Research
Record 1259, 1990.
22. Williams, Ellis G. Design and Construction of Large Stone HMA
Bases in Kentucky. Hot Mix Asphalt Technology, Winter 1988.
23. Kandhal, P.S. Large Stone Asphalt Mixes: Design and Evaluation.
1990.
24. Herrin, Moreland. Bituminous Patching Mixtures. Transportation
Research Board, NCHRP Synthesis 64, 1979.
25. Kandhal, P.S. and D.B. Mellott. A Rational Approach to the Design
of Bituminous Stockpile Mixtures. Transportation Research Board,
26. Epps, J.A., D.N. Little, R.J. Holmgreen, R.L. Terrel and WB.
Ledbetter. Guidelines for Recycling Pavement Materials.
Transportation Research Board NCHRP~eport 224-; · September
1980.
27. Kandhal, P.S., E.R. Brown, and Steve Cross. Guidelines for Hot Mix
Recycling in State of Georgia, Georgia Department ofTransportation,
1989.
28. Brown, E.R. Insuring Quality in Hot Mix Recycling, T~ortation

Research Board, Transportation Research Record 885, 1982.
29. Terrel, R.L. and J.A. Epp. Using Additives and Modifiers in Hot Mix
Asphalt (Part A). National Asphalt Pavement Associatio,n, Quality
Improvement Series QIP 114 A, 1989.
30. Epps, J.A. Asphalt Pavement Modifiers. Civil Engineering Magazine,
April, 1986.
31. McGennis, Bob. Asphalt Modifiers Are Here to Stay. Asphalt
Contractor Magazine, April, 1995.
32. Terrel, R.L. and J.L. Walter. Modified Asphalt Pavement Materials,
The European Experience. Proceedings, Association of Asphalt
Paving Technologists, Vol. 55, 1986.
33. Brown, E.R., J.L. McRae, and A. Crawley. Effect·of Aggregates on
Performance of Bituminous Concrete. ASTM Special Technical
34. Kandhal, P.S. Evaluation of Baghouse Fines in Bituminous Paving
Mixtures. Proceedings, Association of Asphalt Paving Technologists,
Vol. 50, 1981.
35. Anderson, D. A. Guidelines on the Use of Baghouse Fines. National
Asphalt Pavement Association, Information Series 101, 1987.
36. Yoo, Z. And C.L. Monismith. Behavior of Asphalt Mixtures with
Carbon Black Reinforcement. Proceedings, Association of Asphalt
37. Khosla, N.P. and S.Z. Zahran. Behavior of Asphalt Mixtures with
Carbon Black Reinforcement. Proceedings, Association of Asphalt
38. Kandhal, P.S., D.B. Mellott, and G.L. Hoffman. Laboratory and
Field Characterization of Sulphlex as a Paving Binder. ASTM Special
Technical Publication 807, October, 1983.
39. McBee, WC. State-of-the-Art Guideline Manual for Design,
Quality Control; and Construction of SEA Pavements. Federal
Highway Administration Report No. FHWA-TS-80-240, August,
1980.
40. Kandhal, P.S. Evaluation of Sulphur Extended Asphalt Binders in
Bituminous Paving Mixtures. Proceedings, Association of Asphalt
PavingTechnologists, Vol. 51, 1982.
---4-I-.Fromm;--H:J:-Su:Iphur=Aspira:l:rPavement Performance and Recyclmg.
1981.
42. Al-Otaishan, A.T. And R.L. Terrel. Material Characteristics and
Predicted Performance of Sulphur-Asphalt Mixtures from In-Service
--
Pavements. Proceedings, Association of Asphalt Paving Technologists,

Vol. 49, 1980.
43. Terrel, R.L. and S. Rimsritong. Wood Lignins Used as Extenders for
Asphalt in Bituminous Pavements. Proceedings, Association of
Asphalt Paving Technologists, Vol. 48, 1979.
44. Hines, M.L. Asphalt Cement Performance Improved by Styrelf-
Laboratory and Field Data. Koch Materials Company, 1993.
45. Heitzman, M.S. State of the Practice-Design and Construction of
Asphalt Paving Materials with Crumb Rubber Modifier. Federal
Highway Administration Report No. FHWA-SA-92-022, May, 1992.
46. Roberts, F.L., P.S. Kandhal, E.R. Brown, and R.L. Dunning.
Investigation and Evaluation of Ground Tire Rubber in Hot Mix
Asphalt. National Center for Asphalt Technology Report No. 89-3,
August, 1989.
47. Guidelines for Materials, Production and Placement of Stone Matrix
Asphalt (SMA). National Asphalt Pavement Association, Information
Series IS 118, August, 1994.
48. ASTM Standard D4552-92 on Practice for Classifying Hot-Mix
Recycling Agents. ASTM Annual Book of Standards, Vol. 04.03, 1995.
49. Little, D.N., R.J. Holmgreen, and J.A. Epps. Effect of Recycling
Agents on the Structural Performance of Recycled Asphalt Concrete
Materials. Proceedings, Association of Asphalt Paving Technologists,
Vol. 50, 1981.
50. Epps, J.A., D.N. Little, R.J. Holmgreen, and R.L. Terrel. Guidelines
for Recycling Pavement Materials. Transportation Research Board,
NCHRP Report 224, September 1980.
51. Kandhal, P.S. and D.B. Mellott. Pennsylvania's Experience with the
Design, Construction, and Performance of Gussasphalt. Proceedings,
Association of Asphalt Paving Technologists, Vol. 46, 1977.
52. Kandhal, P.S. Moisture Susceptibility of HMA Mixes: Identification
of ;Problem and Recommended Solutions. National Asphalt Pave-
ment Association, Quality Improvement Publication. QIP 119,
December, 1992.
53. Tunicliff, D.G. and R.E. Root. Antistripping Additives in Asphalt
----------'Co.n..cr..e_t_e_: State-of-the-Art Report. Proceedings,~ Association of
Asphalt Paving Technologists, Vol. 51, 1982.
54. Kandhal, P.S. Field and Laboratory Evaluation of Stripping in Asphalt
Pavements: State-of-the-Art Report. Transportation Research
Record 1454, 1994.
476 Hot Mix Asphalt Materials; Mixture Design and Construction
55. Hicks, G.R. Moisture Damage in Asphalt Concrete. Transportation

Research Board, NCHRP Synthesis of Highway Practice No. 175,
October, 1991.
56. Kennedy, T.W. Use of Hydrated Lime in Asphalt Paving Mixtures.
National Lime Association, Bulletin 325, 1984.
57. Kandhal, P.S. Waste materials in Hot Mix Asphalt-An Overview.
ASTM Special Technical Publication 1193, 1993.
58. Brock, J.D. and D. Shaw. From Roofing Shingles to Roads. Technical
Paper T-120, Astec Industries, Chattanooga, TN, 1989.
59. Paulsen, G., M. Stroup-Gardiner, and J.A. Epps. Roofing Waste in
Asphalt Paving Mixtures. Center for Construction Materials Research,
University of Nevada, Reno, 1988.
60. Larsen, D.A. Feasibility of Utilizin._g Waste Glass in Pavements.
Connecticut Department of Transportation, Report No. 343-21-
89-6, June, 1989.
61. Hughes, C.S. Feasibility of Using Recycled Glass in Asphalt Mixes.
Virginia Transportation Research Council, Report No. VTRC 90-R3,
March, 1990.
62. Murphy, K.H., R.C. West, and G.C. Page. Evaluation of Crushed Glass
in Asphalt Paving Mixtures. Florida Department of Transportation,
Research Report No. FL/DOT/SMN91-388, April 1991.
63. Stuart, K.D. and W.S. Mogawer. Laboratory Evaluation of Verglim.it
and Plusride. Federal Highway Administration Report FHWA-RD-
91-013, March, 1991.
CHAPTER 8. PERFORMANCE/
DISTRESS OF HMA
SERVICEABILITY/PERFORMANCE CONCEPTS
The need for a quantifiable measure of pavement performance became

evident after the Western Association of State Highway Officials
(WASHO) failed to agree on the definition of failure conditions for their
road tests in the early 1950s (1). When the American Association of State
Highway Officials (MSHO) Road Test was being designed in the 1960s,
the statisticians recognized that some objective measurement of perfor-
mance was required if pavement design equations were to be developed. To
solve this problem, Carey and Irick (2) developed the following concepts:
1. Highways are designed and built for the comfort and convenience
of the travelling public. Therefore, a good highway is one that is
smooth and safe.
2. One user's opinion about how well a highway is serving its function
is subjective. ·
3. There are objective characteristics of the road which, when mea-
sured and properly combined, can be correlated to the average 1?Ub-
jective evaluation of the highway.
4. The serviceability of a road can be adequately represented by the
arithmetic average of opinions of a group of highway users. Single
rating values should not be used since honest differences in opin-
i:0n are too divergent to produce a reliable rating.
5. Performance is defined as the area under a serviceability-time curve
from the time of construction to the time performance is being
evaluated.
The serviceability/performance concept was first applied at the

MSHO Road Test. Reference 2 describes the performance relationships
developed between the average serviceability rating of a panel of raters and
roughness measured by the MSHO profilometer and the Bureau of Public
Roads (BPR) roughometer. These studies showed that about 95 percent of

the information about the serviceability of a pavement is contributed by
the roughness of the surface profile (1). This means that only five percent
of the information explained by all the objective measurements is con-
tained in terms other than roughness. In the original performance equa-
tions, emphasis was placed on roughness. The original regression equation
relating present serviceability index to measurements made on the pave-
ment was (2):
PSI = S.03 - 1.91 log (1 + SP) - 1.38 RD2 - 0.01 ~c + p
where,
PSI = present serviceability index;

SV = slope variance, the variance of slopes measured over a
6-inch wheel base using the CHLOE profilometer;
RD = average rut depth, inches;
C = pavement cracking in feet/I 000 square feet of
pavement surface; and
P = patching in square feet/ I 000 square feet of pavement
surface.
To determine the serviceability of a pavement at a particular point in

time, one only needs to go to the pavement site and collect the objective
data included in the PSI equation and then substitute those values into the
equation to determine the PSI. As stated previously, PSI is an estimate of
the average rating that the raters would have given the pavement surface
had they been able to rate it on the same day that the objective data were
collected. Generally, a new pavement has a PSI between 4 and 5, and repair
is usually needed when the PSI is between 1.5 and 2.5.
This type of relationship has been useful in presenting the pe.efor-
mance concept as shown in Figure 8:._l. Notice that serviceability (PSI) is
plotted on the ordinate and cumulative traffic is plotted on the abscissa,
although time could also be plotted on the abscissa. The serviceability
begins at a }:iigh level, typically between 4.0 and 4.5, remains fairly con-
stant for a period of time and then begins to drop off fairly quickly until it
--reaches-a:-selecred--rernrrn-a:J.-level wnere some rehab1htat1on action is
required, typically a HMA overlay. With the overlay, the serviceability level
is restored to some higher value shown by the dashed line in Figure 8-1,
and the second performance period begins. The performance for the first
Chapter 8-Performance / Distress of HMA 479
ACCUMULATED TRAFFIC LOADS--

► -
Figu.re 8-1. Typical Relationship Between PSI and Cumulative Traffic
period is represented by the area under the PSI versus cumulative traffic
curve. Comparisons between competing thickness design combinations
can be made using the estimated or observed historical performance.
The performance area under the curve during a particular period is sig-
nificantly affected by the initial PSI for that period. Therefore, consider-
able emphasis has been placed on the initial PSI by many state agencies. In
fact, many states have adopted roughness or smoothness specifications that
specify the maximum allowable roughness that a roadway can have for the
contractor to receive 100 percent pay. Many of these same agencies have a
maximum permissible roughness level. If this level is exceeded, the con-
tractor may be required to repair the surface to meet specification require-
ments. Many of these same specifications pay the contractor a bonus if the
smoothness significantly exceeds the requirements.
Pavement Roughness
Pavement roughness is produced by surface deviations that produce a

-----';esp0nse-i-I1-t--he-s.u.spc;.I1.si.Q.n-S-y-St-e-m--0Lthe-v:ehicles-travelling.over the road.
Pavement roughness is important because this is the one pavement prop-
erty that is most noticeable to the travelling public.
Vertical and horizontal alignment of the highway are design parame-
ters established by the geometric design standards for the roadway, and
these designed deviations are not usually considered to be components of

roughness even though they induce body motion in the vehicle. However,
other variations in the road profile in both the longitudinal and transverse
direction are very important. Roughness in the longitudinal direction is the
most critical (2). The other offender is the roll component which occurs as
a result of differential transverse elevations between the inside and outside
wheel paths. A uniform or smooth changing cross-slope is not normally
considered to be objectionable, rather it is the random variations in differ-
ential elevation which cause objectionable roughness.
Since most vehicles travel in well defined wheel paths, roughness mea-
1 surements are typically made in either or both of these wheel paths.
Additionally, comparisons between the wheel paths can give some indica-
tion of the cross-slope variations for a pavement section. The road eleva-
tion profile can be defined by performing a rod and level survey or by run-
Vehicle Response Function

2.--------------------....
C:
C I,
(!)
----~0-=l'--+----+--+--+---t---11---+--+---+---+--+----+---------
o 10 20
Frequency (Hz)
Figure 8-2. Frequency Response Curve for a Mays Road Meter
Chapter 8-Perfonnance / Distress of HMA 481
ning a profile measuring device over the test section (1). Using one of these
techniques, the profile elevations can be defined, and using standard math-
ematical techniques such as Fourier analysis, the wave length of the profile
can be determined. This wave length is the input that the road provides to
a vehicle travelling over it. Different vehicles have significant differences in
wheel base suspension characteristics as well as different tire and wheel
response characteristics. Also, the way each of these vehicles responds to a
unit amplitude of road output varies with the frequency of the input. For
example, Figure 8-2 shows a response curve for a vehicle containing a
Mays Road Meter, a type of device that measures roughness by determin-
ing the amount of relative vertical movement between the rear axle of the
vehicle and the rear deck of the body of the vehicle. Notice in Figure 8-2
that the ordinate is the relative amplitude which represents the vertical
response of the vehicle to a unit of vertical input. In some frequency ranges
a unit of input produces a unit of output. Typically however, the vehicle
attenuates any vertical input by 20 to 80 percent. However, in other fre-
quency ranges, a unit of input produces significantly more than a unit of
output. These regions of response are near the natural frequency of some
component of the vehicle, either the whole vehicle or the tire wheel assem-
bly or the spring-shock absorber-axle assembly.
Since the road profile is static, it should be recognized that as a vehi-
cle travels over a fixed profile the frequency of a particular wave length
input decreases as speed increases. This relationship is demonstrated in
Figure 8-3 which shows that the typical resonant frequency range of most
automobiles lies between 1 and 10 cps. This plot shows that at any speed
there is a road profile wave length that can induce the natural frequency of
the vehicle. If the amplitude of that frequency is large, the quality of the
ride can be significantly affected.
Road Roughness Measurement Methods
A wide variety of roughness measuring devices have been developed

and utilized over the years. State-of-the-art summaries have been prepared
periodically, describing the available devices: Hveem in 1960 (i), Hudson,
et al., in 1968 (.6,), Roberts in 1975 (_Z), Wambolt, et al., in 1981 (.@, and
Hudson, et al., in 1987 (2). From a review of the literature and construc-
tion specifications of individual states, it appears that the roughness devices
most commonly used for esnmatmg serv1cea51hty and acceptance of con-
struction are:
1. Rolling straightedge;
2. Car road meters, either the PCA or Mays type; and

3. Surface Dynamics Profilometer.
Each of these devices is described briefly in the following sections.
60
50
5: 40 IO CPS Resonant Frequency

Cl.
~
e,~
30
I CPS Resonant Frequency
0::
l3 20
/0
0
0 /0 20 30 40 50 60 70 80
WAVE LENGTH (FeefJ
Figure 8-3. Relationships Among Resonant Frequencies of Passenger

Vehicles, Speed, and Profile Wave Length (after Haas and
Hudson, J)
The rolling straightedge or profilograph is used by a large number of

states to measure the initial roughness of new construction and overlay
projects. The device employs a set of bogey wheels that are from 10 to 30
feet (3 to 9m) apart to provide a reference plane. Deviations from this
plane are measured by a recording wheel located at the profilograph mid-
pofat. -Tile-vertical deviations are accumulated over the project length and
converted into the number of inches of vertical displacement per mile
called the roughness index. Because of its geometry, this device cannot
---detecr-rou:ghrress-wi.th-wave--lengths--rha-r-are-orre=ha±f,--on:e-=-qmi:rter;--m1e-.,...- - - - - - -
eighth, etc. of the wheel base length of the rolling straightedge. The devices
are used primarily on construction projects or overlays not opened to traf-
fic because of their low operating speed. Most of these devices are pushed
Chapter 8-Performance / Distress of HMAs 483
by hand or towed by a vehicle at less than 10 miles per hour. Therefore,

they are not suitable for use in any roadway serving high speed traffic.
Car road.meters (CRM) have become very popular roughness devices
both for measuring initial roughness of newly constructed surfaces and for
monitoring the serviceability level of in-service pavements. There are two
major versions of the CRM, the. PCA device developed by Brokaw (l.Q)
and the Mays Road Meter (MRM) originally developed by Ivan Mays,
described by Phillips and Swift (11), and manufactured by Rainhart Co.
(12).
The PCA version of the road meter is usually mounted in a typical pas-
senger car or station wagon. The device is an electromechanical system that
measures both the number and magnitude of vertical deviations (inch-
es/mile) between the body of the vehicle and the center of the rear axle.
The MRM basically measures the same relative displacement between the
body of the vehicle and the rear axle as does the PCA meter. However, the
measuring systems are completely different. The MRM uses a photocell
sensing system with 0.1-inch resolution to measure the vertical excursions.
These car road meters offer a number of advantages:
1. They operate at normal highway speeds.

2. They are relatively inexpensive, simple, and easy to operate.
3. They are very portable and are easily moved from one site to an-
other.
4. They can be efficiently used to collect inventory type data on the
whole highway system.
5. They provide reasonably good repeatability, especially if the average
of several runs is used to represent the condition of the roadway.
However, these devices have a number of very serious drawbacks that

may limit their usefulness for some applications:
1. They require fairly frequent calibration.

2. They require that operational precautions be carefully observed.
3. They require recalibration with any change in the vehicle suspen-
sion system (such as changing tires, or shock absorbers).
4. They are not sensitive to long wave length components of roughness
as shown in Figure 8-2.
Most state agencies utilize these devices because they believe the advantages
outweigh the disadvantages.
The Surface Dynamics Profilometer (SDP) was originally developed
by General Motors and described by Spangler and Kelley in 1962 W).
The original device measured road profile in only one wheelpath; how-
ever, it has since bee::n modified to measure the profile in both wheelpaths
simultaneously.
Figure 8-4 shows a schematic diagram of how the SDP works. In the
early devices, a set of wheels were used to provide contact with the road
surface. These small wheels were held in contact with the road surface by
a set of torsion bars which applied a 300 lb. (1350 N) force to each wheel.
A potentiometer was mounted between a yoke above the road follower
wheel and the body of the panel truck or van in which the device was locat-
ed. The potentiometer output measured the relative displacement between
the road follower wheel and the body of the vehicle. An accelerometer was
mounted directly above the potentiometer base plate inside the vehicle.
The accelerometer senses the vertical acceleration of the body of the vehi-
cle. This vertical acceleration signal is integrated twice by an onboard com-
puter to produce the vertical movement of the vehicle body. Because of the
sensitivity of the accelerometer to long wave length inputs which would
saturate or overcharge the accelerometer, a set of signal conditioning filters
were built into the computer to attenuate the amplitude of the long wave
length signals which correspond to grade changes or topography.
Therefore, the computer double integrates the filtered accelerometer signal
and adds the potentiometer signal to it to produce a representation of the
road profile.
The profile data is in either analog or digital form and can be
processed using sophisticated techniques such as power spectral density
techniques. Exce::pt for onboard processing techniques, the summarization
of the road profile data is complicated and time consuming and requires
considerable development work to analyze and select the types of profile
summaries that work best for the user. Tl-iere are significant advantages to
using the SDP; however, the high capital and operating personnel costs,
the complexity of the system, and the high cost of data analysis have severe-
ly limited the use of this device within highway agencies.
Several significant advantages of the SDP are listed below:
1. Actual road profiles are a direct output of the device.

2. The operating system is time stable, i.e., the output from the device
does not change as the carrier vehicle deteriorates with use.
3. Repeatability from run to run is excellent.
4. Long wave lengths can be detected and measured and then analyzed.
-----5·-:-The operatmg speed 1s high enough to be funct10nal rorinany types
of roadways and to cover a reasonable mileage of roadway in a days
time.
6. The device provides an excellent time stable means of calibrating
CRM devices.
0::
w 0::
I- w
w I-
:::E w
0 :::E
0::
w 0
_J I-
E w
u
z
3: w
u I-
n <(
- 0
a.
-
-0
-0
N
\
t:::
+
E E
NI !N NI
~ ~
z
0
I-
<(
(!) I-
a:::,
_JQ.
<( :::E
zo
<(U
Figure 8-4. Schematic Describing the Measurement System of the

Surface Dynamic Profilometer (after Spangler and
Kelley, 1.3)
Since the mid-70s, several other devices have been developed and mar-
keted that collect other roadway information in addition to roughness;
among these devices are the ARAN (ldJ, the APL (15) and GERPH O (15)
units. These devices typically collect condition survey data in addition to
roughness. Some have special capabilities for taking pictures of the road
surface or have options to mark in the computer records the occurrence of
bridges, culverts, intersections, etc. A discussion of these devices is outside
the scope of this book.
Smoothness Requirements for Construction
The widespread acceptance of the serviceability/performance concept

led to the use of roughness data as the primary measure of public satisfac-
tion with the highway system. The increased emphasis now being placed
on pavement management activities has produced a heightened interest in
smoothness specifications and measurements. As a result of this interest,
the AASHTO Highway Subcommittee on Construction developed and
transmitted a series of questionnaires on the use of rideability specifica-
tions. Questionnaires were sent out in 1981, 1984, and 1987 (1§). The
primary thrust of the 1987 questionnaire was to distinguish between ride-
ability and bump specifications, to determine what roughness equipment
was used in the specifications, and to determine the extent of use of incen-
tive and disincentive clauses.
Bump specifications are those which specify roughness as a maximum
vertical deviation from a plane surface of specified length. This type of
specification has been used for many years to control bumps or ridges at
joints and other discontinuities in the pavement. These humps are nor-
mally measured with a 10-ft. (3m) straightedge. A straightedge is laid on
the pavement surface and the maximum deviation of the pavement surface
from this straightedge is measured and recorded. Rideability specifications
are those which specify roughness over longer distances using some type of
profilograph or response type roadmeter. Table 8-1 contains a summary of
the number of states either using or considering the use of rideability spec-
ifications from the 1981, 1984 and 1987 AASHTO surveys. In the 1987
survey, there were 8 states which employed only a bump specification,
while all those states which included a rideability specification also includ-
--~d-a-bump-s~ci.f1Gati-0n.-It-is-0-l:i-v-:iQUs-Fh.-at-rhe-speei.f-iea:1:ion-wri-ters-view~------
that both types of specifications serve a useful and somewhat different pur-
pose. As shown in Table 8-2, it is also apparent that smoothness is more
often required for concrete pavements than for HMA pavement. The pri-
mary reason for this is that historically, concrete pavements have been
rougher pavements than HMA immediately after construction. In 1987,

25 percent of the states using smoothness specifications applied them to
both concrete and HMA surfaces.
The most commonly used bump specifications for HMA are 1/8 inch
in 10 feet (3mm in 3m) (used by 15 states) and 3/16 inch in 10 feet (5mm
in 3m) (used by 7 states). There is less consensus on the limit values for
rideability. There are two major reasons for the variations in rideability
specifications; first, several different devices are used to measure roughness,
and second, some states use bonuses or other payment adjustments for dif-
ferent roughness levels. Of the 32 agencies using a rideability specification
in 1987, approximately 1/3 had neither a bonus nor a reduction-in-pay
clause, and approximately 1/3 had only a reduction-in-pay clause. There is
considerable variation in the acceptance limits established as part of the
specifications. The specification limit used by most states was set between
7 and 12 inches per mile. These limits were selected based on engineering
judgement.
Table 8-1. Summary of States Using or Considering Using Rideability

Specifications to Control Construction Smoothness (1.i)
Rideability Specifications
Survey
Year Using . Considering
1981 17 13
1984 24 12
1987 32 5
'--
The equipment used to measure rideability varies substantially from
state to state. In the 1987 survey, 3 different types of equipment were used:
l. Profilographs
California-21 states
Rainhart--4 states
2. SDP profilometers-2 states
3. Mays Road Meters (MRM)-2 states
The specifications of the individual states vary considerably concern-

_____,··-ng-wlw-p-rovid:es-the-roughnes-s-mea-sttri-ng-equipment,performs the tests,
and reduces the data. In 1987 fourteen states provided the equipment, per-
formed the tests, and reduced the data. Eleven states required the contrac-
tor to provide the equipment; however, state personnel made the measure-
ments and, in all but three, reduced the data. Five other states used either
state or contractor supplied equipment, but in all five cases, the state per-
sonnel reduced the data.
Reference 16 notes that no state indicated that a study had been con-
ducted to quantify benefits of a smoother pavement as a basis for estab-
lishing bonus thresholds: However, the results of a study conducted by
Janoff (12) indicated that annual savings of up to $600 per lane mile could
result if the initial pavement smoothness was reduced from 35 to 5 inches
per mile (555mm to 80mm per km) as measured with the Mays Road
Meter. The data set used by Janoff was obtained from the states of Arizona
and Pennsylvania and contained data on pavement condition, environ-
ment, traffic, pavement cross-section, and maintenance cost. Janoff con-
ducted a statistical analysis of the data and reported that the pavements
with lower initial roughness (1) had lower roughness after 10 years of ser-
vice than'did pavements with higher initial roughness, (2) had lower crack-
ing levels after 10 years of traffic than did pavements with higher initial
roughness, and (3) required lower average annual maintenance costs after
10 years of traffic than did pavements with higher initial roughness.
Table 8-2. Pavement Type for which the Rideability Specifications

Apply(M)
Rideability Specifications Applied To:

Survey
Year Hot Mix Asphalt
Concrete Only Only Both
1981 11 1 5
1984 11 1 12
1987 24 0 8
Janoff concluded that from a cost-benefit standpoint, owner agencies

could afford to spend a considerable amount of additional funds to build
smoother highways and still accrue significant savings in reduced annual
maintenance cost. Figure 8-5 shows the relationship between initial pave-
ment roughness and roughness after 8 years of traffic. This data basically
shows that for initial roughness values ofless than 35, the roughness after
8 years increases by approximately 4 to 6 points. Therefore; a smoother
road immediately after construction will carry more traffic before a pre-
---~s=c='ri:bed roughness level is reached. However, m some cases so much empha-
sis has been placed on smoothness that mix designs have been modified to
produce a tighter, more workable mixture, sometimes resulting in loss in
performance due to low in-place air voids and small maximum aggregate
size.
In Figure 8-6 Janoff showed the effects of increased roughness on

annual maintenance costs using low (conservative) and best (high) cost
estimates based on the strategies used to correct the load associated crack-
ing distress. Notice that for initial pavement roughness levels below about
23 inches per mile (365 mm per km), the high and low repair cost esti-
mates are about the same. However, above 23 they begin to diverge dra-
matically and at 35 inches per mile (555 mm per km) the difference is
about two-fold. Significant cost savings accrue as a result of reduced initial
roughness, suggesting that owner agencies could economically justify
smoothness bonuses.
Table 8-3 contains the basics of a smoothness specification being con-
sidered by AASHTO for HMA. The adjustments in Table 8-3a apply to
payments for those roadway segments where the owner agency does not
use incentive payments for smoothness. It should be applied to pavements
that had profile index values measured with a California profilograph.
Table 8-36 is provided as one example of a smoothness specification
which provides a bonus for high quality work. While it is very good to have
the option of being paid a bonus for very smooth roads, achieving this level
of smoothness is very difficult and requires exceptional care by all members
of the paving crew. The NAPA has prepared guidance (J.ffi to assist the
contractor with improving the smoothness of his paving operations. The
AASHTO Smoothness Specification shown in Table 8-3 is under review
and thus had not been formally adopted at the time this book was pre-
pared. The construction factors which affect pavement smoothness are dis-'
cussed in Chapter 6.
DISTRESSES IN HMA PAVEMENTS

HMA, like all other paving materials, is subject to a variety of dis-
tresses. The development of these distresses should not be viewed with
alarm unless the distresses occur early in the life of the pavement. As pave-
ments approach their design life, distresses are expected to occur as a result
of the environment and repeated traffic loads. It is important to recognize
that a distress that occurs at the surface can have a number of different
causes which must be properly identified before the proper corrective
action can be taken. For example, if rutting has occurred in a HMA sur-
face, the engineer must determine which layer(s) in the pavement structure
caused the rutting to occur before a repair method can be designed.
Rutting can occur due to unstable mixes, consolidation under traffic load-
mg, stnppmg m the fl.MA:, or higher wheel loads or tire pressures than
those considered in design. The visible manifestation of the problem at the
surface may be the same; however, the solution for each cause may be dif-
ferent. Therefore, if the remedial action is to be effective, the cause of the
problem must be identified and corrected.
--------------------,~
0 -E
I")
--
C:
IO U)
C\J U)
w
z
I
0
C\J
r-
0
0
~
U)
IO
_J
-<i
r-
0 z
0 IO O 10 0 IO O IO O IO
---'~~ st___,q:__~_!D--~-~'---------------
(!W/U!) SS3NH9n0H ~H3.l 9N01
Figure 8-5. Initial Pavement Roughness Versus Roughness after
Serving Traffic for 8 Years (after Janoff, 12)
--------------------~
-... _ IO
rt)
o-
--
1") E
C
IO Cf)
C\J Cf)
w
z
I
0 1-
C\J 0
0
::E
IO Cf)
_J
<{
1-
0 z
IO
o...____..___o""!!-------------------.......-----...i..-~o
0
rt)
~ ¢
0
0
m
~
~
~
W
2 g g
IO rt) C\J
O
rt)
0
Figure 8-6. Initial Pavement Roughness Versus Annual Maintenance

Costs after Serving Traffic for 8 Years (after Janoff, 1LJ
Table 8-3. AASHTO Suggested Smoothness Specifications and Pay

Adjustment Factors Based on Measurements with a
California Profilograph
a. Price Adjustments with no Smoothness Bonus and for Pavements Which Have Had No
Smoothness Corrective Work
Prome Index Contract Unit Price Adjustment

Inches/mile per 0.1-mile section Percent of pavement unit bid price
Less than 10 100
Over 10 to 11 98
Over 11 to 12 96
Over 12 to 13 94
Over 13 to 14 ~- 92
Over 14 to 15 90
' Over 15 Corrective Work Required
b. Price Adjustments-with Bonus Payments for Smoothness
Prome Index Contract Unit Price Adjustment

Inches/mile per 0.1 mile section Percent of pavement unit bid price
3 or less 105
Over 3 to 4 104
Over 4 to 5 103
Over 5 to 6 102
Over 6 to 7 101
Over 7 to 10 100
Over 10 to 11 98
Over 11 to 12 96
Over 12 to 13 94
Over 13 to 14 92
Over 14 to 15 90
Over 15 Corrective Work Required
In 1979 Kennedy, et al., U2J developed the following set of definitions

_ ____,l;fo.ulexible_p.a.Y.e.m.en.t..dis.tr.e.s.s_e.s.:_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __
1. Distress is a condition of the pavement structure that reduces ser-

viceability or leads to a reduction in serviceability.
2. Distress manifestations are the visible consequences of various dis-
tress mechanisms, which usually lead to a reduction in serviceabil-

ity.
3. Structural failure is a fracture or distortion that may or may not
cause an immediate reduction in serviceability but leads to a future
loss of serviceability.
4. Fracture is the state of a pavement material that is breaking.
5. Distortion is a permanent change in the shape of the pavement or
pavement component.
6. Disintegration is the decomposing or abrading of a pavement into
its constitutive elements.
7. Reflection cracks occur in the surface course of a pavement and
coincide with and are caused by the relative movement of cracks
or joints in underlying layers.
8. Low-temperature cracks are generally transverse cracks that are
caused when tensile stresses caused by shrinkage of the HMA
exceed the tensile strength of the surface material.
9. Ravelling is the progressive disintegration of a HMA layer from the
surface downward by the dislodgement of aggregate particles. This
can be caused by insufficient binder in the mix, hardening of the
asphalt binder, wet or dirty aggregate, aggregate with a smooth sur-
face texture, or insufficient compaction.
10. Ruts are longitudinal depressions that form in the wheel paths of
flexible or composite pavements and result from compaction or lat-
, eral migration of one or more of the pavement layer materials under
the action of traffic and environment.
11. Polished aggregates are surface aggregate particles that have smooth,
rounded surfaces with fine microtexture, either in their original
condition or after abrasive wear by traffic.
12. Fatigue cracks in a pavement layer are caused by the combination of
repetitive strains and apparent reduction of tensile strength caused
by fatigue of the layer material. The repetitive strains that cause
fatigue are usually the result of traffic but may include thermally-
induced strains or other types of strains.
Using this set of definitions, Welborn @ selected seven distress mod-

els which included load-associated cracking, nonload associated cracking,
reflection cracking, distortion (shoving, rutting, and slippage), disintegra-
tion, reduced skid resistance, and roughness. Many of these factors were
included in a flowchart, prepared by Hveem, et al in 1962 (21), that was
designed to assist with the analysis of HMA pavement failures for which
the HMA is primarily responsible.
1.
I
Cracking
Cracks occur from a variety of causes including stresses from axle

loads, temperature changes in the HMA layer, or moisture and tempera-
ture changes in an underlying layer. It is important to accurately identify
the type of cracking which a pavement exhibits in order to accurately assess
the causes for the cracking and s4-bsequendy the proper repair techniques.
The types of cracking discussed in the next sections include fatigue, low
temperature, longitudinal, block, reflection, and slippage.
Fatigue Cracking. Fatigue cracking (Figure 8-7) is often called alliga-
tor cracking because this closely spaced crack pattern is similar to the pat-
tern on an alligator's back. This type of failure generally occurs when the
pavement has been stressed to the limit of iJs fatigue life by repetitive axle
load applications. Fatigue cracking is often associated with loads which are
too heavy for the pavement structure or more repetitions of a given load
than provided for in design. The problem is often made worse by inade-
quate pavement drainage which contributes to this distress by allowing the
pavement layers to become saturated and lose strength. The HMA layers
experience high strains when the underlying layers are weakened by excess
moisture and fail prematurely in fatigue. Fatigue cracking is also often
caused by repetitive passes with overweight trucks and/ or inadequate pave-
ment thickness due to poor quality control during construction.
Figure 8-7. Typical Case of Fatigue (Alligator) Cracking in a HMA

Pavement
Fatigue cracking can lead to the development of potholes when the

individual pieces of HMA physically separate from the adjacent material
and are dislodged from the pavement surface by the action of traffic.
Potholes generally occur when fatigue cracking is in the advanced stages
and when relatively thin layers of HMA comprise the bound portion of the
pavement.
Repair strategies for fatigue cracking typically involve removal and
replacement. The fatigue cracking is often caused by high moisture con-
tent, poor subgrade, or some other local problems that can be repaired
without major reconstruction. In these instances, the poor material is
removed and replaced with good material and drainage improved. If the
failed areas are extensive, one typical repair strategy is to place a HMA
overlay over the entire surface. Fatigue cracking is a load-associated failure
and hence the thickness of the overlay must be structurally designed to
carry the number of loads anticipated during the design life of the pave-
ment. Too often, thin overlays are used as a repair strategy for fatigue crack-
ing with little consideration given to the traffic loads or to the condition of
the underlying pavement layers. Consequently, many pavements that
exhibit fatigue cracking and receive thin overlays fail prernaturely because
the underlying layers are too weak and the thin surface does not provide
enough thickness to withstand the high stresses produced by truck load-
mgs.
An invest~gation is needed prior to repair of a pavement with fatigue.
cracking. This 1nvestigation should determine the cause of the problem so
that a proper solution can be selected. An investigation may involve dig-
ging a pit to determine the thickness of all pavement layers. Samples of the
layers can be obtained to determine if the material quality is sufficient to
carry the anticipated loads. The investigation should determine if subsur-
face water is a problem. Once the problem is isolated, a solution is select-
ed that involves removal or stabilization of the weak layer, or that provides
enough pavement thickness over the weak layer to protect it from the
expected traffic.
Low Temperature (Thermal) Cracking. Low temperature cracks are

transverse cracks which generally run perpendicular to the roadway cen-
terline and are often approximately equally spaced (Figure 8-8). These
---~tr.ansverse cracks normally occur when the temperature at the_~ll,rface drops
sufficiently to produce a thermally induced shrinkage stress in the HMA
layer that exceeds the tensile strength of the asphalt mixture. These cracks
usually initiate at the top of the HMA and propagate downward through
the mixtures. As discussed in Chapter 2, some of the principal factors avail-
able to the engineer to control the occurrence of thermal cracking are pen-
etration at 77°F (25°C), viscosity at 140°F (60°C), and temperature sus-
ceptibility. Increasing the asphalt penetration is the most often suggested
designer contr9lled factor for minimizing this type of cracking. However,
as pointed out in Chapter 2, asphalts with high temperature susceptibili-
ties should be avoided as these asphalts become extremely stiff at low tem-
peratures regardless of the penetration at 77°F (25°C) or viscosity at I 40°F
(60°C) The asphalt must be soft enough at the lowest anticipated service
temperature to provide some flexibility; otherwise, cracking will occur.
Low temperature cracks are typically repaired by sealing with a liquid
asphalt or other type of sealing material (ASTM D5078 and D3405).
Sealing the cracks prevents moisture from entering the base course and
subgrade and helps to minimize ravelling adj-acent to the cracks. HMA can
provide years of satisfactory service after developing transverse cracks if the
cracks are kept sealed. When the sides of the crack are at different eleva-
tions, grinding has been performed to minimize the bump at the crack,
thereby improving the ride quality. If cracking becomes too severe, the
HMA layer may have to be removed and replaced with an overlay. If all the
cracked pavement layer is not removed before the overlay, it is likely that
the underlying cracks will reflect though the overlay during the first or sec-
ond winter.
Longitudinal Cracking. Longitudinal cracks are individual cracks that

basically run parallel to the centerline of the roadway as shown in Figure
Figure 8-8. Typical Spacing of Low Temperature (Thermal) Cracking

in a HMA Pavement
8-9. They most often occur at the joint between adjacent lanes of asphalt
mixture or at the edges of the wheel paths in a rutted pavement. These
cracks allow water to penetrate into the underlying layers, possibly soften-
ing nonstabilized layers and accelerating the development of fatigue cracks
radiating outward from the longitudinal crack. Ravelling of HMA materi-
al adjacent to the longitudinal crack often widens the crack and accelerates
the overall damage to the pavement structure.
The longitudinal cracks between adjacent lanes can be induced by low
temperature, since the density at the joint between paving lanes is lowest,
resulting in low tensile strength. Typically the density of the HMA near a
longitudinal joint is at least 2 to 3 percent lower than the remaining pave-
ment. Low tensile strength, coupled with the possibility of oxidative hard-
ening of the asphalt due to high voids in the mixture in the joint area,
makes that material particularly susceptible to cracking and subsequent
ravelling. Poor adhesion between the asphalt and the aggregate can aggra-
vate the problem adjacent to longitudinal cracking and cause more rapid
deterioration.
Longitudinal cracks at the edges of the wheel path of a tutted pave-
ment are most often caused by heavy loads or high tire pressures applied in
the wheel path. As the HMA in the wheel path is densified by traffic, resid-
ual stresses develop in the HMA adjacent to the wheel path. "When these
stresses exceed the tensile strength of the HMA, cracking occurs. This
problem is aggravated in colder climates when heavy loads are applied to a
Figure 8-9. Typical Case of Longitudinal Cracking at the Centerline

Between Two Adjacent Paving Lanes
498 Hot Mix Asphalt Materials; Mixture Design and Construction
coid asphalt mixture over a soft subgrade. Under such circumstances, a rel-
atively high deflection can occur, causing a longitudinal crack to develop
in the brittle asphalt mixture.
During early stages, these cracks are most often repaired by sealing.
After ravelling begins to develop, the repair is more extensive, since the
materials adja~ent to the crack must be removed and replaced. By the time
a HMA pavement gets into this condition, an overlay is usually the best
solution.
Block Cracking. \Xlhen a HMA layer cracks both longitudinally and

transversely in approximately square shapes, this is referred to as block
cracking. Figure 8-10 show~ a typical case of block cracking in which the
blocks are about 10 to 12 feet (3 to 3.6m)·fong on each side. Typically,
these cracks are caused by the same factors which cause low temperature
cracking. The block cracking pattern most often develops on facilities
which carry low traffic volumes. Because of the low traffic volume, the
asphalt material has the opportunity to exhibit thixotropic hardening, i.e.,
it sets up and develops a type of internal structure which exhibits brittle
behavior under thermal loading. The HMA exposed to low traffic volumes
does not densify sufficiently; hence, the HMA has a relatively high void
content. This high void content results in more oxidation and hardening
-·- J
Figure 8-10. Typical Case of Block Cracking in a HMA Pavement
of the asphalt, leading to thermal cracking. Block cracking can also be

caused by oxidative hardening of the asphalt if mixed too long in the pug-
mill of a HMA batch facility, mixed too hot, or stored too long in silos. All
these mechanisms make the asphalt cement especially susceptible to tensile
strains which can exceed the tensile strain capacity of the asphalt mixture
and cause the block cracking pattern.
Block cracking is more often seen in large paved areas, such as parking
lots or airfield pavements, than on roads and streets. It can be very serious,
especially if the cracks begin to exhibit ravelling and other advanced stages
of deterioration, such as development of secondary cracks. Because the
cracks can be closely spaced, the underlying layers can be exposed to sig-
nificant quantities of infiltrating surface water. These cracks should be
sealed to prevent or minimize intrusion of water.
In advanced stages of block cracking with ravelling, repair can be very
expensive. Often the old surface is removed by milling and overlaid with a
new or recycled HMA material. Cutting out the damaged material at the
cracks and replacing with virgin HMA is seldom satisfactory since the
pavement shrinks and cracks in the old pavement material with the next
cycle of cold temperatures.Therefore, less costly repair strategies are usual-
ly used until the entire surface can be replaced. Either a thick overlay is
placed on the existing surface or the old material is removed and replaced
with a new HMA surface.
Reflection Cracking. Reflection cracks are caused by discontinuities in

the underlying layers which propagate through the HMA surface due to
movement at the crack (Figure 8-11). Reflection cracks may be caused by:
1. Cracks or joints in an underlying concrete pavement;

2. Low temperature cracks in the old HMA surface;
3. Block cracks induced by the old HMA surface or those induced by
subgrade soil cracking due to shrinkage whether stabilized or not;
4. Longitudinal cracks in the old surface; or
5. Fatigue cracks in the old surface.
The name reflection cracking indicates that the crack is initiated by action
in the underlying layers that produce stresses in the HMA surface which
-----·XGeed-the-str-e.ng-th-0£-the-m-a-ter.iaL-Ther.efore,-if-the-new-surface .is bond-
ed to the old surface using a standard tack coat, cracks in the underlying
layer almost always propagate through the new surface within 1-2 years.
Some engineers believe that placement of a thin stress or strain absorb-
ing layer over the old surface absorbs much of the horizontal movement in
500 Hot Mix Asphalt Materials,'Mixture Design and Construction
Figure 8-11. Reflective Cracking in a HMA Surface Over a Jointed

Concrete Pavement
the underlying layer, thereby attenuating the crack before it penetrates the
top layer of HMA. Asphalt rubber and geotextiles have been used as this
intermediate layer. Another repair technique for old HMA surfaces
involves reworking a portion, the top 2 to 4 inches (51 to 102mm), of the
old surface using recycling techniques. This reworking of the old surface
eliminates the cracking pattern in at least a portion of the old pavement
and prevents, or at least retards, the development of reflection cracking.
If the reflection cracking is caused by horizontal and vertical move-
ment of two adjacent portions of a cracked concrete pavement, reflection
cracks through the HMA surface can hardly be prevented. One technique
that has been successfully used to prevent the recurrence of these cracks is
breaking up the underlying slabs into small pieces and then seating them
into the subbase layer using heavy rollers. This cracking and seating process
is followed by a thick (usually equal to or greater than 4 inches or 102mm)
HMA overlay. Another technique that has been successfully used to treat a
concrete pavement prior to overlay is to rubblize the existing concrete. This
method requires a thicker overlay than the crack and seat method but is
not as likely to develop reflective cracks. A third process that has been used
----ro treat reflective··cracks overconcrete pavemertts·ts-to saw cut fnrotignfn 6
'e~------
HMA overlay directly over the existing joints in the concrete. These cuts
are made shortly after construction and before the reflection cracks devel-
op. This produces a straight cut in the HMA that can be easily sealed with
a material such as that specified in ASTM D3405 instead of a wandering
crack that is difficult to maintain. This technique cannot be applied to ran-

dom cracks in the concrete pavement. These techniques are discussed in
detail in Chapter 9.
Figure 8-12. Slippage Cracking in a HMA Surface
Slippage Cracking. Slippage cracks typically occur as a result of poor

bond between the HMA surface and the underlying layer. These cracks
most often show up first in areas where vehicles brake, turn, or accelerate
(Figure 8-12). Slippage cracks form a distinctive U-shape with the top of
the U always pointing in the direction that the forces are being applied.
When these cracks are caused by braking, the end of the slippage cracks
(top of the U) points in the direction of travel. If a slippage crack is caused
by acceleration, the ends of the crack (top of the U) will point toward the
rear of the vehicle. The only repair for areas with slippage cracks is removal
and replacement with new material, being careful to ensure that a proper
bond between the old surface and the new HMA material is obtained.
Removal of the existing slipped material with a milling machine results in
- - - - - . a rough .surface.texture_ that.should_help._pLO:v.ide_go_od_hQn__d_when overlaid.
The material must be removed to a depth slightly below the point at which
slippage has occurred.
A lack of bond may also show up quickly in areas free from braking or
acceleration. A surface layer that is not bonded is not an integral part of the
502 Hot Mil( Asphalt Materials, Mixture Design and Construction
pavement structure and is easily overstressed by normal truck loading and

temperature changes. Therefore, premature cracking develops since there is
no bond between the layers, and the material breaks loose creating a hole
which grows quickly, especially in cold, wet weather. Removal and replace-
ment of the material is the only viable solution to this problem.
Distortion
Distortion distresses are characterized by a permanent change in the

shape of the pavement or pavement layer. The most common type of dis-
tortion on rural highways is wheelpath rutting. However, in urbanized
areas and at intersections, where heavy vehicles move slowly or stop fre-
quently, both rutting and shoving can occur. ]n some cases when unstable
HMA mixes have been placed in intersections, corrugations or waves have
developed. Each of these distress manifestations is discussed in the follow-
ing sections.
Rutting. Ruts are depressions which occur in the pavement's wheel

path as a result of traffic loads. Some negligible amount of rutting occurs
in HMA surfaces due to the continued densification under traffic after ini-
tial compaction during construction. In fact, it is quite common for void
contents of HMA surfaces to be reduced from 7 or 8 percent after con-
struction down to 4 or 5 percent after the first 2 or 3 summers of traffic,
due to densification. In a 4--inch (102mm) thick HMA pavement, this
densification results in a rut depth of approximately 0.12 inches (3mm).
This amount of rutting is insignificant. However, rutting in a pavement
structure can be much larger than 0.12 inches (3mm) if the HMA layer,
underlying layers, or the subgrade soil is overstressed and significant den-
sification or shear failures occur. Therefore, when an unacceptable level of
rutting occurs in a pavement structure, it is imperative that the engineer
determine which layer or layers in the pavement structure are contributing
to the rutting before a selected repair strategy can be successful.
How much rutting is too much? To answer that question, one must
deal with the issue of serviceability. The purpose of a pavement is to pro-
vide a safe, smooth riding surface for vehicular travel. Therefore, when rut-
ting interferes with this function, it has become excessive. Typically ruts are
--~"-o=iform_depxessionsin-the.....wheeLpa.th,-as-Sh-0-wH-~P-ig-u-r:e-8-l-J,a-Btl-<l0--_ _ _ _ ___,
not contribute significantly to the longitudinal roughness. In fact, if the
PSI equation given earlier in this chapter is recalled, the rutting term is
-1.38 RD 2 where rut depth is in inches. Therefore, a 0.5 inch (13mm) rut
would only reduce serviceability by about 0.35 units; however, a 0.5 inch
Figure 8-13. Typical Rutting
(13mm) rut may be entirely unacceptable from a safety standpoint. In con-

sidering the safety issue, the important factor is cross drainage of surface
water. As long as the rut is not deep enough to pond water in the wheel
path, it usually poses no significant safety problem. However, when water
begins to pond in the wheel path, hydroplaning of fast moving vehicles
becomes a possibility, and the safety of the motoring public is jeopardized.
In cold climates this water may freeze, creating a slick condition.
Therefore, the cross-slope of the pavement section is the controlling factor
in determining when a rut is deep enough to accumulate water.
As mentioned previously, before excessive rutting can be effectively
repaired so that it will not reoccur, the cause or causes must be established.
For example, if rutting is caused by overstressing the subgrade soil because
the pavement structure is too weak for the loads, levelling in the wheel
paths and a thin overlay does not permanently solve the rutting problem.
The only solution to such a problem is a complete structural evaluation
followed by a redesign of the structural section and construction of a pave-
ment thick enough to reduce the stresses in the subgrade and other pave-
- - - - - ment layers toa tolerablelev:eL~-- ___________
Problems of overstressing pavement layers can be aggravated by poor
drainage in and along the pavement section. Many pavements have been
constructed with little consideration given to drainage of water that pene-
trates the pavement structure. For these pavements, there are few alterna-
tives which can provide positive drainage of water from within the pave-
ment structure. However, the surface cracks can be kept sealed and the
drainage ditches adjacent to the shoulders can be kept free flowing in order
to quickly remove surface water to prevent intrusion of excess water into
the foundation of the pavement structure. Since most soils and many gran-
ular materials lose strength at high water contents, it is good practice to
keep moisture levels as low as practical in order to maintain pavement
strength. ·
Rutting can be caused by densification in these same pavement mate-
rials because of insufficient initial compaction. The pavement layers may
be compacted to meet specifications; however, the specified density may be
too low. An NCHRP study on granular bases conducted by Rostron, et al.,
(22) showed that many states used AASHTO"T99 compaction efforts on
granular bases even as late as 1971. Base materials compacted to this refer-
ence density are susceptible to significant additional densification under
traffic. Therefore, it is imperative that materials of all types be compacted
during construction to levels very near those densities which will be
reachea after serving traffic for several years.
Perhaps the most common recent cause of rutting is associated with
the HMA layer, especially on routes with heavy loads and high tire pres-
sures. Much of this rutting can be attributed to improper mix design. Some
of the most common mistakes when designing heavy duty HMA mixtures
are:
1. Selection of an asphalt content that is too high;

2. Use of excessive filler material (minus 200 material); and
3. Use of too many rounded particles in both coarse and fine aggre-
gates in the HMA.
Probably the single largest contributor to rutting in HMA is excessive

asphalt content. There are several items which may occur individually or
collectively that contribute to the selection of a design asphalt content that
is too high (21). Some of these items are discussed below.
One cause of a high design asphalt content in HMA is insufficient
compactive effort during mix design and quality control testing. When the
Marshall design method was developed, the selected laboratory compactive
effort was selected to produce a density in the laboratory approximately
equal to that ultimately obtamed m the field under traffic.lt,.......,th_e_c_o_m_-- - - - - -
pactive effort in the laboratory produces a density lower than the ultimate
in-service density, then the asphalt content selected will likely be too high.
A study by Brown and Cross showed that in-place density of HMA often
exceeds the laboratory density (21).
Most results have shown that for high volume roads the laboratory
compactive effort should be at least equivalent to 75 blows per side with
the Marshall manual hammer. Some field studies indicate that even 75
blows with a manual hammer do not produce the density ultimately
achieved in the field under traffic. The obvious result is overfilling the
voids of the in-place HMA mixture with asphalt cement which leads to
plastic flow of the HMA and rutting.
Another problem that has resulted in rutting on a number of projects
has been the use of an excessive amount of material passing the No. 200
sieve (filler). Typically, filler material fills the voids in an asphalt mixture
and lowers the optimum asphalt content. Some amount of filler is neces-
sary to obtain the desired stability, but excess filler results in a mixture at
optimum asphalt content that is brittle and which tends to crack. The
asphalt content must be adjusted for higher filler contents; otherwise, rut-
ting will likely occur.
Mineral filler characteristics vary with the gradation of the filler. If the
size of the mineral filler particles is smaller than about 10 microns, the filler
acts as an extender of the asphalt cement since the thickness of most
asphalt films in dense-graded HMA is less than 10 microns. If the miner-
al filler size is larger than 10 microns, it acts more like an aggregate. If an
excessive amount of this larger sized mineral filler is present, the asphalt
demand may increase because of increased VMA. Therefore, care must be
taken to consider not only the amount of mineral filler, but also its size in
evaluating whether an excessive amount of fine material is present in a mix
being designed or controlled in the field. There is an interaction between
the mineralogy of the fines and the asphalt cement which can cause the
behavior of some material combinations to vary. Discussion of these fac-
tors can be found in numerous reports in the literature by Anderson (25,
26, 22), Craus, J., et al. (2.i), Eick and Shook (2.2), and from several high-
way departments including California (312), West Virginia (3.1),
Washington (32), and Pennsylvania (33).
Mixtures designed with high amounts of natural sands and uncrushed
gravels have low stabilities and are prone to excessive rutting. Many states
require that the coarse aggregates be sufficiently crushed to incxease
Marshall stability above a minimum level. Natural sand is often rounded
or subrounded, which leads to low mixture stability and aggravates the rut-
----~ri~o-g. p.r.o.blem--..Na:t:.ural sands sometimes contain cfay _balls and organic
material which lead to an overall reduction in performance. Many states
limit the amount of natural sand in heavy duty HMA mixes to 10 to 25
percent. The project specifications should specify the percent of fractured
faces for both the fine and coarse aggregate.· Typical requirements for
crushed coarse aggregate call for at least 75 percent of the particles to have
two or more fractured faces and that at least 90 percent of particles to have
at least one fractured face. Typical requirements for crushed fine aggregate
call for at least 90 percent of the particles to have two or more fractured•
faces not including the natural sand.
Shoving and Corrugations. Shoving and corrugations are caused by

shear flow of the mixture or slippage between the layers. The shear flow
may be a result of an unstable mixture, i.e., void content too low for stable
behavior or the applied loads too high forcing aggregate particles to move
over each other, producing shoving or corrugations. A typical case of shov-
ing is shown in Figure 8-14 for an intersection in an urbanized area. Most
cases of unstable mixtures show up first in areas of the pavement where
traffic moves slowly or where stopping occurs. Therefore, in rural areas, ,
truck lanes where the commercial vehicles slow down significantly below
highway speed are often the first or, in some cases, the only area where
shoving occurs. If there is a horizontal curve, these areas will experience
shoving as a result of the horizontal friction forces between the road sur-
face and the tire which are required for the vehicle to negotiate the curve.·
In urbanized areas intersections are typically the first place where excessive
rutting, shoving, or corrugations occur because of the slow moving traffic
and the additional horizontal friction forces caused by braking and accel-
Figure 8-14. Typical Shoving at an Intersection

Chapter 8-Performance I Distress of HMA 507
eration. Bus stops are often the first locations to exhibit premature distor-
tion in the form of shoving. Shoving can be easily detected at edge lane
markings where these lines bend outward in the vicinity of the shoving.
Corrugations typically show up in intersection areas on the side where
vehicles stop, but can occur on the acceleration side if high acceleration
forces occur. This unstable HMA forms ridges of material which are per-
pendicular to the direction of the applied shear forces which cause the flow.
A typical corrugation pattern is shown in Figure 8-15.
Shoving and corrugations occur when the applied forces exceed the
shear strength of the HMA or underlying layers. Shoving can occur when
thin HMA surfaces are placed over granular bases and subbases. Shoving
and corrugations are primarily due to unstable HMA layers. Most of these
problems are similar to the same problems discussed under rutting.
Figure 8-15. Typical Corrugation Pattern
Disintegration
Disintegration is the breakup of the pavement structure. It usually

involves loss of individual pieces or the separation of the individual com-
ponents of the HMA from each other. Disintegration typically takes the
-----foi:rn.s-of..r.a¥ellio.g-,-wear.loss,---s.tripping,__and.p.otholes.__ _
Ravelling. Ravelling is the progressive disintegration of a HMA layer

from the surface downward as a result of the dislodgement of aggregate
particles (Figure 8-16). The opportunity for dislodgement is precipitated
'508 Hot Mix Asphalt Materials, Mixture Design and Construction
by a loss of bond between the aggregate particle and the asphalt coating. A
number of conditions can lead to ravelling:
1. A coating of fine dust on the aggregate thick enough that the

asphalt film sticks to the dust rather than to the aggregate. Surface
traction forces abrade the asphalt film and then dislodge the aggre-
gate.
2. Segregated spots in the surface layer where most of the fine aggre-
gate is absent. In these areas the coarse aggregates are in contact with
each other, but there are only a few contact points where the asphalt
cement bonds the aggregate matrix. With time the asphalt cement
may harden, stripping may weaken the bond between the asphalt
cement and the aggregate, freezing water in the voids may rupture
the bonds, or traction forces may be high enough to break the indi-
vidual particles out of the segregated area. Figure 8-17 shows ravel-
ling due to segregation.
3. Low in-place density in the asphalt surface course. High density is
required to provide sufficient cohesion in the asphalt mixture. The
aggregate is easily dislodged from the mixture when it has been
inadequately compacted.
Figure 8-16. Ravelling of a HMA Pavement Surface

Figure 8-17. Ravelling in a Segregated HMA
A badly ravelled pavement can pose a safety problem if the abraded

material leaves a depression in the pavement surface deep enough to pond
water, which could cause hydroplaning. Additional problems associated
with loose debris on the highway may have to be dealt with, depending on
the rate of progression of the ravelling problem. Loose aggregate on the
surface reduces skid resistance and can be picked up by the tires of moving
vehicles and thrown into the path of passing or trailing vehicles.
Wear Loss. Wear can either be gradual or it can occur very quickly in
some instances. After asphalt cement has oxidized and is eroded, the shear
stresses between tire and surface may be sufficient to remove aggregate par-
ticles. This type of wear is accentuated by studded tires and can become
catastrophic on surface treatments. Brittle materials are particularly sus-
ceptible to this type of wear.
Generally, wear is of the gradual, continuous type, at least as far as
individual aggregate particles are concerned. This wear is greatly accelerat-
ed by loose particles on the surface which serve as an abrasive. Studies of
wear of various minerals were made (~ by subjecting ten mineral sam-
ples, predominantly oxides, to abrasion from dry fine abrasives.
Microscopic photographs of worn surfaces revealed two phenomena,
- - - - scratching and pitting~ Rapidly wearirii m1nerals suffered both types of
damage while slow wearing minerals displayed no scratching and only a
small amount of pitting. Rapid wear occurred when the abrasive was hard-
er than the mineral and was found to be proportional to load and inverse-
ly proportional to mineral hardness. The slow wear of minerals softer than

the abrasive was independent of load. Pitting was related to energy con-
cepts while scratching was related to hardness. Mineral hardness was found
to be the most important parameter in the rapid wear of homogeneous
minerals.
From a skid resistance standpoint, differences in the wear rate of com-
ponents of surface courses are desirable. Different wear rates enhance sur-
face texture, thus improving traction between the tire and pavement sur-
face.
Strippin~. Stripping is a distress that is characterized by the loss of

bond between the aggregates and the asphalt cement which typically begins
at the bottom of the HMA layer and progresses upward. When stripping
starts at the surface and progresses downward it results in ravelling.
Stripping is one of the most difficult distresses to identify because the sur-
face manifestations can take numerous forms: rutting, shoving, corruga-
tions, ravelling, or cracking. Therefore, the only accurate way to determine
if stripping is the cause of the observed distress is to open up the pavement
structure and look at the material removed from the cross-section. If strip-
ping has occurred, there will be visible evidence in the form of partially
coated or uncoated aggregates. In many cases, the stripped layer loses so
much of its structural integrity that cores cannot be removed intact from
the pavement. In some cases all the asphalt cement strips from the aggre-
gate in highly strip prone mixtures.
Research conducted at the National Center for Asphalt Technology
(NCAT) under the SHRP A-003B Project has shown that the physico-
chemical surface properties of mineral aggregate are more important for
moisture induced stripping compared to the properties of asphalt cement
binder. Some mineral aggregates are inherently very susceptible to strip-
ping. However, in many cases external factors and/or in-place properties of
HMA pavements induce premature stripping in HMA pavements (.12). A
proper knowledge of these factors is essential in identifying and solving the
stripping problem. A discussion of these factors follows:
I. Inadequate Pavement Drainage:
Inadequate surface and/or subsurface drainage provides water or mois-

ture vapor which is the necessary ingredient for inducing stripping. If
excessive water or moisture is present in the pavement system the HMA
pavement can strip prematurely. Kandhal et al 00 have reported case his-
tories where the stripping was not a general phenomenon occurring on the
entire project but rather a localized phenomenon in areas of the project

over-saturated with water and/or water vapor due to inadequate subsurface
drainage conditions.
Water can enter the HMA pavement layers in different ways. It can
enter as run-off through the road surface, primarily through surface cracks.
It can enter from the sides and bottom as seepage from ditches and high
water table in the cut areas.
The most common water movement is upward by capillarity under a
pavement. Above the capillary fringe water moves as a vapor. Many sub-
bases or subgrades in the existing highway system lack the desired perme-
ability, and therefore, are saturated with the capillary moisture. The con-
struction of multilane highways (or widening) to greater widths, gentler
slopes and milder curves in all kinds of terrain has compounded the sub-
surface drainage problem. Doubling the road width, for example, makes
drainage about four times as difficult as before (.32). Quite often, a four-
lane highway is rehabilitated by paving the median and shoulders with
HMA resulting in a fully paved width of72-78 feet which is equivalent to
a six-lane highway without any increase in the subsurface drainage capa-
bility.
Figure 8-18. Typical Surface Distress Resulting from Stripping

Air voids in the HMA pavement may become saturated with water
even from vapor condensation due to water in the subgrade or subbase. A
temperature rise after this saturation can cause expansion of the water
trapped in the mixture voids resulting in significant void pressure when the
voids are saturated. The pore pressure from stresses induced by traffic can
also cause the failure of the binder-aggregate bond. Initially, the traffic
stresses may further compact the mixture and trap or greatly reduce the
internal water drainage. Therefore, the internal water is in frequent motion
(cyclic) and considerable pore pressure is built up under the traffic action.
Tell-tale signs of water damage to HMA overlays (over concrete pave-
m.ents) have been described by Kandhal et al (35). They observed wet spots
on the HMA overlay surface scattered throughout the project. Usually at
these wet spots water oozed out during hot ah@rnoons. Some of the wet
spots contained fines suspended in the water which were tracked on the
pavement by the traffic and appeared as white spots. Most white spots
turned into fatty areas (resulting from asphalt stripping and migrating to
the surface) which usually preceded the formation of potholes. Figure 8-18
shows all three stages: white spots, fatty areas, and potholes on a four-lane
highway. Figure 8-19 shows severely stripped aggregate in a pothole.
Figure 8-19. Cfose-Up of Pothole Showing Severely Stripped

Aggregate
Usually stripping in a four-lane highway facility occurs first in the slow

traffic lane because it carries more and heavier traffic compared to the pass-
ing lane. Typically, stripping starts at the bottom of HMA layer and pro-
gresses upwards.
It is evident from the preceding discussion that inadequate subsurface
drainage is one of the primary factors inducing premature stripping in
HMA pavements.
2. Inadequate Compaction:
Inadequate compaction of HMA mat is probably the most common

construction-related factor responsible for premature stripping. Studies
have shown that at less than 4-5 percent air void content in the HMA the
voids are generally not interconnected and thus almost impervious to
water. Most HMA mixes are designed to have 3 to 5 percent air void con-
tents. When constructed, a maximum air void content of 8 percent (at least
92 percent of the theoretical maximum specific gravity) is specified by
most agencies. It is assumed that the pavement will get densified to the
design air void content under 2-3 years traffic. However, some agencies do
not exercise good compaction control resulting in air voids content higher
than 8 percent at the time of construction, which makes the HMA pave-
ment pervious to water. If the HMA pavement remains pervious for an
extended period of time, stripping is likely to occur due to ingress of water
and hydraulic pore pressures induced by the traffic.
3. Excessive Dust Coating on Aggregate:
The presence of dust and clay coatings on the aggregate can inhibit an
intimate contact between the asphalt cement and aggregate, and provide
channels for penetrating water. The asphalt cement coats the dust coating
and is not in contact with the aggregate surface. It has also been hypothe-
sized that some very fine clayey material may cause stripping by emulsify-
ing the· asphalt cement binder in presence of water.
4. Use of Open-Graded Asphalt Friction Course:
-------0-ev-er-aLst-ates-i.n-the..south~tem-Uni-ted-States-experienced stripping
in the HMA course underlying open-graded asphalt friction course
(OGFC) during the late 1970s. It has been hypothesized that the OGFC
retains moisture for a longer time and does not dry out after rain as fast as
a conventional dense-graded HMA surface. The water in OGFC is also
pressed into the underlying course by the truck tires initiating the stripping
action which can cause flushing, rutting or shoving at the surface. Several
states suspended the use of OGFC in early 1980s. Some studies have also
shown that the stripping in the layers underlying OGFC resulted from
their high air void content (lack of adequate compaction). Evidently, it is
all the more desirable to have an impervious HMA course below the
OGFC to minimize stripping. If the OGFC must be used, the underlying
HMA course should be tested for moisture susceptibility.
5. Inadequate Drying ofAggregates:

Laboratory studies (311) have shown that _high residual moisture con-
tent in the mineral aggregate prior to mixing,:with asphalt cement binder
increases the potential for stripping. Undoubtedly, a dry aggregate surface
will have increased adhesion with the asphalt cement co,mpared with a
moist or wet surface.
6. Weak and Friable Aggregate:

If weak and friable aggregates are used in the HMA mix, degradation
takes place during rolling and subsequently under heavy traffic.
Degradation or delamination exposes new uncoated aggregate surfaces
which can readily absorb water and initiate the stripping phenomenon in
the mix. Also, if not observed carefully, these uncoated aggregate surfaces
can mistakenly be deemed as stripped aggregate particles. Obviously, use of
sound and durable aggregate in the HMA is recommended.
7. Overlays on Deteriorated Concrete Pavements:

Many concrete pavements of interstate and primary highways are dete-
riorating before the end of their design lives. Recent years have seen
increased use of HMA overlays over these existing concrete pavements,
some of which have faulted, spalled, cracked, and water-pumping slabs.
Dense-graded subbase _material under concrete pavements can hold con-
siderable amounts· of water which escape through cracks, longitudinal and
transverse joints. Once the concrete pavement is overlaid. with an impei,-vi-
ous HMA course the water is trapped underneath. Excessive pore pressure
--..iS-bu.ilt-Under-the-tr-affic;-initiat-i-ng-stFippiag-arui-s1:1:bsequen-1cly-potholing.-·--·--·"
Whenever a concrete pavement isI due to be overlaid for the first time, it is
necessary to evaluate the existing drainage conditions. If necessary, the pro-
ject must include installation of a positive drainage system. Unless this is
done, the problem of stripping and potholing will persist indefinitely.
Chapter 8--Performance / Distress of HMAs 515
If the existing concrete pavement is badly deteriorated, cracked, and

pumping water due to inadequate subsurface drainage, it is recommended
to provide a 4-inch drainage layer of open-graded asphalt treated perme-
able material (ATPM) directly above it prior to placing the dense-graded
HMA overlay (32, ~- This drainage layer should be connected to the
edge drain(s). The ATPM will not only drain the water very efficiently, it
will prevent any moisture vapor buildup in the pavement system. The
ATPM has been used successfully in such applications. It will also help to
minimize reflection cracking emanating from the concrete pavement. If
required, the ATPM can also be placed over concrete pavements which
have been subjected to crack and seat, break and seat, and rubblizing oper-
ations.
8. Waterproofing Membranes and Seal Coats:

If the source of moisture is from beneath the pavement, which is usu-
ally the case, then sealing of the road surface can be detrimental. Use of
some waterproofing membranes (such as stress absorbing membranes to
minimize reflection cracking) and seal coats between the pavement cours-
es or at the surface acts like a vapor seal or a vapor barrier.
The entrapped moisture or vapor can induce stripping in the pave-
ment layers underlying waterproofing membranes and seal coats. The
potential for stripping should, therefore, be considered whenever such sys-
tems are used.
Antistripping Agents (AS):

AS agents might be needed if it has been established that a particular
HMA mix is inherently prone to stripping based on the results of the field
investigations and moisture susceptibility tests.
Liquid Antistripping Additives. Most of the liquid AS agents are sur-

face active agents which, when mixed with asphalt cement, reduce surface
tension and, therefore, promote increased adhesion to aggregate. The
chemical composition of most commercially produced AS agents is pro-
prietary. However, the majority of AS agents currently in use are chemical
compounds that contain amines (1.1). These AS agents must be "heat sta-
-----<-·te"-;-th:i.ris;-tlrey-should.rorl.-osecheireffectiverress-when-the·· modified
asphalt cement is stored at high temperatures for a prolonged period of
time.
The simplest and most economical way is to mix the AS agent with the
asphalt cement in a liquid state prior to mixing the asphalt cement with the
aggregate. The amount of AS agent to be used is important. Too little may

not be effective and too much may be detrimental to the HMA mix.
Lime Additives. Unlike liquid AS agents which are usually added to

the asphalt cement, lime is added to the aggregate prior to mixing with
asphalt cement. Many studies indicate that lime is a very effective anti-
stripping agent. However, its antistripping mechanism is not well under-
stood. Various mechanisms have been postulated: (a) lime interacts with
acids in the asphalt cement that are readily adsorbed on the aggregate sur-
face, (b) lime provides calcium ions which can replace hydrogen, sodium,
potassium and other cations on the aggregate surface, and (c) lime reacts
with most silicate aggregates to form a calcium silicate crust which has a
strong bond to the aggregate and has sufficient porosity to allow penetra-
tion of the asphalt cement to form another strong bond.
Both hydrated lime Ca(OH) 2 and quick lime CaO are effective,
although the former is most commonly used. Dolomitic limes (both Type
S and N) have also been used as antistripping additives. Generally, I t.o
I 1/2 percent of lime by weight of dry aggregate is used. Finer aggregates
may require higher percentages because of increased aggregate surface area.
Aggregates have been treated with lime by the following four methods:
I. Dry hydrated lime: The main problem in using dry lime is to

maintain its coating on the aggregate surface until it is coated with
asphalt cement. It is more critical in drum mixers which tend to
pick up some of the lime in the exhaust gas flow. However,
Georgia DOT has successfully used dry hydrated lime in drum
mixers by injecting lime into the drum just ahead of the asphalt
cement;
2. Hydrated lime slurry: This method requires additional water to
be added to the aggregates which results in increased fuel costs and
reduced HMA production rates. Additional mixing equipment is
also needed.
3. Dry hydrated lime to wet aggregate: In this method dry hydrated
lime is added to wet aggregate, usually containing 3-5 percent
water, and then mixed in a pugmill or tumble mixer to obtain a
homogeneous mix. Dry lime can also be added to dry or moist
aggregate and then sprayed with water. Unless the water content is
-------+,0w,i:n.ereased-fu:el-oost-s-and-reduced-HM:A-production-rates-will
result.
4. Hot (Quicklime) slurry: The use of quicklime (CaO) has at least
two advantages: (a) its cost is equal to that of hydrated lime but
when slaked the hydrated lime yield is 25 percent greater, and (b)
the heat froin slaking results in an elevated temperature which

helps in the evaporation of the added moisture. It should be han-
dled with caution because it can cause burns on humans.
Comparative laboratory and field studies involving the preceding four

treatments have been generally inconclusive and therefore, increased fuel
and equipment costs and decreased HMA production rates associated with
the wet process may not be justified at the present time.
Potholes. Potholes are relatively small holes that penetrate all the way
through the surface of HMA and down into the base course. Potholes
occur most frequently on roadways with relatively thin HMA surfaces, typ-
ically 1 to 2 inches (25 to 50mm), and seldom occur on roadways with
HMA surfaces more than 4 inches (100mm).
The occurrence of potholes often coincides with a period of heavy
rainfall during which water penetrates the HMA surface through cracks,
usually closely spaced alligator cracks, and softens the granular base course.
In northern climates, potholes generally occur during the spring thaw. If
the cracked pieces of the pavement lose their bond to the surrounding
HMA material, they are removed by the action of the tire as it rolls over
the material. Fine material is pumped through the cracks so that the under-
lying base support is weakened, resulting in removal of the adjacent mate-
rial by traffic. Once the first piece is dislodged, the pothole grows rapidly
since all other pieces are more easily dislodged than the first piece.
Potholes can develop along any crack that occurs in a pavement. For
example, either longitudinal or transverse cracks can deteriorate at loca-
tions where the wheel continually crosses the crack. Since all unsealed
cracks allow water to penetrate, base softening occurs which leads to either
ravelling adjacent to the crack or fatigue cracking on the edges of the crack.
Once ravelling or fatigue cracks begin to occur, pothole development is a
real possibility.
Pothole development can be minimized if good surface and subsurface
drainage is provided. Potholes are normally repaired by patching with
either cold mix or HMA. In either case, proper patching techniques are
absolutely essential if the integrity of the pavement is to be restored in the
pothole area. Discussion of proper maintenance techniques is contained in
Chapter 9.
------~~--------------------- ----- ------ -
Loss, of Frictional Resistance
Frictional resistance is defined as the relationship between the vertical

force and the horizontal force developed as a tire slides along the pavement
surface. To the vehicle operator, frictional resistance is a measure of how

quickly a vehicle can be stopped. To the design engineer, frictional resis-
tance is an important safety-related property of the pavement surface that
must be accounted for through proper selection of materials, design, and
construction. In terms of pavement management, frictional resistance is a
measure of serviceability. The deterioration of frictional resistance below a
minimum acceptable (safe) level prevents the pavement from serving its
desired function. In a life cycle analysis of pavement performance, the cost
of restoring frictional resistance at some point in the pavement's life must
be considered. Therefore, predictive methods for the deterioration of fric-
tional resistance are important to both the pavement designer and the
owner agency.
Considerable research has been devoted·1:o the measurement of fric-
tional resistance and the development of devices for making the measure-
ments. Friction resistance (commonly called skid resistance) is defined
according to the American Society for Testing and Materials (ASTM) as:
FN = 100 FIL
where,
FN = friction number;
F = frictional resistance to motion at the pavement
surface; and
L = load normal to the pavement surface.
The friction resistance can be determined by a number of methods which

can be categorized as:
1. Locked-wheel-trailer methods. Usually the ASTM Method E 274

is prescribed in which the tire, the method of applying water to the
pavement, the speed (40 mph or 64 km/hour), etc. are specified.
2. Yaw mode. This is the use of tires in other than the locked-wheel
mode, e.g., the Yaw mode, to determine a sideways friction factor.
Examples of devices which measure frictional resistance in the Yaw
mode are the SCRIM machine and the Mu meter.
3. Portable friction testers. These are portable devices which can be
used both in the laboratory and the field to measure frictional resis-
tance. Examples are the British Portable Tester, the California Skid
----------+·es-1:er-,----a-flcl-1:h-e-Keyst-one-B-rag----Tester:-'Fhese-devrces-rypi:caliy-a1<·----· ·····--
most sensitive to microtexture.
4. Automobile methods. Automobiles equipped for braking with a
diagonal pair of wheels measure the distance travelled from a spe-
cific speed to a full stop.
The majority of measurements of in-service pavements are made by the

locked-wheel-trailer method or the Yaw mode procedure while laboratory
measurements are made using the portable friction testers.
Frictional resistance characteristics that are desirable in a good pave-
ment surface are (12):
I. High frictional resistance. Ideally the frictional resistance when wet

should be almost as high as that of the dry pavement.
2. Little or no decrease of the frictional resistance with increasing
speed. The frictional resistance of dry pavements is nearly indepen-
dent of speed, but this is not the case for wet pavements.
3. No reduction of frictional resistance with time, from polishing or
other causes.
4. Resistance to wear by abrasion of aggregate, attrition of binder or
mortar, or loss of particles.
5. Structural durability. Resistance to compaction, ravelling, breakup,
etc.
6. Low tire noise generation.
7. Low cost. Not necessarily low first cost, but cost per year of service
with acceptable frictional resistance.
The frictional resistance of a pavement surface is a function of the sur-

face texture which is divided into two components (13_, 44, 15), microtex~
ture and macrotexture. The microtexture provides a gritty surface to pene-
trate thin water films and produce frictional resistance through good fric-
tion between the tire and pavement surface. The macrotexture provides
drainage channels for water expulsion between the tire and the roadway
which allows better tire contact with the pavement to improve frictional
resistance and prevent hydroplaning. The microtexture contributes to fric-
tional resistance at all speeds, but it is the dominating influence at speeds
less than 30 mph. In contrast, the macrotexture is less important at low
speeds, but is essential at high speeds in wet conditions. Average texture
depths up to 0.5 mm are generally classified as microtexture and larger tex-
ture depths are classified as macrotexture. The pavement microtexture
develops frictional resistance through friction between the tire and the
roadway: The sharp, fine particles in the surface penetrate thin _water films
and permit contact between the tire and the roadway. Microtexture
depends largely on the mineral composition and the rugosity of the aggre-
gates. The fine, hard grains in the surface of the coarse aggregates or fine
aggregates provide the microtexture.
Macro texture can be obtained by controlling the gradation (46) of the

aggregates in HMA. The open-graded friction course (OGFC) is a good
example of a mix with a satisfactorily designed macrotexture. The aggre-
gates must not polish readily if good frictional resistance is to be main-
tained throughout the life of the pavement. Ideally, aggregates for HMA
should be composed of hard, coarse, angular minerals well bonded into a
softer matrix so that gradual differential wear occurs.
Characteristics of aggregates that influence frictional resistance include
type, size (gradation), shape, texture (micro and macro), hardness (suscep-
tibility to wear and resistance to polishing), and mineral composition.
Suitable aggregates may be either natural or synthetic. The most common-
ly used natural aggregates are those obtained by quarrying and crushing
rocks such as sandstone, granite and diabase-. They may also consist of
stream and bank gravels that are obtained by dredging, washing, and
screening. These. gravels are usually crushed to improve their angularity
since shape is critical in providing good frictional resistance and good mix-
ture stability.
Two types of synthetic aggregates have been used to improve friction-
al resistance. The first type, which is produced by heating natural clays and
shales, is called "expanded lightweight aggregate." The second type, called
slag, is produced as a by-product from blast furnaces in the manufacture of
iron and steel. Synthetic aggregates have texture and chemical characteris-
tics decidedly different from those of natural aggregates such as limestone,
sandstones, and gravels, and therefore have widely differing frictional resis-
tance qualities. In general, however, they are less susceptible to polishing
than are the natural aggregates, but they may tend to abrade more rapidly.
The ability of an aggregate to resist the polish-wear action of traffic has
long been recognized as an important characteristic (42). When an aggre-
gate becomes smooth, it has poor frictional resistance. Also, if it wears
(abrades) and polishes too rapidly, the pavement loses its texture and
becomes very slippery when wet.
The shape of an aggregate particle also affects its frictional resistance
properties. Aggregate shape depends on many of the same factors that
influence its texture. These factors include hardness of grains, strength of
the matrix, and overall resistance of the aggregate to abrasion. Also, pro-
cessing procedures govern the shape of both natural and synthetic aggre-
gates. Crushed aggregates exhibit an angular shape, whereas those from
--st--r-eam-beds-ri:erma:l-ly-exh-i-b-i-t-a-rou:nd:ed-shape:-'Fhe-a:ngularity-oran·aggre~-----
gate contributes to its frictional resistance qualities as long as it remains
angular. Angularity relates to the manufacturing process, but the retention
of angularity depends on such characteristics as mineralogical composition
and the amount of polish-wear produced by traffic.
Polishing is the reduction in microtexture, resulting in the smoothing

and rounding of exposed aggregates. This process is caused by particle wear
on a microscopic scale and is difficult to quantify. Low-speed friction mea-
surements, such as the British Portable Friction Tester (Figure 8-20), have
been used in an attempt to quantify polishing (dZ). Such measurements
are, of course, made with the surface wetted, but the rubber shoe of the
pendulum tester displaces enough water so that the hydrodynamic effects
(which influence friction at higher speeds and which are controlled by the
coarser features of the surface) are practically absent. The rubber responds
essentially only to microtexture.
Figure 8-20. British Portable Friction Tester
On the road, most of the polishing probably takes place while the
pavement is dry because most pavements are dry 96-97 percent of the time
_ _ _ ______,.4@__The-number-0£passes,_w.hethe.c...0£.a_pacLor-a...tire,-go:v:ernuhe. polish-
ing rate, but typically it takes two or more years of service before the sur-
face has been polished to its final state. The frictional resistance may, how-
ever, continue to drop because wear continues to reduce the macrotexture,
decreasing the channels available for clearing the water from under the tire.
REFERENCES
1. The WASHTO Road Test, Part 2: Test Data Analyses and
Findings. Highway Research Board Special Report 22, 1955.
2. Carey, W.N. and P.E. Irick. The Pavement Serviceability -
Performance Concept, Highway Research Board Bulletin 250,
1960.
3. Haas, R. and WR. Hudson. Pavement Management Systems.
McGraw-Hill Book Company, New York, NY, 1978.
4. Spangler, E.B. and WJ. Kelley. GMR Road Profilometer: A
Method for Measuring Road Profile. General Motors Corporation,
December 1964.
5. Hveem, EN. Devices for Recording and Evaluating Pavement
Roughness. Highway Research Board, Bulletin 264, 1960.
6. Hudson, WR., WE. Teske, K.H. ·Dunn, and E.B. Spangler. State
of the Art of Pavement Condition Evaluation. Highway Research
Board Special Report 95, 1968.
7. Roberts, EL. State of the Art of Estimating Pavement Serviceability
Using Roughness Measurements. Proceedings, American Society of
Civil Engineers, Specialty Conference in Pavement Design for
Practicing Engineers, Atlanta, GA, June 1975.
8. Wambolt, J.D. State of the Art of Measurement and Analysis of
Road Roughness. Transportation Research Board, Transportation
9. Hudson, WR., G.E. Elkins, W Uddin, and K.T. Reilly. Improved
Methods and Equipment to Conduct Pavement Distress Surveys.
Federal Highway Administration Report FHWA-TS-87-213,
1987.
10. Brokaw, M.P. Development of the PCA Road Meter: A Rapid
Method for Measuring Slope Variance. Highway Research Board,
Highway Research Record 189, 1967.
11. Phillips, M.B. and G. Swift. A Comparison of Four Roughness
Measuring Sy$tems. Highway Research Board, Highway Research
Record 291, 1969.
12. Rainhart Co. Mays Ride Meter Booklet, Austin, Texas, 1972.
---~U._S_p_angler,_E...R_and__W...J_.__Kelle.y_Semo.:S.eis.m-i-C-Method-0£_ _ _ __
Measuring the Road Profile. Highway Research Board, Bulletin
328, 1962.
14. Automatic Road Analyzer Mobile Data Acquisition Vehicle.
Product Bulletin, Highway Products International, Inc., 1985.
Chapter 8-Perforrnance / Distress of HMAs 523
15. DeWilder, M. Proceedings, Roadway Evaluation Equipment

Workshop- GERPHO and the APL. Federal Highway
Administration Report FHWA-TS-85-210, May 1985.
16. Summary Results of the 1987 MSHTO Rideability Survey.
MSHTO Highway Subcommittee on Construction, MSHTO,
1988.
17. Janoff, M.S. The Effect oflncreased Initial Pavement Smoothness
on Long Term Pavement Performance and Annual Pavement
Maintenance Costs. Presented at NAPA Annual Meeting,
Honolulu, HI, January 1990.
18. Foster, C.R. Pavement Smoothness. National Asphalt Pavement
Association, Information Series 53, 1982.
19. Kennedy, T.W, EL. Roberts, and J.B. Rauhut. Distresses and
Related Material Properties for Premium Pavements.
Transportation Research Board, Transportation Research Record
715, 1979.
20. Welborn, J.Y. State of the Art in Asphalt Pavement Specification.
Federal Highway Administration Report No. FHWA/RD-84/075,
July 1984.
21. Hveem, E, E. Zube, and J. Skog. Proposed New Tests and
Specifications for Paving Grade Asphalts. Proceedings, Association
of Asphalt Paving Technologists, Volume 32, 1963.
22. Rostron, J.P., EL. Roberts, and W Baron. Density Standards for
Field Compaction of Granular Bases and Subbases. Transportation
Research Board, NCHRP Report 176, 1972.
23. Brown, E.R. Design of Asphalt Mixtures for High Tire Pressures.
Proceedings, National Conference on Effect of High Tire Pressures,
Austin, Texas, 1987.
24. Brown, E.R. and Steve Cross. Comparison of Laboratory and Field
Density of Asphalt Mixtures. Paper presented at Transportation
Research Board, January 1991.
25. Anderson, D.A. and J.P. Tarris. Effect of Baghouse Fines on
Mixture Design Properties. National Asphalt Pavement
Association, QIP 102, 1982.
26. Anderson, D.A. Guidelines on the Use of Baghouse Fines.
National Asphalt Pavement Association, Information Series 101,
--------.....N-T-o-v-em__,.b_e_rr9·8 .
27. Anderson, D.A. and J.P. Tarris. Adding Dust Collection Fines to
Asphalt Paving Mixtures. Transportation Research Board, NCHRP
Report 252, 1982.
/28. Craus, J., I. Ishai, and A. Sides. Some Physio-Chemical Aspects of

the Filler Effects and the Role of the Filler in Bituminous Paving
Mixtures. Proceedings, Association of Asphalt Paving
29. Eich; J.M. and J .F. Shook. The Effects of Baghouse Fines on
Asphalt Mixtures. The Asphalt Institute, Research Report
78-3(RR-78-3), November 1978.
30. Scrimsher, T. Baghouse Dust and Its Effects on Asphaltic Mixtures.
California Department of Transportation, Report CA-DOT-TL-
3140-1-76-50, October 1976.
31. Ward, R.G. and J.M. McDougal. Bituminous Concrete Plant Dust
Collection System - Effects of Using Recovered Dust in Paving
Mix. West Virginia Department of Highways, Research Report
FHWNWV-79-003, December 1979.
32. Gietz, R.H. Interim Report: Mineral Fines Effect on Asphalt
Viscosity. Washington State Department of Transportation, Report
No. 164, April 1980.
33. Kandhal, P.S. Evaluation of Baghouse Fines in Bituminous Paving
Mixtures. Proceedings, Association of Asphalt Paving
34. Stiffler, A.K. Relationship Between Wear and Physical Properties of
Roadstones. Highway Research Board, Special Report 101, 1969.
35. Kandhal, P.S. Field and Laboratory Investigation of Stripping in
Asphalt Pavements: State-of-the-Art Report. Transportation
Research Re~ord 1454, 1994.
36. Kandhal, P. S., C. W. Lubold, and F. L. Roberts. Water Damage to
Asphalt Overlays: Case Histories. Proc. Assoc. of Asphalt Paving
37. Cedergren, H.R. and W. R. Lovering. The Economics and
Practicality of Layered Drains for Road Beds. Highway Research
Record 215, 1968.
38. Parker, F. Field Study of Stripping Potential of Asphalt Concrete
Mixtures. Alabama Highway Department, Report ST 2019-6,
Aug. 1989.
39. Cedergren, H. R., J. A. Arman, and K. H. O'Brien. Development
- - - - - e > : -f_G.uidel.i-r:1eS-fo.r-the-D.es.ig:n-0£...£.ubsu-r-fa{;e---D-rainage-..£y.stems .- - - - - -
FHWA, Report RD-73-14, Feb. 1973.

40. Forsyth, R. A. Asphalt Treated Permeable Material - Its Evolution.
and Application. National Asphalt Pavement Association, QIP
series 117, 1991.
Chapter 8-Performance I Distress of HMAs 525
41. Tunnicliff, D. G. and R. E. Root. Antistripping Additives in

Asphalt Concrete: State-of-the-Art Report. Proc. Assoc. of Asphalt
42. Skid Resistance. NCHRP Synthesis 14, 1972.
43. Balmer, G.G., and R.R. Hegmon. Recent Developments in
Pavement Texture. Transportation Research Board, Transportation
44. Balmer, G.G. Pavement Texture: Its Significance and Development.
666, 1978.
45. Leu, M.C., and J.J. Henry. Prediction of Skid Resistance as a
Function of Speed From Pavement Texture Measurements.
666, 1978.
46. Kummer, H.W, and WE. Meyer. Tenative Skid-Resistance
Requirements for Main Rural Highways. NCHRP Report 37,
1967.
47. Kandhal, P.S., F. Parker, and E.A. Bishara. Evaluation of Alabama
Limestone Aggregates for Asphalt Wearing Courses.
1418, 1993.
48. Safety Effectiveness Evaluation: Selected State Highway Skid
Resistance Programs. National Transportation Safety Board,
Report NTSB-SEE-80-6, Sept. 1980. .
L
CHAPTER 9. MAINTENANCE,
REHABILITATION, AND
RECONSTRUCTION OF HMA
INTRODUCTION
The performance of a pavement is affected by the type, time of appli-

cation, and quality of the maintenance it receives. Preventive timely main-
tenance slows the rate of pavement deterioration due to traffic and envi-
ronmentally applied loads. Delays in maintenance and deferred mainte-
nance increase the quantity of defects and their severity so that, when cor-
rected, the cost of repair is greater. Continued deferral of maintenance and
rehabilitation actions shortens the time between overlays and reconstruc-
tion, and thus increases the life cycle costs of a pavement considerably.
A 1979 report (D on the economic impacts of deferred maintenance
showed that rehabilitation with an overlay before failure resulted in a sav-
ings of 3¼ inches (95 mm) in the required thickness of the overlay
(Figures 9-1 and 9-2) with an attendant cost savings of over $100,000 per
mile (Figure 9-3). Notice that the difference in the amount of traffic
served between Figure 9-1 and 9-2 is quite small and often translates into
only one or two years of additional service before failure. However, the cost
consequences are very significant and indicate the critical need for observ-
ing distress development in pavements that carry large volumes of heavy
vehicles. This compounding effect of deferred maintenance has been
expressed in a more general way by Stevens (2) and is shown in Figure 9--4.
Notice in the figure that for the first 75 percent of the life of the pavement,
the quality varies from very good to fair, which constitutes about 40 per-
cent of the quality. Notice that the rate of deterioration of quality increas-
es dramatically so that the next 40 percent of the quality is consumed dur-
----·-ng-Fae-fl.e:X-t-1-7--per-c-eftt-0f-t-h.-e--t-ifl."le;---It-i-s-obvious-then-that-it-is prudent to
execute the needed pavement maintenance and repairs to keep the quality
in the upper 40 percent region, especially since each dollar spent then saves
4 to 5 dollars in the next 2 to 3 years. Therefore, it seems apparent that the
most critical point in the life of a pavement is when it enters the fair region,
1310
:r:
<!lu
z~
u J 8
<{ Cl)
LJ...C/)
a::w
~~ 6
Cl)!:!
:::,:C
o1-
z1- 4
:EZ
~'.j
iii~ 2
5
0
w 0
1,000,000 2,000;000 3,000,000

ACCUMULATED 18,000 LB. EQUIVALENT AXLES
Figure 9-1. Overlay Thickness Requirements when Repaired before

· Failure (1)
T5-3/4"
1
1,000,000 2,000,000 3,000,000
ACCUMULATED 18,000 LB. EQUIVALENT AXLES
Figure 9-2. Overlay Thickness Requirements when Applied after

Failure (1)
1
which corresponds to a PSI level between 2 and 3. Maintenance decisions
made at this time appear to be very cost effective, especially if the rate of
deterioration relative to traffic is high.
----A-f:1:1:1:r-ry--e-f-aet-iv-i-ty-b-y-me-st-at-es-clt1:r-i-flg-1:he-l-980s-i-n-cliea:tes--a-ree0gn-....
· -----
tion that timing of maintenance activities is critically important.
7-
Additionally, many states have attempted to identify the specific types of
distresses which should have priority in maintenance because of the high
consequences of deferring maintenance on those types of distresses. The
Chapter 9-Maintenance, Rehabilitation, and Reconstruction of HMA 529
COST PER MILE OF $ [61,000

PAVEMENT 40' WIDE
,..,.,.,..,=== $ 56,000
ili\~1111
Before Failure After Failure
Figure 9-3. Cost Comparison for Overlay Strategies Shown in Figures
. 9-1 and 9-2 (1)
VERYriiiiiiiiiiiiiiiiiiii7~5.;%;T:l:ME~--~:t
GOOD
40 % EACH $1.00 OF
QUALITY DROP RENOVATION
GOOD
!~ COST HERE •••
FAIR
40 % ~ILL COST $4 TO
·- -;JIii-
POOR QUALITY DROP $5 IF DELAYED
TO HERE
VERY
POOR
4 8 12 16
YEARS
Figure 9-4. Pavement Deterioration/Rehabilitation Relationship (2)
relative effects of deferring certain types of maintenance on safety, condi-

tion of the facility, liability, level of service, energy consumed, and soci-
ety-environmental impacts have been established. Given today's social and
legal environment;safetyand lfa.bilitymay likely 15e at the top of the pri-
ority list, followed by condition of facility and level of service. The type of
facility plus traffic volume and composition of the traffic stream will like-
ly determine whether condition of facility or level of service is given the
higher priority. It seems apparent that several of these factors are affected
by the same group of conditions such as roadway type, traffic volume and
makeup, federal aid available, and expectations of the public.
The importance of effective maintenance is reflected in the fact that
while, nationwide, the number of miles of the roads in poor condition sta,-
bilized by 1977, the number of miles deteriorating from good to fair con-
dition increased. The FHWA estimated that the amount of total highway
expenditure used nationally for physical maintenance in 1975 conJinued
to rise at a rate of about $300 million per year and amounted to almost
one-third of total highway expenditures. By the late 1980s the rate of
increase in maintenance funds grew to almost $600 million per year and
amounted to over half of total highway expenditures. This change in
expenditure category from a capital intensive program to a maintenance
intensive program reflects the maturing of tli:e highway system, especially
as the construction o(the interstate system has been completed and the
emphasis changed to rehabilitation and reconstruction of the oldest sec-
tions of the interstate system.
The proper identification of pavement defects and the determination
of their causes is crucial to the selection of proper maintenance treatments
and repair strategies. In addition, a recognition that different pavements
deteriorate at different rates is important in selecting the proper time for
maintenance. Factors which affect pavement deterioration rates are traffic
loads, weather, materials, thickness, construction quality, and effectiveness
of previous maintenance. Generally, the rate of deterioration increases with
use and age.
The primary reasons for maintenance are to correct deficiencies pro-
duced by distress, to protect the investment, and to provide a smooth, safe
road for the user. Therefore, it is important to determine which distresses
relate most to these factors. Darter and Shahin (3) identified what they
believed to be the principal factors which initiated maintenance and the
types of distresses that were related to those factors (shown in Figure 9-5).
Only alligator cracking, ravelling/weathering, and rutting show up under
all three of the principal factors which initiate maintenance.
MSHTO (.1'J has identified 17 types of HMA pavement distresses,
which are listed in Table 9-1. Table 9-2 relating distresses and mainte-
nance/ rehabilitation methods prepared by Hudson et al. (i) should be use-
ful in familiarizing the reader with the elements of typical maintenance and
---;ehabil-i1:a1:i-0-R-1:eehn-i-q_-1:1-@s.-'I'al,le~-9-gi:ves-t-he-p0ssie-le-eauses--of--l-7--El.-ist.r~s,,,_____
types and rehabilitation alternatives recommended by the FHWA (.6).
Various maintenance, rehabilitation, and reconstruction techniques gener-
ally used by the highway agencies are discussed in the following sections.
Relative Need for Asphalt Pavement Rehabilitation
User-Related Factors Maintenance Needs Protection of Investment

• Roughness All distress types at Cracking
Corrugations medium to high • Alligator
• Bumps severity levels • Block
• Depressions
• Potholes Ravelling/Weathering
• Swells
Rutting
Skid
• Bleeding
Polished
Aggregates
• Rutting
• Pothole Potential
• Cracking
Weathering/
Ravelling
• Traffic Volume
Figure 9-5. Distress Types Having a Primary Effect on Factors which

Initiated Maintenance Action (J)
MAINTENANCE
Generally, maintenance activities are divided into two categories, pre-
ventive and corrective maintenance. Preventive maintenance is that group
of activities performed to protect the pavement and decrease the rate of
deterioration of the pavement quality. Corrective maintenance is that
group of activities performed to correct a specific pavement failure or area
of distress. Some maintenance methods serve both functions. The follow-
ing sections include descriptions of the most common maintenance meth-
ods, distresses which are repaired by each method, and typical life
expectancies for each maintenance method.
Preventive Maintenance
Surface seals and crack sealing are discussed in this section. Surface
seals are those maintenance activities consisting of applications of asphalt
alone or asphalt and aggregates which are applied continuously to the
whole surface of a traffic lane. These surface seals are most often applied to
Table 9-1. General Categories of Types of Asphalt Pavement

Distress (i)
Primarily Traffic Primarily

Distress Type Load Caused Climate/Material
Caused
1. Alligator or Fatigue Cracking X
2. Bleeding X
3. Block Cracking X
4. Corrugation X
5. Depression X
6. Joint Reflection Cracking from PCC Slab X
,,__
7. Lane/Shoulder Dropoff or Heave X
8. Lane Shoulder Separation X
9. Longitudinal and Transverse Cracking X
10. Patch Deterioration X
~
11. Polished Aggregate X

12. Potholes X
13. Pumping and Water Bleeding X X
14. Ravelling and Weathering X
15. Rutting X
16. Slippage Cracking X
17. Swell X
Note: It is difficult to attribute all of a particular distress to only one cause. It is recognized that
in the above list, some types of distress might arguably have more than one cause.
(1) rejuvenate or retard the oxidation of asphalt at the surface; (2) restore
skid resistance of the surface; (3) seal fine cracks which have appeared at
the surface; (4) prevent the intrusion of water into the pavement structure
through cracks which penetrate through the HMA layer; and (5) retard the
ravelling of aggregates from a weathered, disintegrating surface. The most
common types of seals applied to HMA surfaces are (1) fog seals, (2) reju-
venators, (3) chip seals or surface treatments, and (4) slurry seals (.2). Each
type of surface seal is discussed separately.
-------1Fog-Seals.-fug-seal-s-inv:0-l¥e-a-light-ap-pliGat:i0:F1-0£-a-slew-se1:1ing--eFnul-------
sion to an oxidized pavement surface. The primary objective is to restore
or rejuvenate the asphalt cement at the surface of the roadway. Often the
aged surface has a light gray rather than dark gray or black appearance with
fine cracks apparent and, in many cases, there is evidence of ravelling. The
Table 9-2. · Candidate Maintenance and Rehabilitation Methods for

Asphalt Pavement Distress (j)
Distress Maintenance/Rehabilitation Methods

Alligator Cracking Full-Depth Spot Repairs
Overlay
Recycle
Bleeding Apply Hot Sand
Chip Seal with less binder
Block Cracking Slurry Seal
Chip Seal
Overlay
Recycle
Seal Cracks
Depressions Level-Up Overlay
Polished Aggregate Chip Seal
Slurry Seal
Open-Graded Overlay
Potholes Full-Depth Repairs
Patch
Recycle
Ravelling and Weathering Chip Seal
Slurry Seal
Fog Seal
Sand Seal
Stress Absorbing Membrane
Rutting Overlay
Col,d Mill
Hot Recycle
Swell Removal and Replace
treatment is suitable only for low-volume roads which can be closed to

traffic for the four to six hours needed for the emulsion to break and set up
sufficiently i:o provide adequate friction between the road surface and the
tire for safe sto'pping and cornering maneuvers.
A fog seal consists of a single application of a slow setting emulsion
w):iich has been diluted by an equal amount of water. This diluted liquid is
sprayed on the surface at a rate of 0.1 to 0.2 gallons per square yard (0.3
to 0.6 liters/m2), the higher quantity being used for coarse-textured, dry
surfaces. The primary attraction of the fog seal is its low cost and ability to
postpone the need for a surface treatment for a year or two. However, care
must be taken when applying fog seals to ensure that excess asphalt is not
applied; otherwise, loss of skid resistance occurs. Some sand should be kept
on hand for use as a blotter in case such a situation develops.
u,
~ ~
-
0"'
-- _--...
n
PoulbleCau11 RehabUltalloa
\0
Con-- , &
Dlllnn Type Stntdural
Fallun
Mb
....
Compod·
Temp. or
MolUun
Chana;e1
lloa
Suda«
Reqdlna;
Thlft
Ottrlay
. .,....
Op,,,-
Gnd<d
..,_uni
"-lay 11,q<llq
~~
a~g an
1. AJligalor or Fatigue Cracking X X X .x
2. Blttding X X X X
3. Block Cracking X X X X X
4. Com.1gation X X X X x' X X X
~~
....
5. Deprn1ion X X X X n
6. Joint ReOed:ion Cracking X X X ~9 ::.:
from PCC Slab
;. ~
S;::
~
7. Lane/Shoulder Dropoff or Heave X X X
;;;·
8. Lane/Shoulder Separation X X X
"' ►
9. longiludlna.1 and Transverse Cracking X X X .... 'g.
JO. Palcb Decerioralioa X X X X = ~
S;::
11. Polished Aurcgale X X X X X
12. Polho~ X X X
i
X ~ ~
8.
i:r
~
,,,
e?..
-........
X X
13. Pumping and Waler BJc.cding
14. Ravelling and Weathering

X
X
X
X X X
.... S;::
;;;·
X X X X xi X X X
.... ij
~
15. Rulling
(I)
16. Slippage Cracking X
0 ti
17. Swell X X X = (I)
[!l.
{IQ
:::,
....,
0
8.
Notes: I. If problem is extensive enough (')
2. When rutting is minor
~
0
:::,
3. Over milled surface
'"g.. g
e.
....
~-
0
0
:::,
Rejuvenators. Many proprietory products are available in the market

which, according to the suppliers, rejuvenate the oxidized asphalt at the
surface. Some products contain maltene (oil) fractions of asphalt cement.
Rejuvenators are also made from tar products to make the surface resistant
to fuel spillage. Application is usually made with asphalt distributors.
Rejuvenators must be able to penetrate the HMA surface to the desired
depth (at least 1/4 inch) (6.4 mm) within a reasonable time. If the HMA
surface is not sufficiently porous or the viscosity of the rejuvenator is high
or too much rejuvenator is applied, the rejuvenator tends to form a thin
layer at the surface, resulting in slippery and hazardous conditions.
Therefore, rejuvenators are mostly used on low volume, low speed roads or
parking lots. A test patch is recommended to determine the application
rate. The durability of rejuvenating applications has not been well estab-
lished. Like fog seals, a rejuvenator might be able to postpone the need for
surface treatment for a year or two.
Chip Seals or Surface Treatments. This type of surface consists of

single or multiple applications of asphalt and aggregate over a weathered
surface or a prepared base course as the original surface (B, 2) . Asphalt sur-
face treatment is a broad term encompassing several types of asphalt and
asphalt-aggregate applications to the existing surface. The thickness of
such applications is limited up to 1 inch. Some states.consider a thin HMA
overlay as surface treatment also.
A single surface treatment usually called chip seal involves spraying
asphalt material (asphalt cement or emulsified asphalt) followed immedi-
ately by a thin (one stone thick) aggregate cover, which is rolled as soon as
possible. Cutback asphalts have been used in the past for this purpose but
are not used much anymore due to environmental concerns. These single
surface treatments are often given different names because of the size or
type of aggregate used to cover the asphalt. For example, sand seals are sin-
gle surface treatments which use sand as the cover aggregate. Sand seals are
low cost treatments and are appropriate for use on low volume rural roads
or residential streets. Suggested asphalt and cover aggregate quantities for
single surface treatments are given in Table 9-4. Multiple surface treat-
ments are obtained by repeating the process a second or even a third time.
However, the maximum size of the cover aggregate for each successive
application is usually reduced by 50 percent. The cost of multiple surface
treatments approaches that of a thm HMA surface, and the public gener-
ally prefers the smooth, quiet ride of the HMA. A properly designed thin
HMA surface is likely to be more durable than multiple surface treatments.
Conventional surface treatments may not be suitable for use on high
volume, high speed roads, because there is a distinct possibility that the
public will complain about loose aggregates on the surface which are
picked up by vehicle wheels and can chip paint and break windshields.
Surface treatments add little load-carrying strength to the pavement struc-
ture. However, when properly designed and constructed, they are eco-
nomical, easy to place, and reasonably durable. Not only do they provide
a waterproof cover over the existing pavement structure, they resist surface
abrasion by traffic. Surface treatment is also used in skin patches which are
placed directly on localized distressed areas.
Cover aggregate should be clean; otherwise, it does not adhere to the
asphalt material. The ideal shape of the aggregate is· cubical. Flat or elon-
gated particles are not desirable because they lie flat on the surface and nor-
mal amounts of asphalt binder covers them, resulting in bleeding or flush-
ing surface conditions. A one-size cover aggregate is preferred for surface
treatment. However, many states use a graded cover aggregate because
one-size aggregate is not economical.
Asphalt binder can be asphalt cement, cutback asphalt, or emulsified
asphalt. Most states prefer emulsified asphalts (rapid setting grades)
because they can be used with damp aggregate, set up more rapidly than
cutback asphalts, and eliminate the fire hazard that is associated with the
use of cutback asphalt. Asphalt cement is used in hot climate areas.
However, the aggregate must be dry and very clean, and be applied before
the asphalt cement cools down to ambient temperature. Cutback asphalts
Table 9-4. Suggested Asphalt and Aggregate Quantities for Single

Surface Treatments and Seal Coats {10)
Quantities
Nominal AASHTO
Aggregate Aggregate Aggregate 1.l Asphalt 1,3
Size Range, in. Designation lb/sy gal/sy
3/4 to 3/8 6 40-50 0.40-0.50
1/2 to No. 4 7 25-30 0.30-0.45
3/8 to No. 8 8 20-25 0.20-0.35
No. 4 to No. 16 9 15-20 0.15-0.25
Sand M-6 10-15 0.15-0.20
1Toe lower application rates of asphalt shown are for aggregates on the fine side of the spec limits.
----lli+iligher-application-rates-are-for-aggregates-on-the-coarse-side·ofthe,pec-limi,-.,.~_- - - - - - - - - -
2Weights are based on an aggregate specific gravity of 2.65. If the specific gravity is greater than
2.75 or less than 2.55,' multiply the table value by the aggregate specific gravity divided by 2.65.
3
Asphalt is an emulsion of acceptable grade and type. Consult Reference 2 or state specifications
for recommendations.
are normally not used due to environmental problems, cost, and safety
problems (fire hazard).
Although Table 9-4 gives a range of application rates for asphalt
binder and cover aggregate for specific aggregate sizes, it is recommended
to use a design method to compute the application rates. A vast majority
of highway agencies still use the quantities of binder and cover aggregate
determined by experience and/or precedent. This may result in surface
treatments that have poor performance characteristics, but past experience
with particular aggregates and asphalts must be used to ensure best perfor-
mance.
An excellent review of surface treatment design procedures existing
prior to 1968 was made by Herrin, et al. (JJ.). These include the following
design methods: (a) Hanson, (b) California, (c) Nevitt, (d) modified
Kearby, (e) Lovering spread modules, (f) European, (g) McLeod, (h)
Mackintosh, (i) American Bitumel, and (j) Asphalt Institute.
Design formulas for simple surface treatments using one-size and
graded aggregate are given in Reference 2- Reference 10 gives McLeod's
method in detail. His method is a modification of the method developed
by the County Roads Board of Victoria, Australia. The method is based on
the following principles:
1. After rolling, the voids between one-size cover aggregate particles

are about 30 percent.
2. The voids are reduced to about 20 percent when the traffic reori-
ents the particles into their densest positions with all particles lying
on their flattest sides.
3. The average thickness of the surface treatment is equal to the over-
all average smallest dimension of the aggregate particles, which is
called the "average least dimension" (ALD) of the cover aggregate.
4. If the traffic is moderate, the quantity of asphalt binder used should
fill about 70 percent of the 20 percent void space to obtain good
performance.
Many design methods, incuding the McLeod's method, involve time-

consuming or complex test procedures and/ or computations. This has dis-
-----~c~o~u=ra=-ged their use, especially for low volume, low cost _r.9a,ds. However,
simplified nomographic methods such as used by the Pennsylvania DOT
(1.J_) are available, which are reasonably accurate for low volume roads. The
design of surface treatments is outside the scope of this book, but design
information can be obtained from the various cited references.
The construction sequence for single surface treatment is as follows:
1. Set up traffic control. A detour is preferred.

2. Clean the :Surface to remove dirt and other loose materials. This.is
an extremely important step since the asphalt will bond to whatev-
er is at the surface. If the chip seal is to have a long life, it is imper-
ative that the asphalt form a strong bond with the road surface and
not the debris on the surface.
3. Apply the asphalt binder to the surface at the specified rate and
temperature using a calibrated asphalt distributor.
4 . Spread the aggregate at the specified rate evenly over the surface
immediately after the asphalt binder is applied. The aggregate
spreader should be properly calibrated prior to starting work. Avoid
excessive application of aggregate because of the tendency of traffic
to roll the excess aggregate against that aggregate penetrating· the
asphalt layer and dislodging it. This dislodgement problem is espe-
cially acute during the first few hours that the treated road is open
to traffic.
5. Immediately roll the aggregate to push it through the asphalt
binder and to seat it firmly against the underlying layer. The aggre-
gate should be rolled preferably with a pneumatic roller; however,
a steel roller can be used. Spreading and rolling of the aggregate
should be completed before the emulsified asphalt, if used, breaks
to ensure adequate bond to retain the aggregate. The pneumatic
roller is better suited for pressing the aggregate against the under-
lying layer than the steel-wheeled roller, especially if the surface
being repaired has high spots, since the steel-wheeled roller bridges
over the low spots and does not compact aggregates below the
plane surface between the high spots. The steel-wheeled roller may
also tend to break the aggregate particles.
While some aggregate loss always occurs, most of the typical problems
' can be alleviated by using well maintained equipment and knowledgeable
construction crews. Typical construction problems and their solutions
include:
1. Streaking of asphalt. Streaking occurs when the distributor is not

supp ymg asp i:ilt binder at a umform pressure to the spray bar, the
spray nozzles are not adjusted at the proper angle (see Figure 9-6),
or the spray bar is not set at the proper height to provide either
double or triple overlap as required by the specification (see Figure
9-7).
--NOZZLE ANGLE SETT~G, 15 TO 30 DEGREES
SPRAY BAR AXIS

Figure 9-6. Proper Nozzle Angle Setting to Achieve Proper Fan of
Asphalt Spray for Uniform Coverage (1JJ)
Figure 9-7. Spray Bar Height Adjustment t~ Produce Uniform

Coverage (1JJ)
2. Low application rate of asphalt binder. Insufficient binder can be

caused by improper settings on the distributor truck, an improper
evaluation and adjustment for the dryness (asphalt demand) of the
existing surface, or the use of a more absorptive aggregate than
anticipated in design.
3. Excessive aggregate loss. Aggregate loss can result from (a) applying
too much aggregate, (b) delaying application of aggregate, (c)
applying dusty, dirty aggregate, (d) using damp or wet aggregate
with asphalt cement binder, (e) applying insufficient amount of
asphalt bindei; to hold the aggregate, (f) delaying compaction and
(g) opening the treated roadway to traffic too soon.
To minimize whip-offloss and increase stone retentfon~-many organi-

zations require the use of precoated cover aggregate and/ or the use of some
type of polymer additive to the asphalt binder to increase its stickiness.
Cover aggregate can be precoated with paving grade asphalt cement in a
HMA facility. Using 0.6 to 1.2 percent asphalt cement by weight of the
aggregate is sufficient for precoating. The precoated aggregate should be
free flowing so that it can be. applied with an aggregate spreader.
Slurry Seals. This type of surface treatment is distinctly different from

all other types which have been discussed because the components, proce-
dures, and construction equipment are different. A slurry seal is a homo-
geneous mixture of emul,sified asphalt, water, well-graded fine aggregate,
and mineral filler which has a creamy, fluid-like appearance when mixed
in proper proportions. These components are typically mixed in a special
truck-mounted, continuous mixing chamber (Figure 9-8) which dumps
the slurry into a spreader box (Figure 9-9) that spreads the slurry into a
smooth, thin layer. At the rear of the spreader,hox is a series of flexible rub-
ber squeegees which are mounted around the perimeter of the spreader
box. These squeegees keep the slurry in the spreader box while the rear
strike-off squeegee controls the thickness of the slurry seal layer (Figure
9-10). The layer thickness is approximately equal to the maximum aggre-
gate size.
~ There are generally three gradations used for slurry seals. The selection
is based on the function served by the slurry seal. For example, the Type I
slurry ofTable 9-5 is used to fill fine surface cracks and provides a very thin
(typically slightly thicker than the maximum aggregate size of 1/8 inch
(3.2 mm)) covering on the road surface. The Type I slurry has the finest
gradation with most of the aggregate smaller than the No. 8 sieve (2.36 mm).
SCREW REVERSES PITCH

APPROXIMATELY HERE
EMULSION
CONVEYOR
-------------,; ~:~>----------~
!' SLURRY ~;X/ .
Figure 9-8. Flow Diagram of a Continuous Slurry Mixer (1f!)
Figure 9-9. Continuous Mix Slurry Machine and Spreader Box

(Corps of Engineers)
Figure 9-10. Spreader Box Strike-off Squeegee (Corps of Engineers)

Table 9-5. Slurry Seal Aggregate Gradations Recommended by the

International Slurry Seal Association
Type of Slurry I II III

1st or 2nd Application,
Crack Filling General Seal on Medium 2 Course Slurry on
General Use and Fine Seal Textured Surfaces Highly Textured
Surfaces
Sieve Size % Passing % Passing % Passing
3/8 in. I - 100 100
#4 100 90-100 70-90
#8 90-100 65-90 45-70
#16 65-90 45-70 -. 28-50
#30 40-60 30-50 19-34
#5·o 25-42 18-30 12-25
#100 15-30 10-21 7-18
#200 10-20 5-15 5-15
Residual Asphalt
Content, % by
Weight of
Aggregate 10-16 7.5-13.5 6.5-12
Application Rate,
lb/yd2 6-10 10-15 15 or more
The Type I slurry is sometimes used as a preparation for a HMA overlay or

chip seal. This type slurry has been used on low traffic areas such as park-
ing lots and paved shoulders where sealing from moisture and prevention
of additional oxidation and loss of surface materials are the primary func-
tions of the seal.
The Type II seal (Table 9-5) is slightly coarser t~an the Type I seal,
having a maximum aggregate size of 1/4 inch (6.4 mm); therefore, a typi-
cal thickness of slightly greater than 1/4 inch (6.4 mm). The Type II slur-
ry seal is the most extensively used of the mix types. It is suggested for use
when moderate to severe ravelling has occurred due to oxidation of the
asphalt and to restore or improve skid resistance. When improved skid
resistance is an important criterion, the aggregate should be angular and
polish resistant; therefore, the use of natural sands may be prohibited since
---most-oLthese-mat.ei:i<il.s-ha:v:e.-r.ounde-d.-0-r-sHb-FeuRae-a-shap@s-an-a-p00-ic-.-----
microtexture.
The Type III seal is typically reserved for pavements which have devel-
oped severe surface irregularities requiring a thicker sealer with larger
aggregate (Table 9-5). This Type III seal is recommended for use as the
first course in a multicourse application and to restore skid resistance.

Because of its thickness, it can fill slight depressions to prevent water pond-
ing and reduce the probability of hydroplaning.
Slow setting (SS) emulsified asphalts of all types are used in slurry seal
applications and, when the roads must be quickly opened to traffic, the
quick set (QS) emulsified asphalts are used. Additives such as limestone
dust, portland cement, and fly ash are sometimes added to the slurry to
produce particular characteristics in the slurry. Since mixture design of
these slurry seals is outside the scope of this book, the reader is referred to
publications of the International Slurry Seal Association U.J., I 4, 15).
ASTM D3910 contains standard practices for design, testing, and con-
struction of slurry seals.
Crack Sealing. Crack sealing is a routine maintenance activity that

basically involves cleaning out the cracks which are wide enough to clean,
and then filling the clean cracks with a sealant in order to prevent water
and noncompressibles from entering the pavement structure. The most
common types of cracks which are sealed include longitudinal cracks,
transverse cracks, reflection cracks, edge cracks, and block cracks. Fatigue
(alligator) cracks are seldom routed and sealed individually because the
intensity of the crack patterns usually warrants an area treatment rather
than individual crack treatment. The treatments may involve removal and
replacement, overlay, or seal coat for low severity fatigue cracking.
Generally, cracks which have opened up more than 1/4 to 1/2 inch
(6.4 to 12.7 mm) are candidates for sealing. Typical sealing procedures
involve the following steps:
1. Place proper traffic control devices in the work zone.
2. Rout the cracks to remove debris and any deteriorated crack
sealing material.
3. Clean the cracks with compressed air or a stiff bristle broom. If
compressed air is used, the connection should not be at a port with
automatic oiling since oil on the sides of the crack prevents a good
bond between the sealant and the sides of the crack.
4. Pour hot crack filling material (the most commonly used) into the
crack to within 1/4 inch (6 mm) below the surface. This prevents
tracking of the sealant and the need for blotting sand. Some agen-
cies completely fill the cracks, squeegee the excess f!1aterial on the
surface, and blot with sand. This procedure produces a much
messier appearance and provides opportunity for significant track-
ing of material. Therefore, it is not recommended for areas of heavy
pedestrian traffic.
5. After allowing the sealant to cool, broom off the excess sand,
remove traffic control devices, and open to traffic.
A wide variety of crack filler materials are in use today. However, rub-
berized asphalt crack sealers have been found quite effective (.1.§J.
Specifications for such crack sealers are contained in ASTM D5078. The
prepackaged material is heated and applied using a special piece of equip-
ment that heats and dispenses the material. Since the heating kettles used
by most manufacturers involve flame heating, all appropriate safety proce-
dures should be followed, including the use of goggles and heavy protec-
tive clothing by the sealing crew. Special precautions are required to keep
water away from hot asphalt since hot asphalt will foam dramatically
(approximately 15 times its volume) when it comes in contact with water.
Serious burns and fires can result when handling and safety procedures are
ignored.
Corrective Maintenance
Corrective maintenance generally consists of patches, chip seals, and
thin HMA overlays. Since chip seals have already been discussed, only
patching and thin overlays are included in this section.
Patching. Patching is one the most common methods for repair of

localized areas of intensive cracking, whether the cracking is load associat-
ed (alligator) or environmental or construction related (transverse or lon-
gitudinal). If the cracking has deteriorated to the point that disintegration
of pavement materials around the cracks is occurring, the defective mater-
ial must be removed and replaced.
Patching can be either partial or full depth. Partial depth repairs usu-
ally involve removing the surface layer and replacing it with HMA. Full
depth repairs involve removal of the complete pavement down to the sub-
grade or to an intermediate subbase layer that is intact. Typical repair in
flexible pavements for cracked areas involves full depth repairs. Partial
depth repairs usually are reserved for slippage cracking due to poor bond
to the underlying layer or for shoving and corrugations.
Full depth patching is the most common procedure used for the repair
of localized alligator cracking and potholes. A patch consisting of HMA
can be considered a permanent repair and should last for years. Emerg~Q.CY--------•~--··--
repairs of potholes frequently involve the use of cold-mix materials under
severe environmental conditions. This type of repair is considered to be
temporary and serves as an expedient measure in order to keep traffic mov-
ing until a permanent repair can be made.
The proper procedure for making a permanent patch in a HMA sur-

face involves the following steps (2):
I. Set up traffic control consistent with the class of highway.

2. Mark off the area to be removed using paint, chalk, or keel.
Rectangular areas are preferred; however, a polygon with straight
sides is acceptable. Be sure that the lines are at least 1 foot_(300 mm)
into sound pavement material surrounding the distressed area to be
removed.
3. Cut through the pavement around the perimeter of the p~lygon
using a jackhammer or saw. Vertical faces on the cuts are very
important to ensure that the patch material bonds and interlocks
adequately with the sound adjacent pavement.
4. Remove the loose material down to a solid base. Remove all loose
matter, including water and saturated base, subbase or subgrade
materials. If subgrade material is removed, crushed stone or other
suitable base material may be used to backfill up to the top of the
subgrade. For small patches it is usually easier and better to use
HMA full depth. Compact all materials used as backfill as well as
the top of the layer at the bottom of the patch.
5. Apply a tack coat to the sides and bottom of the hole when patch-
ing with HMA. Do not allow the tack coat to puddle in the bot-
tom of the hole.
6. Place 4-inch (100 mm) lifts of HMA and compact each layer.
Proper compaction is one of the most important steps in the patch-
ing procedure. In fact, if proper compaction does not occur, the life
of the HMA patch can be reduced significantly and may be no
longer than that of a cold mix patch. The final uncompacted lift
should be extended above the surrounding pavement by about 1/4
inch (6 mm) per inch of the compacted height of the last lift to
ensure that the compacted patch will have adequate density and be
at the same height as or slightly higher than the original pavement.
Hand tamping does not provide adequate compactive effort;
rather, plate compactors or rollers are essential to achieving ade-
quate density in the HMA patching material.
7. Using a straightedge, check around the perimeter of the patched
area to ensure that the patch is approximately level with the sur-
---------0u-aa-i-ag-pa:.v:~m~nt.-I-f..the-edg~S-are-not-le:vd,-the riding quality
will be adversely affected, especially if several potholes are being
repaired in close proximity.
8. This, step is considered optional by many agencies but should be
used if long life is required from the patch. Seal the edges of the
patch using a sand or chip seal which involves a light application of

liquid asphalt blotted with sand or aggregate chips. This strip seal
needs to be 6 to 8 inches (150 to 200 mm) wide to provide a water-
proof seal at the edge of the patch.
9. Sweep all loose material from the road surface and remove to a suit-
able site; then remove traffic control devices and open to traffic.
The patching procedure described above achieves the best results when
repairs are made during warm, dry weather. Unfortunately, not all repair
can be postponed until after inclement weather. Therefore, it is possible
that some patching must be done during cold, wet weather. In such cases,
temporary repairs are made using cold mix asphalt patching materials and
emergency, temporary patching procedures.,_In such cases, the following
patching procedure is often followed until permanent repairs can be made:
1. Set up traffic control devices.

2. Remove all large, loose material from the hole and as much water
as is practical by brooming. No reshaping of the hole is attempted
at this time.
3. Fill the hole with cold mix patching material and compact by hand
tamping or preferably with the truck wheels.
4. Sweep all loose material from the roadway surface, remove traffic
control devices, and open to traffic. ·
Pothole repair is a special case of patching, not because it is different

in methodology, but because it has such high visibility to the public. In
fact, pothole patching is the most common form of patching in many juris~
dictions because it is a continuous operation, especially in the spring of the
year. Therefore, pothole patching is discussed next. In preparing these
materials, extensive use has been made of the pothole primer (1ZJ which
was prepared by the Corps of Engineers Cold Regions Research and
Engineering Laboratory (CRREL) in 1981. Similar procedures are also
included in MS-16 of the Asphalt Institute (1.i).
The permanent pothole repair procedure suggested in Reference ll
follows precisely the steps outlined above for any permanent patch. The
procedure is shown pictorially in Figure 9-11 with a brief description of
each picture on the right.
As with the general patching _erocedure, pothole repair must some-
times be performed at less than optimal conditions. In those cases, the tem-
porary repair procedure described above can be used for emergency pot-
hole repair. However, the temporary repair procedure is too often the only
repair procedure conducted for potholes, resulting in very costly conse-
Tack coal applied lo

sides 8. bottom of hole
4 ,,~··:;,·,·:=:.:~,:-.; :·!'.=~·~·-,·-o'.:,
1
inu ~o1:pih b~~ ~i~~f!~i~d
%,·~-:~:~·o:; "::"!···; :"-:·:6. ::~. :•..:·,f;·_ ~~o.··~.;#q
Figure 9-11. Permanent Pothole Repair Procedure (11)
quences. Reference 1Z reports the results from a value engineering study

which compared costs for pothole repairs using various repair methods. A
comparison of the costs per ton of material placed along with the average
life expectancy of the repairs produced the data shown in Table 9-6.
Table 9-6 shows that the pothole procedures can be divided into three
different groups as a function of the types of materials and equipment used
and the intensity of the effort applied to make the patch permanent. The
groups are as follows:
1. Group I-Methods No. I and 2 ofTable 9-6

2. Group 2-Methods No. 3 and 4 of Table 9-6
3. Group 3-Method No. 5 of Table 9-6.
Group I, which involves the use of cold mix material, utilizes no effort
to remove loose debris from the hole and little effort to compact the cold
mix placed in the hole. For method No. 1, compaction is simply whacking
the loose material twice with the backside of a shovel; for method No. 2,
compaction is with the truck wheels. The life of these patches is short, I to
2 months, and the cost is very high at about $300/ton and $200/ton annu-
ally for methods I and 2 respectively.
Groups 2 and 3 of Table 9-6 include the use of HMA, tacking the
sides of the hole with asphalt, and compaction with power equipment.
Methods 3 and 4 include shaping the hole with hand equipment, remov-
ing loose material from the hole, and compacting with a wacker tamper for
Method 3, and with a pup roller for Method 4. These improvements in
procedure produce significant increases in life expectancy of the patch
from 1 or 2 months (for methods 1 and 2) up to 12 months at about 1/3
the annual cost per ton of material. Group 3 (Method 5) is similar to
Group 2 except that the patched area is shaped with power equipment
which slightly increases the cost per ton of material, but extends the aver-
age life of the patch.
The use of Method 1 is warranted only in the most extreme circum-
stances because the costs per ton of mix are so high. All maintenance forces
should be trained in the procedures described in Methods 3, 4 and 5 and
be told the cost consequences of using Methods 1 and 2.
Thin HMA Overlays. The traditional method for protecting a deteri-

orating pavement, reducing roughness, restoring skid resistance, and
strengthening the pavement structure of a flexible pavement is with an
overlay of HMA. It is the typical practice of most state DOTs to place thin
HMA overlays up to I-inch (25 mm) thick when considering the need for
restoring skid resistance and protecting a deteriorating pavement. Some
DOTs even use thin overlays in an attempt to reduce roughness and
improve rideability, and urider some circumstances, thin overlays perform
satisfactorily. However, if there are components to the roughness that pro-
duce variations of even a few inches iu..25Je_eL(7~6_m)_,_a..th.in..mi:erlay_s,_,_· i = - -- - - - - ~
ply does not remove that roughness. Therefore, the total overlay thickness
should vary depending on the type of roughness components present in the
road profile in order to allow greater overlay depths when warranted by a
roughness survey. Many DOTs utilize levelling courses in an attempt to
~
(')
-
::r
O"' ~
n ....ft
i
\0
Cost per ton
Tons In place
per shift
or material
in place
ure or
patch
Annual cost
per ton
~
ei.
No, .. I l Description or Method >X Equipment Required (7.5 hrs) ($) (months) ($)
~n a
(>
::,
~ ~ §
Fm hole in one lift with mixture, and
corpact by hitting the patch with the back
Dump truck
shovels
18.0 25.64 1 307.68
e.,... 0
~ shovel twice. No effort made to clean "'n ·"~

or shape the hole and no tacking of the
osed surface of the hole. §a~
(>
::r
g.
'
2 I Sate as method no. 1, except compaction Dump truck 12.0 31.80 2 190.80 ~
....
fl>
is erformed with the tire of the dump shovels 0 F
truck.
=
fl> §
p.
3 I Sh~pe the area to be patched with an axe Dump truck, pickup, 6.0 63.29 1• 63.29 ~
anr sledge, remove loose asphalt with heating kettle, wacker, . 8"'
>'t
(>
0
0
mhtock, sweep area clean, tack the exposed axe, sledge, brooms, ::,
.,,
su aces of patch area, shovel in material,
anlt level with lute. Compact with wacker
mattocks
....t:, g
~
0
i:t.
(vi~rat_ory compactor) and seal edges with 0
ta~k 011 and # 1B stone. g
....
::,
....,
0
~
4 I Sa~e as method no. 3, except a pup roller Dump truck, pickup, 7.0 61.41 7• 61.41
is rsed for compaction. heating kettle, pup roller,
5
I
I Saitle as method no. 4, except the area-to
be: patched is shaped with a pavement
axe, sledge, brooms, mattocks
2 pickups, dump truck,

heating kettle, air
7.0 65.22 7• 65.22
f
0
,=...
fl>
>
breaker. compressor and pavement

breaker, Pup roller, Brooms 0
"+)
e-~
•For methods 3, 4 and 5 it was assumed that failure would occur after 12 months, though in ail probability patch life would extend beyond one year.
-
0
n "'
.ia,.
\C
smooth out a profile that exhibits severe distortion in either the horizontal
or transverse profile. In many cases it is more satisfactory and economical
to use a milling machine to mill the old pavement to a plane surface rather
than to use levelling courses.
Where strengthening the pavement structure is the primary objective
for the overlay, thin overlays are seldom cost effective, even if several are
placed over a period of years, primarily because thin overlays cool quickly
and therefore are difficult to compact to the required air void content
which yields optimum strength and durability. Also, thin HMA overlays
are generally prone to debonding from the existing surface. Therefore, it is
imperative, when pavement strengthening is needed, that the engineer per-
form an adequate analysis to establish the cg_ndition of the existing pave-
ment; determine the strength properties of the in situ material; determine
the length of the performance period required for service and estimate the
traffic to be served; and determine the required overlay thickness using one
of the overlay design methods included in References 1., 12, 20, or the
appropriate procedure of the agency. Only by using good engineering
design procedures and sound life-cycle economic cost analyses can the
most cost effective combination of overlay design thickness and future
maintenance activities be determined.
If thin HMA overlays must be used, it should be ensured that (a) the
maximum aggregate size in HMA mix is appropriate for the overlay thick-
ness, (b) proper tack coat is applied, (c) work is carried out in warm weath-
er so that desired compaction level is obtained, and (d) good HMA con-
struction quality control is maintained.
REHABILITATION
As mentioned earlier, the rate of deterioration of pavements increases

with use (traffic loads) and age (weathering). Although preventive and cor-
rective maintenance helps in prolonging their useful life, pavements need
to be rehabilitated sooner or later. Quite often, more traffic than estimat-
ed warrants the rehabilitation to. be carried out sooner than anticipated.
Table 9-3 gives the various rehabilitation techniques used by the highway
---agel'leies,depend-ing-on-t:he-distress-typetSJ:-Hor-surface-reqdirrg;-thin
overlays, open-graded friction courses, milling, structural recycling, struc-
tural overlays, crack relief layers or treatments, and cracking and seating of
concrete pavements are some rehabilitation alternatives. Thin overlays have
already been discussed.
Hot Surface Recycling

Oxidation or aging of asphalt cement in HMA pavements occurs most
rapidly at the pavement surface because of exposure to the elements. This
oxidation and aging can induce ravelling and surface cracks. Hot surface
recycling is an effective and innovative technique which has been used to
renew and rejuvenate aged, brittle HMA pavements up to about 1-inch
depth from the surface. It has also been used to correct other surface defi-
ciencies such as minor corrugations, polished aggregate, and bleeding sur-
faces. Surface recycling is not effective if the pavement surface distress such
as alligator cracking and potholes is caused by a poor base.
Hot surface recycling is usually done in conjunction with a new HMA
overlay which provides the new riding surface. The application of heat and
scarification of the existing surface mix inhibits the occurrence of reflection
cracking in the new HMA overlay. Also, the process provides a strong bond
between the old and new HMA overlay.
The heater-overlay process can be accomplished by either a paving
train or an integral (single-pass) method. The following primary steps are
involved:
Heating. One or more machine-mounted high-intensity heaters trav-

el over the road at a slow speed to heat the existing HMA pavement to the
desired depth. The heat should be just enough to soften the HMA mix so
that it can be scarified without breaking aggregate particles. Excessive heat
causes charring of the HMA pavement, damages the asphalt cement
binder, and produces undesirable smoke. Asphalt cement on the surface at
locations such as sealed cracks also produces smoke.
Scarifying. The scarifying unit loosens the existing surface up to the
desired depth and redistributes the aggregate particles to some degree. This
process breaks up the existing crack pattern and helps prevent reflective
cracks.
Rejuvenating. A rejuvenating or recycling agent is applied to the loos-

ened mix if the asphalt cement in the existing HMA pavement is very hard
and brittle. Selection of a rejuvenator or recycling agent is made after prop-
er testing of the aged asphalt cement.
Placing the HMA Overlay. New HMA is placed over the scarified
(and rejuvenated if necessary) mix. If the purpose of surface recycling is to
remove only surface deficiencies in a structurally adequate pavement, a
thin HMA overlay will suffice. On the other hand, properly designed
thicker HMA overlays must be used when the pavement needs to be

strengthened for present and future traffic.
Compacting the Overlay. Normal compaction procedures are used to

,compact the scarified and new HMA overlay. The roller(s) must follow the
paving train very closely to compact the mat while it is still hot.
Various types of equipment are now available to accomplish the

heater-overlay process. Some integrated units are capable of blending the
scarified HMA with virgin HMA mix in a pugmill. This blending process
is useful to improve the composition (gradation and asphalt content) of the
existing mix. For example, if the existing HMA pavement is bleeding or
rutting because of excessive asphalt content, ii can be blended with a lean
virgin mix so that the blended mix has the desired asphalt content.
, Similarly, mix gradation can also be improved.
Guidelines and specifications for hot surface recycling (2J) and other
recycling processes are available from the Asphalt Recycling and
Reclaiming Association.
Open-Graded Friction Courses
A detailed discussion of open-graded friction courses (OGFC) is con-

tained in Chapter 7. OGFC is a high-void HMA wearing course which
can be used to rehabilitate asphalt pavements which have lost their skid
resistance due to aggregate polishing and/or flushing. OGFC contains a
large percentage of one-sized coarse aggregate to ensure a high void con-
tent which makes it highly permeable to water. It is very effective in min;-
imizing hydroplaning and reducing wet pavement accidents. In Europe it
is used extensively in urban areas to reduce road noise. It is important to
correct all structural deficiencies and seal all cracks prior to the construc-
tion ofOGFC.
Cold Milling
Cold milling is a process in which a milling machine containing a

---t:0-tati-ng-d-r-1.1-m-wi-th-Gar:bid~bi.t-s-s-t1:.ikes-th~M-.MA-s1:1-F-faEe-a:B:d-rnm-0-ves--the
material to a prescribed cutting depth. Milling machines were improved
during the 1980s to the point that they are now the primary method for
removing old surface material for recycling or prior to overlay. These
milling machines have automatic grade control and can be used to reestab-
lish both longitudinal and transverse grade in preparation for overlaying

operations. In fact, most pavement distortions can be removed using mod-
ern milling machines. About the only type of distortion which may be too
severe for restoration by a milling machine is that resulting from swelling
subgrade soils. Therefore, milling machines can be used to eliminate major
and minor rutting, bumps, deteriorated surface material, and complete lay-
ers which are experiencing shear flow behavior or stripping. These
machines remove material without harming the underlying material, and
leave a rough surface which has good skid resistance and provides an excel-
lent bond with the overlay. After milling, the overlay is placed on a plane
surface; hence, the overlay thickness is uniform. The overlay can be com-
pacted uniformly and can be placed to meet even very restrictive smooth-
ness requirements.
Some of the primary advantages of using milling machines include:
1. Eliminating the need for wedge-shaped levelling courses and elim-

inating the problems of estimating quantities required to establish
plane sections and compaction of this material of varying width
and thickness (see Figure 9-12).
2. Providing reclaimed asphalt pavement (RAP) material for recycling
operations.
3. Allowing efficient removal of deteriorated HMA material that is
unsuitable for retention in the overlaid pavement, especially that
material exhibiting severe shear flow or stripping.
4. Providing a highly skid resistant surface suitable for temporary use
by traffic until the final surface is placed. Milling can also be used
to restore skid resistance to a slippery pavement such as one expe-
riencing bleeding or aggregate polishing.
5. Allowing curb and gutter lines to be maintained or reestablished
before overlaying with HMA.
6. Providing an efficient technique for removal of material near over-
head structures in order to maintain overhead clearances for bridge
structures, traffic signals, and overhead utilities.
Milling machines have become more efficient during the last few years
______a_s_m_an-ufacturers have mcreased the size of the power umts, simplified the
hydraulic systems used to control the machines, made many parts inter-
changeable hetween model lines, and made improvements in tooth design,
including tooth shape and materials to allow more tonnage removed per
tooth. These improvements have resulted in the typical milling costs

shown below as a function of depth of cut:
Depth of cut1, in. Typical 1989 costs2, Dollars/Ton

1 1.86-1.88
2 1.39-1.40
3 1.04.:..1.13
4 0.93-0.90
1
Single pass cuts
Costs include depreciation, maintenance, machine operating cost,
2
and 2-man crew
'21,11t111!ll(1,1YP
(a) Wedge sections used to correct a depression.
~ 2 n d layer
I st layer
(bl Wedges used to correct excessive crown.
Figure 9-12. Problems of Levelling Eliminated by Using Milling

Machines to Establish Slopes and Crowns (Asphalt
Institute)
It is believed that the development and advances in technology of the
cold milling machine and reductions in unit cost is one of the most signif-
icant equipment advances since the late 70s. The versatility of these
machines in helping to solve some of the problems associated with build-
up of HMA materials on an aging highway system has been remarkable.
Structural Recycling
Unlike hot surface recycling, which is usually limited to the top 1-inch
of HMA pavement, structural recycling can be accomplished to any
r-
desired depth. It involves recycling one or more load-bearing structural

layers of the flexible pavement. Recycling is done for various reasons. For
example, some untreated pavement layer needs to be upgraded in quality
and strength, or some stripped HMA layer needs to be removed and
replaced. In the past it was a common practice to replace old materials with
new materials. However, the need in recent years for conservation of ener-
gy and materials has encouraged many agencies to adopt pavement recy-
cling strategies.
The first step in structural recycling is to remove the desired structur-
al layer(s) which is generally accomplished with the cold milling described
earlier. Reclaimed asphalt pavement (RAP) and/ or reclaimed aggregate
material (RAM) obtained from the existing pavement can then be recycled
in a HMA facility. Hot mix recycling has been discussed in detail in
Chapter 7. Many agencies are permitting contractors to substitute recycled
HMA in lieu of virgin HMA because both are considered equal in quality
and structural strength. Some states do not allow recycled HMA in wear-
. ing courses primarily because there is lack of performance data for the
wearmg course.
Recycling of concrete pavements has also been attempted successfully
by crushing the concrete pavement chunks into aggregates and mixing
with asphalt cement to produce HMA mix. However, hot mix recycling is
the predominant method of structural recycling.
Structural Overlays
Pavements become structurally deficient when the traffic loads exceed

the design loads used in the initial structural design analysis. These defi-
ciencies require additional strength from overlays to carry the increased
loads. In most cases structural overlays consist of HMA surface courses.
Overlays reduce maintenance costs, increase pavement life, provide a
smooth ride to the motoring public, and improve pavement surface skid
resistance. Unlike concrete pavements, HMA overlays can be placed with
minimal disruption to the traffic.
The structural design of HMA overlays is based on the evaluation of
the existing pavement structure using three techniques: deflection analysis,
-------omponent-a-na:lys-is,ancl.-mec-ha-n-istie-an.tlysis~----- - - -
The deflection procedures involve measuring pavement deflection and
analyzing them with respect to traffic. The in-place, load-carrying capa-
bility of the existing pavement structure is measured and used for estimat-
ing fu~ure requirements. The component analysis procedure converts the
thickness of each pavement layer into an equivalent HMA layer thickness

based on laboratory measurements or past experience from road tests such
as the AASHO road test. Then, the relationship between subgrade
strength, pavement structure, and traffic loadings is used to estimate the
required total thickness. The mechanistic (analytical) technique involves
evaluation of in-site elastic and viscoelastic properties of the materials in
the structure and uses damage criteria for pavement distresses such as rut-
ting and fatigue cracking. As mentioned earlier, Refere:Uces 1., 12., and 20
give the overlay design methods. Reference 22 gives an excellent synthesis
of asphalt overlay design procedures used in the United States and abroad.
Crack Relief Layers and Systems
Reducing or minimizing the occurrence of reflection cracking is a high

priority in extending the serviceable life of overlays of HMA on concrete
or HMA surfaces. Reference 23. summarizes information related to mini-
mizing reflection cracking. Most problems with reflection cracking occur
when concrete pavements are overlaid with HMA. Almost any practical
joint spacing on concrete p~vement produces high stresses in HMA over-
lays adjacent to the joints because of temperature induced movements. The
joint movements may be associated either with temperature, which pro-
duces horizontal movement and curling and warping, or load, which pro-
duces vertical deformation at joints, cracks and edges.
Figure 9-13 shows the load-induced stresses in an overlay due to a
moving wheel load on the HMA immediately adjacent to and over a crack
in the old surface. Notice that there are three places where the stress reach-
es a maximum in the HMA overlay:
I . Bending stresses increase as the load gets closer to the crack and
reach a peak when the load is directly over the crack, point B, and
then reduce as the load moves away.
2. Shear forces build up in the HMA overlay as the load approaches
the crack.
(a) The shear force reaches a maximum just before the load reach-
es the crack, point A, and then reduces to zero immediately as
--------he-lo-ad-crosses-over-tlre-crac1~ . - - - - - - - - - - - - - - - - - -
(6) The shear force then builds up rapidly in the opposite direc-
tion as the load approaches point C and then gradually
reduces toward zero as the vehicle approaches the center of the
concrete slab.
TIP OF THE CRACK
OVERLAY
OLD SURF ACE COURSE
~en
a. Zen
j:: 15~
w:.:: QI-
:cu men
I-<
1-5
<w A
en:c
WI-
en ~en
en LL ~en
WO
a:::
a:::w
I-
<a:::
w1-
en V5en
POSITION OF WHEEL LOAD

Figure 9-13. Traffic Induced Stresses and Resulting Crack Growth in
HMA Overlays (25)
The time period over which this loading sequence occurs is fairly
short, approaching 0.05 seconds at highway speeds. At these loading rates,
the stiffness of both the concrete and the HMA material is very high.
------~~ohably_the__m_Qs_t_s.ignific.an.t.foadmgJ.sJl.o_uhaLd_u_e_tQ_whed loads,
but those induced by temperature cycling. Figure 9-14 shows schemati-
cally the effects of temperature on the thermal contraction and curling of
a concrete slab and the effect of that movement on the deflected shape of
the BMA overlay.
These temperature changes occur at a very slow loading rate, as com-

pared to the loading rate from axle load associated stresses. They occur over
a period of several hours or even a major portion of a day. Under these very
slow loading rates, the stiffness of the HMA material is quite low, perhaps
1,000 to 10,000 times lower than the modulus exhibited by these same
materials under traffic-induced loading rates.
With each load cycle (axle or temperature), a reflection crack that has
formed in the overlay pavement grows a little longer. One major desire of
engineers is that they be able to select an overlay material with a set of
properties that can retard or prevent the growth of these reflective cracks.
However, there has been little theoretical work done to develop a mathe-
matical model for the reflection cracking phenomenon to determine the
material properties which affect the develbpment of those cracks. The work
by Lytton and Germann (21.) and Lytton and Pickett (25.) provides one of
the earliest, clearly developed procedures for modeling the reflection crack-
ing phenomenon along with the development of laboratory equipment
needed to determine the material properties required for the theory to be
successfully applied.
The use of Lytton's theory is in sharp contrast to the literature relative
to evaluations of various techniques to retard the development of reflection .
i✓
I
cracking. Generally, field techniques have been experimental in that vari-
ous procedures and methods thought to retard reflection cracking were
I actually tried. The developed cracking patterns were compared with a field
control section to determine effectiveness. The success of these attempts
has varied greatly with many conflicting reports on the effectiveness of dif-
ferent techniques. The techniques typically tried in the field can generally
CRACK GROWTH DUE
TO THERMAL STRESS
IN OVERLAY
TEMPERATURE DECREASE
OVERLAY
OLD SURFACE COURSE
CRACK GROWTH DUE TO

THERMAL CONTRACTION
-t------------'l-f-----------------ANE>-el:JRl::ING-0F-+HE.------
OLD PAVEMENT SURFACE
LAYER
Figure 9-14. Crack Growth Due to Temperature Induced Stresses in

HMA Overlays (25)
be divided into 4 groups: (1) increasing HMA overlay thickness, (2) per-
forming special treatments on the existing pavement, (3) performing spe-
cial treatments only on the cracks and/or joints in the existing pavement,
and (4) special consideration of the HMA overlay design.
Thick Overlays. Overlays of HMA are sometimes designed to add

structural strength to the existing pavement; however, they are most often
used to correct surface deficiencies, improve skid resistance, or prevent
water penetration through discontinuities in the existing surface. In such
cases, HMA overlays are seldom thicker than 2 or 3 inches. Attempts to
retard reflection cracking by increasing the thickness of the overlay have
proven only partially successful. Increasing the thickness delays the pro-
gression of reflection cracks only in those cases where the additional thick-
ness produced significant improvements in the pavement strength (2QJ.
Retarding the rate of crack progression in thick overlays is primarily
due to a combination of: (1) the additional material the crack must pro-
pogate through; (2) the additional strength of the pavement reducing the
strain at the tip of the crack which is induced by the load; and (3) the sta-
bilizing in oxidative hardening of the material in the lower portion because
the overlay protects that material from light and oxygen (2.6). It seems
apparent that the effects of items (1) and (2) produce the major impact in
retarding the development of reflection cracks.
Treating the Existing Pavement. The treatments included under this

category involve either breaking up the cracking pattern in the existing
pavement or covering the whole pavement with a layer of material which
will intercept the crack and prevent it from progressing through that layer.
Destroying the cracking pattern in the existing pavement typically involves
(1) breaking up a concrete slab into smaller pieces, which will act like
small, individual pieces rather than slabs of much larger size; (2) pulver-
ization of an existing HMA pavement, addition of a rejuvenating agent,
and compaction of the layer; and (3) heater scarification of badly cracked
HMA surfaces to various depths, typically 3/4 to 1 inch (20-25 mm),
addition of emulsions or rejuvenation agents, and compaction of the layer.
Breaking of concrete slabs and then seating the pieces with a heavy
roller has become a popular treatment to minimize reflection cracking. A
dt!tail-ed-d:iscussion-of-this-a:1-terrrative-app-eardaterin-rhis -chapter.
Some attempts have been made to place relatively thick layers of
crushed stone and asphalt stabilized layers between a badly deteriorated
concrete pavement and a subsequent overlay. Typically, a crushed stone
layer 4 to 6 inches (100-150 mm) thick is placed and compacted with a
roller. The Arkansas DOT has experimented with the use of an asphalt sta-
bilized, thick (up to 6 inches) (150 mm) interlayer oflarge, single-sized (2
to 3-inch) (50-75 mm) aggregate before overlaying with HMA. This mix,
called the Arkansas mix, has been tried with mixed results (2!SJ in that the
Arkansas mix has delayed reflection cracking but not prevented it.
One other technique which has been used to retard reflection cracking
is to coat the existing surface with a thin layer of asphalt, asphalt rubber,
asphalt and fabric, or a seal coat of asphalt and aggregate. This layer is often
called a stress-absorbing membrane interlayer (SAM!) since the objective
of the treatment is to provide a very flexible, soft layer of asphalt or asphalt
rubber which will not readily transmit the stress produced by horizontal
movement of the underlying layer to the HMA overlay. This soft, strain:..
absorbing layer prevents propogation of the ;rack tip from the old surface
to the HMA overlay because it is expected to absorb high strains witho~t
cracking. Field experience with these layers has also been varied with vari-
able results reported from different full scale trial sections.· Perhaps one of
the basic problems with many of these techniques is that there is no mate-
rial property measurable in a laboratory which can be related to a theoret-
ical model which can predict the rate of crack propogation from an under-
lying layer. As a result, it is difficult to design or select a material with par-
ticular properties that can resist propogation of reflection cracks.
Treatment of Existing Cracks and Joints. Since approximately 1960,

a number of treatments have been tried in an attempt to reduce reflection
cracking by treating the pavement in the immediate vicinity of the crack or
joint. Typically, larger cracks or joints are cleaned of foreign matter, filled
with a compressible material, and then covered with a strip of bond break-
ing material applied as a narrow strip over the crack and for several inches
on either side of the crack or joint. The operating theory is that the bond
breaking material keeps the HMA overlay from bonding to the old surface
in the vicinity of the crack, thereby increasing the gage length over which
the overlay can absorb the strain produced by movement of the underlying
layer. Bond breaking materials which have been used in field installations
include sheet metal, asphalt saturated building paper, wax paper, fine stone
dust, agricultural lime, aluminum foil, and asphalt and asphalt-rubber
· .,.,....regnated-fabr:i:e~, :26t.--Gonst:Fl:l:et:ion-El:iffieal-t:ies,-c-one@m-over-lat~i:-~---
+ - - - - -....
al dislodgement during the braking action of tires, the frequent occurrence

of multiple cracks instead of a single reflection crack, and the development
and introduction of stress-relieving interlayers have limited the use of these
treatments.
Special Overlay Design Considerations. Since the movements of the

underlying layer cannot be controlled, research has been directed toward
designing HMA mixtures which are more tolerant of the strains induced
by pavement movement. Typical methods of design modifications involve
the use of softer asphalts and rubber or polymer modified asphalts as the
binder in HMA. One side effect from the use of larger quantities of soft
asphalt is reduced stability, which may result in bleeding. Reduced stabili-
ty and bleeding reduce the safety for the travelling public. This method has
proven to be generally unsuitable for use on high-volume roadways.
Engineers have investigated the use of additives and reinforcement
interlayers within the HMA overlay. The use of additives has generally con-
centrated on the use of materials which increase the ductility, resilience,
and toughness of the mixture. Additives which produce these property
changes include ground scrap tire rubber and various polymer materials.
There has been much activity within many of the state DOTs to evaluate
these materials both in the laboratory and in field installations. Most of the
field trials had not been serving traffic long enough by 1990 to provide
definitive answers on which materials work best and under what circum-
stances.
Adding reinforcement to the HMA overlay has been used for years to
increase the tensile strength and make it less susceptible to reflection crack-
ing. Early efforts, which included steel reinforcement in the form of weld-
ed wire fabric or expanded metal mesh and the addition of asbestos fibers,
were largely unsuccessful. Asbestos fibers are not used anymore for envi-
ronmental reasons. Later synthetic fibers of polypropylene or polyester
were added in small percentages during the mixing process. Though exper-
imentation with fiber reinforcement is relatively new and results inconclu-
sive, some potential advantages have been observed (23, 25, 2(i).
15 in.
r-----------------==-Fabric
Moveable Plate
Figure 9-15. Schematic Diagram of the TTI Overlay Tester with a

Specimen Mounted and a Crack Being Propogated
~22)
There has been a significant emphasis on the use of synthetic fabrics as

reflection crack arrestors. These fabrics are either woven or non-woven fab-
rics of polyester, polypropylene, or nylon, and are applied in pavement
overlays to increase the tensile strength, improve the waterproofing func-
tion of the overlay, and reduce the severity and extent of reflection crack-
ing. As mentioned previously, these fabrics began to be used as a treatment
directly on the existing pavement, but have been used as a reinforcement
for the overlay by being placed at some point within the overlay to serve as .
a tensile load carrying member. The Federal Highway Administration has
encouraged the trial use of these materials, and states across the U.S. have
constructed field sections to evaluate their effectiveness (23). Many reports
in the literature deal primarily with field installation problems and proce-
dures while treating lightly.the condition of the roadway used in the exper-
iment and the variables which might affect the performance of the instal-
lation. Nevertheless, many of the field installations show promise of
improved performance. What is needed is a methodology for evaluating
the effect of pavement condition and its properties on the material prop-
erties needed for the fabric i11~~rlayer to serve as a crack attenuation and
strengthening material. The 6nly laboratory device which appears capable
of serving that purpose is the TTI overlay tester developed and reported by
Lytton (24, 25) and shown schematically in Figure 9-15.
The TTI overlay tester is a device which can simulate the effect of the
horizontal movement of an underlying pavement on an overlay bonded to
the plates of the tester. A 3-inch by 3-inch (76.2 by 76.2 mm) beam of
HMA is prepared in the laboratory, and after curing, is glued to the over-
lay tester plate when the moveable plate is adjacent to the fixed plate. The
moveable plate is hydraulically moved an amount equal to the anticipated
horizontal movement of the crack while the force required to open the
crack is measured, and the distance that the crack propagates with each
cycle is measured. The simulated joint movement of the overlay tester is
cycled until the crack propogates to the surface and failure occurs. From
the resulting measurements, the fracture properties of the HMA material
can be determined.
Comparisons can be made between a standard overlay and various fab-
ric interlayer treatments by inserting fabrics in the beam mold at the
appropriate depth during beam compaction. The effects of various strength
----fabncs, tack coat quaqtic1es, and depths or embedment have been studre••.1------
by Lytton (25). Results from Lytton's work show that fabrics used as a rein-
forcement interlayer should prove to be cost effective in the field. However,
field installations have not yet been in service long enough to make a def-
inite assessment.
Techniques for Cracking Concrete Slabs to Reduce

Reflection Cracking
As mentioned earlier in this chapter, preventing the reflection of cracks
from a concrete pavement through HMA overlays is desirable to extend the
serviceable life of the overlay. In addition to the methods previously dis-
cussed for retarding growth of reflection cracks, engineers for over 30 years
have attempted to break up the slab action of concrete pavements as a
method of preventing reflection cracking. Since the mid 1970s a consider-
able amount of field work has been done by the individual state DOTs to
determine the best technique for retarding and controlling reflection crack-
ing. The methods most often compared and evaluated are:
1. Cracking and Seating;

2. Breaking and Seating;
3. Rubblizing; and
4. Saw and Sealing.
The Federal Highway Administration has developed a set of defini-

tions related to these concrete pavement rehabilitation techniques which
are reported by Witczak (2.2) as follows:
Cracking. Tight cracks that permit load transfer across the crack through
aggregate interlock with little loss of structural value, generally applicable
to plain jointed concrete pavement CTCP).
Breaking. Results in broken reinforcement or loss of concrete-to-steel

bond, generally applicable , to jointed reinforced concrete pavement
(TRCP).
Rubblizing. Complete destruction of slab action, applicable to all types of

concrete pavements.
Seating. Accomplished with 2 to 3 passes of a 35 to 50 ton (32 to 45 Mg)

rubber tire roller over pavements previously subjected to cracking or break-
ing. Rubblized pavements are generally compacted (seated) with 2 passes
of a 10-ton (9 Mg) vibratory roller.
Saw and Sealing. Involves sawing joints in the HMA overlay directly
above joints and working cracks in the existing concrete pavement and
sealing the joints, generally applicable to joint reinforced concrete. pave-
ments with long joint spacings in good condition. ·
Table 9-7. Guide to Selection of Appropriate Treatment Techniques

for Concrete Pavements Before Overlaying with
HMA(22)
Type of Concrete Pavement

Condition of Concrete Pavement JCPl JCRP2 CRCP3
structurally sound but roughness is saw and seal
high and/ or additional strength
needed
structurally sound but roughness is crack and seat break and seat
high, and significant patching and
spalling is present
deteriorated so that there is little rubblize and rubblize and rubblize and
potential to retain slab integrity compact -compact compact
and structural capacity
~ointed concrete pavement

toint reinforced concrete pavement
continuously reinforced concrete pavement
Using these definitions and the description (in Reference 2.Z) of the
pavement conditions for which these techniques have been applied, a guide
for selection of an appropriate treatment technique for concrete pavements
before overlaying has been prepared and shown in Table 9-7. Notice that
crack and seat is appropriate only for JCP, while break and seat is appro-
priate only for JRCP. Saw and seal could be used for both JCP and JRCP,
but the joint spacing on most JCP is so short that it makes the cost of saw
and seal impractical for JCP. Rubblizing and compacting is appropriate for
any type of severely deteriorated concrete pavement.
Equipment for Cracking, Breaking, and Ruhhlizing Concrete

Pavement. This equipment can be broken into the following five general
categories:
1. Crane with wrecking ball;

2. Pavement breakers (drop hammers, guillotine, and hydraulic/pneu-
matic impact hammers);
3. Pile-driving hammer;
4. Whip hammer; and
5. Resonant pavement breaker.
A brief description of each type of equipment is included; however, see

References 27. and 28 for more detailed descriptions. A summary of the
equipment types, characteristics, and productivity is found in Table 9-8

(22).
Successful use of the crane and wrecking ball for producing the speci-
fied cracking pattern is wholly dependent on the skill of the crane opera-
tor in controlling the drop height, swing and roll of the wrecking ball, and
orientation of the drop. Production rates are only moderate and results
vary among operators. Use of this device must be carefully evaluated where
traffic is to be maintained on adjacent lanes. Chances of producing accept-
able cracking patterns with this device are minimal.
Several types of pavement breakers have been used successfully to crack
concrete slabs. Drop hammer units are self-propelled units which raise a
heavy mass several feet above the pavement and then release the weight
which then falls and strikes the surface of the slab. The geometry of the
breaking head which contacts the concrete slab is important because, if it
is too thin, the concrete will shatter, producing rubble. But if it is too wide,
the energy is dissipated and either no crack will form or the crack will not
penetrate through the slab. Therefore, some experimentation with differ-
ent head geometries may be necessary to select a head which will produce
the desired crack pattern. Figure 9-16 shows one type of falling weight
drop hammer.
The guillotine drop hammer is a second device which has become very
popular. Commonly, this drop hammer uses a 7-ton weight which is
automatically lifted to a preset, adjustable height and released (see Figure
9-17). The device is self propelled and can have its speed adjusted to pro-
duce the desired impact spacing required to achieve the specified cracking
pattern. The bottom surface of the weight has a bevel shape with a foot-
print of 4 by 60.5 inches. The energy range can be varied to produce the
desired cracking by adjusting the height
The third type of pavement breaker is the impact hammer, which is
typically either a hydraulically or pneumatically actuated device mounted
at the rear of a backhoe or excavator (see Figure 9-18). The energy sup-
plied to the pavement surface is through a bit that is typically shaped like
a spade or cone. In current configurations, this device appears to exces-
sively damage the concrete materials around the crack. Additional work is
therefore warranted to improve on the design of bit types and operating
methods.
1Tecl-ri-v~in_g_h~am--m-e-rs-(Figure 9-19) nave also oeensuccessfully used to
crack concrete pavements. The impact energy of the driver is determined
by controlling the amount of fuel which goes to the hammer. Increased
fuel flow increases impact energy and vice versa. Typically, the hammer rate
remains constant at approximately 50 blows per minute, and the impact
-~=-=--/--
S3 ~
f
Energy Range Production Rate
"ti
i
V)
EqulpmJ.nt Type Applications (ft. lbs.) (lane-miles/day) Advantages I
I
-Umitatlons
aa
ball l
Crane and Jrecking small areas where
specialized equipment not
.25 10 .75 Equipment generally
available
Requires skilled crane
operator ta
r, Ill
justified
JCP I 3:~
Whiphammef 1.00 High productivity covers Not effective on JRCP
4o
full lane width
~,.,..
Drop ~amm¢r JCP and wire 'mesh 4,800 • 15,000 .15 to .75 Effective with JCP I Developed unusual
&~ ~
~
reinforced JRCP crack pattern with mesh
reinforced JRCP.
Several passes required
=
.;·
Guillotine JCP and JRCP 12,000 - 120,000 .75 to 4.0 Versatile - effective with
JCP and JRCP;
a
n,
=
.... i
preferred by several
states; high;productivity;
~ f
2,000
covers full lane width
Rubblized portions of
'"O
~-
J
~
Impact hamter
concrete below cracks
Pile-driving I JRCP 8,000 • 115,000 .27 Covers full lane width I Low productivity
(j
=i-
8
@
~
Resonant
breaker
paljement JCP, JRCP, CRCP <2,000 .so Excellent for rubblizing I Not effective for
making transverse
cracks i........
.,
....fl>
~-
8.
()
I
(')
fl>
s
i::i...
spacing is determined by the tractor speed. The breaking pattern is estab-

lished by varying the fuel and observing the effect of the impact on crack
development and then choosing a throttle setting to achieve the desired
result.
The whip hammer (Figure 9-20) consists of a flexible arm made from
a series ofleaf springs of different lengths with a flat cracking head attached
to the impact end. The device is mounted on the rear of a truck which pro-
vides motive power and hydraulic power for lifting the arm to its striking
position. In the striking position, the leaf springs are bent and, when
released, the hammer swings down to strike the pavement surface
with an impact force due to the whipping action of the arm. The striking
plate typically has a diamond, square, or rectangular shape.
The resonant breaker works on the principle that the frequency of a
vibratory force can be varied until the resonant frequency of the body
being vibrated can be determined. At resonant frequency, the effect of the
energy supplied through the vibration is greatly increased because at the
resonant frequency of a body, the ratio of the response amplitude to the
input amplitude increases without bound. Therefore, a resonant breaker
supplies power (vibration energy) to the slab through a bevelled plate on a
12.5 foot steel beam. The natural frequency of a concrete slab is about 44
Figure 9-16. Falling-weight Drop Hammer Used for Cracking

Concrete Pavements
Figure 9-17. Guillotine Type Falling-weight Drop Hammer Used for

Cracking Concrete Pavements
cps, and the input amplitude can be varied from 1.25 to 1.5 inches at an
impact force of about 2000 pounds. This force system breaks the concrete
slab apart and destroys the bond between any concrete and steel, enhanc-
ing the removal of the reinforcing steel or wire fabric if required. The steel
is often left in place, but protruding bars are cut. Joint sealers are easily
removed manually.
Crack Pattern and Piece Size. Information collected as part of the

· studies reported by Crawford (22) and Kilareski and Stoffels C2fi.) shows
that many states require cracked piece sizes on the order of 1.5 to 2 feet
(460-610 mm) on a side. There is considerable variation, however, within
the requirements of each of the states. This variation is believed to be a
result of experimentation to determine which piece size works best in pre-
venting reflection cracking. For example, the state of Kentucky has exper-
imented with cracked pieceAirri~nsions froJ!!3 to 6 inches (75 to 150 mm)'------
e-----u-p to -i 5 to 3 feet (760-910 mm). States requiring pieces larger than 3 by
3 feet (910 by 910 mm) include California, Iowa, Kansas, Minnesota, and
Pennsylvania. Some states specify a transverse crack spacing.
Once the concrete pavement slabs have been cracked, some determi-
nation must be made to ascertain whether the cracking pattern meets spec-
Figure 9-18. Hydraulic Impact Hammer Used for Cracking Concrete

Pavements
ifications and if the cracks extend all the way through the slabs. The prop-
er cracking technique results in a series of fine cracks which run the full
depth of the slab. Some states report that a fine spray of water is needed to
render these fine cracks visible to the inspector. There is some question as
to whether cracks so difficult to see extend all the way through the slab.
Therefore, some states use the visual cracks survey plus a deflection survey,
conducted both before and after the slab cracking is complete, in an
attempt to determine indirectly from the deflection ratio whether the
cracking extends all the way through the slab. Deflection ratio is the deflec-
tion after cracking and seating divided by the deflection before cracking
and seating and is therefore an indication of the breakup of the slab action.
Additional insight should be provided when the results from the study
described in Reference 2.Z. are compiled and presented to the project spon-
------,<.sb-ts:-This study mvolved the ident:ification and collection of partially com-
plete data on 487 field projects in 34 states, which was followed by the
selection of 118 projects for further field investigation. These 118 projects
included concrete pavements having slabs subjected to (a) break and seat,
(6) crack and seat, (c) rubblization or (d) HMA overlay that was sawed
above all joints and working cracks and then sealed.
Figure 9-19. Pile Driving Equipment Used for Cracking

Concrete Pavements
The field investigations involved visual distress surveys made in accor-

dance with the Pavement Condition Index (PCI) highway methodology
and non-destructive deflection testing using a falling weight deflectometer
(FWD). The FWD data were used to backcalculate the effective modulus
of the concrete layer in each test section. Other pertinent data were secured
from the individual highway departments to produce a complete set of
data for each test section.
Seating of Cracked Slabs. Once cracked, the concrete pavement

pieces must be firmly seated against the subbase layer. If proper seating is
not achieved, the individual pieces may rock under load and cause reflec-
tion cracking. Therefore, just as the concrete slab is broken to deter the
~
horizontal movement which causes reflection cracking, the pieces must be

in a stable position to prevent reflection cracking due to vertical move-
ments under loads. Generally, heavy rubber-tired rollers have been used
most successfully to seat the concrete pavement pieces. Steel-wheeled
rollers do not achieve the desired result because they bridge over low spots
and therefore cannot apply load to every broken piece of the concrete slab.
The best seating results have been obtained with either 35 or 50-ton
(32-46 Mg) rubber-tired rollers as shown in Figure 9-21. Two passes with
a 50-ton (46 Mg) roller have given good results as well as 7 or 8 passes with
a 35-ton (32 Mg) roller. Generally, the 50-ton (46 Mg) roller should be
used with the longer crack spacings or bigger pieces, and the 35-ton (32
Mg) roller used with shorter .crack spacings or smaller pieces.
Other Considerations in Cracking and Seating. Before making a

decision to crack or break and seat an old concrete pavement, a detailed
condition survey should be performed to determine the condition of the
joints and working cracks on the job. If there are too many badly deterio-
rated joints which require extensive amounts of undersealing to fill voids,
it might be more economical to rubblize the concrete slab. Alternative
thickness designs should be prepared comparing the crack and seat, break
and seat, and the rubblize-plus-overlay, using a life-cycle cost analysis.
Many states install edge drains at the time that the crack and seat or
break and seat operation is underway. Positive drainage improves the per-
Figure 9-20. W'hiphammer Used for Cracking Concrete Pavements

formance of these sections and provides an opportunity to view the effi-

ciency of the cracking operation. at the pavement edge. An effective crack-
ing procedure should produce dear visual evidence that the crack has pro-
gressed through the slab a~ shown in Figure 9-22.
Traditionally, overlay thicknesses on crack or break and seat pavement,
as well as for rubblized concrete pavements, have been based on experience
'ii, ,
'·I!
and engineering judgement. References 22 and 3Q contain suggestions for
estimating appropriate structural layer coefficients for the crack-and-seat
or break-and-seat slabs. When a concrete pavement has been reduced to
rubble, it basically becomes an aggregate base and should be treated as such
in the pavement design process to determine how much HMA is required
to carry the traffic during the design period. For additional guidance on ·
thickness design for overlays on concrete pavements or on rubblized con-
crete, refer to the 1986 AASHTO Guide (1), MS-17 of the Asphalt
Institute U2), or Principles ofPavement Design, (2Q).
Saw and Seal. The basic concept behind saw and seal is that, since
reflective cracks occur in an HMA overlay on concrete pavement, it is
easier to seal and maintain a sawed straight groove than a narrow, mean-
dering crack. Kilareski and Bionda (3.1) report that saw and seal was first
suggested in 1954 but was not tried in the field for several years. The pro-
cedures first used and those in use today appear to be very similar with the
Figure 9-21. Heavy Rubber Tire Roller Used to Seat Broken Concrete
Slabs
standard procedure for saw and seal shown in Figure 9-23. A review of the
literature by Kilareski and Bionda and a summary of the results from their
study indicate that:
1. The bulk of the field experience with saw and seal is in the north-
eastern U.S.; however, in recent years several other states have elect-
ed to try this method. States with some experience include Arizona
(G), Connecticut (G), Louisiana (M), Maine (G), Massachusetts
(G), Michigan (M), New Jersey (G), New York (G), North Carolina
(G), Ohio (G), Pennsylvania (G), and Rhode Island (G). (G means
good effectiveness while M means marginal effectiveness of the
method) ,
2. It is absolutely crucial that the joint or crack be accurately located
before the saw cut is made; if not correct to within 1 inch, sec-
ondary reflection cracking occurs and causes serious roughness and
maintenance problems.
3. Thick overlays (5 inches or more) perform better than thin overlays
(2.5 inches or less).
4. Compared to the control sections, properly constructed saw and
seal joints reduce roughness by 20 percent and transverse reflection
cracking by 64 percent. The saw and seal technique cannot be
applied to random cracks in concrete pavements.
Figure 9-22. Cutaway Section Showing the Crack Pattern Produced

by a Crack and Seat Operation
RECONSTRUCTION
Reconstruction of a flexible pavement becomes necessary when the

pavement is not -rehabilitated in time and it begins to deteriorate rapidly.
Some distress types which may indicate the need for reconstruction are alli-
gator or fatigue_ cracking, corrugations, depressions, pumping and water
bleeding, and rutting. Most of these distresses originate from inadequate
base support, so the pavement must be reconstructed from the bottom up.
Quite often, inadequate subsurface drainage is the primary cause of prob-
lems.
Pavement reconstruction was accomplished with all new materials in
the past. However, recycling of existing material is increasingly being per-
formed by agencies to conserve energy and marerials. These recycled struc-
tural layers are then overlaid with new or recycled HMA overlays to achieve
the desired total thickness of the reconstructed pavement.
Cold Mix Recycling with Overlay

Cold mix recycling is a process in which the reclaimed asphalt pave-
ment {RAP), reclaimed aggregate material (RAM), or both, are combined
with asphalt (generally emulsified asphalt), and/or recycling agents in place
or at a central plant to produce cold-mix base mixtures. This process is
suitable for low-volume roads because the cold mix is not structurally as
strong as HMA (32.).
JOINT JOINT SEALANT

RESERVOIR INITIAL SAWCUT
REFLECTION
CRACK
OLD PCC
PAVEMENT
---t - - - OPENING
i
VERTICAL
DIFFERENTIAL
DEFLECTION
NOTE• The sawed joint should be within I in 125.4 mml of

the underlying PCC joint to prevent secondary
cracking.
Figure 9-23. Schematic of Typical Saw and Seal Method of
Controlling Reflection Cracking in a HMA Overlay (.J1)
Since cold recycled mix does not have adequate resistance to abrasion
by traffic and moisture induced damage, it must be capped with a wearing
course ofHMA or a single or double surface treatment. In-place cold mix
recycling is also suitable for roads which have been built over the years with
layers of surface treatments and/or cold mixes. Recycled cold mix in such
cases provides a reasonably good, uniform base course which can be over-
laid with HMA to provide a durable riding surface.
Hot Mix Recycling with Overlay

Hot mix recycling of the RAP (and RAM, if available) is recommend-
ed when reconstructing pavements carrying medium to heavy traffic.
Recycled courses are then provided with a HMA overlay. Hot mix recycling
was discussed earlier in Chapter 7.
Drainage
Inadequate subsurface drainage is frequently a primary cause of the
pavement distresses which necessitate reconstruction. Therefore, when
diagnosing the cause(s) of pavement distresses, drainage must be investi-
gated thoroughly. If drainage is found to be inadequate, it must be
improved as an integral part of the reconstruction project, including
structural overlays.
Aggregate base courses or subbases are most often not free-draining as
intended. Some dense-graded granular bases may actually become saturat-
ed with water that cannot drain from the base. Subbases under some old
concrete pavements get saturated with subsurface water, and concrete slabs
start to pump water through the joints and cracks. If the concrete pave-
ment has a HMA overlay, the water under pressure can cause stripping and
potholing of the overlay (33). Asphalt treated permeable material (ATPM),
as discussed in Chapter 7, can be used as a drainage layer between the con-
crete pavement and HMA overlays.
Longitudinal drains are required only along the lower edge of travelled
lanes in flat terrain. Lateral drains must be provided in rolling or hilly ter-
rain to intercept water in the drainage layer because water is likely to flow
longitudinally on steep grades.
Reconstruction gives a unique opportunity to improve the existing
drairrage-sysrem;-irrdrrdirrg-additiona:l-rrreasures-such-as transverse drains
cut into the pavement structure and changes in the drainage pattern.
Reference 3d gives details on the hydrology and hydraulics of both surface
and subsurface drainage and describes methods for the design and con-
struction of drainage systems.
REFERENCES
I. Consequences of Deferred Maintenance. NCHRP Synthesis of

Highway Practice 58, National Cooperative Highway Research
Program, National Research Council, May 1979.
2. Stevens, L.B. Road Surface Management for Local Governments-
Resource Notebook. Federal Highway Administration, Report No.
DOT-I-85-37, May 1985.
3. Darter, M.I. and M.Y. Shahin. Pavement Rehabilitation: Identifying
the Need. Transportation Engineering Journal, ASCE 106 (1), 1980.
4. MSHTO Guide for Design of Pavement Structures 1986. American
Association of State Highway and Transp.9rtation Officials, 1986.
5. Hudson, S.W, B.F. McCullough, and RF. Carmichael, III. Surface
Design and Rehabilitation Guidelines for Low-Volume Roads. Federal
Highway Administration, Report No. FHWNST-87-225, June 1987.
6. FHWA Pavement Rehabilitation Manual. Federal Highway
Administration, Report No. FHWA-ED-88-025, March 1988.
7. Brown, E.R. Preventive Maintenance of Asphalt Concrete Pavements.
Transportation Research Board, Transportation Research Record 1205,
1989.
8. Herrin, M., C.R. Marek, and K. Majidzadeh. Surface Treatments:
State-of-the-Art. Highway Research Board, Special Report 96, 1968.
9. Asphalt Surface Treatments and Asphalt Penetration Macadam.
Asphalt Institute, Manual Series MS-13, November 1969.
10. A Basic Asphalt Emulsion Manual-Volume I-Understanding and
Using Emulsions. Federal Highway Administration, Report No.
FHWA-IP-79-1, January 1979.
11. Kandhal, P.S. Simplified Design Approach to Surface Treatments for
Low-Volume Roads. Transportation Research Board, Transportation
12. Kandhal, P.S. and J.B. Motter. Criteria for Accepting Precoated
Aggregate for Seal Coats and Surface Treatments. Paper presented at
the 1991 Annual Meeting of the Transportation Research Board,
January 1991.
13. Report on Slurry Seal Applications. International Slurry Seal
----------r>·SS0€-ia1ci-0fl,-I-n-for-m-a:t--i0fl.-Ser-ies-N-o:--l,R:evi-sed-Editio·...,_.- - - - - - - - - -
14. Benedict, C.R. Slurry Seal Systems for Pavement Conservation. Paper
presented to Toledo Section, ASCE/University of Toledo Seminar,
April 12, 1982, paper available from: International Slurry Seal
Association, Washington, D.C.
--
15. Design Technical Bulletin. International Slurry Seal Association. 1978.

16. Peterson, D.E. Resealing Joints and Cracks in Rigid and Flexible
Pavements. Transportation Research Board, NCHRP Synthesis 98,
December 1982.
17. Pothole Primer-A Public Administrators Guide to Understanding
and Managing· the Pothole Problem. U.S. Army Corps of Engineers,
Cold Regions Research and Engineering Laboratory, Special Report
81-21, September 1981.
18. Asphalt in Pavement Maintenance. The Asphalt Institute. Manual
Series No. 16 (MS-16), March 1983.
19. Asphalt Overlays for Highway and Street Rehabilitation. The Asphalt
Institute. Manual Series No. 17 (MS-17), Second Edition, 1983.
20. Yoder, E.J., and M.W. Witczak. Principles of Pavement Design,
Second Edition, John Wiley and Sons, Inc., New York, NY., 1975.
21. Guideline Specifications for Hot Surface Recycling. Asphalt Recycling
and Reclaiming Association, July 1986.
22. Finn, F.N. and C.L. Monismith. Asphalt Overlay Design Procedures.
Transportation Research Board, NCHRP Synthesis 116, December
1984.
23. Sherman, G. Minimizing Reflection Cracking of Pavement Overlays.
Transportation Research Board, NCHRP Synthesis 92, September
1982. .
24. Germann, F.P. and R.L. Lytton. Methodology for Predicting the
Reflection Cracking Life of Asphalt Concrete Overlays. Texas
Transportation Institute, Report No. 205-5, Texas A&M University,
January 1977.
25. Picket, D.L. and R.L Lytton. Laboratory Evaluation of Selected Fabrics
for Reinforcement of Asphaltic Concrete Overlays. Texas
Transportation Institute, Report 261-1, Texas A&M University,
August 1983.'
26. McLaughlin, A.L. Reflection Cracking of Bituminous Overlays for
Airport Pavements, A State of the Art. U.S. Federal Aviation
Administration, Report No. FAA-RD-79-57, May 1979.
27. Guidelines and Methodologies for the Rehabilitation of Rigid
Pavements using HMA Overlays. Pavement Consulting Services, Draft
- - - - - - - ·..e~:Ft-t.e-Nat-itinal-Asphal.t-12avement-Association, December 1989.
28. K.ilareski, W.P. and S.M. Stoffels. Structural Overlay Strategies for
Jointed Concrete Pavements-Volume 11 Cracking and Seating of
Concrete Slabs Prior to AC Overlay. Federal Highway Administration,
Draft Report No. FHWA-RD-89-143, April 1990.
29. Crawford, C. Cracking and Seating of PCC Pavements Prior to

Overlaying with Hot Mix Asphalt-State of the Art. National Asphalt
Pavement Association, Information series 98/87, 1987.
30. Lukanen, E.O. Structural Evaluation of Cracked and Seated PCC
Pavements for Overlaying with Hot Mix Asphalt. National Asphalt
Pavement Association, Information Series 100/87, September 1987.
31. Kilareski, W.P. and RA. Bionda. Structural Overlay Strategies for
Jointed Concrete Pavements-Volume I-Sawing and Sealing of
Joints in AC Overlays of Concrete Pavements. Federal Highway
Administration, Draft Report No. FHWA-RD-89-142, April 1990.
32. Kandhal, P.S. and W.C. Koehler. Cold Recycling of Asphalt Pavements
on Low Volume Roads. Transportation Research Board, Transportation
Research Record 1106, Vol. 2, 1987. _
33. Kandhal, P.S., C.W. Lubold, and F.L. Roberts. Water Damage to
Asphalt Overlays: Case Histories. Proceedings, Association of Asphalt
PavingTechnologist, Vol. 58, 1989.
34. Drainage of, Asphalt Pavement Structures. The Asphalt Institute,
Manual Series No. MS-15, September, 1984.
Index 579
INDEX
AASHTO Road Test, 477-479 Air voids, 192-204, 215-216

Abrasion resistance of aggregate, 138- Aging:
139 age hardening, 42-45
Absorption (by aggregate), 147-153, aging index, 44-45
203-204 aging tests, 29-31, 75-76, 283-284
Additives, 448-471 Alligator cracking, 64-66, 97-98, 494-
Aggregate: 495
abrasion resistance, 138-139 Antioxidant, 463-464
absorption, 147-153 Antistripping agents, 71, 466-468,
blending, 160-169 515-517
chemical properties, 134-135 Asphalt, 7-23
cleanliness, 143-145 cutback asphalt, 19-23
crusher, 123-128. emulsified asphalt, 17-19
dryer, 327-328 historical background, 8
durability, 139-140 natural asphalt, 8
feed, 318, 385-389 petroleum asphalt, 8
feeder calibration, 387-389 refining, 8-15
Fuller's curve, 154-155 sampling, 390
gradation, 153-170, 239-244 see alro asphalt cement
maximum density line, 156, 185- Asphalt absorption, 203-204
187 Asphalt cement:
maximum size, 155-157 absolute viscosity, 23-25
mineralogy, 129-134 aging tests, 29-31, 75-76, 283-284
particle shape, 140-143, 235-238 amphoterics, 114
physical properties, 136-158 AR viscosity grading, 41
plasticity index, 145 bending beam rheometer, 86-92
production, 123-128 chemistry, 102-114
quarry operations, 123-128 composition, 103-106
sampling, 128-129, 385-387 consistency tests, 23-29
sand equivalent test, 144 direct tension tester, 92-93
sieve analysis, 157-158 ductility test, 28-29, 47-48
size, 153-170, 520 dynamic shear rheometer, 78-86
soundness, 139-140 flash point, 31-32
specific gravity, 145-153, 192-204 functional groups, 104-105
surface texture, 140-143, 520-521 grading systems, 33-41, 93-102
surface area, 169-170 kinematic viscosity, 25-26
-------ynthetie,l-2.Z-,5-2tt0--------moleeular-strueture, 103-104
toughness, 138-139 molecular size distribution, 111-
uncompacted voids, 235-238 113
Air-blowing, 13-14 penetration, 26-27, 46-47
Air quality control system, 331-337 penetration grading, 35-37
penetration index (PI), 50-51 Bump specifications, 486-489

pen-vis number (PVN), 51-53 Burner, 321-325
performance grading (PG), 93-
102 Centrifuge kerosene equivalent, 219-
physical tests, 23-33, 73-93 220
polar groups, 104-105 Chemistry of asphalts, 102-114
pressure aging vessel, 75-76 Chip seals, 535-540
purity tests, 31 Cleanliness (aggregate), 143-145
rolling thin film oven test, 30-31, Cold milling, 552-554
73-75 Cold mix recycling, 574-575
rotational viscometer, 76-78 Compaction,
safety tests, 31-32 control strip, 376-377
shear susceptibility, 54 cooling of mat, 365-366
softening point, 27-28 equipment, 356-365
specific gravity, 32 field, 354-379
spot test, 32-33 gyratory, 244-249, 304-305
stiffuess, 54-64 laboratory, 189.:.191
Superpave physical tests, 73-93 pneumatic-tired roller, 360-363
Superpave specifications, 93-102 rolling pattern, 366-369
temperature susceptibility, 49-54 specifications, 369-379, 382-383
thin film oven test, 29-30, 283- steel-wheeled roller, 359-360
284 Superpave gyratory, 244-249
viscosity grading, 37-41 vibratory roller, 363-365
viscosity, 23-26, 48-49 Complex modulus, 289-291
viscosity-temperature Composition of asphalt cements,
susceptibility (VTS), 53-54 103-106
Asphalt content, 400-401 Control strip, 376-377
Asphalt paving, see paving Corrective maintenance, 544~550
Asphalt plant, see facility Crack propagation, 561-562
Asphalt treated permeable material Corrugations, 506-507
(ATPM), 411-413, 575 Crack relieflayers, 556-559
Crack sealing, 54J-544
Baghouse, 335-337 Cracking, 64-70
Base mixes, 411-413 Cracking and seating concrete slabs,
Batch facility, 316-318, 327-331, 563-572
444-447 Creep, 301-304
Bidding, 381 Creep compliance, 302
Bleeding, 276, 532-534 Crude petroleum, 8-11
Blending (aggregate), 160-169 Crusher, 123-128
Block cracking, 498-499, 532-534
---.Boffing test, 306-307 Deicing calcium chloride, 471
Bond breaking material, 559-562 Density, 216-217, 369-379
Breaking and seating concrete slabs, Distillation, 10-12
563-572 Distortion (pavement), 502-507
British portable skid tester, 521 Distresses, 489-521
Index 581
block cracking, 498-499 weigh hopper, 330

corrugations, 506-507 wet scrubbers, 333-335
fatigue cracking, 64-66, 97-98, Fatigue cracking, 64-66, 97-98, 494-
494-495 495
longitudinal cracking, 496-498 Fatigue resistance, 182, 292-298
polishing, 517-521 Feeder calibration, 387-389
potholes, 517 Fibers, 461-462
ravelling, 64, 276, 507-509 Field management, 402-408
reflection cracking, 499-501, 563- Filler, 451-453, 505
573 Film thickness, 187-189
rutting, 70,276,370-371, 502- Fines feeder system, 326-327
506 Flash point, 31-32
shoving, 506-507 Flexural fatigue, 292-298
skid resistance, 517-521 Fog seals, 532-534
slippage cracking, 501-502 Fractured faces, 142
stripping, 70-71, 510-517 Friction course (open-graded), 413-
thermal cracking, 66-70, 98, 495- 415, 513-514
496 Friction resistance, 252-254, 517-521
wear loss, 509-510 Fuller's curve, 154-155
Drainage, 575
Drum mix facility, 313-316, 318- Glass (waste), 469-470
327, 446-447 Grading asphalt cements, 33-41, 93-
Drying and heating, 323-325 102
Ductility, 28-29, 47-48 AR viscosity grading, 41
Durability, 29-31, 75-76 penetration grading, 35-37
Superpave grading, 93-102
Equipment, see facility _ viscosity grading, 37-41
Extenders, 454-455 Gradation (aggregate), 153-170, 239-
244
Fabrics, 560 Gyratory shear index, 305
Facility (plant): Gyratory testing machine, 304-305
aggregate feed, 318
air quality control system, 331- Hardening agents, 465-466
337 Hot bins, 328-330
baghouse, 335-337 Hot mix recycling, 250-252, 439-448
batch, 316-318, 327-331 Hot surface recycling, 551-552
burner, 321-325 Hubbard-Field Mix design method,
drum mix, 313-316, 318-327 175-176
dryer, 327-328 Hveem cohesiometer, 177, 286
drying and heating, 323-325 Hveem mix design method, 176-177,
- - - - - ~fines feeder system, 326-327 218-234, 262-263
hot bins, 328-330 Hveem stability, 176-177, 230-231,
primary dust collectors, 332-333 285
pugmill, 330-331 Hydraulic structure mixtures, 439
screens, 328-330 Hydrocarbons, 464-466
Igneous rock, 121 characterization, 275-312

Immersion compression test, 308 complex modulus, 289-291
Indirect tension, 298-300 creep test, 301-304
Intersections, 506 design, 173-274
flexure test, 292-298
Laboratory wheel tracking test, 305- hydraulic structure mixtures, 439
306 indirect tension, 298-300
Large stone mixes, 174-175, 423-430 large scone mixes, 174-175, 423-
Laydown operations, 344-354 430
Lignin, 452, 454-455 maximum density curves, 156,
Lime, 467-468 185-187
Loaded wheel tester, 305-306 modulus of elasticity, 286
Longitudinal cracking, 496-498, 532- moisture susceptibility, 182, 249-
534 250, 306-310
Lottman test, 307-310 open-graded base mixes, 411-413
Low temperature cracking, 66-70, 98, open-graded friction course
495-496 (OGFC), 413-415, 513-514
patching mixtures, 430-439
Maintenance, 527-550 permanent deformation, 97, 301-
chip seals, 535-540 304
corrective maintenance, 544-550 recycled mixtures, 439-448
crack sealing, 543-544 resilient modulus, 286-289
fog seals, 532-534 sampling, 390
patching, 544-548 segregation, 379-380, 508-509
preventive maintenance, 531-544 specifications, 382-385
rejuvenators, 535 stiffness, 54-64
slurry seals, 540-543 .variability, 277-278
surface treatments, 535-540 Mixture design:
thin overlays, 548-5 50 history, 173-181
Marshall flow, 209-210, 218 Hubbard-Field method, 175-176
Marshall mixture design method, Hveem method, 176-177, 218-
177-181, 204-218, 255-262 234, 262-263
Marshall stability, 209-210, 217-218, Marshall method, 177-181, 204-
284-285 218, 255-262
Marshall stiffness index, 285 modified Marshall method, 425-
Material variability, 277-278 429
Maximum density line, 156, 185-187 objectives, 181-185
Mays road meter, 480-481 Superpave method, 181, 234-249,
Metamorphic rock, 122 263-273
Milling, 552-554 Modifiers, 448-471
-----------iviineral filler, 451-453, 505 Modulus of Elasticity, 286
Mineralogy of aggregate, 129-134 Moisture Susceptibility, 182, 249-
Mixture: 250, 306-310
additives, 448-471 Molecular size distribution, 111-113
asphalt content, 400-401
Index 583
Nuclear gauge, 378-379 Quarry operations, 123-128
Open-graded base mixes, 411-413 Ravelling, 64, 276, 507-509

Open-graded friction course Reclaimed asphalt pavement (RAP)
(OGFC), 413-415, 513-514 material, 440-448
Overlays, 514-515 Reconstruction:
Oxidants, 463 cold mix recycling, 574-575
drainage, 575
Particle shape, 140-143, 235-238 hot mix recycling, 250-252, 439-
Patching, 544-548 448
Patching mixtures, 430-439 Recycled mixture design, 439-448
Pavement: Recycling, 250-252, 439-448
performance, 42-71, 477-525 Recycling agents, 464-465
roughness, 479-489 Refining (asphalt), 8
Pavers, 344-354 Reflection cracking, 499-501, 563-
Paving, 344-354 573
Paving screed, 344-352. Rehabilitation, 550-573
Paving specifications, 353-354, 369- bond breaking material, 559-562
379 breaking and seating concrete
Penetration, 26-27, 46-47 slabs, 563-572
Penetration grading, 35-37 cold milling, 552-554
Penetration index, 50-51 crack relieflayers, 556-559
Pen-vis number, 51-53 cracking and seating concrete
Performance (pavements), 42-71, slabs, 563-572
477-525 open-graded friction courses, 552
Permanent deformation, 97, 301-304 reflection cracking, 563-573
Permeability, 370-373 rubblizing, 563-573
Plant, see facility sawing and sealing, 563, 572-573
Plastic, 460-461 stress-absorbing membrane inter-
Plasticity index (aggregate), 145 layer (SAMI), 560-562
Poisson's ratio, 287-288 structural recycling, 554-555
Polar groups, 104-105 structural overlays, 555-556
Polishing, 517-521 surface recycling, 551-552
Polymers, 455-461 thick overlays, 559
Potholes, 517, 546-548 Rejuvenators, 535
Present serviceability index (PSI), Resilient modulus, 286-289
478-479 Rheological properties, 42-71
Preventive maintenance, 531-544 Rollers, 356-365
Primary dust collectors, 332-333 Rolling patterns, 366-369
_____P_ro'--fi_ilo-'--m--'-et--'e-'rs.:...,-"'47'--'9'--_4_8'---9_ _ _ _ _ _---"-It=o-"'ofi=m=g shingles, 469 __
Pugmill, 317, 330-331 Root-Tunnicliff method, 308
Roughness of pavements, 479-489
Quality assurance, 399-408 Rubber, 456-460
Quality control, 399-408 Rubblizing, 563-572
Quality control charts, 404-408 Rut depth, 478
Rutting, 70,276,507-509 Stockpile patching mixtures, 430-439

Stone matrix asphalt (SMA), 415-423
Safety tests (asphalt cement), 31-32 Storage silo, 341-344
Sampling: Strategic Highway Research Program,
aggregate, 128-129, 385-387 72, 181, 234-235
asphalt, 390 Stress absorbing membrane interlayer
mixture, 390 (SAMI), 560-562
Sand equivalent test, 144 Stripping, 70-71, 510-517
Saw and seal, 563, 572-573 Structural design, 555-556
Scrap rubber, 457-460, 469 Structural overlays, 555-556
Screed, 344-353 Subsurface drainage, 510-513
Sedimentary rock, 121 Superpave:
Segregation, 379-380, 508-509 binder tests, 72-93
Shear susceptibility, 54 buiaerspecification, 93-102
Seal coats, 515 mix design, 181, 234-249, 263-
Seive analysis (aggregate), 157-158 273
Serviceability, 477-489 Sulfur, 452, 454-455
Shingles (roofing), 469 Surface area (aggregate), 169-170
Shoving, 506-507 Surface capacity, 220-221
Silicones, 470-471 Surface recycling, 551-552
Slag, 122 Surface texture (aggregate), 140-143,
Slippage cracking, 501-502 520-521
Slurry seals, 540-543 Surface treatments, 535-540
Smoothness of pavements, 486-489 Surge silo, 314-316, 341-344
Softening point, 27-28
Technician qualifications, 404
Solubility test, 31
Temperature susceptibility, 49-54
Solvent deasphalting, 12
Tender mixes, 157, 187
Solvent extraction (ROSE), 12-13
Testing frequency, 404
Soundness of aggregate, 139-140
Thermal cracking, 66-70, 98, 495-
Specific gravity:
496
asphalt cement, 32
Thin overlays, 548~550
aggregate specific gravity, 145-
Toughness (aggregate), 138-139
153, 192-204
Transportation operations, 338-344
apparent specific gravity, 145-151,
Transportation vehicles, 338-341
196
Trucks, 338-341
bulk specific gravity, 146-153, 196
effective specific gravity, 146-147,
196 Viscosity, 23-26, 48-49
absolute viscosity, 23-25
t h eoretical maximum specific
------gravity;t97::'.204-;-374-=3•iG'-------ki=·=n=em=a.tic_vjs_c_Qsiry;.25-=2_6,_ _ _ _ _ _ __
Specifications (HMA), 382-385, 390- Viscosity grading, 37-41
394, 479-482 Viscosity-temperature susceptibility,
53 4
Statistical concepts, 390-398 -5
Stiffness, 54-64 Voids, 192-204, 215-216
Voids filled, 192-204, 248
--
lnde:,: 585
Voids in mineral aggregate, 192-204,

214-215, 248
Volumetric properties, 402-408
Waste glass, 469-470

Waste materials, 468-470
Wear loss, 509-510
Weigh hopper, 330
Wet scrubbers, 333-335
Workability, 185

Hot Mix Asphalt Materials Mixture Design and Construction

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Hot Mix Asphalt Materials Mixture Design and Construction

Uploaded by

Copyright:

Available Formats

• C

HOT MIX ASPHALT

National Asphalt Pavement Association

This book or any part thereof shall not be reproduced without

Library of Congress Catalog Number: 91-60083

Printed in the United States of America

National Asphalt Pavement Association Research

PREFACE TO FIRST EDITION

All typing, editing, xeroxing, and drafting were coordinated by Joyce

Many people reviewed either portions of the textbook or the entire

Those individuals reviewing several chapters included Richard Davis,

Thanks to Brown Printing Company for working closely with the

ORIGIN OF BOOK ................................ 1

IMPORTANCE OF SUBJECT ............. , ......... .3

DESCRIPTION OF CONTENTS ..................... .4

CHAPTER 2. ASPHALT REFINING, USES, AND

HISTORICAL BACKGROUND AND REFINING ......... 7

ASPHALT TYPES AND USES ....................... 15

ASPHALT CEMENT PHYSICAL TESTS .............. .23

Safety Tests .................................... 31

ASPHALT CEMENT GRADING SYSTEM ............. .33

RHEOLOGICAL PROPERTIES AND PAVEMENT

SUPERPAVE ASPHALT BINDER TESTS AND

ASPHALT CHEMISTRY ........................... 102

REFERENCES ........ , ......................... .114

CHAPTER 3. AGGREGATES ......................... 121

INTRODUCTION ............................... 121

AGGREGATE PRODUCTION ...................... 123

AGGREGATE SAMPLING ......................... 128

MINERALOGY & CHEMICAL PROPERTIES ......... 129

PHYSICAL PROPERTIES OF AGGREGATES .......... 136

CHAPTER 4. HOT MIX ASPHALT MIXTURE DESIGN

HISTORY OF MIX DESIGN ....................... 173

MARSHALL MIXTURE DESIGN METHOD ..........204

HVEEM MIXTURE DESIGN METHOD .............218

SUPERPAVE MIX DESIGN METHOD .............. ,.234

OTHER CONSIDERATIONS .......................249

EXAMPLES OF MIXTURE DESIGN ................. 254

INTRODUCTION ............................... 275

MATERIAL CHARACTERIZATION TECHNIQUES ....276

CRITERIA FOR ASPHALT MIXTURE TESTS ......... 277

CURRENT TESTS ...............................281

REFERENCES ................................... 310

INTRODUCTION ............................... 313

HOT MIX ASPHALT FACILITIES .................. .313

TRANSPORTATION AND LAYDOWN OPERATIONS . .337

MIXTURE SEGREGATION ........................ 379

CONTRACT AND MATERIAL SPECIFICATIONS ...... 380

STATISTICAI:;-Ce>NeEPTS ........................ 3:9·1c1--- - - -

QUALITY CONTROL/QUALITY ASSURANCE

REFERENCES .................................. .408

CHAPTER 7. SPECIAL MIXTURES, RECYCLING, AND

OPEN-GRADED HMAMIXES .................... .411

DENSE-GRADED LARGE STONE MIXES ........... .423,

ASPHALT PATCHING MIXTURES ................. .430

HMA MIXTURES FOR HYDRAULIC STRUCTURES .. .439

ADDITIVES AND MODIFIERS IN HMA ............ .448

CHAPTER 8. PERFORMANCE/DISTRESS OF HMA ... .477

REFERENCES ............... " .................. .522

CHAPTER 9. MAINTENANCE, REHABILITATION, AND

INTRODUCTION ............................... 527

MAINTENANCE ............................... .531

- - - - - - - ~ 0 N .................. ~~., ........ 550

RECONSTRUCTION ............................ .574