Viscosity Testing of Asphalt and Experience With Viscosity Graded Specifications STP532-EB.29769

Copyright by ASTM Int'l (all rights reserved); Tue Feb 20 02:01:08 EST 2018
Downloaded/printed by
Indian Inst of Tech Kharagpur (Indian Inst of Tech Kharagpur) pursuant to License Agreement. No further reproductions authorized.
VISCOSITY TESTING OF ASPHALT
AND EXPERIENCE WITH
VISCOSITY GRADED
SPECIFICATIONS
A symposium
presented at the
Seventy-fifth Annual Meeting
AMERICAN SOCIETY FOR
TESTING AND MATERIALS
Los Angeles, Calif., 2 5 - 3 0 June 1972
ASTM SPECIAL TECHNICAL PUBLICATION 532

W . G. Gunderman, symposium chairman
04-532000-08
AMERICAN SOCIETY FOR TESTING AND MATERIALS

Ammûir 1916 Race Street, Philadelphia, Pa. 19103
© BY AMERICAN SOCIETY FOR TESTING AND MATERIALS 1973
Library of Congress Catalog Card Number: 72-97872
NOTE
The Society is not responsible, as a body,
for the statements and opinions
advanced in this publication.
Printed in BaltinK>re, Md.

June 1973
Printed in New Carlisle, Obio

March 1 9 8 1
Foreword
The Symposium on Viscosity Testing of Asphalt and Experience with
Viscosity Graded Specifications was presented at the Seventy-fifth Annual
Meeting of the American Society for Testing and Materials held in Los
Angeles, Calif., 25-30 June 1972. Committee D-4 on Road and Paving
Materials sponsored the symposium. W. G. Gunderman, Highway Research
Board, National Academy of Sciences, presided as symposium chairman.
Related
ASTM Publications
Rapid Test Methods for the Determination of Bitumen

Content in Bituminous Mixtures, STP 461
(1969), $3.75 (04-461000-08)
Fatigue of Compacted Bituminous Aggregate Mix-
tures, STP 508 (1972), $13.75 (04-508000-08)
Contents
Introduction 1
Rheology of Asphalt Cements at 60 C (140 F)—H. E. SCHWEYER, J. P. MOORE,
AND J. K. LING 3
Temperature Susceptibility 4
Experimental 7
Results 9
Discussion 12
Conclusions 14
Study of Asphalt Viscosity Testing at 60 F—j. v. EVANS AND J. W. GORMAN 16
Experimental 17
Data and Analysis 18
Conclusions and Recommendations 25
Problems in Determining Viscosity of Asphalts at Low Temperatures with
Shell Sliding Plate Microviscometer—E. O. BUSBY AND L. F. RADER 28
Range of Microviscometer 29
Need for Low-Temperature Viscosities 29
Summary 30
Florida Studies with the Puzinauskas Field Viscosity Cup for Asphalt Cements—
C. F. POTTS 31
Base Chart 32
Puzinauskas-Asphalt Institute Field Viscosity Cup (P-AI) 34
Calculations 35
Experimental Data 36
Acceptability Diagram 37
Viscosity Characterization of Asphalt Cement—L. W. CORBETT and H. E.
SCHWEYER 40
Examination of Typical Asphalt Sources 41
Viscosity at 140 F Versus Penetration at 77 F 41
Identifying Grade While Defining Character 45
Viscosity at 275 F as Dependent Upon Viscosity at 140 F 49
Summary 50
The Rolling Thin-Film Circulating Oven—An Improved Rolling Thin-Film
Oven Test—R. J. SCHMIDT 52
Present Situation 53
Description of Oven 58
Comparison of Temperature Control 58
Test Procedure 58
Comparison of Results 58
Collaborative Test Program 61
Comparison of RTF-C and TFO Exposure by Using Collaborative Test
Results 63
Conclusions and Recommendations 63
vi CONTENTS
Asphalt Viscosity-Related Properties of In-Service Pavements in Pennsyl-

vania—P. S. KANDHAL, L. D. SANDVIG, W. C. KOEHLER, AND M. E.
WENGER 66
Design and Construction of Test Pavements 67
Performance of Test Sections 67
Rheological Properties of Asphalts 69
Discussion of Results 73
Conclusions 77
Laboratory Measurement of the Durability of Paving Asphalts—R. J. SCHMIDT 79
Laboratory-Hardened AC Specimens 80
Rolling Microfilm Oven Test (RMFO) 82
Influence of Mineral Filler and Temperature 84
Extra High Air Circulation During R M F O Exposure 85
Capillary Venting of R M F O Exposure Bottle—Preliminary R M F - C
Procedure 86
Effect of Simulating Hot-Mix Hardening Prior to Simulating Long-
Term Durability 87
Effect of Increasing Exposure Time—Final R M F - C Procedure 88
Comparison of Hardening Obtained by Laboratory Aging of AC Speci-
mens with R M F - C Aging of a Variety of Asphalts 89
Comparison of R M F - C Exposure Hardening with Life of Zaca-
Wigmore Test Pavement Project 89
Comparison of R M F - C Exposure with Hardening in Recent California
Division of Highways Field Tests 90
Precision of R M F - C Test 94
Labor Required to Perform R M F - C Test 94
Summary and Conclusions 94
Asphalt Durability Tests and Their Relationship to Field Hardening—c. R.
KEMP 100
Field Test Sections 101
Laboratory Test Procedures 104
Laboratory Test Results 106
Analysis of Test Section Data 106
Correlation of Laboratory and Field Test Results 115
Findings to Date 117
Conclusions 121
STP532-EB/Jun. 1973
Introduction
During the past ten years or so there has been a rather concerted drive
to change asphalt cement specifications and tests from the old empirical
or penetration basis to the newer fundamental unit or viscosity basis.
A materials engineer requires knowledge of the consistency of asphalt
cement during mixing, transporting, placing, compaction, and loading
under traffic in order to minimize or avoid failures. Knowledge of the con-
sistency under summer temperatures and shear properties under winter
temperatures is a requisite to providing durable pavements.
Asphalt cement can go from a liquid during mixing, transporting, and
placing to a plastic during compaction and service under traffic, and even
to a solid at low service temperatures. Satisfactory performance requires a
knowledge of the asphalt consistency under all conditions.
It can be readily seen that all asphalts will not meet all of the perform-
ance criteria required and that with asphalts from a wide geographic range
available to a single producer there is and has been a need for better
methods of characterizing, testing and specifying asphalt cements.
This Symposium on Viscosity Testing of Asphalt and Experience with
Viscosity Graded Specifications was organized to bring to the attention of
the entire country the latest viscosity testing which, is somewhat region-
alized, and the experience gained by a few states which have used viscosity
graded asphalt specifications.
Although not intended to be a state of the art, it serves that kind of
purpose in these areas which are still somewhat limited in the scope of
their applications.
This special technical publication should be of use to those individuals
in asphalt testing, to materials engineers or supervisors, and to those con-
cerned with writing specifications.
The information included in this publication illustrates that viscosity
testing of asphalts is very much in a stage of transition with a shift from
Thin-Film Oven Testing to Rolling Thin-Film Oven Testing, and the small
amount of information concerning low-temperature viscosity testing is an
indication of the need for more work and better test methods and equip-
ment.
Copyright 1973 b y A S F M International www.astm.org

2 VISCOSITY TESTING OF ASPHALT
It also becomes readily apparent that, although much has been said and
written on the subject of viscosity graded specifications, there is still only
limited experience in their use.
W. G. Gunderman
Engineer of Materials and Construction,
Highway Research Board,
National Academy of Sciences,
Washington, D. C , 20418;
symposium chairman.
H. E. Schweyer,^ J. P. Moore,^ and J. K. Ling^
Rheology of Asphalt Cements at 60 C (140 F)
REFERENCE: Schweyer, H. E., Moore, J. P., and Ling, J. K. "Rhelogy of Asphalt

Cements at 60 C (140 V),"Viscosity Testing of Asphalt and Experience with Vis-
cosity Graded Specifications, ASTM STP 532, American Society for Testing
Materials, 1973, pp. 3-15.
ABSTRACT: Studies on the viscosity of 18 asphalts at 60 C (140 F) showed that
the result depends upon the apparatus and procedure used. The Florida capillary
method at constant shear rate gives significantly higher values than does the
Cannon-Manning method. Furthermore, it was shown that, although shear thin-
ning may result for certain asphalts when run in Cannon-Pressure instruments,
there may also be compensating effects of pressure which can raise the viscosity.
These variations are attributed to both differences in the flow mode (transient or
steady state) and to effects of the imposed pressures.
KEY WORDS: asphalts, viscosity, pressure, Florida capillary method, Cannon-

Manning method, pressure viscosity, asphalt viscosity, bitumens
As theological studies of asphalt cements have become more sophisti-

cated, greater knowledge of their behavior is being found. Many earlier
investigators using a limited number of materials proposed relations that
later were found invalid as generalizations because exceptions were found.
This problem was recognized by Stanfield and Hubbard [1^ describing
temperature susceptibility of asphalts, and by Traxler et al [2] on Theo-
logical studies.
Lewis, Halstead, Welborn and other co-workers at the U.S. Bureau of
Public Roads [3-6] presented data on asphalt cements from a wide variety
of sources that emphasized the need to recognize diflferences depending
upon source stocks.
In Ref 3 considerable information on empirical tests on a wide variety
of asphalt source materials was given. In Ref 4 data were shown relating
kinematic viscosity to temperature on a variety of asphalts, and some
peculiarities of 60 C (140 F) were noted.
Reference 5 describes properties of a set of 85/100 asphalt samples, and
comparable information was given in Ref 6 for other penetration grades.
Sisko [7] provided a comprehensive analysis of asphalt-temperature data
• Professor and graduate students, respectively. Department of Chemical Engineering,
University of Florida, Gainsville, Fla. 32601.
' The itahc numbers in brackets refer to the list of references appended to this paper.
Copyright 1973 b y A S T M International www.astm.org

using shift factors that showed flattening of the viscosity-temperature

curves below about 60 C (140 F) with greater effect at the higher shear rates.
Welborn et al [8] presented considerable viscosity data on a new set of
asphalt samples in their vis-grading studies. They also showed viscosity
"offset" phenomena in the region of 60 C (140 F) and lower temperatures.
Halstead and Welborn [9] presented a review of the development of'
viscosity grading specifications that summarizes certain viscosity data at
60 C (140 F) with respect to specification limits.
Certain pertinent information reported by Fink and Griffin [10] on com-
parative viscosity of asphalt cements at 60 C (140 F) showed the sliding
plate data to be about 15 percent higher than capillary vacuum methods
data. In the discussion Gzemski suggested shear susceptibility as a poten-
tial explanation.
The discontinuity of linear viscosity-temperature plots below about 60 C
(140 F) was shown by Griffith and Puzinauskas [77] on a plot of log-log
viscosity versus log absolute (degrees Rankine) temperature for an asphalt
described as a "gel" type.
Similar peculiar effects at 60 C (140 F) were observed for some data on
an ASTM temperature-viscosity plot by Lefebvre and Robertson [12].
In a paper by Schweyer [75] a significant difference in vacuum capillary
viscosities at 60 C (140 F) was noted for shear susceptible asphalts. As a
result of the anomalies noted, a further study of viscosity at 60 C (140 F)
was undertaken as reported herein.
Temperature Susceptibility
One outgrowth of the preceding anomalies noted in the literature was an
in-depth study of the temperature susceptibilities of asphalt cements above
the range of 60 C (140 F) in order to quantitize the effect of temperature
variations at 60 C (140 F).
Stanfield [7] studied susceptibility and reviewed the various proposed
evaluations to that date. Crouch and Cameron [14] presented a review of
seven viscosity-temperature equations for lubricants. More recently Jonge-
pier and Kuilman [75] updated such studies, and numerous investigators
have considered correlations based on activation flow energies and shift
factors. Davis [76] has presented data on coal tars and asphalts. Similar
analyses have been summarized by Ling [17]. The ASTM Standard Test
method [18] has been found generally applicable as a quantitative graphical
representation of the temperature relation.
Certain data for asphalt cements are plotted on the ASTM chart in Fig. 1
for asphalts identified in Table 1, and the analysis can be quantitized as
follows:
In (log v) = M In T + In B (1)
SCHWEYER ET AL ON RHEOLOGY OF ASPHALT CEMENTS 5
347°F 383°F
175°c 195°C
TEMPERATURE
FIG. 1—Temperature susceptibility of certain asphalts.
where
ri = viscosity, centipoises
T = absolute temperature, deg R = 459.7 + deg F
M, B = constants derived from analytical geometry
The equation is confirmed by experimental data as shown in Fig. 1 for
data at 140 F and above. Equation 1 may be expressed as
In [(log n)/B\ = MlnT = lnTM (2)
and
log rj = BTM (3)
Thus
•q = e2.303BrW (4)
TABLE 1—Asphalts studied.
P25 Susceptibility Parameters

Penetration
Identification 25 C (77 F) -M B/10 1 0
1. R70-50 Arkansas 89 3.454 2.1169

2. R70-51 Air blown 89 3.623 6.0573
3. R70-55 Mexican 90 3.321 0.9277
4. R70-56 Californian A 89
5. R70-58 Californian B 96 3.931 41.612
6. S71-24 Air blown 88
7. S64-47 BPR sample 19 3.528 4.5146
8. S71-28 Roofing asphalt 33 3.494 3.3481
9. S68-2 Special 79 3.574 3.9403
10. S71-10 Canadian A 92 3.400 10.455
11. S71-3 Venezuelan B 91 3.379 1.2987
12. S71-4 Midcontinent 87 3.473 3.0007
13. S71-5 Venezuelan A 300+
14. S71-7 Venezuelan C 91
15. S71-8 Venezuelan A 62
16. S72-2 Canadian B Soft
17. S72-3 Canadian C Soft
18. S71-29 Roofing asphalt 22 3.527 5.0498
This form, it is recognized, desensitizes the data but does relate the results
empirically as well as any other relation.
The differential rate of change of viscosity with temperature is the
derivative of Eq 4:
dv/dT = (2.3035M)7 w - 1 e 2 - 303BrM (5)
which is seen to be a function of both the constants as well as the absolute
temperature.
To evaluate M, the slope of the viscosity-temperature chart, the two-
point formula is used to evaluate both B and M. If y is the left side of Eq 1,
and x = In T:
(y - yd = - — - O - *0 = *t(x - *o (6)
M — X\
and
In B = - M In T + In (log v) (7)
at 7\ = 1:
B = log T/i (8)
The parameters have been evaluated for the asphalts as listed in Table 1.
In Table 2, the effect of small temperature variations at 60 C is given for
three asphalts with high and low susceptibilities. These results show about
a 7 percent variation in viscosity for an 0.5 C (0.9 F) change in tempera-
ture. These numbers provide a quantitative value for the temperature
SCHWEYER ET AL ON RHEOLOGY OF ASPHALT CEMENTS
TABLE 2- -Calculated viscosity variations with small temperature changes at 60 C

(140 F).
Calculated Viscosities
Parameter, M Viscosity-
Asphalt Table 1 degC (deg F) Kilopoises
No. 1 Arkansas -3.454 59.5 (139.1) 2.55

59.9 (139.8) 2.42
60.0 (140) 2.39
60.1 (140.2) 2.36
60.5 (140.9) 2.24
No. 5 Californian -3.931 59.5 (139.1) 1.038
59.9 (139.8) 0.983
60.0 (140) 0.970
60.1 (140.2) 0.957
60.5 (140.9) 0.906
No. 8 Roofing -3.494 59.5 (139.1) 41.3
59.9 (139.8) 38.7
60.0 (140) 38.1
60.1 (140.2) 37.5
60.5 (140.9) 35.2
variation in testing. They also give a quantitative measure of how asphalts

vary by comparing the slope values for M.
Experimental
The experimental work consisted of measurements of absolute viscosities
under varying shear rates and conditions at 60 C (140 F), including some
information at other temperatures.
The apparatus used were the Cannon-Pressure, Cannon-Manning, and
Cannon-Zeitfuchs instruments shown in Fig. 2, with their control panel
and bath arrangement as shown in Fig. 3, and the Florida capillary arrange-
ment [13] as illustrated in Fig. 4.
It should be recognized that there are two basic differences in the modus
operandi of these methods. These differences in flow regimes produce
different rheological results. In the Cannon-Manning instrument the flow is
more or less transient, depending upon the relative size and length of the
flow reservoirs compared with the flow length of the constricted flow tube.
In the Cannon-Pressure instrument the major flow resistance is through an
orifice into a flow tube and reservoir of large diameter. The transient flow
effects here depend upon the relative pressure drops in the two flow sections
(orifice and tube) of the apparatus. In the Florida capillary method, the
transient flow effects are minimized by selection of the apparatus dimen-
sions so that the major portion of the pressure drop occurs in the capillary
tube section.
In all of these apparatus the complex flow patterns where the material
Manning Zeitfuchs Pressure
FIG. 2—Cannon instruments.
FIG. 3—Transient mode apparatus arrangement.
COMPRESSION
Uu
LOAD —
CELL
MOVING
CROSSHEAD "
R M Up—
LOAD
CELL
EXTENSION
SELF
RHEOMETER ALIGNING
PLUNGER SUPPORT
FIXED
SUPPORT
--E }
RHEOHETER
BARREL
ASPHALT
SAMPLE
RHEOMETER
SAMPLE TUBE
FIG. 4—Rheometer assembly schematic.
enters the capillary or orifice result in the material in the center of the
flow path being subject to extension forces, which lead to elongational
viscosity considerations; see Cogswell and Lamb [19].
Results
Comparative viscosity results at 60 C (140 F) on a variety of asphalts
are given in Table 3. The following points are observed and should be
considered in light of the remarks of the preceding section.
In general, there appears to be a trend of higher viscosities at higher
pressure drops in the Cannon-Manning data. This is not true for every
set of data. Data at about 0.70-atm driving force for the Cannon-Manning
and Cannon-Pressure apparatus appear to be in reasonable agreement. In
S5
T j n o o - - N W2 5 ci — <N f> O <*1 Tf O © <
2£
O 00 __ — ^ r ^ © O © 0 0 © C ^
°.T :3^3 oo ^f rf <r
o T 00tSO\O\TinaON
I
.S
w © o" © o —• o — o — — © o d © o © — o
r- wi TT r^ ô — oo (N«o*o • o © ô
gvo — ^ - r n ^ - — — oo * t M O fN O —« m
O
• r^ oo vo • * ro •
• »nCT\~- ro 1^ TJ- • oo r- (N • <N oo i
(N - ^ m ^ ^ F - I M
©
_ " i. .cs
. oo .r- O
_ . r-
*o o cs © r— o> t^
*-*oomo^"000'e3"t^^f>0<NOvr-)
I c
o 3 r-
- 3d
«ri
>
E
C 3 <"1
N N T f - - H ^ t ^
• © OS <S
• oo so -<r
-J
CO
< U 6 o .no.- O r j - H O - n m o O i
3
«i .O
13
wet firs
2
more
E E
o
Z <
P." a." o£
>.
o jy
< a
Ul
•o
1>
8
"V 7 7 7 7 N t N M 2 m Tf in M » M n ?J
°f oc
CGOSfiSQS«C/3Cfl(/5t/3OTt/5C/5WC/3C/3(/5Cflc/3
a
SJ••?
< Q
^HVir!TJ<n^drôdo>d^n'*i^t'o^Jr^x
the Cannon-Pressure instruments there is some indication of shear thinning,

but this is not a generalization because much of the variation may be within
the limits of experimental error. The results are not conclusive, and real
differences are attributed to differences in flow regime (orifice versus
capillary).
The Florida capillary data in Table 3 may show results that are 50 to
100 percent higher than the vacuum method results. There are a few ex-
amples where somewhat similar results are found (Nos. 4, 5, 13, 16, 17).
Most of these are soft asphalts where anomalous flow phenomena are
minimal. In studying the Florida capillary data, consideration must be
given to the fact that the average imposed pressures are of the order of 1000
up to 10 000 psi (70 to 700 atm), which are somewhat higher than the ranges
for the Cannon-Pressure instruments, where the maximum upstream
pressure is only about 10 atm at 150-psi imposed pressure (Table 4).
TABLE 4—P<50 Data on two asphalts, 60 C (140 F)
(Cannon vacuum and Cannon pressure instruments).
Avg. Air Blown Canadian

Nominal Absolute No. 6 No. 10
Pressure Drop, Pressure," Viscosity,6 Viscosity,6 •'
psig atm atm kilopoises kilopoises
CANNON-MANNING VACUUM
VISCOMETERS
-10.5 -0.73 0.64 1.69" 0.694
- 8.8 -0.40 0.80 1.72'' 0.708"
- 2 -0.13 0.98 1.72" 0.748
CANNON-PRESSURE VISCOMETERS
2 0.12 1.06 2.07 1.25
4 0.24 1.12 1.85 1.03
8 0.54 1.27 1.87 0.976
10 0.68 1.34 1.80 0.972
25 1.70 1.85 1.77 0.904
50 3.4 2.7 1.58 0.890''
100 6.8 4.4 1.45 0.907"
150 10.2 6.1 1.28 0.622"
" Arithmetic average of upstream and downstream pressures.

6
All vacuum data are average of two or more runs. All pressure data in a No. 55 orifice
with a No. 28 tube except highest pressure, which used a No. 45 orifice with a No. 88 tube.
* All pressure data in a No. 45 orifice with a No. 124 tube except highest pressure, which
used a No. 32 orifice with a No. 118 tube.
" Average of two or more.
As a matter of interest, certain limited data for measurements at 40 C

(104 F) were obtained in two instruments as given in Table 5. This table
also demonstrates the higher results obtained in the Florida capillary
method compared with those for the orifice-type flow of the Cannon-
Pressure apparatus.
TABLE 5—Comparative viscosity data at 40 C (104 F) in kilopoises.
Florida Capillary
Cannon Viscosity
Pressure Complex -
Flow 0.05 1.0 10' ergs
Asphalts 25 psi Index, C s' 1 s-' per s-cm*
1. R70-50 62.0 0.96 79 74 74

2. R70-51 52.5 0.78 111 60 57
3. R70-55 99.2 0.86 128 89 89
4. R70-56 23.9 1.01 34 35 35
6. S71-24 55.1 0.81 106 62.8 59.8
Discussion
These differences in viscosity results at 60 C (140 F) are not surprising
when the rheological problems are recognized. For example, it is well known
that asphalts exhibit anomalous or complex flow behavior. They may not
follow the power law over all ranges of stress. Furthermore, in the Florida
Asphalt Laboratory, Lodge [20] has demonstrated that the viscosity of
asphalts at 25 C (77 F) increases with pressure, which obviously can affect
the results of the type presented.
If the data are viewed from a knowledge of these effects, the results seem
to be in order. Thus, as the absolute pressure increases for the transient
vacuum method, the viscosity increases slightly.
For the Cannon-Pressure instrument transient data, the decrease in
viscosity with increased pressure drop for the same instrument is in general
attributed to shear thinning effects by anomalous flow. This results in
lower viscosity at increased pressure drops as shown in Table 4 and Fig. 5.
4.0
• NO. 6 A I R BLOWN
3.0 . 0 ^.v,u.~.
VISCOSITY, KILOPOI
•••
-_#___^
^ - *
3)1
£o? A
«"
2 3 4 5
AVERAGE IMPOSED PRESSURE, ATMOSPHERES
FIG. 5— Measurement mode effects on V60, viscosity at 60 C (140 F).

The transient mode gives a generally lower average pressure than that for
a steady-state flow over the same flow path length, which tends to limit the
effect of imposed pressure where the latter causes an increase in observed
viscosity.
Since the data in Table 3 for the Florida capillary were taken at a con-
stant rate of shear, it was considered desirable also to compare results at
constant stress by a new method under development by the first author [21].
Some data are shown in Table 6, where it is observed that the Florida capil-
lary constant stress data at 150-psi upstream gage pressure and the constant
shear rate data are about the same magnitude but differ from the results
TABLE 6—Comparative Florida capillary data at 60 C (140 F) at constant stress

(I atm downstream; all data in kilopoises).
Pressure No. 6 Asphalt No. 10 Asphalt

Conditions ;
6
Florida Capillary Florida Capillary1"
Up-
stream, Average," Cannon Constant Constant Cannon Constant Constant
psig atm Pressure Stress Shear Pressure Stress Shear
10 1.34 1.80 4.95 4.06 0.972 2.46 1.93

25 1.85 1.77 4.65 3.45 0.904 2.43 1.91
150 6.1 1.28 4.54 2.42 0.622 2.31 1.85
" Average of absolute upstream and downstream pressures.

b
All data corrected for wall effect with appropriate values of y read for proper upstream
pressures in the first column.
for the transient mode of the Cannon-Pressure method. Furthermore, it

will be observed that the average absolute pressure for the upstream and
downstream pressures is relatively low (only 6.1 atm at 150 psig).
The data for the Florida capillary method in Table 3 show generally
higher results, which is attributed to the pressure effects since high pressure
drops (maybe up to 10 000 psi) are imposed for the constant shear rate
mode. It is known [20] that increased imposed pressure does increase the
viscosity of asphalts, although they might show essentially Newtonian flow
at smaller pressure drops. At very high shear rates, asphalts sometimes show
evidence of the melt fracture phenomenon that occurs with polymers and
is related to dilatancy (opposite of shear thinning). Some limited data on
this pressure effect are given in Table 7.
The results in the right-hand column of Table 7 compared with the other
data show the great effect of pressure on the viscosity result. This is not an
uncommon finding among other hydrocarbons and polymers. The five- to
tenfold change in viscosity when the material is under a high pressure drop
at different downstream pressures introduces still another transient condi-
tion in the test procedure that will vary with the geometry of the apparatus
and the system under study. Such factors as length and diameter of the
TABLE 7—Effect of pressure on viscosity at 60 C (140 F).
Viscosity, kilopoises
Cannon Florida Capillary at 1 Os" 1

b
Asphalt Manning* Pressure 1 atm c 1 atnr' 400 a t n r
2. R70-51 1.86 1.85 4.04 3.61 23.9

3. R70-55 4.50 3.32 9.60 4.85 32.8
4. R70-56 1.19 1.08 1.48 1.46 16.8
5. R70-58 1.10 1.48 1.80 2.00 20.9
7. S64-47 5.55 6.07 13.4 13.4 99.0
8. S71-28 40.4 42.8 49.4 29.1 160.0
9. S68-2 1.90 1.76 3.16 3.51 19.8
12. S71-4 1.70 2.00 4.06 3.40 13.8
°6 At 30-cm Hg vacuum.
c
At 10 psig.
d
Regular capillary method, one-atm downstream pressure.
High-pressure method extrapolated to a 1-atm downstream pressure.
' High-pressure run at 400-atm downstream pressure, run at constant rate of shear mode
and read at one reciprocal second.
capillary and sample tube, the imposed pressures, the compressibility of the
material, and its composition must be considered. The first author hopes
to work on a few of these problems in the future.
Conclusions
It is important to recognize that differences in asphalt viscosity results
can be obtained in order that nonbiased inferences are made in asphalt
technology. By exploring the basic underlying physical phenomena the
differences can be explained, and from this information correct conclusions
can be drawn.
The problems associated with the results shown in the paper should not
discourage the use of viscosity grading specifications for asphalts. It is
important first to recognize any problems that might exist, to resolve them
with available knowledge, and then to reconcile the results by different
methods. In general, resolution of the problems will be accomplished
readily for asphalt rheology by standardizing the most appropriate appara-
tus and test procedures used.
References
[/] Stanfield, K. E. and Hubbard, R. L., "Asphalts from Rocky Mountain Crude Oils,"
U.S. Bureau of Mines, Technical Paper 717, 1949.
[2] Traxler, R. H., Schweyer, H. E., and Romberg, J. W., Industrial and Engineering
Chemistry, Vol. 36, 1944, p. 823.
[3] Lewis, R. H. and Welborn, J. Y., Public Roads, Vol. 21, 1940, p. 1.
[4] Lewis, R. H. and Halstead, W. J., Public Roads, Vol. 21, 1940, p. 127.
[5] Welborn, J. Y. and Halstead, W. J., Proceedings, Association of Asphalt Paving

Technologists, Vol. 28, 1959, p. 242.
[6] Welborn, J. Y., Halstead, W. J., and Boone, J. G., Proceedings, Association of Asphalt
Paving Technologists, Vol. 29, 1960, p. 216.
[7] Sisko, A. W., Highway Research Record No. 67, 1964.
[8] Welborn, J. Y., Oglio, E. R., and Zenewitz, J. A., Proceedings, Association of Asphalt
[9] Halstead, W. J. and Welborn, J. Y., Highway Research Record No. 350, 1971.
[10] Fink, D. F. and Griffin, R. L. in Papers on Road and Paving Materials and Symposium
on Microviscometry, ASTM STP 309, American Society for Testing and Materials,
1961, p. 79.
[11] Griffith, J. M. and Puzinauskas, V. P. in Fundamental Viscosity of Bituminous
Materials, ASTM STP 328, American Society for Testing Materials, 1962, p. 20.
[12] Lefebvre, J. A. and Robertson, W. D., Proceedings, Canadian Technical Asphalt
Association, Vol. 15, 1970, p. 101.
[13] Schweyer, H. E., Highway Research Record No. 404, 1972, p. 86.
[14] Crouch, R. F. and Cameron, A., Journal of the Institution of Petroleum Technologists,
Vol. 47, 1961, p. 307.
[15] Jongepier, R. and Kuilman, B., Rheologica Acta, Vol. 9, 1970, p. 460.
[16] Davis, B. E., "Viscosity-Temperature Relationships of Certain Bituminous Ma-
terials," M.S. thesis, University of Wyoming, Laramie, Wyo., 1966.
[17] Ling, J. K., "Temperature Dependence of Asphalt Viscosity," M.S. Engineering
thesis, University of Florida, Gainesville, Fla., 1972.
[18] ASTM Standard Viscosity-Temperature Chart for Asphalts D 2493-68, 1971 Annual
Book of ASTM Standards, Part 11, American Society for Testing and Materials.
[19] Cogswell, F. N. and Lamb, P., Plastics and Polymers, Oct. 1970, p. 331.
[20] Lodge, R. W., "Effect of Pressure in Asphalt Viscometry," M.S. Engineering thesis,
University of Florida, Gainesville, Fla., 1971. (A paper is being prepared for publi-
cation.)
[21] Schweyer, H. E., unpublished privileged information.
J. V. Evans^ and J. W. Gorman^
Study of Asphalt Viscosity Testing at 6 0 F
REFERENCE: Evans, J. V. and Gorman, J. W., "Study of Asphalt Viscosity

Testing at 60 F," Viscosity Testing of Asphalt and Experience with Viscosity
Graded Specifications, ASTM STP 532, American Society for Testing and
ABSTRACT: A cooperative study among six laboratories has defined test con-
ditions to minimize diflferences in results obtained by different laboratories using
four different instruments for determining the viscosity of asphalt at 60 F (16 C).
The methods—cone-plate, falling plunger, sliding plate, and sliding plate at con-
trolled rate of shear—are all currently proposed for the Book of ASTM Standards,
Part 11.
The study shows that adhering closely to specified temperatures and times in
sample preparation and using a uniform viscosity loading sequence will decrease
the variability in results among laboratories and methods. A loading sequence
going from low to high gives viscosities about 8 to 9 percent higher than those
obtained with a reverse sequence. These results are the groundwork for a future
study of precision.
KEY WORDS: asphalts, viscosity, viscosity testing, viscometers, bituminous
materials, bituminous concrete, bituminous pavings
In recent years, asphalt technology has emphasized tests for measuring

fundamental mechanical properties of asphalts and asphalt concretes. Such
information should increase understanding of the performance of bitu-
minous materials and improve our ability to engineer asphalt uses and
applications. Interest has focused on specifying asphalt cement viscosities
at 60 F (16 C) and 0.05 s~' shear rate. However, asphalt product property
surveys, sponsored by the Bureau of Public Roads [ly and the Asphalt
Institute [2], showed wide differences in viscosity measurements between
laboratories. For example, results on one asphalt, averaging 47 megapoises,
varied from 31 to 61 as reported by eight laboratories.
Of the several methods for determining low-temperature viscosities of
paving asphalt, the four most widely used are: sliding plate (regular and
with controlled shear), falling plunger, and cone-plate [3]. To minimize
differences between such methods and between laboratories, a task force
' Director, Asphalt Technical Service, American Oil Company, Whiting, Ind. 46394.
- Senior research engineer. Research and Development Department, American Oil Com-
pany, Whiting, Ind. 46394.
' The italic numbers in brackets refer to the list of references appended to this paper.
Downloaded/printed by 16

EVANS AND GORMAN ON VISCOSITY TESTING AT 60 F 17
organized by the Asphalt Institute has conducted a cooperative study in

which six laboratories participated. All of them measured the viscosity of
two commercial paving asphalts by one or more of the instruments. Sample-
handling techniques and shear-loading sequences were specified for all
measurements, and the results were analyzed statistically.
Experimental
The six laboratories and the instruments they used are given in Table 1.
Except for the falling plunger, each instrument was used by two labora-
tories; one laboratory used two instruments. The nominal viscosities of the
two asphalts (designated A and B) are listed in Table 2.
TABLE 1—Laboratories participating and instruments used in study.
Sliding Plate
Cone- Falling Controlled

Laboratory Plate Plunger Shear Regular
American Oil Co. x

(Whiting, Ind. 46394)
Asphalt Institute x x
(College Park, Md. 20740)
Gulf Oil Canada Limited x
(Toronto, Ontario, Canada)
Bureau of Public Roads x
(Washington, D.C. 20235)
Douglas Oil Co. x
(Paramount, Calif. 90723)
Shell Development Co. X
(Emeryville, Calif. 94608)
Each laboratory made seven measurements on each asphalt; three meas-

urements were with the laboratory's normal sample-handling procedure
and shear-loading sequence. Two used a rigorous sample-handling pro-
cedure and an increasing loading sequence; and two involved the same
rigorous sample handling and a decreasing loading sequence. Details of the
procedures and sequences are described in Table 3.
TABLE 2—Nominal viscosities of asphalts tested.
Viscosity
Viscosity, poises at 60 F,
megapoises
Asphalt 275 F 140 F (0.05 s"1)
A 3.64 1265 5.3

B 2.94 1334 53.0
TABLE 3—Sample handling procedures and loading sequences.
Rigorous Handling"
1. Make sure 3-oz tin cover is in place but the lid is readily removable (some portions of
steps, 2, 3, 4, and 5 are concurrent).
2. Place tin in oven maintained at 275 F + 5 F.
3. Keep tin in 275 F oven for 60 min ± 3 min.
4. Place portion of viscometer equipment related to holding the sample in oven at 140 F
+ 5F.
5. Keep equipment in 140 F over for 30 min ± 3 min.
6. Remove equipment from 140 F oven; remove 3-oz tin from 275 F oven; remove lid
from tin and pour sample. This should be accomplished within 3 min and done in a
room at an ambient temperature between 70 and 80 F.
7. Allow equipment to cool in room for 45 ± 3 min.
8. Place equipment in 60 F bath ± 0.02 F using a calibrated thermometer. Allow to
remain in bath before start of testing for 30 min ± 3 min.
9. Start testing (total lapsed time to this point should be 138 ± 9 min).
Loading Sequence
Increasing—Start with low load to yield approximately 1 X 10~3 rate of shear at equi-
librium and record data necessary for viscosity calculation. Increase load to yield five
shear rates over the range 1 X 10 -3 to 1 X 10_1 s~>.
Decreasing—Reverse of above starting with a load to produce a 1 X 10_1 rate of shear
and decreasing load to yield five shear rates over the range 1 X 10~3 to 1 X 10-' s -1 .
Normal
Use your normal laboratory technique in its entirety. Please describe if it differs from
that published in ASTM.
" The intent of this technique is to require as close attention to (1) sample-handling times
and (2) apparatus and sample heating as is compatible with the different instruments.
The seven measurements on each asphalt were run in r a n d o m order at

the rate of one measurement per day. Eight 3-oz samples of each asphalt
were sent to each laboratory. A fresh sample could be used for each
viscosity determination with one spare in case of trouble. The results were
reported on data forms and viscosity graphs as illustrated for a typical
test in Figs. 1 and 2. The original data reports and graphs, as prepared and
submitted by each laboratory, are retained at the American Oil Labora-
tories. Because of the limited testing time available, no effort was made t o
separate the effects of the sample-handling procedures from those of the
loading sequences.
Data and Analysis
Table 4 summarizes the viscosities at 60 F at a shear rate of 0.05 s _ 1 as

reported by each laboratory. The data analysis utilized the logarithms of
the viscosities, which presumes a constant percentage error [4]. The vis-
cosities of the two asphalts differed by a factor of ten; by using logarithms,
their results could be analyzed jointly. A standard deviation, 5, in terms of
logarithms (base 10) multiplied by 230.3 gives, approximately, a coefficient
of variation—a standard deviation expressed as a percentage of the mean.
1. Viscometer Instrument - Cone Plate, Medium Cone
2. Test No. 14
3. Date of Test 2/2/66
4. Sample Material Designation (Circle)
5. Sample Handling Procedure (Circle) ( Rigorous ) Normal
6. Loading Sequence (Circle) ( Increasing) Decreasing Normal
7. Time of placement of sample in 275 F oven 1Q;45

Room temperature a t t h i s time 74.Q F
8. Time of removal of sample from 275 F oven 11:46
Room temperature a t t h i s time 74.2 F
9. Time of placement of equipment in 140 F oven 11:15
10. Time of removal of equipment from 140 F oven 11:45
11. 'Time of completion of pouring sample in equipment 11:49
12. Time of placement of equipment and sample in 60 F bath 12; 34
Room temperature at t h i s time 73.8 F
13. Time when t e s t i n g s t a r t e d 1:04
14. Testing d e t a i l s
Viscometer*
Record in order of t e s t Time Constant Constant Shear
Load Displacement (Sees.) (Viscosity) Viscosity (Shear Rate) Rate
grams degrees dynes/cm
500 1 1576.0 128,910 1.26 x 10° 1.615 .001025
1000 1 749.3 257,910 1.20 x 10 8 1.615 .002156
2000 1 305.4 515,910 9.76 x 10 7 1.615 .005288
6000 1 62.5 1,547,910 5.99 x 10 7 1.615 .02584
16000 6 59.8 4,127,910 2.55 x 10 7 1.615 .1620
15. Time at end of t e s t i n g . 2:50

16. Plot viscosity vs shear r a t e and attach p l o t , i n t e r p o l a t e , and report
viscosity at 0.05 second-*-. 47.0 megapoises at 0.05 sec .
*Shear Stress
FIG. 1—Typical data report.
There were too few laboratories for any one instrument to give reliable
estimates of instrument repeatability and reproducibility. The data are also
really not suitable for an overall variance estimate.
Within-Laboratory Standard Deviations
Table 5 shows the influence of the two sample-handling procedures on
the within-laboratory standard deviations. Generally, the rigorous pro-
cedure is statistically equivalent to or better than the normal. The two
exceptions are the Bureau of Public Roads and the Asphalt Institute.
20 VISCOSITY TESTING O F ASPHALT
oo
c ^fO «"» ©OOOO© ©"">"•» ""><N
~- o o r-- «n — o o oo
— r~* w-i r^ — r- ô
\© r}- v> "3- ^J" Tf ^ "/i Tf »/-) <N m mm
|.S oo" ©"
»c -* fs
" vT *n"
t- r-
m" " oo" o "
ô r*
©~ o " ^ r--" ©"
t— r - oo »n v£> oo m
CO 5 § v> »o in Tfr c i *3" ^fr Tf ^ t oo m <N m *tf-
E Z j rT ©" " v{ " >o" " oo" wf «o"" ©" ©" oo" ©"
Ul »CO t TfM ^D^t
© \0 *N <N ON r- O
o
a Z
C^ »n <o »n TJ- ^v-, W-JTJ- co m CN m -<T
o
d
3 ©»n Q Q O
sO © O
O
^ ">0
rj- »r> >o © © Ov ^«
s
ca
II in TJ- ^
oo"©"
m vo in © r^
^Tf
in v i
" ©"in"
\ o »o^1"
Tj-Tf
©" "
r^ oo
^t-^t
• m ©
TJ-^
r^f <n"
Tfr r t
't 't
©
vo
©"
>n
t*-
\ o m oo r -
m m mmj
©*" ©" m" vd"
fN "*o >n oo
Tjm m m
acfl
3
O CO TtO m in © -rf (N© v-) © O MOO
E © «N m m »n m oo
'tfl "O \ o r ^- ^C © © f s ô
\o v~> in in <n »n TJ- rt^ ^r T *3- ^ n
oo" vT " in* in" rT " ©"o" ©" " o " ^ in'irT
>r> m rN ^ t 0\ m ON
m o\ o r j ON i— ON
in Tf in in T" T 't m m c* TJ- m ^j m
CO
O u cd <u O o
HI?!
OS d
3% * d
o rz;
U
P 13 X5
3
n
2
Ii 3 « o •C«o
(1.
3 CS
urea
£ => =3 3
a ^ O
< < O <55 ca Q < < O < 1/1 cau
g EC g EC
I V
B
.r2
«
CO
X 8
M
> "
«"S
CO ~
J3
u
00
B
3
—
.
H
CO
O
w"8
CO
u ii CO = 3
H 5 -3
a -S -3 3 •3.8
«
a
"?•
oo in
•B.2 a
"?•
oo e a 00 00 00 c
Cone-
Fallin
Slidin
Slidin
Slidin
Slidin
Fallin
(CO
Cone
(co
a
E
1| o
o >
Ut
3
°05 6o
I &r8
.0980
,0147
.0338
,0246
I S 55
OTQ
© o o o
•i
Hi
o >•
oo ^
o
83
•3
I
o 8-* I
QHH
o
a
I
U
•o C
u O •a- r-
c8
T3 ta ^ oo gj
c o 8!
X a1
C/2
2 38 o"
H
a
J
(0
2
I
•S-a
d) O
£
o
•?! CO
73 ^ S: o
U
. o
to a
8|
III 15 O 05
III} r U co
< </3
jl
i o
Q
^
H
SHSIOd 'A1ISOOSTA
Interestingly, these laboratories were just concluding an extensive testing

program [3,4] in conjunction with a national survey of paving asphalt
properties, which included a large number of low-temperature viscosity
measurements. Consequently, their normal procedures were probably so
well practiced that a switch to a new procedure, even though it is more
rigorously defined, could have resulted in greater within-laboratory varia-
tion. The normal procedures varied between laboratories and influenced
the within-laboratory standard deviation as shown in Table 5. However,
these differences (within instrument types) disappeared when both labora-
tories used the rigorous sample-handling procedures. Except for the falling-
plunger test, the within-laboratory standard deviations for the rigorous
procedure do not differ significantly between instruments. It should be kept
in mind that the rigorous procedure was used in conjunction with a spe-
cified loading sequence (increasing or decreasing) which also may have
influenced the within-laboratory error.
The falling-plunger results, which were typically obtained at shear rates
of 0.0003 to 0.004 s _1 , appear deceptively poor in Table 5 because it was
necessary to extrapolate, rather than interpolate, the raw data to obtain
the viscosity at 0.05 s _1 . A reexamination of the individual laboratory
measurement plots showed that had the viscosities been reported at 0.001
s~', the falling-plunger within-laboratory coefficients of variation would
have been 5 to 8 percent, more in line with those for the other viscometers.
This emphasizes the need for specifying that viscosity data be taken in the
immediate neighborhood of the desired shear rate.
The graphs of measured viscosities and shear rates for estimating the
viscosity at the specific shear rate requested were made with a variety of
scales. Also, the number of measurements varied from 5 to 11 per test.
These variations in technique may also have contributed to the variation
in results reported.
Variability Between Laboratories and Instruments

Figure 3, scatter plots of the data gathered by the three "handling-
loading" combinations, illustrates that the rigorous procedures result in
significantly better agreement between laboratory-instrument combina-
tions. Each laboratory-instrument combination appears as a point whose
coordinates are the average values for Sample A (abscissa) and Sample B
(ordinate). A tighter grouping of the points shows better agreement. Again,
the falling-plunger results appear poor because of the abnormal data
extrapolation. There is no significant difference in scatter between the two
loading sequences.
The change from normal to rigorous technique did not seem to change
the relative values obtained by the different laboratories using the cone-
plate and regular sliding-plate viscometers (see Table 6). The sliding-plate
A. NORMAL
SLIDING PLATE (REG) U4
MEGAP OIS
D SHELL
• ASPHALT INST.
C O N E PLATE
0 AMERICAN t
VISCOS
• ASPHALT INST.
SLIDING PLATE (CONT.)

BD
A DOUGLAS
m
A BPR
2
4
FALLING PLUNGER v>
X GULF OIL CANADA

3 4 5 6 7 8
SAMPLE A. VISCOSITY MEGAPOISE
B. RIGOROUS C. RIGOROUS
(INCREASING) (DECREASING)
3 4 5 6 7 8 3 4 5 6 7 8
SAMPLE A. MEGAPOISE SAMPLE A, MEGAPOISE
FIG. 3—Scatter plots of laboratory averages.
(controlled shear) test did not demonstrate this effect. There is an indication
in the data that this effect is sample-related which implies that additional
procedural controls are possible.
TABLE 6—Relative values obtained with cone-plate and

sliding-plate {regular) viscometers.
Handling
Laboratory
Comparison Normal Rigorous
American Oil Co.

Cone-plate 1.177 1.158
Asphalt Institute
Shell Oil Co.
Sliding-plate (regular) 0.870 0.948
Asphalt Institute
Table 7 attempts to further indicate (1) the effects of loading sequence

and (2) the reduction in variance associated with instruments, which can be
attributed to more uniform handling-loading procedures. The averages of
all results, by the different techniques, for each instrument are summarized.
Either of the uniform handling-loading procedures significantly decreases
the range of values obtained by each instrument. The ranges of the means
for the rigorous procedure are \i\ox/i that for the normal procedure. This
indicates a substantial improvement in agreement between the three instru-
ments. Also it should be noted that the two loading sequences result in
different viscosity values. If the initial load is low and loads are increased,
then the viscosity results are 8 to 9 percent higher for either material than
if the load sequence is high to low. The loading procedure significantly
affects the level of the results but not the concordance. The choice of one
loading sequence could be based on agreement with the average for normal
procedures. A decreasing load sequence comes closer to the average of the
normal procedures.
Conclusions and Recommendations

This study shows that concordance of results among laboratories and
instruments can be improved. Before ASTM undertakes further inter-
laboratory studies:
1. It would be highly desirable to provide uniform, stringent sample-

handling procedures prior to testing, either in the current ASTM proce-
dures for low-temperature viscosities, or as a supplementary standardized
technique.
2. A uniform load-application sequence, compatible with all instru-
ments, should be provided. A decreasing sequence would be in accord with
normal practice.
Agreement between instrument types is substantially improved by

bracketing the measurements about the desired shear rate so that the
desired viscosity is an interpolated value. At least one of the measurements
should be close to the desired shear rate. A standard graphical or numerical
interpolation procedure for determining the viscosity at the particular shear
rate should be considered.
Acknowledgments
The authors wish to thank the participants in this program: J. Y. Wel-
born and W. J. Halstead, Bureau of Public Roads; V. P. Puzinauskas,
The Asphalt Institute; H. R. Hawthorne, Gulf Oil Canada Ltd; C. C.
Evans, Shell Development Co.; R. L. Dunning and R. S. Hodgson,
Douglas Oil Co.; and L. C. Brunstrum and A. W. Sisko, American Oil Co.
CO
f l CN
E
o oo e>
o
Z Z
«5
r-- m vo
\c ^ o
8
Q
o
oo
r n o o >o O
VO •* ^ W
CO
E
o O
! z Z
•O m rf
w-> O O
u 3 Q
CO O
<
W Q N O
<N ^ m —
i f l <n <n «~)
g J?
00
c
3
-o
O
-J
•s m
= 3 -o '"S
T3 C3
o O
!i.
Mli
C
CO
. J . -3
X 00 00
e e
a
E CO 00 =.
£ c oo .£
_>
References
[1] Welborn, F. Y., Oglio, E. R., and Zenewitz, J. A., Proceedings, Association of Asphalt
Paving Technologists, Feb. 1966, p. 19.
[2] Puzinauskas, V. P., Proceedings, Association of Asphalt Paving Technologists, Feb.,
1967, p. 489.
[3] ASTM Standards, Part 11, March 1969, pp. 871-896.
[4] Hald, A., Statistical Theory and Engineering Applications, Wiley, New York, 1952,
p. 164.
E. O. Busby^ andL. F. Rader^
Problems in Determining Viscosity of

Asphalts at Low Temperatures with
Shell Sliding Plate Microviscometer
REFERENCE: Busby, E. O. and Rader, L. F., "Problems in Determining Viscosity

of Asphalts at Low Temperatures with Shell Sliding Plate Microviscometer,"
Viscosity Testing of Asphalt and Experience with Viscosity Grade Specifications.
ASTM STP 532, American Society for Testing and Materials, 1973, pp. 28-30.
ABSTRACT: A Shell sliding plate microviscometer with improved steel plates
was used for low-temperature testing of asphalt but was found to be inadequate
for the stresses imposed. There is need for low-temperature viscosity values to
establish properties influential in determining resistance to thermal cracking of
asphalt concrete pavements. Manufacturers should redesign a microviscometer to
withstand the stresses imposed by low-temperature testing.
KEY WORDS: asphaUs, concrete pavements, bituminous concretes, low tempera-
ture tests, microviscometer, instruments, sliding plate microviscometer, thermal
cracking, viscosity, bitumens
At the University of Wisconsin for a number of years, we have been
doing research work on resistance to thermal cracking of asphalt concrete
pavements and on physical properties of asphalt cements and asphalt con-
crete mixtures at low temperatures. We have found the Model 1113A Shell
sliding plate microviscometer to be a satisfactory instrument for the deter-
mination of viscosity of asphalt at 77 F (25 C), but we have experienced
difficulty in determining viscosity of asphalts at 39.2 F (4 C) with this
instrument.
We have followed the procedure in the paper by Evans and Griffin^ for
the modified instrument with a Varian G-10 graphic recorder. Special
stainless steel viscometer plates were purchased from the manufacturer
(Hallikainen Instruments) and they proved to be much more satisfactory
than glass plates. The width of the steel plates has been reduced in order
to decrease the loads, but the length of 30 mm has been retained so that the
maximum allowable displacement would remain unchanged. The steel
plates d o not chip under the higher loads applied in the low-temperature
' Dean of engineering. University of Wisconsin-Platteville, Platteville, Wis. 53818.
2 Professor of civil engineering. University of Wisconsin, Madison, Wis. 53706.
' Evans, C. C. and Griffin, R. L., Proceedings, Association of Asphalt Paving Technolo-
gists, Vol. 32, 1963, p. 64.

BUSBY AND RADER ON PROBLEMS AT LOW TEMPERATURES 29
tests as glass plates do. Film thicknesses of 100, 300, and 500 nm are used
on the steel plates depending on the grade of asphalt. Control of film thick-
ness is set by improved spacers. The procedure of testing is easier than with
glass viscometer plates. Since film thickness can be accurately chosen, it is
not necessary to weigh the plates, and the calculations are simplified. Also,
since the procedure with steel plates eliminates certain steps, operator time
and chance for error are reduced.
The linkage between the movable plate holder and the balance was
modified. The standard model instrument employs a chain connected to a
steel rod. With the large loads required for measurements at low tempera-
tures (for example, 39.2 F (4 C)), unexpected large deviations can occur
owing to deformation of the chain. This was partly corrected by replacing
the chain and rod with a phosphor bronze ribbon obtained from the manu-
facturer. The modified linkage and the modified steel viscometer plates
extend the range and precision of the microviscometer.
Despite the reports of superior repeatability by using the steel plates and
phosphor bronze ribbon, it should be pointed out that loads of over 1000
g required for testing at 39.2 F (4 C) overload the beam far beyond its
designed capability. The spring that holds the fixed viscometer plate is
overstressed and frequently gives way. The microviscometer should be com-
pletely redesigned to give satisfactory results at 39.2 F (4 C) and to with-
stand the stresses imposed on the instrument in testing asphalt cement at
this low temperature.
Range of Microviscometer
Evans and Griffin3 stated that the modified microviscometer has a work-
ing range of 10 to 104 megapoises. In spite of the difficulties just described,
Busby4 has verified this upper working range, obtaining a value of 15 000
megapoises on asphalt cement marked "C" at 39.2 F (4 C) (See Table 1),
but he recommends redesign of the instrument.
Need for Low-Temperature Viscosities

Why determine viscosity of asphalt cement at 39.2 F (4 C)? Our answer
is that viscosity at some low temperature (for example, 39.2 F (4 C)) is
needed (as well as the viscosities at normal and at elevated temperatures)
to establish those low-temperature properties influential in determining the
resistance to thermal cracking of asphalt concrete pavements. This informa-
tion is of importance in designing asphalt concrete pavements in the northern
tier of states in the United States and in Canada.
* Busby, E. O. and Rader, L. F., Proceedings, Canadian Technical Asphalt Association,
Vol. 16, 1971, p. 289.
TABLE 1—Rheological properties of viscosity graded asphalt cements.
Asphalt Cement
Property
Asphalt Designation A B C
Viscosity at 39.2 F (4 C), poises 50 X 10« 970 X 10« 15,000 X 10"

Viscosity at 77 F (25 C), poises 0.3 X 10« 1.1 X 10« 26 X 10«
Viscosity at 140 F (60 C), poises 834 976 886
Viscosity at 275 F (135 C), poises 3.64 2.59 2.19
Penetration index -0.2 -0.8 -1.5
Penetration at 77 F (25 C) 156 99 62
Penetration at 39.2 F (4 C) 34 18
Penetration at 32 F (4 C) 24 15
Ductility at 77 F (25 C), cm J50+ 150+ 150+
Ductility at 39.2 F (4 C), cm 34 0
Viscosity grade AC-10 AC-10 AC-10
For example, Busby4 has reported viscosities at 39.2 F (4 C) and at 77 F

(25 C) on three AC-10 asphalt cements with approximately the same vis-
cosity at 140 F (60 C) but with different penetration indices. See Table 1.
The producers furnished the additional test data. Note that the magnitude
of the viscosity values is reversed between Asphalts A and C for 275 F
(135 C) on the one hand and 39.2 F (4 C) and 77 F (25 C) on the other hand.
This illustrates the need for viscosity values at low temperature.
Summary
It is recommended that consideration be given by the manufacturer to
redesign the sliding plate microviscometer instrument to withstand the
stresses imposed by low-temperature testing. This recommendation is based
on the belief that asphalt paving technologists should be able to determine
viscosity of asphalt cements at a low temperature such as 39.2 F (4 C) in
an accurate and satisfactory manner.
C. F. Potts^
Florida Studies with the Puzinauskas

Field Viscosity Cup for Asphalt Cements
REFERENCE: Potts, C. F., "Florida Studies with the Pnzjnauslcas Field Viscosity
Cup for Asphalt Cements," Viscosity Testing of Asphalt and Experience with
Viscosity Graded Specifications. ASTM STP 532, American Society for Testing
and Materials, 1973, pp. 31-39.
ABSTRACT: An ASTM base chart is shown in the paper to confirm the linearity
of the temperature-viscosity relation forfivedifferent asphalts over the temperature
range of 60 C (140 F) to 195 C (383 F). These data were run in the laboratory
using the procedures in ASTM methods D 2170 and D 2171. All data were cor-
rected to poises for plotting, using densities calculated at the test temperatures.
The viscosity temperature characteristics for different grades of asphalt from
several producers are shown in separate charts. Kinematic viscosities, as measured
with the Puzinauskas-Asphalt Institute (P-AI) field cup, in the laboratory, are
superimposed on these charts to illustrate the agreement of the (P-AI) data with
carefully controlled data from ASTM Methods D 2170 and D 2171.
Field data at actual construction sites were compared with check data rerun on
the field samples in the laboratory for agreement. It was found that for field tests
it is necessary for the cup and the asphalt cement in the cup to be at the test
temperature read on the thermometer. With experienced technicians, the field tests
readily agree with laboratory data.
KEY WORDS: viscosity, asphalts, bituminous cements, penetration, absolute
viscosity, kinematic viscosity, viscometer cup, specifications, bitumens
Viscosity grading specifications for asphalt cements as a replacement for

penetration has led to the desirability of field control of viscosity in place
of penetration.
In implementing the new specifications, several methods for control in
the field are available. A m o n g the most exact would be measurement of the
viscosity ¥60^ (at 140 F) (60 C) in poises. This has disadvantages of time t o
equilibrate the temperature of the sample and the need t o clean the appa-
ratus. The Puzinauskas field viscosity cup (P-AI field) obviates these dis-
advantages t o some extent. It permits sampling directly from the t a n k trans-
port or supply, or from a 5-gal sample can. However, it does require that
' Bituminous materials and research engineer. Office of Materials and Research, Florida
Department of Transportation, Gainesville, Fla. 32601.
2 The code form V60 refers to viscosity at 60 C (140 F) and V135 to viscosity at 135 C
(275 F) for purposes of simplification.

a record of the test temperature be made since the temperature suscep-

tibility may show considerable variation from situation to situation and
from supplier to supplier.
Base Chart
The viscosity grading specification determines an allowable range of
V60, viscosity at 140 F (60 C) in poises, but stipulates only a minimum
V135,2 kinematic viscosity at 275 F (135 C). Accordingly, it is necessary to
establish a band of viscosities at the temperatures at which a field viscosity
will be within acceptable limits. Furthermore, this band may vary for each
supplier. In addition, it is desirable to have a provisional band where final
approval would depend upon verification tests at V60 (140 F) (60 C).
To establish these acceptable bands, it is necessary to develop base data
on typical asphalts such as in Fig. 1. The samples were in Cannon-Manning
vacuum viscometers according to ASTM Method D 2171-66—Absolute
Viscosity of Asphalts— at the lower temperatures and by ASTM Method
D 2170-67—Kinematic Viscosity of Asphalts—for the higher temperatures.
It will be noted that all viscosity data are in poises, which means that the
kinematic viscosity data as measured in stokes at the higher temperatures
have been multiplied by the density at the test temperature. The equation
o. il 1 1 i i i i i i i i 1 1—
140 160 180 200 220 240 260 280 300 320 340 360 380
TEMPERATURE, deg F
FIG. 1—Temperature susceptibility of various asphalt cements.
POTTS ON PUZINAUSKAS FIELD CUP STUDIES 33
for the density relation including the calculation for viscosity is as follows
(see also ASTM method D 2170-67):
Viscosity in poises = (kinematic viscosity in stokes)
(P - 0.61 AT X 10-3)
(1)
where
p = density at a known temperature and
AT1 = incremental temperature difference in degrees Celsius above
temperature at which the density is known. Celsius is absolute
temperature on a centigrade scale = 273 + deg C.
For example, if p is 1.0147 g/cm3 at 25 C (77 F), and the test temperature
is 135 C (275 F), then 135 C = 408 deg Celsius. For a kinematic viscosity
often stokes, the viscosity in poises becomes (since AT = 408 — 298 = 110
deg Celsius):
V135 = 10(1.0147 - 0.61 X 110 X lO"3) = 10 X 0.948
= 9.48 poises
For any asphalt, if the density at some temperature is known, it is substi-
tuted in Eq 1 and the appropriate temperature increment used. Where
the density of the asphalt is not known, a value of 1.0147 g/cm3 at 25 C
(77 F) may be assumed. For this assumed density at 25 C (77 F), correction
factors for density at other temperatures may be taken from Table 1 or Fig. 2,
_i i i f-K i i i i i i i
150 175 200 225 250 275 300 325 350 375
TEMPERATURE, deg F
FIG. 2—Density-calibration plots for asphalt cements.
which agree with the procedure of ASTM D 2170-67. The last column is
the correction density factor for an assumed value of 1.0000 g/cm3 at 25 C
(77 F), which may be of some use.
TABLE 1—Correction factors and densities for assumed conditions.
Corrected Density, g/cm'
Correction Base: 1.0147 Base: 1.0000
25 C ( 77 F) 0.0000 1.0147 1.0000

60 C (140 F) 0.0214 0.9933 0.9786
80 C (176 F) 0.0336 0.9811 0.9664
100 C (212 F) 0.0458 0.9689 0.9542
135 C (275 F) 0.0671 0.9476 0.9329
150 C (302 F) 0.0762 0.9385 0.9238
175 C (347 F) 0.0915 0.9232 0.9085
195 C (383 F) 0.1017 0.9110 0.8963
Puzinauskas-Asphalt Institute Field Viscosity Cup (P-AI)

The P-AI field viscosity cup consists of a special brass cup containing a
tubular discharge which extends into the cup. The cup is surrounded by an
annular holding space with a special handle. The instrument is shown in
Fig. 3 and is available from the Cannon-Instrument Company, State
College, Pennsylvania.
(a) Side view of cup. (6) In operation.

FIG. 3—Photographs of Puzinauskas viscosity cup.
POnS ON PUZINAUSKAS FIELD CUP STUDIES 35
The instrument is dipped into any suitable container of asphalt (trans-

port, tank, tank car, drum, or can) and allowed to drain, and is dumped
several times to bring the cup to the temperature of the asphalt and also to
clean the cup walls of any residual material left from a previous sample.
For measurement, a thermometer is placed in the asphalt container and
the temperature read during the measurement of efflux time. To make a
measurement, the preconditioned cup is dipped below the surface and then
withdrawn with a smooth vertical motion. When the cup breaks the surface
of the asphalt in the container, a stopwatch is started. The watch is stopped
when the top of the tubular outlet becomes visible as the asphalt drains
from the cup.
From the measured efflux time, the kinematic viscosity is computed by
multiplying by the viscosity cup constant. The absolute viscosity to use in
Fig. 1 is calculated by the appropriate factor from Eq 1 or Table 1. (Note
that both factors can be combined into a single constant where measure-
ments are made over a small temperature range, say ten degrees, without
greatly affecting the results.)
Calculations
The P-AI field cup as received has a precalibrated constant that, when
multiplied by the efflux time, gives the kinematic viscosity in centistokes
or stokes. However, the desired results are in absolute units or poises, which
requires understanding the basis relations that are readily clarified.
If the instrument constant is C, in stokes per second, then the absolute
viscosity is:
r\p = C,pt, poises (2)
where
t = efflux time, s and
p = density at temperature T, g/cm3.
For a limited temperature range where an average density, pr, may be
considered constant, a combined constant for absolute viscosity, Cp, in
poises per second, may be used so that
•tjj, = Cpt, poises (3)
where the absolute viscosity constant Cp = C,pT- Thus, efflux times are
converted directly to absolute viscosity (in a given temperature range) by
multiplying the time by the C„ constant. (Note: 1 poise = 100 centipoises
and 1 stoke = 100 centistokes.)
The operator measures the efflux time for an "on-site" sample and
multiplies the time by the appropriate constant depending upon the data
desired at the measured temperature. This provides the operator with a
measure of the viscosity and whether or not it has a high probability of
meeting specifications. In making measurements it is assumed that a proper
technique is used to ensure that the cup is at temperature and that the
measured temperature is correct for the material in the cup at the time of
test.
Experimental Data
Data obtained on a 5-gal can sample heated in an oven and tested peri-
odically over a range of temperatures are shown in Fig. 4 for asphalts of
two different grades. Experimental laboratory data were obtained in
Cannon viscometers by ASTM Methods D 2170 and D 2171 over a tem-
perature range of 140 F (60 C) up to 383 F (195 C). Data obtained with
the P-AI cup are superimposed on these lines and show excellent agreement
except for several outliers. In all cases, the outliers are explained by in-
sufficient temperature equilibration time or inexperienced operators. The
lines shown in Fig. 4 were arbitrarily drawn parallel through the points
and essentially define the slope to be expected for this source. It is necessary
to know this expected slope from any supplier in order that data at high
temperature in the field may be extrapolated to 140 F (60 C).
Results for one set of field test data are shown in Fig. 5. Good agreement
of the field results with laboratory results on a retained sample are shown.
The outlier shown at the high temperature is attributed to some test error.
The outliers shown at the lower temperature are attributed to cooling of
the sample and viscometer cup during the longer flow times at these tem-
oi I i i i i i i 1 1 i 1 1 1
140 160 180 200 220 240 260 280 300 320 340 360 380 400
TEMPERATURE, deg.F
FIG. 4—Laboratory data with P-AI viscosity cup.
l(ft .
02 -
0,1 I I 1 I I ' i ' ' ' I I I

140 160 180 200 220 240 260 280 300 320 340 360 380 400
TEMPERATURE, deg F
FIG. 5—Field and laboratory data on an asphalt cement.
peratures. It is recommended that the test be run at temperatures above

275 F (135 C) in the field to avoid these effects.
Acceptability Diagram
Based on the above data and other information, a proposed acceptability
basis diagram has been drawn up for one grade (AC-20) of asphalt. The
control chart is in the exploratory stage and must be considered only in a
preliminary manner as demonstrating how field control can be exercised by
use of the P-AI instrument. Any acceptability diagram will vary with the
specification ranges and the suppliers' materials. Accordingly, they must
be designed separately depending upon individual requirements.
A test diagram is shown in Fig. 6 for various situations assuming that a
top limit for V135, viscosity at 275 F (135 C), is twice the minimum of 3.00
stokes (2.84 poises) or 5.68 poises as shown for an AC-20 where V60,
viscosity at 140 F (60 C), is 2000 ± 400 poises.
Line A is for minimum limits, which establishes one temperature suscep-
tibility value (the rate of change in viscosity with temperature) that is
evaluated from the slope of Line A. Line B is drawn as the maximum limits
for the assumed maximum V135 viscosity (275 F) (135 C). Line C is parallel
to Line A and drawn through the maximum of 2400 poises at 140 F. Line D
is drawn through minimum viscosity V135 (275 F) (135 C) and maximum
allowable viscosity at 140 F (V60). It represents a maximum slope. The
line for E was drawn through the minimum V60 (140 F) (60 C) and the
200 220 240 260 280 300 340

TEMPERATURE, deg-F
FIG. 6—Selected range limits for AC-20 asphalts.
maximum assumed V135 (275 F) (135 C) and represents the minimum

slope situation.
Using these viscosity ranges, two lines are drawn in Fig. 7 on expanded
scales. The lower line is the maximum allowable slope drawn through 2400
poises at 60 C (140 F) (60 C). The upper line is the minimum allowable
slope drawn through 1600 poises at 60 C (140 F). The area between the
lines delineates the acceptable area for viscosity measured at the desired
field temperatures. By using this blown-up area in the range where the
P-AI cup is being used for plotting the field measurements as shown in
Fig. 7, it is then possible to evaluate compliance with specifications by
immediate inspection.
It should be emphasized that the procedure described is only a suggestion
based on one set of conditions. For example, a wider range of area suscep-
tibility might be employed for tentative acceptance, with a penalty arrange-
ment where a retained sample did not meet specifications after laboratory
checking.
With these guidelines it would be possible to select acceptability criteria
for three situations such as the following:
Case I. Approved: Based on at least five successive acceptable tests and
a known slope for the supplier or the asphalt source. In general these
would be expected to fall in some middle range of the central area of Fig. 7.
Case II. Conditionally Approved: Based on five successive acceptable
tests and an unestablished slope, or any point outside some acceptable
^^
^
^ ^ ^
1.0
0.8 ~
^v. LINE E.MIN. SLOPE
o.e
0.4
0.3 - L I N E D , MAX.SLOPE
1 1 1 1 1 1 1 1 1 1 I I
0.2
270 280 290 300 310 320 330 340 350 360 370 380 390 400
TEMPERATURE, d . g F
FIG. 7—Control chart for AC-20 asphalts.
central area of Fig. 7, as established for Case I, but not outside the total
area of Fig. 7.
Case III. Rejection: Any confirmed set of multiple points that fall out-
side of the total area would be sufficient cause for rejection.
Acknowledgment
Some portions of the results reported in this paper were obtained under
a research contract sponsored by the Florida Department of Transportation
at the University of Florida's Engineering and Industrial Experiment
Station with H. E. Schweyer as principal investigator.
L. W. Corbett^ and H. E. Schweyer^
Viscosity Characterization of Asphalt Cement
REFERENCE: Corbett, L. W. and Schweyer, H. E., "Viscosity Characterization

of Asphalt Cement," Viscosity Testing of Asphalt and Experience with Viscosity
ABSTRACT: There exists an inherent relationship between the commonly used
consistency tests (viscosity at 140 F (60 C), viscosity at 275 F (135C), and pene-
tration at 77 F (25 C)) as applied to asphalt cements used in paving. Based on
data accumulated from 84 worldwide crude sources of asphalt, it is shown that
the viscosity at 140 F is related to each of the other two tests by power law
logarithmic function, and that the overall relationship is represented by a nesting
of lines. With the viscosity at 140 F as the independent variable, there exist
significant differences in the penetration (77 F) and viscosity (275 F) levels but
with only minor differences in the slope of the lines. Thus, the relationships
between this viscosity and the other two consistency tests serve to identify the
viscosity grade as well as to indicate the source of the asphalt, which in turn
relates to temperature susceptibility properties.
KEY WORDS: viscosity, bituminous cements, bitumens, viscosity tests, asphalts,

kinematic viscosity, absolute viscosity, penetration
There are numerous methods and relationships by which both the con-
sistency level and the general rheological properties of an asphalt have
been measured, each of which has served a useful purpose. With increasing
use today of standard methods for determining absolute viscosity at 140 F
(60 C) [If and kinematic viscosity at 275 F (135 C) [2], there are better
opportunities for defining the engineering properties of asphalt cement
binders [3,4,5]. Perhaps one of the biggest contributions that viscosity
testing has made is that it is revealing rheological properties that previously
were hidden under the empiricism of the older methods. As a result, new
efforts are being made to grade on viscosity as well as to define related
rheological properties using high- and low-temperature viscosity and pene-
tration [6].
It is the objective of this paper to point out some of the inherent rela-
tionships involving three of the common standard consistency tests being
' Research associate, Esso Research and Engineering Co., Linden, N. J. 07036.
^ Research professor. Department of Chemical Engineering, University of Florida,
Gainesville, Fla. 32601.
' The italic numbers in brackets refer to the list of references appended to this paper.
Copyright® 1973 b y A S T M International www.astm.org

CORBETT AND SCHWEYER ON ASPHALT CEMENT 41
applied today, namely, viscosity at 140 F, viscosity at 275 F, and penetra-

tion at 77 F (25 C). To support this, data on asphalts from a large number
of sources have been accumulated to work out, both graphically and by
equation, the relationships between these two viscosity tests and the pene-
tration test. It is a further objective of this paper to bring out the normal
variation in these test properties, as attributed to both consistency level and
the source of the asphalt, and to suggest means by which specification
quality of an asphalt cement may be more realistically defined.
Examination of Typical Asphalt Sources

As a basis for this work, data were accumulated on asphalt cements
derived from 84 worldwide crude sources (see Table 1). This represents a
comprehensive cross section of sources, most of which are commonly used
in the United States. From each of these sources, a series of 3 to 7 asphalt
cements was obtained by straight distillation or air-blowing. On each of
these asphalt cements the viscosities (140 F and 275 F) and penetration (77
F) were determined. Of the total sources involved, 63 were straight reduced
from a single crude source, 16 were straight reduced from crude or base
stock blends, and 5 were prepared by mild air-blowing treatment. Each
series of asphalt cements from a single source was plotted linearly on
log-log coordinates.
Viscosity at 140 F Versus Penetration at 77 F

Several other authors [7,8,9] have reported the use of the linear relation-
ship between the log of viscosity (140 F) and the log of penetration (77 F),
especially as pertains to a series of asphalts from a single source. This
relation may be written as follows:
log (P/Px) = Mlog(V/V0 (1)
where
P = penetration (77 F), decimillimeters,
V = viscosity (140 F), poises corresponding to P,
Px = penetration (77 F) at another level,
Vi = viscosity (140 F), poises corresponding to Pu and
M = slope of the log-log plot = A log V/A log P.
Using Eq 1, the best line was fitted to each of the 84 different series of
asphalt cements using a least-squares regression. For each of the 84 series
of asphalts the following was determined:
1. Penetration at a reference viscosity level of 500 poises.
4. Viscosity at a reference penetration level of 90 decimillimeters.
5. Slope factor M of Eq 1.
TABLE 1—Calculated values of penetration intercept, viscosity intercept,

and slope M for viscosity (140 F) versus penetration (77 F).
Asphalt Pen @ Pen @ Pen @ Vis @

Number V-M KJOOO K«>oo Pn Slope M
1 195 90 61 1993 -0.59

2 200 89 60 1974 -0.58
3 174 75 49 1470 -0.61
4 165 70 45 1324 -0.62
5 233 82 49 1764 -0.75
6 214 83 51 1767 -0.69
7 207 84 54 1811 -0.65
8 210 79 48 1654 -0.71
9 228 81 48 1730 -0.75
10 96 33 19 542 -0.78
11 194 86 57 1845 -0.59
12 216 90 58 2013 -0.63
13 130 60 41 971 -0.56
14 224 92 58 2056 -0.65
15 96 33 19 542 -0.78
16 144 49 29 914 -0.78
17 204 91 60 2021 -0.59
18 137 48 28 873 -0.76
19 123 41 24 742 -0.79
20 119 39 22 709 -0.80
21 122 45 27 764 -0.73
22 116 43 27 716 -0.71
23 182 99 73 2503 -0.64
24 142 47 27 886 -0.79
25 190 72 44 1451 -0.70
26 191 68 41 1377 -0.74
27 208 90 59 1997 -0.60
28 204 90 60 2017 -0.59
29 208 77 47 1619 -0.71
30 197 78 50 1625 -0.66
31 199 77 47 1581 -0.69
32 256 109 72 2755 -0.61
33 210 98 67 2345 -0.58
34 147 63 42 1121 -0.61
35 170 78 53 1560 -0.56
36 174 67 42 1310 -0.69
37 192 85 57 1830 -0.58
38 95 34 20 539 -0.74
39 285 126 83 3523 -0.59
40 185 70 43 1393 -0.71
41 153 64 41 1159 -0.64
42 199 71 42 1453 -0.74
43 167 63 39 1199 -0.70
44 153 71 52 1706 -0.60
45 226 96 62 2216 -0.62
46 217 78 47 1652 -0.74
47 235 92 57 2051 -0.68
48 165 74 49 1415 -0.58
49 233 100 66 2380 -0.61
50 173 68 43 1326 -0.67
51 112 51 34 729 -0.57
52 165 55 32 1073 -0.79
TABLE 1—Continued.
Asphalt Pen @ Pen @ Pen @ Vis®

Number ^500 K2000 K4000 P90 Slope M
53 174 57 33 1143 -0.80

54 201 77 48 1593 -0.69
55 234 82 49 1780 -0.75
57 217 82 50 1644 -0.64
58 193 79 51 1338 -0.78
59 196 66 38 1259 -0.72
60 176 64 39 1161 -0.80
61 177 58 33 914 -0.76
62 177 62 37 2281 -0.78
63 144 49 29 2325 -0.77
64 210 97 66 2375 -0.58
65 195 97 68 745 -0.56
66 231 100 66 2375 -0.61
67 115 49 32 745 -0.62
68 202 92 62 2069 -0.57
69 174 65 40 1276 -0.71
70 237 92 58 2075 -0.68
71 155 60 38 1110 -0.69
72 118 38 214 692 -0.81
73 155 54 315 1018 -0.77
74 154 62 39 1138 -0.66
75 248 98 62 2277 -0.67
76 200 82 52 1717 -0.65
77 156 50 28 976 -0.82
78 183 67 41 1337 -0.72
79 188 77 50 1582 -0.64
80 191 64 37 1294 -0.79
81 182 75 48 1504 -0.64
82 234 96 62 2216 -0.64
83 151 55 34 1020 -0.72
84 168 58 35 1139 -0.76
NOTE—Pen @ F500 = penetration at 77 F (100/5) at viscosity (140 F) level of 500 poises.

Vis @ P90 = Viscosity (140 F) in poises at penetration level of 90 decimilli-
meters.
Slope M = defined by Eq 1.
The data on a typical series of asphalt cements separated by distillation

from a single source illustrate the linear relationship between log viscosity
(140 F) and log penetration (77 F), see Fig. 1. Because of the growing
interest in viscosity grading, the viscosity (140 F) logically becomes the
independent variable and therefore is shown on the abscissa for graphical
representation. Penetration (77 F), as the dependent variable, is then
plotted on the ordinate, illustrating the sensitivity of this variable as
dependent on viscosity.
Figure 2 shows a nesting of the least-squares regression lines for 18 typical
asphalt sources within the high and low lines which represent the essential
—T 1 r 1 1 1 r- 1
1
500
400 - "
300 -
V.
K 200 ^v -
o
o
U_
-
•
<
o 100 - -
< \
£ 80
UJ
Z
•v.
UJ
a.
60
40
V
-
30 - -
20
1- -I 1 1 '
200 300 400 600 8 0 0 1000 1500 2000 3000 4 0 0 0 5000
VISCOSITY AT 140 F, poises
FIG. 1—From a single source the relationship between viscosity at 140 F (60 C) and
penetration at 77 F (25 C) may be defined by log-log straight line.
outside limits for this relationship. These observed limits are the lines for
maximum and minimum penetration levels (asphalt Nos. 39 and 10) with
their corresponding slopes. Each of the 84 asphalt series can be plotted
within these outside lines. There also appears to be some degree of uni-
formity in slope (varying from 0.56 to 0.82) with a very slight convergence
of all lines in the direction of lower viscosity. In general, it is noted that
asphalt sources yielding cements with higher penetrations have the lowest
slope factors, and those yielding lower penetration cements have the highest
slope factors.
Figure 3 was prepared to illustrate further the relationships of penetra-
tion with viscosity. For a given source and given viscosity, the penetration
value for that asphalt cement can be calculated from Eq 1, or it can be
determined by plotting the line and determining the penetration graphically.
200 300 400 600 800 1000 1500 2000 3000 4000 5000
FIG. 2—Most asphalt cements fall within outside parameters of viscosity at 140 F (60 C)
versus penetration at 77 F (25 C).
Identifying Grade While Defining Character

Specifications for viscosity-graded asphalt cement are being written
today [16] by specifying a flat minimum penetration for each grade, as
illustrated by the block areas in Fig. 4. The inconsistency of this flat mini-
mum control limit is apparent when one considers the actual observed slope
relationships between the two variables, viscosity (140 F) and penetration
(77 F), which have just been explained. As illustrated by the broken lines
in Fig. 4, an asphalt cement from source A is acceptable only within a
viscosity range of 800 and 1000 poises for an AC-10, and an asphalt cement
from source B is acceptable only within a viscosity range between 1600 and
2000 poises for an AC-20, according to current specifications. However,
published literature [12,13] show that if an asphalt cement derived from
either source A or B is satisfactory for the lower half of the viscosity range,
I . , , , I , , , , ]
200 300 400 600 800 1000 1500 2000 3000 4000 5000
FIG. 3—Essential parameters of viscosity at 140 F (60 C) versus penetration at 77 F

{25 C) are defined by intercept and slope.
it should also be satisfactory for the upper half of the viscosity range as
well, since only its viscosity level has been changed. Because the viscosity
(140 F) in conjunction with penetration (77 F) largely defines the viscosity-
temperature or the stiffness-temperature properties, both the grade and the
temperature susceptibility of asphalt cement binder can be covered by this
single simple relationship.
These considerations can lead to a better description of asphalt cement
by defining the minimum penetration as a function of the variation of
viscosity level, as illustrated in Fig. 5. This would provide for realistic
variation in penetration as dependent on the viscosity level, thus identifying
the grade while more accurately defining its viscosity-penetration relation-
ship. For the manufacturer or purchaser to determine compliance with
specification would require only a single test for viscosity and a single test
1 1 T 1 1 1 1 I 1
AC-2|
*— Range*-
FIG. 4—Some specification requirements are not consistent with normal characteristics of
asphalt cement.
for penetration on any one asphalt cement in a given grade and from a
given source. The point that is stressed here, however, is that minimal values
of a dependent variable (penetration) should be related to an identified
level of the independent variable (viscosity) in order to be consistent with
normal viscosity-penetration characteristics. This can be accomplished by
specifying the minimum penetration at the midviscosity level of each grade,
which essentially governs the penetration at both upper and lower limits
for the grade. For example, the specification for an AC-10 asphalt cement
could read, "Penetration (77 F) shall be a minimum of 70 at the viscosity
(140 F) level of 1000 poises."
Several asphalt temperature-susceptibility characteristics can be deter-
mined when computing from a basic reference level; for example, viscosity
at the 90 penetration level or the penetration at the 2000 viscosity level.
AC-21
••— Range*.
I . 1 1 1 I I . 1 1 1——J
200 300 400 600 800 1000 1500 2000 3000 4000 5000
FIG. 5—Penetration when identified with viscosity level is more consistent with normal
characteristics of asphalt cement.
As shown in Table 2, these intercept factors appear to be more significant

than slope factors. With the typical data given in the table, it is possible to
characterize a given asphalt cement relative to the high or low viscosity
limits, as judged by either viscosity (140 F) at 90 penetration or penetration
at 2000 viscosity (140 F).
TABLE 2—Typical intercept and slope factors for high- and low-viscosity asphalts.
Vis (140 F) at Pen at Slope

Pen 90 Vis 2000 M
Highest-viscosity asphalt 3520 130 -0.59

Average of all data 1400 72 -0.67
Lowest-viscosity asphalt 550 33 —0.78
Viscosity at 275 F as Dependent Upon Viscosity at 140 F

When these two parameters for a given source are plotted on a log-log
chart, a linear relationship also results (Fig. 6). This means that a similar
equation may be used, changing only the symbols as follows:
log (K/K) = M'los(V/Vd (2)
where V and Vi are viscosities at 140 F at two levels, and K and Ki are
corresponding viscosities at 275 F. Because viscosities at 275 F are deter-
mined in centistokes, it may be desired to convert them to centipoises using
the ASTM factor of 0.948 X KVcSt = DVcp [2]. This would have both
viscosity scales in the same basic units, although the differences are hardly
significant and not enough to see graphically. A computer calculation simi-
lar to that used before, and applied to the same series of asphalt cements
described previously, can be used to determine:
1. Viscosity (275 F) at reference level of 500 poises (140 F).
T t
200 300 400 500 600 800 1000 1500 2000 3000 4000 5000
FIG. 6—Relationship between viscosity at 140 F (60 C) and viscosity at 275 F (135 C) may
also serve in the characterization of asphalt cement.

4. Viscosity (140 F) at reference level of 300 centistokes (275 F).
5. Slope factor M'.
Again, when plotting the lines as in Fig. 6, there is found a similar nesting
of lines, but in this case the higher-viscosity asphalts have higher slope
values and the lower-viscosity asphalts have lower slope values. The range
of characteristics as shown in Table 3 again suggests that the intercept
characteristic is of more significance than the slope factors.
TABLE 3—Typical intercept and slope factors for high- and low-viscosity asphalts
when based on viscosity-viscosity relationships.
Vis (140 F) at Vis (275 F) at

Vis 300 (275 F) Vis 2000 (140 F) Slope M>
Highest-viscosity asphalt 720 490 0.47

Average of all data 1310 340 0.45
Lowest-viscosity asphalt 2850 250 0.39
The viscosity (140 F) at the 300 centistokes level (275 F) as well as the
viscosity (275 F) at the 2000 poise level (140 F) both appear to have possi-
bilities for identifying temperature-susceptibility qualities in the same
fashion, as when dealing with viscosity (140 F) and penetration (77 F). Of
these two, the viscosity (275 F) at the 2000 poise (140 F) level, or at any
other grade level, appears to best satisfy the need for identifying the grade
while describing the viscosity-susceptibility properties.
Summary
1. The relationship between viscosity (140 F) and penetration (77 F) as
well as between viscosity (140 F) and viscosity (275 F) can be described
by power-law logarithmic functions when applied to asphalt cements of
varying consistency level derived from a single source.
2. Asphalt cements from each source therefore can be characterized by
penetration intercept and slope factors using these functions. When con-
sidering asphalt cements prepared from 84 sources, it is found that the
overall relationship of viscosity to penetration is represented by a nesting
of lines with significant differences in penetration level but with only minor
differences in slope.
3. Because the relationship between viscosity (140 F) and penetration
(77 F) is characteristic of the source of an asphalt cement, these two tests
can be used to identify the viscosity grade while indicating temperature
susceptibility as measured by the penetration.
4. The viscosity (275 F) at a given viscosity (140 F) level also appears
to have possibilities for identifying the viscosity grade, as above, while
describing the viscosity-susceptibility properties as related to the source
of the asphalt cement.
References
[I] ASTM Standard Method D 2171-66.
[2] ASTM Standard Method D 2170-67.
[3] Fundamental Viscosity of Bituminous Materials, ASTM STP 328, American Society
for Testing and Materials, 1962.
[4] Grading of Paving Asphalts by Viscosity at 140 F Versus Penetration at 77 F, ASTM
STP 424, American Society for Testing and Materials, 1966.
[5] Highway Research Record, No. 350, 1971.
[6] AASHO Viscosity Graded Asphalt Cement Specification M 226-70, American Asso-
ciation of State Highway Officials.
[7] Halstead, W. J. and Welborn, J. Y., Highway Research Record, Vol. 350, 1971.
[8] Puzinauskas, V. P., Proceedings, Association of Asphalt Paving Technologists, Vol. 36,
1967, p. 489.
[10] Corbett, L. W., Proceedings, Association of Asphalt Paving Technologists, Vol. 39,
1970, p. 481.
[11] LeFebvre, J. A., Proceedings, Association of Asphalt Paving Technologists, Vol. 39,
1970, p. 443.
[12] Fromm, H. J. and Phang, W. A., Highway Research Record, Vol. 350, 1971, p. 30.
[13] Burgess, R. A., Kopvillem, O., and Young, F. D., Proceedings, Association of Asphalt
[14] Duthie, J. L., Proceedings, Association of Asphalt Paving Technologists, Vol. 41,1972.
[15] Christison, J. T., Murray, D. W., and Anderson, K. O., Proceedings, Association of
Asphalt Paving Technologists, Vol. 41, 1972.
[16] New Jersey, New York, and New England States Testing Engineers Association modi-
fication of AASHO Viscosity Graded Asphalt Cement Specification M 226-70.
R. J. Schmidt^
The Rolling Thin-Film Circulating Oven—

An Improved Rolling Thin-Film Oven Test
REFERENCE: Schmidt, R. J., "The Rolling Thin-Film Circulating Oven—An

Improved Rolling Thin-Film Oven Test," Viscosity Testing of Asphalt and Experi-
ence with Viscosity Graded Specifications, ASTM STP 532, American Society for
Testing and Materials, 1973, pp. 52-65.
ABSTRACT: The thin-film oven test (TFO), ASTM D 1754-69, is well established
and functional; however, it requires 5 h exposure time. The rolling thin-film oven
test (RTFO), ASTM D 2872-70, is also functional. Its precision is better, and it is
faster to do. Unfortunately, the two exposures differ enough so they should not be
used interchangeably without separate specification limits for each test. A good
solution to this dilemma would be another oven exposure which has the advan-
tages of the RTFO (that is short exposure time and high precision) and which, at
the same time, affects asphalts in the same way as the TFO. This has been sub-
stantially accomplised by converting the RTF oven from a convection oven to a
forced-draft oven. The added circulation causes more volatile loss to occur. This
rolling thin-film circulation (RTF-C) oven test not only gives exposures very
close to the well-accepted TFO; but it is just as fast as the RTFO to perform. It is
suggested that this new oven exposure be considered as a replacement for the
RTFO and interchangeable with the TFO in specifications.
KEY WORDS: asphalt, oxidation, specifications, durability, thin films, bitu-

mens, viscosity, viscosity tests
Many different methods for determining the loss on heating of asphalts

were used before 1911 when the A S T M issued a provisional method for
determining " T h e Loss on Heating of Oils and Asphaltic C o m p o u n d s . "
In 1916 this test was made the standard test method, A S T M D 6-16. The
present version is A S T M D 6-67.
Lewis and Welborn [7],^ in 1940, proposed a thin-film oven test (TFO)
for asphalts. They demonstrated that the T F O exposure (now A S T M Test
for Effect of Heat and Air on Asphaltic Materials, (D 1754-69)) was closely
related t o the average hot-mix hardening experienced in practice; whereas,
the A S T M D 6 loss on heating test was not a good measure of this effect.
Since 1940, others have supported not only the relationship of the T F O
hardening to hot-mix hardening but also to a limited extent to long-term
' Senior research associate. Chevron Research Company, Richmond, Calif. 94802.
^ The italic numbers in brackets refer to the list of references appended to this paper.

SCHMIDT ON A ROLLING THIN-FILM OVEN TEST 53
hardening in pavements [2-6]. Although the TFO exposure is known to be

reasonably functional, it requires 5-h oven exposure.
An improved thin-film aging test (now ASTM Test for Effect of Heat and
Air on a Moving Film of Asphalt, (D 2872-70)) was proposed in 1959 by
the California Division of Highways [7-9]. It required only 85-min oven
exposure and had significantly better precision than the TFO [10]. Correla-
tion of this rolling thin-film oven (RTFO) test with hot-mix hardening and
to long-term asphalt durability was shown by the California Division of
Highways [7-9] and by Santucci and Schmidt [6,11], respectively. The
latter authors [12] also related the properties of RTFO exposed asphalts
to the setting rate of asphalt concrete.
Several of these studies compared the TFO and RTFO exposures. Al-
though neither of these tests was suggested as being more functional than
the other, one study demonstrated the greater precision of the RTFO.
Also, the RTFO requires only 85 min exposure, whereas the TFO requires
5 h. Because of these advantages, the Uniform Pacific Coast Paving
Asphalt Specifications were modified in 1967 to permit the alternate use
of either the TFO or RTFO exposures. The same specification limits were
retained for either test. In addition, the Federal Highway Administration
considers the two tests as acceptable alternates [13].
Present Situation
A recent cooperative study by Pacific Coast Asphalt Producers indicated
these two oven exposures to be less equivalent than previously believed. In
this study, each cooperator determined the effect of TFO and RTFO ex-
posures on his own asphalts. Some of the findings of this study are shown in
Figs. 1 and 2. The TFO weight loss, as shown in Fig. 1, is considerably
greater than the RTFO weight loss. This trend is also found in Fig. 3, which
shows the TFO weight loss versus the RTFO weight loss for the asphalts
described in Tables 1 and 2. Also shown in Fig. 3 for the asphalts listed in
Table 2 are the weight losses obtained by the Materials and Research
Laboratory of the California Division of Highways (CHD). On both Figs.
1 and 3 the weight loss differences for the asphalts having the higher losses
are especially large.
Differences found by the cooperative study on percent penetration re-
tained (PPR) are shown in Fig. 2. The difference between the PPR on the
TFO and RTFO exposures is shown in groups corresponding to the pene-
tration grades of the original asphalts. With few exceptions, the TFO
causes considerably greater hardening, especially for the softer asphalts
where the weight losses are usually greater. This tendency also is shown in
Fig. 4 for the Tables 1 and 2 asphalts. This same tendency is also shown in
Fig. 4 for the TFO and RTFO exposures determined by the CDH on the
asphalts in Table 2. The CDH data show the TFO to harden several of the
asphalts considerably more than the RTFO exposure.
54 VISCOSITY TESTING O F ASPHALT
oo rt r* SO 00 —
CM ( N ( N O f*> -<t
© CM fS
<N <N oo r - r - \o i-» r-» m (S r- O © ©
I <*) r-i f ^
ffl SO so SO
(N(NTJ-
o © O <N so so c-)
— »n O
03 r- r» oo
O N 0
o
— o ~-
I + (N
T— .— ^
pa
b
264
_*. ^-. f*">
OS
O <N o
o o
VO v o _ - " " """ i o f- " > ^»"
1
(£
K 03
*;
T3
C
a
O O
047
fc„ (N
so <o >
SO O so
n C\ "O © o o
1
8K 03
+
*!
*&
J0«
ii
-204
»n r*i so © © o
oo "<* —•>
03
5 +
w
ffl
< s r f r-i -H O O so © © ©
I r - (N * -
so r^ r^
a +
s. 3
-* G
G
u 0W o *n
ô
tt. '
§£ c
k,
o c
fc.
>o c
. tl. tt, u —
O "O Si
<NO c 6 g
N O
O >
rÔ c
I — <N
zo
• _ 4
<D 1
•S x — u
C
1
•S •« u
5Z O>
a 5 . > 3 o >
« Z
z
oO «> Z oO Z o O
o' . u d .V 5 o .V
35 Vi L
tt, o tt. •3 a-Ofi. •? tt. o ^ tt. O ft<
HU.H Htt,H
>> Oi H OS p hK uH hOS .N
M s hoSfE-c hOS OS H OS
SCHMIDT ON A ROILING THIN-FILM OVEN TEST 55
m n (V
«/"> oi o NO
<N OO <N m oo »o i n m
53, 52,
.27,2.40
"^ rr o r-- ** vT d • *
do
295, 30
N N M i n wi \ o ©' o
§ 00°,
0.56,
o\ o
0.51,
0.46,
I
<N
^r NO
<N
930, 929, 885

• * d
TT
•<f 00 v>
,— O „
ON ON ON
^ CO
1.3
7.5
9.2
1.4
O M ^ 1
NO
© <nfgo\ ©
\
43,
oo «o •q- T t*- v i l—i
< •^ «-) t - O 0O
L25,
0.4,
OH
ON o\
o
NO
~ CO O 00 f \ l •O O M J l
00
© N-H 00 "O <r> l~- . piriri* 0\ r-
ON O <ri •* •* r ; 8 m <s ' d d
86,0.
< fN cs m m n
OH
00 ~ NO OO o
"^ "^1
5 NO
- " t o o
«' NO O o f> ^r
a o o o K -w>^
«I en •o r-- NO NO
"°
NO
o o
3
45, 45, 44
s 00 -M -H
° 3 o o
ro' TT Tj- •
50
46
42
o 00 NO 00
S 9 ON ©
d—
• *
£ 3
5 C), cs
G G G
C),kp
5C)
a
h,
8
•^mS
*-1
- 2 2/ j> b. k"».^
b ~~ — t^ a § c K? e e
£fcfc ts ss g " SS S > £ fi > s s
f ^ O *rt 1 O Si g
5b"
8.
on --H cs
rn -W -w 1?NOO
o" 1 >
C 1 «o o- NoJ Jc 1 Oo "£ ^£ S
O SS
<NOO C vlo
o gS
«oos
fes d i N . s;
f
5 CO «
Q
Z o oO 1S Z o o O 1 ? Z.6 d o «r Z o o o
1
FO, N
FO, N
0, N
O, N
F-Ci
>•* L.
l-( (« HO o" . .V 5
FT-C
FO,
TFO,
1 1 8 8 ftl p-OOfc tL,do"fe

o HBHH,H o •5 H [In tt- H
H fcfcH s>
S S •& .9 s;
o £ > > (2|
gHHC* Ss ( ^ H H P i ^ BiHHo!
i£ ? ««HhBt
S! ^
1.0
0.8
0.6
0.4-
0.2- AS B D
B
Xo 0 S S C P G°
o C s r B
=> o>
i I o. i _
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
TFO W e i g h t L o s s , *
FIG. 1—Comparison of TFO and RTFO weight losses obtained in cooperative test program.
Although these tests are shown to be unequivalent, the literature seems

to indicate that both test exposures are equally acceptable from the stand-
point of functionality in simulating hot-mix hardening. Similar equivalency
is indicated for their functionality in indicating long-term durability.
Nevertheless, the rather substantial differences shown in the exposure
severities negate their interchangeable use unless separate specification limits
20
™o 10
*- u-
a> ,_
t= cc
<u i
a- o
u- M
#i- 0 ^-G-
K -$r
AF D
F s H 0
E M
A
BC
°o"
O .
*% -10 C
J
B°
O
u
S
A
-20
40/50 60/70 85/100 120/150 200/300
Penetration Grades
FIG. 2—Comparison of differences in percent penetration retained between TFO and

RTFO exposures (cooperative test data).
SCHMIDT O N A ROLLING THIN-FILM O V E N TEST 57
1.0
• Weight Losses on Tables 1 and 2
A Weight Losses Obtained by CDH

for Asphalts on Table 2
0.8
0.6
A k
0.4 -
o
u.
I—
0.2
0.1
~r^r
0.1 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
TFO W e i g h t Loss, %
FIG. 3—Comparison of TFO and RTFO weight loss.
are established for each test exposure. The latter alternative seems im-
practical.
A good solution to this dilemma would be another oven exposure which
affects asphalts in the same way as the TFO. This objective has been sub-
stantially accomplished by converting the RTF oven from a convection
oven to a forced-draft oven. The extra circulation not only improves the
uniformity of oven temperature but also changes the type of exposure so
its effect on asphalt is more nearly like the TFO exposure. The added circu-
lation causes more volatile loss to occur. This modified oven is identified
as the rolling thin-film circulating (RTF-C) oven.
No. 1 TFO Oven vs. No. 2 RTFO

+10r
(Tables 1 and 2 Asphalts)
TFO vs. RTFO ICDH Tests)
(On Table 2 Asphalts)
» "" -10 ••
40/50 60/70 85/100 120/150 200/300

Penetration Grade
FIG. 4—Differences in percent penetration retained between TFO and RTFO exposures.
Description of Oven
The RTF-C oven is shown in Fig. 5. The change involves installation3 of
a blower which takes air directed by a plenum chamber from the bottom
of the oven. The plenum chamber is only a sheet-metal insert, The squirrel-
cage blower is driven by a shaft passing through a pair of high-temperature
ball bearings. This blower shaft is coupled to the motor shaft by a flexible
coupling. Thermistor control elements from the electronic proportional
controller are no longer placed in the center of the oven chamber but, in-
stead, are placed in the center of the blower discharge. Also placement of
the thermometer is moved to the side so it can be readily observed. If left
in the center rear, it would be hidden by the blower.
Comparison of Temperature Control

Better temperature control is obtained in the RTF-C oven with the same
controller. This improvement, illustrated in Table 3, shows the original
TABLE 3—Comparison of oven temperature variation.
RTFO RTF-C
Cycle Cycle
Avg, deg F Spread, deg F Avg, deg F Spread, deg F
Center 326.0 ±3.0 321.7 ±0.8

Right top 327.4 ±2.1
Left top blower wash 320.0 ±i.o
Air jet stream 337.3 ±3.1 325.2 ±1.0
Bottom 345.7 ±3.3 323.0 ±1.0
Left center 330.5 ±4.5
Right center 325.0 ±0.5
Max difference 19.8(11. 0 C ) 5.0(2.8 C)
Max cycle spread ± 4 . 5 ( 2 . 5 C) ±1.0(0.56C)
RTFO to have differences from one location to another as large as 19.8 F

(11.0 C). Also, cyclic variation in any one location is as large as ±4.5 F
(±2.5 C) from the mean. Temperature differences between various loca-
tions in the RTF-C oven are reduced to 5.0 F (2.8 C) maximum along with
a maximum cyclic variation of ± 1 F (±0.56 C). Thus, the temperature
variation in the RTF-C oven is about one fifth as great as in the RTF oven.
This improvement should markedly improve the precision of the test.
Test Procedure
The procedure is identical with the California Division of Highways
Test Method 346-D and ASTM D 2872-70 except that the modified oven is
3
A kit to modify existing RTFO can be obtained from Cox and Sons, P. O. Box 674,
Colfax, Calif. 95713.
/ SI
FIG. 5—RTF circulating oven.
used. It is now identified as the California Division of Highways Test

Method 346-E.
Comparison of Results
In Tables 1 and 2 are shown the results of testing a number of asphalts
by three different test exposures: the TFO, RTFO, and RTF-C. Com-
parison and differences between the TFO and RTF-C weight losses and
percent penetration retained (PPR) are illustrated in Figs. 6 and 7. Weight
losses are shown in Fig. 6 to be very nearly equal for both tests. Our TFO
No. 1 oven is slightly more severe; whereas, our TFO No. 2 oven is slightly
less severe than the RTF-C.
Examination of Fig. 7 shows a similar relationship for PPR. The differ-
ence between our No. 1 TFO PPR and the PPR after RTF-C exposure is
at most 5.3 percentage points. The ASTM repeatability precision of the
• No. 2 TFO Oven vs. RTF-C

A No. 1 TFO Oven vs. RTF-C
0.4 0.6 0.8 1.0 1.4 1.6

TFO Weight Loss. %
FIG. 6—Comparison of TFO with RTF-C weight losses.
PPR on a TFO residue is about six percentage points at this penetration.

When using our No. 2 TFO, the maximum difference shown is 3.0 percent-
age points. This compares with 13.2 percentage points difference between
PPR obtained with our No. 1 TFO and our No. 2 RTF oven.
In Table 2, the values obtained by the RTF-C exposure for weight loss
and PPR are repeated three times. Examination of these repeat runs sug-
20
• No. 2TF0-RTF-C
A No. 1 TFO - RTF-C
10
A
t*
r
A
A
fr A
A
-10
-20
40/50 60/70 85/100 120/150 200/300
Penetration Grade
FIG. 7—Comparison of differences between percent penetration retained between TFO and
RTF-C exposures.
gests that our repeatability on the PPR is as good as our ability to repeat
penetration tests. Similarly, the repeatability of the weight losses appears
to be exceptionally good. Reliable repeatability studies on the RTFO for
PPR and weight losses are not available for comparison. However, our own
experience indicates the RTF-C precision to be better than the RTFO.
Collaborative Test Program

A group of eleven Pacific Coast asphalt user-producer organizations,
listed in Table 4, made a preliminary precision study of the RTF-C test
exposure and its equivalency to the TFO test exposure.
TABLE 4—Collaborators in TFO RTF-C study.
California Division of Highways

Chevron Asphalt Company, both
Western and Eastern Laboratories
Chevron Research Company
Douglas Oil Company of California
Edgington Oil Refineries, Inc.
Golden Bear Oil Division, Witco Chemical Co.
Shell Oil Company
State of Nevada Department of Highways
U.S. Oil and Refining Co.
Washington Department of Highways
Six typical straight-run, quite different, West Coast asphalts were chosen:
two California Valley, two California Coastal, and two Los Angeles Basin.
One member of each pair of asphalts was a 60-70 penetration grade and
the other member a 200-300 penetration grade.
Each laboratory determined, for each asphalt, the penetration and
viscosity at 140 F (60 C) and 275 F (135 C) of the original asphalt and of the
TFO and RTF-C residua. Also determined were the weight losses resulting
from the TFO and RTF-C exposures. Duplicate exposures and tests were
made on all samples exposed to the RTF-C oven. This permitted repeat-
abilities as well as reproducibilities to be determined. Only reproducibilities,
however, could be calculated on the tests on the original asphalts and on
the TFO exposure.
Repeatability and reproducibility of the test methods determined in
accordance with ASTM STP 335 are included in Table 5, in which it is
shown that the reproducibility of test values found on the RTF-C residua
is slightly poorer than is found on the RTFO and TFO residua. This find-
ing is not unexpected because most of the collaborators had never used the
RTF-C oven before the 1971 test program. Since the program, informal
questioning by the author of several of the collaborators revealed cases
where some of them had not fully complied with the test procedures.
Examples were: use of on-off oven controls instead of an electronic pro-
6S
OO 00 ^-s
H
« -H O
6?
8
O
u.
<
oo ^. r-
—< oo rs
5T C ~
^ • ^-H . rs • ~H
— r*> —< * 0
o
U-
H
2
H
CO
<
I
• r—
2
<
u O U
Vi
a S
u. u. ^
u.^
r- § v~t
fS
« "«
s
•ri
• «
si >»
.ti >>
,ts
I
SCO
SCO
net
a. > >
portional control, blower direction reversed, and improperly calibrated

thermometers. On the other hand, both the TFO and RTFO exposures had
been used by all of the collaborators for a number of years.
Repeatability, shown in Table 5, is nearly the same for the TFO, RTFO,
and RTF-C. Correction of the errors mentioned above in the RTF-C pro-
cedure should improve its repeatability and reproducibility.
Comparison of RTF-C and TFO Exposure by

Using Collaborative Test Results
The weight losses obtained by the collaborators on exposure of the test
asphalt by the TFO procedure are plotted in Fig. 8 versus the weight losses
they obtained by the RTF-C exposure. Although there is considerable
scatter about the line of unit slope, there does not seem to be a noticeable
bias. Penetration of residues plotted as a natural logarithm in Fig. 9 also
does not show a bias. A similar plot (Fig. 10) of the viscosity at 140 F
(60 C) of the TFO and RTF-C residue shows that with those asphalts
having the highest residue viscosities, the RTF-C exposure is very slightly
more severe than is the TFO. Viscosities of the same residua at 275 F
(135 C) show a similar trend.
Conclusions and Recommendations

The TFO exposure is well established and functional; however, it re-
quires five hours' exposure time. The RTFO exposure is also functional.
1.5r
0.3 0.6 0.9 1.2

Percent Weight Loss on TFO Exposure
FIG. 8—Comparison of TFO and RTF-C, weight loss (collaborative study).
5.2
u_ «-
P e n e t r a t i o n at 77F (25C) of TFO Residua,

Natural Logarithm
F K J . 9—Comparison of TFO and RTF-C, penetration of residua {collaborative study).
0£. •*-
O
— .c
o -^
Viscosity at 140F (60C) of TFO Residua,

Natural Logarithm of Poises
FIG. 10—Comparison of TFO and RTF-C, viscosity at 140 F (60 Q (collaborative study).
Its precision is better, and it is faster to do. Unfortunately, the two expo-
sures differ enough that they should not be used interchangeably without
separate specification limits for each test. A third oven exposure, the
RTF-C, is proposed which not only gives exposures very close to the well-
accepted TFO, but also seems to be at least as precise as the RTFO. Also,
it is just as fast to perform. It is suggested that this new oven exposure be
considered as a replacement for the RTFO test.
Acknowledgment
Repeatability and reproducibility estimates of the 1971 collaborative
tests and the computer plots in Figs. 8; 9, and 10 were made by M. J.
Hillyer of the Shell Development Company. Their permission to include
these data is appreciated.
References
[/] Lewis, R. H. and Welborn, J. Y., Proceedings, Association of Asphalt Paving Tech-
nologists, Vol. 12, 1940, p. 14.
[2] Schmidt, R. J. and Santucci, L. E., Proceedings, Association of Asphalt Paving
[3] Lewis, R. H. and Welborn, J. Y., Public Roads, Vol. 22, No. 2, 1941.
[4] Lewis, R. H. and Halstead, W. J., Public Roads, Vol. 24, No. 8, 1946.
[5] Pauls, J. T. and Welborn, J. Y., Proceedings, Association of Asphalt Paving Tech-
[6] Santucci, L. E. and Schmidt, R. J., Grading of Paving Asphalts by Viscosity at 140 F
versus Penetration at 77 F, ASTM STP 424, American Society for Testing and
Materials, 1967.
[7] Hveem, F. N., Zube, Ernest, and Skog, John, Proceedings, Association of Asphalt
[8] Skog, John, Proceedings, Association of Asphalt Paving Technologists, Vol. 36, 1967,
p. 387.
[9] Hveem, F. N., Zube, E., and Skog, J., Symposium on Road and Paving Materials,
ASTM STP 277, American Society for Testing and Materials, 1959.
[10] Schmidt, R. J., Painter, L. J., Skog, J. B., and Puzinauskas, V. P., Proceedings, Asso-
ciation of Asphalt Paving Technologists, Vol. 37, 1968, p. 476.
[If] Santucci, L. E. and Schmidt, R. J., Proceedings, Association of Asphalt Paving Tech-
112] Santucci, L. E. and Schmidt, R. J., Highway Research Board Bulletin, No. 333, 1962,
pp. 1-9.
[13] Welborn, J. Y., "AASHO Viscosity Grading System for Asphalt Cement Specifica-
tions," Minutes of the 8th Pacific Coast Users and Producers Conference (1969),
Appendix C, p. 4.
[14] Youden, W. T., "Statistical Techniques for Collaborative Tests," Association of
Official Analytical Chemists, 1967.
[IS] 1971 Annual Book of ASTM Standards, Part 11, American Society for Testing and
Materials.
p. S. Kandhal,^ L. D. Sandvig} W. C. Koehler} and
M. E. Wenger^
Asphalt Viscosity-Related Properties of

In-Service Pavements in Pennsylvania
REFERENCE: Kandhal, P. S., Sandvig, L. D., Koehler, W. C , and Wenger,

M. E., "Asphalt Viscority-ildated Properties of In-Service Pavements in Penn-
sylvania," Viscosity Testing of Asphalt and Experience with Viscosity Graded
Specifications, ASTM STP 532, American Society for Testing and Materials, 1973,
pp. 66-78.
ABSTRACT: Six viscosity-graded asphalts from different sources were used in the
construction of the pavements studied and reported in this paper. Tests to deter-
mine the viscosity-related properties of the original asphalts, as well as the asphalts
recovered from time to time from the pavements in service, have been conducted.
Evaluation of the performance of these test pavements has been carried out by a
team of five engneers.
The aging index based on viscosity at 77 F (25 C) has been found to conform
to the pavement performance rating and thus seems to be more meaningful to
indicate comparative aging and life expectancy of the test pavements. Shear sus-
ceptibility of the asphah cements seems to be one of the major factors affecting
the pavement performance. Present viscosity-graded specifications should be
supplemented with a maximum permissible shear-susceptibility value at 77 F so
as to control the durability of the asphalts. The maximum amount of shear sus-
ceptibility that may develop in a paving-grade asphalt prior to pavement
failure is indicated in the data obtained.
KEY WORDS: asphalts, bituminous cements, compaction, durability, evaluation,
grading (classifying), shear properties, viscosity, bitumens
Experimental road projects constructed in Pennsylvania during the years

1964 and 1965 showed that viscosity, not penetration, was the criterion in
placing, rolling, and compacting operations [/].^ Therefore, viscosity re-
quirements were incorporated in the Pennsylvania Department of Trans-
portation specifications for paving-grade asphalt cements in the year 1966.
Recent research [2-6\ has indicated that absolute viscosity alone cannot
specify the complete rheological behavior of paving-grade asphalts, and
other parameters such as shear susceptibility or temperature susceptibility
' Bituminous testing and research engineer, director and engineer of tests, and materials
engineer, respectively. Bureau of Materials, Testing and Research, Pennsylvania Depart-
ment of Transportation, Harrisburg, Pa. 17105.
2 The italic numbers in brackets refer to the list of references appended to this paper.

KANDHAl ET AL ON PAVEMENTS IN PENNSYLVANIA 67
or both are needed. Hence, it has been the intent to study these suggested
viscosity-related parameters in relation to the performance of the in-service
pavements, to gather more information in order to develop optimum
specifications for paving asphalts.
Design and Construction of Test Pavements

Six test pavements, totalling 3.67 miles in length, are located in Clinton
County on Legislative Route 219 (U. S. 220) between Mill Hall and Beech
Creek, Pennsylvania. The original pavement consisted of 18 to 20-ft-wide,
8-in. reinforced concrete which was constructed in 1929 and 1934. This
pavement was resurfaced with 2 in. of ID-2 binder and 1 in. of ID-2 wear-
ing course during October 1964. This study is limited to an evaluation of
six experimental sections of ID-2 bituminous wearing-course surfaces, each
containing a different type of asphalt. An asphalt of 70 to 85 penetration
grade was used in the binder course throughout the entire project. Average
daily traffic on this road is 4200 vehicles.
Properties of the six asphalts used in the project are shown in Table 1.
Excellent control was maintained throughout the entire project so as to
ensure uniform construction of these six pavements. In this closely con-
trolled research project, the only significant variable is the asphalt type.
Since construction of these six pavements, periodical core samples have
been obtained to determine the percent air voids in the pavements and
rheological properties of the aged asphalts. The last core sampling was
done 78 months after the construction.
This paper is an abridgement of a research report [7] issued by the
Pennsylvania Department of Transportation. Detailed test data and anal-
yses can be seen in this reference.
Performance of Test Sections

When constructed in October 1964 and visually inspected during 1965,
no differences in texture or color tones were observed between the asphalts.
Visual evaluation during April 1967 (after 30 months of service) indicated
that the entire road surface was good with the exception of the Asphalt 1
section, which showed some raveling. The decreasing trend of the air voids
with time under traffic is illustrated graphically in Fig. 1. Olsen et al [8]
have suggested a rating method for pavement evaluation to evaluate the
effect of asphalt aging on pavement condition. This rating method was
used as a guideline to accomplish the visual pavement condition survey
by a team of five engineers. Visual evaluation included riding quality,
raveling, spalling, loss of matrix, rutting, corrugations, cracking (trans-
verse, longitudinal, shrinkage, and polygon) and surface texture. The over-
VISCOSITY TESTING O F ASPHALT
o
Si- XXX X
• * w m < n o o o o f > o ©
« < <r> CS O
<Nw <N —. f N © © O - * 00 , C^ T t — _*
S^ x x x X
AosOOCMONCNtN^
S w©\om-*r<i~-©
(N •< w-i 00
• ON
•
N w — — J S O O O \© 1-* m T-
o o o o
2 XXX X
f i • Q m m >ri vo m 00 O
VO Si 00 >?> — — • * © © r- * -
VO < — ON ~- m • « 0 0 0 •
ov v^ «*•> ON os ••• © © ©
rs ^ X X X
i o
^ r i • | t * ^ O N n
• * r w m v O © w - > ( N © ©
«"> < T 0\ ô >n •
— I O P I N - —. O © O \ o r- ( N TH i—i
o o o o
ts XXX X
• ro — i VO
o• * r- o
00 O <N oo r - o -
C-) »-i — TT O l ^ O M M
o
8 xxx X
o
• o o
VD •< co ^r •
* - oo
,3^
~HOCO— m < N O © ©
E
O
B'E 1
00
B
Wi 'So
(8
•s'S - E
•o o
oa a B _
(8
3" * • § 00 V
2-o CM "
<u
a o ON ft. 8 1> •
8 5 2 ° >>>>>. -'3 ÔS - c o

ft. « [ r . • & * . $
ft.
IN ft.
_
-'£ °- °'S « 'S
'a o 2
B
"si
li- ft,*.*: O Q.
, <N ft. ft. 2 - as C * -• S -Si «
- , co I
o P. s^ O oo S i3
to co
§« 1tfl 1Ui S
.2.2
as
8 8**
^^
IBs ed g 111
,o 'a cS ' S ' J >>
B ° .ti oo
u, u u, L. L.
JJ n
•§ 08 p - S
SSS
J3 J3 J3
CO CO W
2 e c ^s S o
« u u *s s '**
s BU 0 - S 55 2
Ut Ui Cfl
g g.a
a, a. >
IF co cd
S'oo
I'S
to o
KANDAHL ET AL ON PAVEMENTS IN PENNSYLVANIA 69
B 24 30 3§ W
TIME - MONTHS
FIG. 1—Air voids versus time in months.
all ratings obtained on individual pavements, in increasing order, are as

follows:
Asphalt 1 19 (poorest)
Asphalt 2 22
Asphalt 6 23
Asphalt 4 25
Asphalt 3 29V2
Asphalt 5 31 (best)
An ideal pavement according to this performance evaluation would rate
33.
Rheological Properties of Asphalts

The following tests were carried out on the asphalts recovered periodi-
cally from the test pavements:
1. Viscosities at 39.2 F (4 C), 77 F (25 C), and 115 F (45 C) over a range
of shear rates, using the sliding-plate microviscometer designed by Shell
Oil Company (both glass and stainless steel plates were used for these tests).
2. Viscosity at 140 F (60 C) using the Cannon-Manning vacuum viscom-
eter.
Unless specifically mentioned otherwise, viscosity data at 77 F implies
0.05 s _1 shear rate.
Viscosity
Viscosity at 77 F has been determined for shear rates of 0.05 and 0.001
s~\ and also at a constant shear stress of 1.63 X 104 dynes/cm 2 . This
shear-stress value has no particular significance except that so far it has
conveniently avoided any extrapolation of the experimental data at 77 F
for all the test asphalts. This constant-shear-stress system was included in
the study to determine an aging index similar to the one reported by
Chipperfield and Welch [9].
Plots of viscosity at 77 F (shear rate of 0.05 s -1 ) versus time in months
and absolute viscosity at 140 F versus time in months are given in Figs. 2
and 3, respectively.
Aging Index
Aging indexes have been determined considering the changes in viscosity
at 77 F (both 0.05 and 0.001 s _1 shear rates and at constant shear stress)
and viscosity at 140 F, as follows:
viscosity after aging
aging index =
viscosity before aging
40
MONTHS
FIG. 2—Viscosity at 77 F (25 C), 0.05 s'1 versus time in months.
KANDHAl ET AL ON PAVEMENTS IN PENNSYLVANIA 71
SLMÔ—I { i i | J4 Jo j . U'E 1 M
0 20 40 60 80
TIME - MONTHS
FIG. 3—Viscosity at 140 F (60 C) versus time in months.
The use of this index (ratio) eliminated the variability caused by difference
in the viscosities of the original asphalts and gives a clearer picture of the
hardening rate [10].
Plots of these indexes versus time periods of aging were made. A typical
plot is given in Fig. 4.
Shear Susceptibility
The shear susceptibility (or shear index) value as employed in this study
is the tangent of the angle of log shear rate versus log viscosity determined
during performance of the viscosity test using the microviscometer. Shear-
susceptibility values have been determined for the six asphalts after in-
creasing periods of aging. Results are shown graphically in Fig. 5. The
relationship between shear susceptibility and aging index (based on vis-
cosity at 77 F) is shown in Fig. 6.
To define the temperature susceptibility of the asphalt cements, a double
logarithm of viscosity in poises has been plotted against the logarithm of
the absolute temperature expressed in degrees (that is, empirical Walther's
equation). The numerical values for slopes of these lines have been indi-
cated as the temperature susceptibility of the asphalt cements within the
specified temperature ranges. At lower than 140 F temperature, slopes of
72 VISCOSITY TESTING OF ASPHAIT
SS« 0.58
20 40 60
TIME - MONTHS
FIG. 4—Aging index [based on viscosity at 77 F (25 C), 0.05 sr1] versus time in months.
06
£4 80 86 Me
20 40 60 80
TIME-MONTHS
FIG. 5—Shear susceptibility at 77 F (25 C) versus time in months.
KANDHAL ET AL ON PAVEMENTS IN PENNSYLVANIA 73
30
20
in
o
d
t
o
>
a . 78 M
X 42 M
s • 30M
x LOG Y = 0.579 + I.I79X
CORR. COEFF. * 0.893
8
z
0.1 0.2 0.3 0.4 Q5 06 Q7

SHEAR SUSCEPTIBILITY AT 77 F
FIG. 6—Aging index [based on viscosity at 77 F (25 C), 0.05 s'1] versus shear susceptibility.
the lines tend to deviate from slopes as established between 140 F and
higher temperatures, because shear-dependent viscosities are encountered
at lower temperatures. Therefore, the temperature-susceptibility data on
original and aged asphalts have been reported in Table 2 separately for
temperature ranges 39.2 to 77 F, 77 to 140 F, and 140 to 275 F. These data
are based on the average of four cores.
TABLE 2—Temperature-susceptibility data.
39.2 to 77 range 77 to 140 F range 140 to 275 F range
Type Original 78 months Original 78 months Original 78 months
1 4.61 1.75 6.34 4.54 3.66 3.83

2 4.18 2.59 5.29 4.44 3.23 3.40
3 4.59 2.69 5.73 5.11 3.53 3.52
4 3.99 2.35 6.24 4.85 3.39 3.68
5 4.05 2.63 5.73 4.54 3.35 3.44
6 3.91 2.91 5.40 4.38 3.37 3.52
Discussion of Results
While discussing the results, as follows, the pavement evaluation ratings
may be kept in view.
Pavement Performance in Relation to Its Physical Properties

All the test pavements had consistent initial air voids within the per-
missible range as per Marshall design and control criteria. According to
these criteria, the pavements are expected to compact under traffic during
the first one to two years, to achieve air voids less than five percent. For
these six test pavements, the type and gradation of aggregate, asphalt
content, asphalt-filler ratio, traffic, and climate have been identical. How-
ever, it is interesting to note that pavement with Asphalt 1 offered maximum
resistance to compaction under traffic. After 78 months, it still had about
6.2 percent air voids, and no further densification is indicated in Fig. 1.
Asphalt 6 is next to Asphalt 1 in offering resistance to compaction under
traffic. However, pavements using Asphalts 2 through 6 were compacted
under traffic below the 3.5 percent air voids level during the first two years.
On examining the viscosities of the asphalt (after mixing in the pugmill)
at 77 and 140 F, it appears that the viscosity at 77 F has affected, to a larger
extent, the capability of the pavements to compact under traffic. Asphalt 1,
with the highest viscosity of 10.1 megapoises after mixing, offered the most
resistance to such compaction, followed by Asphalt 6 which is second high-
est in viscosity as well as in resistance to compaction. Viscosity at 140 F
does not indicate this trend. Thus, it is observed that the so-called low-
viscosity asphalts (in terms of viscosity at 140 F) may not facilitate this
desirable compaction under traffic. This can be accomplished only by
having low viscosity at 77 F. The acceptable maximum limit for thin-film
oven (TFO) viscosity at 77 F, for this project only, seems to lie somewhere
between 3.5 to 10.1 megapoises. Interpolating between Asphalts 1 and 6
for 5 percent air voids after 24 months, it appears to be 6 megapoises.
However, this limit will vary from one project to another depending upon
the other factors, mentioned earlier, affecting the compactability of the
pavement under traffic.
Comparison of pavement evaluation ratings with the air voids after
78 months (Fig. 1) indicates the general trend established by many re-
searchers that durability of the asphalts is affected by the air voids in the
pavement. However, pavements containing Asphalts 2 through 6 had air
voids within a very narrow range of 2 to 3.5 percent (Fig. 1) after 24
months in service, which can facilitate examination of the other parameters
affecting the durability of these asphalt cements and, thus, the pavement
performance.
Pavement Performance in Relation to Viscosities and Aging Indexes

Viscosities at 77 and 140 F of the aging asphalts in the test pavements
versus age of the pavements are plotted in Figs. 2 and 3, respectively, to
examine their relation to pavement performance. It can be seen that vis-
cosity of the aged asphalt alone may not dictate, necessarily, the pavement
performance. For example, Asphalt 2 has the minimum viscosity at 77 F

after 78 months, but its performance is better than Asphalt 1 only.
Values of aging indexes versus time in months were plotted and studied.
One such plot is given in Fig. 4. After 78 months of aging, the ranking
orders of these indexes are as given in Table 3. It can be observed, as far
TABLE 3—Ranking of aging indexes.
Aging Indexes Based on
Viscosity
Viscosity Viscosity at77F
Performance at77F at77F 1.63 X 10* Viscosity
Rating 0.05 s-1 0.001 s-1 dynes/cm2 at 140 F
1 (poorest) 1 (highest) 1 (highest) 1 (highest) 1 (highest)

2 2 4 6 4
6 6 6 2 6
4 4 2 4 2
3 3 3 3 5
5 (best) 5 (lowest) 5 (lowest) 5 (lowest) 3 (lowest)
as this study is concerned, that aging indexes based on viscosity at 77 F

(0.05 s - 1 shear rate) conform exactly to the pavement performance ratings
and, thus, seem to be more meaningful to indicate comparative aging and
life expectancy of the test pavements. Any durability test for asphalt ce-
ments, therefore, should include consistency measurements preferably at
77 F in place of 140 F.
Sisko and Brunstrum [11] have stated that at traffic stress frequencies
unaged and aged asphalts have essentially the same mechanical properties
at lower temperatures (0 to 20 F); differences due to aging appear at higher
temperatures (60 F and higher). The realistic temperature at which the
consistency properties should be measured is still not known; however,
77 F seems closer than 140 F to such temperature for pavements in Pennsyl-
vania.
Raveling was observed in Asphalt 1 test pavement after 24 months when
the aging index (based on viscosity at 77 F, 0.05 s_1) exceeded 12. Asphalts
2 and 6 are beginning to show slight raveling after 78 months on just
exceeding the aging index of 12 (Fig. 4).
Pavement Performance in Relation to Shear Susceptibility of Asphalts
Having observed that pavement performance is indicated by the aging
index based on viscosity at 77 F, it was attempted to find out whether
shear-susceptibility values of the asphalts affect their aging indexes. Shear-
susceptibility values at 77 F of the six aged asphalts were determined after
30, 42, and 78 month periods and plotted with the corresponding aging
indexes. Very good correlation was obtained between aging indexes (based
on viscosity at 77 F) and the shear susceptibility values for these asphalts
(Fig. 6). This indicates that shear susceptibility is one of the major factors
affecting the pavement performance.
Therefore, it appears necessary to control the gain in shear susceptibility
during pavement service life. Skog [5] had stated that a most important
problem in asphalt specification requirements is the determination of the
maximum amount of shear susceptibility that may develop in a paving-
grade asphalt prior to pavement failure. Since Asphalt 1 is showing signs
of distress, referring to Fig. 5, it seems that this value is 0.58. However,
raveling started to show up in Asphalt 1 pavement after 24 months when
the shear-susceptibility value was around 0.40.
In Fig. 4, three distinct groups of asphalts can be seen according to aging
indexes and pavement performance ratings after 78 months in service:
Group 1
Asphalt 1 High shear susceptibility (>0.55). Performance poor.
Group 2
Asphalt 2
Asphalt 6 Intermediate shear susceptibility (0.35 to 0.45). Perform-
Asphalt 4 ance fair to good.
Group 3
Asphalt 3 Low shear susceptibility (<0.35). Performance good to
Asphalt 5 very good.
Unfortunately, no data are available on shear-susceptibility values for
these six asphalts just after mixing or for TFO residue. However, it appears
from Fig. 5 that except for Asphalt 1 the shear-susceptibility values for
Asphalts 2 through 6 were probably less than 0.20 just after mixing. It is
expected that Asphalts 2 through 6 will still be in serviceable condition for
three more years and thus will have at least 10 years' service life before
failure.
Therefore, based on this study, it is believed that a pavement will be
reasonably durable if the shear susceptibility at 77 F of the asphalt cement
after pugmill mixing is less than 0.20 and the pavement is compacted under
traffic during the first 18 to 24 months to the extent that it will have no
more than 5 percent air voids. Shear-susceptibility requirements for TFO
residue can be substituted in place of the ductility test since a good correla-
tion between these two properties was found by Welborn et al [4].
Temperature susceptibility has been considered as one of the parameters
specifying the rheological behavior of paving asphalts [12]. Many re-
searchers have attempted to relate the temperature susceptibility of the

asphalts with pavement distress in the form of transverse or other types of
cracks.
The data on temperature susceptibility of the asphalts in three tempera-
ture ranges (39.2 to 77 F, 77 to 140 F, and 140 to 275 F) are given in Table 2
for the original and aged asphalts. It will be observed that the temperature
susceptibilities of the asphalts have shown a decreasing trend with aging
in the first two temperature ranges, whereas the increase in the 140 to 275 F
range is not appreciable. It appears that the temperature-susceptibility
parameter may have more significance in case of low-viscosity asphalts.
There is an absence of any crack pattern on these test pavements, which
can be attributed to temperature susceptibility.
Conclusions
Based upon the rheological properties of the six asphalts studied and the
preceding discussions, the following conclusions are drawn:
1. Pavement durability is affected significantly by the extent of air voids

in a pavement. The rate of hardening of asphalts is checked considerably
if the pavements can compact under traffic during the first 18 to 24 months,
so as to have air voids of less than 5 percent.
2. All other factors affecting compaction being the same, the apparent
viscosity at 77 F (after pugmill mixing) seems to control the capability of the
pavements to compact under traffic at ambient temperatures. For the
asphaltic mixtures used in this research project, an apparent viscosity of
6 megapoises or less appears to ensure compaction to the desired level.
Low viscosity at 140 F may not accomplish the desired compaction under
traffic if the viscosity at 77 F is higher than the desired value.
3. Viscosity of the aging asphalt alone does not, necessarily, indicate
the pavement performance. Aging indexes based on viscosity at 77 F and
0.05 s _1 shear rate conform to the pavement performance ratings in this
study and thus seem to be more meaningful to indicate comparative aging
and life expectancy of the test pavements. Raveling was observed on the
test pavements when this aging index exceeded 12.
4. Since a change in apparent viscosity at 77 F does provide an index to
possible failure of the binder, any durability test for paving-grade asphalt
cements should include consistency measurements preferably at 77 F in
place of 140 F.
5. It appears necessary to control the gain in shear susceptibility of the
asphalt cements, since this seems to be one of the major factors affecting
the pavement performance. Performance of one of the pavements was
observed to be poor when shear susceptibility of the aged asphalt exceeded
0.55, whereas the performance was good to very good when the shear
susceptibility was less than 0.35. Durability of an asphaltic pavement will
perhaps be ensured if the shear susceptibility of the asphalt cement, after

pugmill mixing, is less than 0.20 and the pavement is compacted under
traffic during the first 1J^ to 2 years to less than 5 percent air voids. From
the durability standpoint, thus, there is an apparent need to consider shear
susceptibility as a necessary specification requirement in place of a ductility
test.
6. For the six asphalts studied, temperature susceptibility does not seem
to have any significant effect on the pavement performance.
The significance of some of the rheological properties of asphalts indi-
cated by the fundamental viscosity measurements will perhaps be more
clear with the continued evaluation of these test pavements until failure.
It is believed that the asphalts weather in different ways and could present
different forms of pavement failure.
Acknowledgments
The authors wish to express their appreciation to David Streby for con-
ducting all the microviscosity tests, to Richard Basso and June Viozzi for
assistance in compiling the data, and to Edward Macko for preparing the
illustrations. Cooperation of W. S. Myers and I. L. Myers in the per-
formance evaluation of the test pavements is also appreciated.
References
[/] Sandvig, L. D. and Kofalt, J. A., Proceedings, Association of Asphalt Paving Tech-
nologists, Vol. 37, 1968, p. 602.
[2] Gzemski, F. C , Fundamental Viscosity of Bituminous Materials, ASTM STP 328,
American Society Testing and Materials, 1962, p. 80-94.
[3] Hveem, F. N., Zube, E., and Skog, J., Proceedings, Association of Asphalt Paving
p. 407.
[<5] Chipperfield, C. H., Duthie, J. L., and Girdler, R. B., Proceedings, Association of
Asphalt Paving Technologists, Vol. 39, 1970, p. 576.
[7] Kandhal, P. S., Sandvig, L. D., Koehler, W. C , and Wenger, M. E., "Asphalt
Viscosity Related Properties of In-Service Pavements in Pennsylvania," Bureau of
Materials, Testing & Research, Pennsylvania Department of Transportation, Research
Report, Feb. 1972.
[8] Olsen, R. E., Welborn, J. Y., and Vallerga, B. A., Proceedings, Association of Asphalt
[9] Chipperfield, E. H. and Welch, R. R., Proceedings, Association of Asphalt Paving
[10] Traxler, R. N., Proceedings, Association of Asphalt Paving Technologists, Vol. 36,
1967, p. 546.
[11] Sisko, A. W. and Brunstrum, L. C , Proceedings, Association of Asphalt Paving
[12] Dobson, G. R., Proceedings, Association of Asphalt Paving Technologists, Vol. 38,
1969, p. 123.
R. J. Schmidt^
Laboratory Measurement of the

Durability of Paving Asphalts
REFERENCE: Schmidt, R. J., "Laboratory Measurement of the Durability of

Paving Asphalts," Viscosity Testing of Asphalt and Experience with Viscosity
ABSTRACT: The objective of this study was to develop a practical laboratory
test to measure reliably the relative durability of paving asphalts. An appropriate
test must harden a wide variety of asphalts in the same order as dofieldexposures.
The studies reported demonstrate that current asphalt durability tests do not
properly relate to field hardening because they exaggerate unrealistically the
volatile loss to obtain sufficient oxidative hardening. A microfilm exposure con-
ducted in an RTF bottle exposed to the oven atmosphere through a short
capillary is shown to prevent excessive volatile loss while permitting sufficient oxi-
dation to take place. The functionality of this new RMF-C test is supported by
three different correlations: a group of recent California highway test projects,
the Z-W Test Project, and a group of asphalt concrete specimens aged for 100
days at 140 F (60 C). The data support the contention that the exposure is a valid
simulation of the field hardening expected to occur in asphalt concrete pavements.
KEY WORDS: asphahs, durability, microfilm, test roads, specifications, bitu-

minous concretes, viscosity
Public agencies responsible for highways, as well as asphalt producers,

have used a variety of methods to predict the relative hardness of asphalt
in pavements after a number of years' exposure. Unfortunately, no method
yet tried has succeeded in doing this reliably. Field tests have failed to
correlate well with laboratory tests, particularly when a large variety of
asphalts are considered.
One of the most widely used asphalt oven exposure tests, ASTM Test
for Effect of Heat and Air on Asphaltic Materials, (D 1754-69), the thin-
film oven test (TFO), was developed in 1940 by Lewis and Welborn [i].^
Limited support for the validity of this test, as a measure of long-term
durability, has been reported [2-5]. Most of the evidence supports the use
of this test as a measure of the relative hot-mix hardening expected of
asphalts but not as a reliable long-term durability test.
' Senior research associates, Chevron Research Company, Richmond, Calif. 94802.
2 The italic numbers in brackets refer to the list of references appended to this paper,

An improved version of this test was proposed in 1959 by the California

Division of Highways (CHD) [6-8]. This improved ASTM Test for Effect
of Heat and Air on a Moving Film of Asphalt (D 2872-70), the rolling thin-
film oven test (RTFO), was also intended as a measure of the hot-mix
hardening [9-11]. However, limited correlation of this test with hardening
of asphalt in pavements was also reported by Santucci and Schmidt [12].
Still more recently, Schmidt [75] proposed an improved RTFO which pro-
vided better precision and better agreement with the TFO exposure.
In order to simulate film thicknesses similar to those existing in asphalt
concrete (AC) mixes, Griffin and co-workers [14,15] devised a microfilm
durability exposure followed by microviscosity measurements to determine
the extent of hardening which took place at 225 F (107.2 C) in 4 h. Subse-
quently, Hveem et al [6] and Skog [7] proposed a modification of the
Griffin thin-film exposure. They reduced the exposure temperature to 210 F
(98.9 C) and increased the time to 24 h. Hveem et al [6] demonstrated in-
directly the relationship between field hardening and their thin-film plate
durability test (TFP). More recently, Schmidt and Santucci [2] proposed a
modification of the TFP test [the rolling microfilm oven test (RMFO)]
which was similar in exposure conditions but which was easier to perform
and had better precision. These authors also presented correlation studies
of the RMFO with the pavement life of the Zaca-Wigmore test road and
to aged fatigue AC specimens [12]. The TFP and RMFO give substantially
the same correlations of the various asphalts with the field and with
laboratory-aged specimens.
Recently, Beaton and Sherman [16] retracted the earlier CDH conten-
tions of the correlation of the TFP exposure with field exposures in pave-
ments. Results from a group of experimental test roads comparing asphalts,
made to have various durability characteristics, suggested that the TFP test
actually could be misleading. Some experimental asphalts which appeared
to be exceptionally durable by the TFP test (as well as by the RMFO test)
were shown in field tests to be no more durable than the control asphalts
having a moderate durability. Beaton and Sherman [16] reported that the
asphalts in AC exposed at 140 F (60 C) in the laboratory hardened in the
same order as they did in the experimental pavements.
It is the objective of this paper to compare the field hardening of these
special and control asphalts used in field tests by the CDH with variations
of the RMFO exposure that theoretically could correct the shortcomings
of the test.
Laboratory-Hardened AC Specimens
Four asphalts, listed in Table 1, were chosen for the preliminary study
of the influence of varying RMFO exposure conditions on the relative
hardening of asphalts. Asphalt A was made from Smackover crude oil.
SCHMIDT O N LABORATORY MEASUREMENT O F DURABILITY 81
Q
S
H
+ <
CO
'it'*'*
O
u.
H
o
' *-H O Ô O +
^ m T3- o r- —
5
D.
© c-i r- «o
s +
O • * i-H m o
GO 0> fN 0\
oo rn ^ -
_1_ •* m *-1 v>
a
'So
oo "* ^t o
T +
O i H » m O
r- (>o
N ^vo r*
- w-i
_1_ "O ^ fN VI
!
-o
.o
2170-67
2171-66
113-6'
93-66
§, •o
2170
2171
5-65
o
Is J3
5
IS
"a8 2
•+•*
QOQO Q Q Q Q
ASTM
ASTM
ASTM
ASTM
ASTM
ASTM
ASTM
ASTM
Tes
ties
u. a
«
& 3
w
-J o £ !* h ^
« a « § S3
<
t! fl woo
oGu- M
< « t«N
92 aj ed _
[I, ^ - - W < N
.— U-< tL. " ^
•sou"
T3 >
• o »n v I - O -O "*
u
•"a* r- r-
-S a •" <N « (N r- a
S c
CB (S C
.ti . t i a ,
M M
I J3 O
o •& >>>,>,
u
•a
I
I I
. 3 73 r<* S
.2 •
•eo
* 43
1 CJ O <W
t » <W CO
•? £ > > E
6 ££>>Q
S
IP
3 X
This widely used asphalt is usually considered to be one of the most durable
asphalts known. Asphalts C and D are made from the same heavy Cali-
fornia Coastal crude oil. This asphalt has been widely and successfully used
not only in most of the Western states but in Alaska and the Pacific islands
as well. Asphalt B was made by vacuum reducing Asphalt D to a hard
grade and subsequently softening it to an 80 penetration with a heavy
lube oil extract. According to TFP tests, as well as RMFO tests conducted
by the CDH, Asphalt C hardened much more than Asphalt B; yet in the
Benicia, California, test road (see Table 4), Asphalt B hardened slightly
more than Asphalt C.
A number of AC specimens, each containing 5 percent (weight) of one of
these four experimental asphalts, were made from Cache Creek gravel
graded according to the curve shown in Fig. 1. This high-voids grading
(22 percent voids in mineral aggregate (VMA)) was used so that asphalt
hardening would proceed rapidly. These AC specimens were then exposed
in a 140 F (60 C) oven. At appropriate intervals, two specimens from each
group were removed and the asphalt recovered by the Abson extraction
procedure. In these extractions, 15 percent methanol was used along with
the benzene to effect more complete removal of the asphalt from the
simulated weathered specimens. Microviscosities on these residua, inter-
polated to 100 days' laboratory exposure, are shown next and also in Item 1
of Table 2.
Viscosities of Asphalts Recovered from
AC Specimens Exposed for 100 Days at 140 F (60 C)
Asphalt A B C D
Viscosity,3 megapoises 27 95 95 125 Item 1
The following studies on variations on the RMFO procedure were di-

rected toward attaining the same relative asphalt viscosities after harden-
ing as in the exposure just described.
Rolling Microfilm Oven Test (RMFO)

This test, proposed [2] as an improved version of the CDH TFP test, is
conducted by depositing, from a benzene solution, a 20-^m film of asphalt
on the walls of an RTFO bottle. The bottle is subsequently exposed for
24 h in an RTF oven held at 210 F (98.9 C). At the end of this period, the
asphalt is scraped from the interior of the bottle, fused, and placed between
microviscosity plates. The viscosity of this RMFO residue is determined
3
Unless otherwise noted, all viscosities reported are in megapoises measured at 77 F
(25 C) at a stress of 167 g/cm2.
SCHMIDT O N LABORATORY MEASUREMENT OF DURABILITY 83
No. 200 50 30 16 8 4 3/8" 1/2" 3/4" 1"

Sieve S i z e s , Exponential Scale
FIG. 1—Gradation of Cache Creek aggregate used in AC specimens at 140 F (60 CI-
TABLE 2—Summary of viscosities' obtained by variation of RMFO test procedure.
Viscosity, megapoises6
Asphalt
Reference
Item Exposure B D
Control: AC specimens
1 100 days at 140 F (60 C) 27 95 95 125
RMFO conditions
2 Open, 24 h at 210 F (98.9 C) 14 36 194 191
3 Open, filled, 1 /2 h at 300 F (148.9 C), 48 h
at 180 F (82.2 C) 12 24 158 135
In RTF-C forced-air oven, 24 ft at 210 F (98.9 O
4 Open 14 30 394 406
5 0.041-in.—ID capillary 15 25 25 25
6 0.17-in.—ID capillary 15 23 26 25
7 0.041-in.—ID capillary (RTFO residue) 16 31 52 51
8 0.041-in.—ID capillary oxygenflushed(RTFO 18 32 56 55
residue)
In RTF-C forced-air oven, 48 ft at 210 F (98.9 C)
9 0.041-in.—ID capillary 31 73 77 76
" All viscosities are reported in megapoises and were determined at 77F (25 C) with a
uniform stress of 167 g/cm2.
* Results in all columns are the average of at least two exposures.
at 77 F (25 C) under a uniform stress of 167 g/cm 2 . Where necessary, for

certain comparisons, the viscosity was also determined at 0.05 s - 1 shear
rate. The functional basis for using the constant-stress viscosity is sup-
ported in Refs 2 and 12. Better precision and correlation with pavement life
is shown with the constant-stress viscosities than with uniform shear rate
viscosities.
Viscosities obtained on RMFO residua from these four asphalts are
listed next and also in Item 2 of Table 2.
Viscosities of RMFO Residua Exposed

24 H at 210 F (98.9 C)
Asphalt A B C D
Viscosity, megapoises 14 36 194 191 Item 2
The difference between the viscosities in Items 1 and 2 is quite apparent.

The similarity between the viscosity of Asphalts B and C in Item 1 (re-
covered from aged AC specimens) is much the same as observed in field
exposure. The large difference between Asphalts B and C, shown in Item 2,
is similar to the difference found in the same asphalts by the CDH by their
TFP test.
Influence of Mineral Filler and Temperature

The lack of agreement between various laboratory tests and field harden-
ing is often attributed to higher laboratory exposure temperatures and
possibly to the absence of aggregate surface in the laboratory tests. To
clarify the importance of these two factors, the RMFO was modified to
include, in the 20-/xm film of exposed asphalt, 50 percent by weight of fine
filler (—325 mesh) separated from the same Cache Creek aggregate used in
the AC laboratory oven exposures listed in Item 1. Also, after the mixture
of asphalt and filler had been deposited from benzene onto the walls of the
RMFO bottle, a 30-min exposure at 300 F (148.9 C) was given to simulate
hot-mix hardening. Next, the asphalt-filler film was exposed at 180 F
(82.2 C) for 48 h (a temperature sometimes reached in actual pavement
surfaces). These modified exposure conditions should now include two of the
additional factors frequently blamed for the lack of agreement between
field and exposure; that is, the asphalt is first hot-mix hardened and then
exposed in a realistically thin film in the presence of about the same surface
area of mineral as found in real mixes and at a temperature often attained
in pavements.
After exposure, the asphalt was recovered by dissolving it in benzene
containing 15 percent methanol. The filler was removed by centrifuging
the solution and filtering through a 15-Mm pore size glass filter into clean
SCHMIDT ON LABORATORY MEASUREMENT OF DURABILITY 85
RTFO bottles. After removal of the solvent by evaporation at 180 F

(82.2 C) in an RTF bottle, the viscosity of the recovered asphalt was deter-
mined.
Listed next and in Item 3, Table 2, are the results of this exposure on
these four asphalts.
Viscosities of RMFO Residua Exposed in the Presence of Mineral Filler

for 30 Min at 300 F (148.9 C) Followed by 48 H at 180 F (82.2 Q
Asphalt A B C D
The ranking found is much the same as occurred in the original RMFO
exposure, that is, in Item 2. Apparently, neither lowering the exposure
temperature nor adding a realistic amount of mineral surface appears to
improve the agreement of the test exposure with the core exposure shown
in Item 1.
Extra High Air Circulation During RMFO Exposure

To improve the precision of the RMFO (and the RTFO exposure for
which the oven was originally designed), the oven was modified to include
a blower which provides a rapid internal air circulation through a plenum
chamber. This improved oven, illustrated in Fig. 5 in Appendix 1 and in
Ref 16, is identified as the RTF-C oven.
The results of exposure of these same four asphalts by the regular RMFO
procedure in this new RTF-C oven are shown next and in Item 4 of Table 2.
Viscosities and Weight Changes of RMFO Residua Exposed in Forced-Air

Circulating Oven (RTF-C) for 24 H at 210 F (98.9 Q
Asphalt A B C D

Weight Change, % -1.0 -1.7 -8.5 -7.8
Asphalts A and B are seen to harden to almost the same level as in the
original RMFO in Item 2, Table 2. Asphalts C and D harden more than
twice as much as before. The significant difference from the RMFO in
exposure condition is that the bottles revolve into the direct air blast of the
blower. This large air blast entering the bottles is supplementary to the air
delivered into the bottles from the air jet. Also just shown and in Item 4,
Table 3, are the weight losses measured in this modified exposure. Asphalts
A and B are shown to lose less than 2 percent weight; whereas Asphalts C
TABLE 3--Summary of weight changes"

i obtained by variation of RMFO test
procedure.
Refer- Asphalt
Item Exposure A B C D
In RTF-C forced-air oven

4 Open 24 h at 210 F (98.9 C) -1.0 -1.7 -8.5 -7.8
5 0.041-in.—ID capillary 24 h at 210 +0.5 -0.5 + 1.5 -0.04
F (98.9 C)
6 0.17-in.—ID capillary 24 h at 210 +0.3 -0.4 +0.3 -0.4
F (98.9 C)
9 0.041-in.—ID capillary 48 hat 210 +0.2 +0.5 +0.4 +0.3
F (98.9 C)
" Weight change values could not be repeated closer than about 1 % of the sample
weight, so all values except the 8.5% and 7.8 % weight loss for Asphalts C and D, line 4,
are6 considered negligible.
Item numbers correspond to those in Table 2.
and D lose about 8 percent by weight. Asphalts B and D are nearly identical
in composition, except that in Asphalt B the light oil component was re-
moved and replaced with a heavy lube oil extract. Other experiments
demonstrated that the composition of the lube oil extract was not respon-
sible for the reduced hardening rate. In this additional experiment, the low-
volatility lube extract was replaced with a low-volatility oil fraction made
from Asphalt D. The hardening rate was almost identical to those shown
for Asphalt B just described. The conclusion evident from this experiment
is that the difference in apparent durability measured by the RMFO be-
tween Asphalts B and C is a consequence of a difference in volatility be-
tween the two asphalts.
A further possible conclusion is that in actual AC mixes the light oils
are lost to a considerably lesser extent than they are in the various acceler-
ated laboratory exposures. Apparently, in the AC mixes, those oil mole-
cules which vaporize must diffuse through long, narrow passages before
they escape into the atmosphere. Under these circumstances, most of the
vaporized oil is reabsorbed by the asphalt before it escapes. These findings
suggest that simulation of field hardening of asphalts will require a pro-
cedure that provides sufficient oxygen but which inhibits, at the same time,
escape of the light oils from the asphalts during exposure.
Capillary Venting of RMFO Exposure Bottle—

Preliminary RMF-C Procedure
The model suggested in the foregoing paragraph can be simulated by
restricting the opening of the RMF bottle during oven exposure. This is
done with an aluminum-foil-covered cork through which is inserted a 2-in.-
long, 0.041-in.-diameter capillary tube. On exposure in the RTF-C oven
(with the air jet off), oxidation proceeds much as before, but the evapora-
tion is minimized. Listed next and in Item 5, Tables 2 and 3, are the results
of this exposure on these same four asphalts.
Viscosities and Weight Changes of RMF-C Residua

Exposed While Vented Through a 0.041-In.-ID
Capillary for 24 H at 210 F (98.9 C) in RTF-C Oven
Asphalt A B C D

Weight change, % +0.5 -0.5 +1.5 -0.04
Hardening of the low-volatility Asphalts A and B is shown to be about

the same as the hardening obtained in the original RMFO. This suggests
that their oxidation has not been limited by a lack of oxygen. The level of
hardening of the volatile C and D asphalts has been greatly reduced.
Asphalts B and C are nearly identical, which agrees with the hardening
found in the AC specimens listed in Item 1. Weight differences listed above
confirmed that the volatility losses have been reduced by the capillary tube
to nearly negligible levels. Weight-loss measurements could be reproduced
only to within about ± 1 percent of the sample weight, and so values <1.5
percent are all considered negligible.
Listed below and in Item 6, Tables 2 and 3, are the results of a series of
exposures identical in all respects to the exposure used above, except that
the diameter of the capillary tube was increased to 0.17 in. They are quite
similar. The diameter of the tube used does not appear to be critical, at
least in this range of sizes.
Viscosities and Weight Changes of RMF-C Exposed Residua While

Vented Through a 0.17-In.-ID Capillary Tube for
24 H at 210 F (98.9 C) in RTF-C Oven
Asphalt A B C D

Weight change, % +0.3 -0.4 +0.3 -0.4
Effect of Simulating Hot-Mix Hardening Prior to

Simulating Long-Term Durability
The level of hardening obtained at 210 F (98.9 C) for 24 h in an RMFO
bottle containing a capillary tube is too low to separate the various asphalts
into a range large enough with respect to the test errors. The difference be-
tween 15 megapoises for Smackover and 25 megapoises for California
Coastal asphalts is rather narrow considering the precision of the micro-

viscosity test. In practice, as much as half of the total field hardening is the
result of hot-mix hardening. It seemed reasonable that simulation of hot-
mix hardening before simulation of the long-term hardening would be more
realistic and would give a wider spread of results without changes of ranking.
The asphalts were first exposed for 85 min at 325 F (162.8 C) under
RTFO conditions. Following this, a small portion of the RTFO residue
was deposited from benzene in a 20-/im film in an RTFO bottle and exposed
for 24 h while being vented at 210 F (98.9 C) to the atmosphere through a
0.041-in.-diameter, 2-in.-long capillary tube. The results listed next and in
Item 7 of Table 2 show that asphalts are no longer ranked in the same way
as found in Item 1. Apparently, the preliminary RTFO exposure distorts
the simulation.
Viscosities of RMF-C Residua (0.041-in.-lD Capillary) on RTFO Residua,

RTFO Exposure 85 Min at 325 F (162.8 C),
RMF-C Exposure at 210 F (98.9 C)for 24 H
Asphalt A B C D
Also important are the results listed next and in Item 8 of Table 2. This
exposure is identical to the ones described above except that the bottle was
flushed with oxygen just before the capillaries were fitted into place. The
values shown below are essentially the same as those shown above. This
is further evidence that sufficient oxygen is present to provide for the
oxidation needs of the asphalt during exposure. The reaction does not
appear to be oxygen limited by the capillary tube.
Viscosities of RMF-C Residua (0.041-In.-ID Capillary) on RTFO
Exposure 85 Min at 325 F (162.8 C), Exposure at 210 F (98.9 C)
for 24 H, Oxygen Flushed
Asphalt A B C D
Effect of Increasing Exposure Time—Final RMF-C Procedure

To increase the difference between hardening of these four asphalts to a
level large enough with respect to testing errors, the RMF-C exposure used
in Item 5 on the original asphalt was increased from 24 h to 48 h. The re-
sults are listed next and in Item 9 of Tables 2 and 3. Also listed are the
weight changes obtained, which are negligible. The ranking agrees reason-
ably well with Item 1, Table 2.
Viscosities and Weight Changes on RMF-C Residua Exposed

for 48 H at 210 F (98.9 C), 0.041-In.-ID Capillary
(Final Procedure)
Asphalt A B C D

Weight change,% +0.2 +0.5 +0.4 +0.3
Although a larger spread might be desirable, it appears to be sufficient.

RMF-C tests conducted on a wide spectrum of asphalt types and grades
by this procedure show the range to extend from about 1 megapoise for a
more durable 200 to 300 penetration asphalt to 1000 megapoises for a
rapid-hardening 40 to 50 penetration asphalt.
This final variation consists of exposure of a 20-fim film of asphalt
deposited on the walls of an RTFO bottle (vented through a 0.041-in.-
diameter, 2-in.-long capillary tube) in an RTF-C oven for 48 h at 210 F
(98.9 C). A detailed procedure for performing this test, identified hence-
forth only as the RMF-C test, is given in Appendix I.
Comparison of Hardening Obtained by Laboratory Aging of AC

Specimens with RMF-C Aging of a Variety of Asphalts
In addition to the series of four asphalts (A, B, C, and D) previously
used for studying the effect of exposure-condition variations on asphalt
hardening, a larger group of asphalts was exposed to both the RMF-C
conditions and long-term laboratory hardening in AC mixes. The results
of these additional tests, as well as those on the original four asphalts, are
compared in Fig. 2. The two exposures are shown to have excellent correla-
tion for a variety of experimental asphalts, including those experimentally
reconstituted from solvent deasphalter bottoms, extracts, and air-blown
residua. These unusual asphalts were included to test the reliability of the
RMF-C exposure on other than ordinary steam-refined asphalts from vari-
ous crude oils.
Comparison of RMF-C Exposure Hardening with Life of

Zaca-Wigmore Test Pavement Project
The Zaca-Wigmore (Z-W) Test Project was constructed in 1954 by the
CDH to determine the influence of asphalt properties on pavement life.
This project was first described in detail by Hveem et al [8]. Other reports
were made by Skog [17], by Simpson and co-workers [15], and by Schmidt
and Santucci [2]. These latter authors compared the life of the Z-W test
road sections to the hardening obtained on the asphalts used in these
sections when exposed to several asphalt durability tests. Included were
RTFO, TFP, and RMFO exposures. The latter exposure was shown to
give a slightly better correlation with pavement life.
iooor
500-
100-
50-
\ft i i i i i i i i L_LJ i i ! I
10 50 100
Viscosity of R M F - C R e s i d u e , 77F ( 2 5 C ) ,
M e g a p o i s e s , 167 g / c m !
FIG. 2—Comparison of viscosities of RMF-C residua with asphalt residua recovered from
AC cores exposed 100 days at 140 F (60 C).
These same 2-W asphalts were exposed to the RMF-C exposure. The
viscosities of the hardened asphalts are compared in Fig. 3 with the life
of the various test sections to 10 percent cracking in this test project. A
slightly better correlation is obtained with the RMF-C than previously
was obtained with the RMFO exposure, that is, the correlation coefficient
increased from 0.8258 to 0.8961. Also, the standard error decreased from
32 to 23 percent.
Comparison of RMF-C Exposure with Hardening in

Recent California Division of Highways Field Tests
The C D H reported [15,18] interim results of their recent field tests com-
paring a number of experimental asphalts with normally available asphalts.
Extracts of their data are shown as Table 4. All cores extracted were taken
from t h e outer wheel track of the traffic lane. In Fig. 4 are plotted t h e
viscosities at 0.05 s _ 1 (obtained by the C D H ) . These viscosities were o b -
tained on asphalts recovered from the last set of cores in each test project.
They are plotted on Fig. 4 versus the R M F - C viscosity (also calculated at
0.05 s _ 1 ). When examining this figure, one should keep in mind that, except
TABLE 4—Viscosities of asphalts recovered from various California Division of

Highways test projects (data supplied by the CDH).
Viscosity of
Pavement Recovered
Age Asphalt,
When 77 F (25 C),
CDH Cores 0.05 s _1
Identification Removed, Shear Rate,
Project Location and Type Number Supplier month megapoises
Martinez:
Control, 85/100 R-4219 0 3.5
22 23
48 224
Special R-4215 D 0 3.5
22 31
48 197
Los Altos:
Control, 85/100 R^1238 D 0 2.2
19 17
45 28
Special R-4236 D 10 2.3
19 20
45 28
Ludlow:
AC-12 R^1298 0 2.2
20 16
40 28
AC-12 R-4299 0 2.6
20 26
40 45
85/100 R-4300 0 4.3
20 41
40 60
Control, 85/100 R^l301 0 3.5
21 36
40 65
Special R-4305 0 4.6
19 26
38 41
18 15
35 31
Olivehurst:
Control, 85/100 R-4331 0 6.6
15 24
33 25
15 16
33 30
15 15
33 30
(Continued)
TABLE A—Continued.
Viscosity of
Pavement Recovered
Age Asphalt,
When 77 F (25 C),
CDH Cores 0.05 s"1
Identification Removed, Shear Rate,
Project Location and Type Number Supplier month megapoises
Blythe:
Control, 200/300 R-4471 H 8 1.74
25 4.78
Special R-4433 H 6 6.1
24 29
Willits:
Special, 5.5 kilopoises R-4005 D 0 2.8
22 4.9
62 6.6
22 3.8
62 5.7
22 3.2
62 6.4
Control, 85/100 R-4008 G 0 3.6
22 7.2
62 11.2
Folsom:
Control 85/100 R-4110 or C 0 3.4
(Asphalt C in Table 1) R-4111 29 17
52 21
Special R-4099- D 0 4.2
R-4102 29 23
52 25
Benicia:
Control 85/100 R-4159 B 0 4.6
(Asphalt C in Table 1) 30 20
57 26
Special R-4158 D 0 4.9
(Asphalt B in Table 1) 30 27
57 36
for the slope, only the results from any one test road should be considered
at a time because each set of data is obtained from different pavements
existing at widely separate locations. The general relationship appears to
be quite good considering the difficulties inherent in obtaining representa-
tive field samples.
When the data were fitted with the regression model, the slope for In
(recovered viscosity) versus In (viscosity RMF-C) was found to be 0.611
± 0.089. The multiple correlation coefficient for the overall regression was
500r Standard Error 23*

Correlation Coefficient 0.8961
Significance Level > 99.9*
Period H.
Period I
10 100 1000
Viscosity at 77F, 167 g / c m z , Megapoise
F I G . 3—Correlations of Zaca-Wigmore pavement life with rolling microfilm RMF-C

residue viscosities [167 g/cm*. 77 F (25 Q ] .
1000
Martinez
100
o u
Olivehurst
CD Legend of
10 Projects
o Martinez
x Ludlow
+ Olivehurst
7 Los Altos
o Benicia
o Blythe
• Willits
a Folsom
uL _i_
1 10 100
V i s c o s i t y at 77F (25C) Megapoises, 0.05 S e c . - 1
RMF-C Exposure
48 Hr at 210F (98.9C)
F I G . 4—Comparison of asphalt hardening infield test projects with hardening in RMF-C

exposure.
0.98547. The correlation coefficient for the viscosity slope alone was found
to be 0.8790, which is an excellent correlation considering that several test
projects are included. The percentage error in predicting the recovered
viscosity from the RMF-C viscosity is approximately 21 percent.
The slope of 0.611 can be used as follows: If two asphalts have an
RMF-C viscosity ratio of 10:1, then after field exposure they would be ex-
pected to have a (recovered) viscosity ratio of 1 0 0 6 u : l = 4.09:1.
A more complete description of the multiple-regression analysis is given
in Appendix II.
Precision of RMF-C Test

When determined according to the procedure given in Appendix 1, the
ASTM repeatability obtained in our laboratory on the microviscosity of
RMF-C residua is 25.5 percent. This corresponds to an ASTM standard
deviation of 8.8 percent.
The ASTM precision of this test is now being established by a committee
of public highway and oil company representatives.
Labor Required to Perform RMF-C Test

Although the elapsed time for conducting an RMF-C exposure in 48 h,
eight samples can be exposed at a time (all started together or separately).
Each RMF-C test (including duplicate microviscosities on the residua)
requires about 13^ man-hours of work when conducted in groups of four
or more. This time could be reduced somewhat if the tests were run on a
routine basis.
Summary and Conclusions

The functionality of the RMF-C test is supported by three different
correlations: a group of recent California highway test projects, the Z-W
Test Project, and a group of AC specimens aged for 100 days at 140 F
(60 C). This latter group contained a number of asphalts reconstituted
from components not widely used at present. The makeup of this group of
asphalts was chosen to test the reliability of the RMF-C exposure when
most diverse types of asphalts are considered. The data support the con-
tention that the exposure is a valid simulation of the field hardening ex-
pected to occur in AC.
There are certainly circumstances where the volatility is not so inhibited
as it appears to be in the interior of AC mixes. These circumstances might
be near or on exposed surfaces or in very open mixes. Possibly the original
RMFO will be more representative of the asphalt hardening occurring in
these types of field conditions.
Asphalts will harden differently in each location and in each different

kind of mix. However, unless the void content is quite low, the RMF-C
exposure should rank asphalts in the same order as they harden in the field.
Each different field condition would be represented by an individual curve
similar to those shown in Fig. 4.
Acknowledgment
The aid and statistical treatment of the data by L. J. Painter of Chevron
Research Company and the field data, samples, and continued encourage-
ment of John Beaton, George Sherman, John Skog, and Glen Kemp of
the California Division of Highways, Materials and Research Laboratory,
are greatly appreciated by the author.
APPENDIX I
Rolling Microfilm Circulating Oven (RMF-C) Durability Test
1. Weigh 0.500 g of the asphalt into a 150-ml beaker using an analytical balance.
Dissolve the sample with approximately 10 g of benzene (reagent grade). Filter the
dissolved sample through a 10 to 15 ^m sintered glass filter directly into an RTF
sample bottle. Use a small pressure bulb to force solution through filter. Rinse the
beaker and filter clean with additional benzene. The dissolved, filtered asphalt is
now in the RTF bottle dissolved in approximately 15 to 20 g of benzene.
2. Place the RMF-C bottle (from this point forward, the RTF bottle containing
benzene and 0.5-g sample is referred to as an RMF-C bottle) into the RTF-C oven
[13]. (See Fig. 6.) The oven was previously set at 210 F (98.9 C). A210F(98.9C)
thermometer, similar to a 275 F (135 C) ASTM C-13 thermometer,4 is hung from the
center of the oven 1 in. from the side. Turn on vacuum to top vents in oven. (See
Fig. 6.) Turn the power to the oven off when adding the benzene-diluted samples.
Rotate for 15 min to evaporate the benzene; turn oven power "on," and run
evaporation cycle an additional 15 min with heat to achieve a positive benzene
evaporation. Without turning power off, remove bottles (in pairs), and insert
aluminum-foil-covered cork stoppers through which extend 2-in. long, 0.041 in.
inside diameter glass tubes. Turn power on and run for 48 h. The 0.5-g residue gives
a film thickness on the inside of the RMF bottle of approximately 20 ^m. The
oven should be equipped with the electronic proportional controller specified in the
RTF-C procedure, CDH Test Method 346-E.
3. At the conclusion of the 48-h RMF-C exposure, remove the bottle and allow
to cool (about 10 min). Scrape out all residue (by means of a razor blade attached
to a handle) from the wall (ignore the asphalt on the ends), and place in one of the
wells of a porcelain spot plate (Coors Size 00, Van Waters and Rogers Catalog
No. 53636) using a small spatula to clean residue from scraper. Within 30 min,
heat the spot plate on a steam or hot plate until the asphalt is fluid [approximately
250 F (121.1 C)] and mix well with the spatula.
4
These thermometers can be ordered from the Brooklyn Thermometer Company.
4. Make up microviscosity plates as soon as the asphalt is fluid. Use metal plates
supplied by Hallikainen or Cannon. We have incorporated several new pro-
cedures in the preparation of microviscosity plates. They include:
(a) Use numbered paired plates— measure the thickness at each end of the pair
set together without shims over the central part with a metric micrometer capable
of measuring to 1 ^m.
(b) Prepare a duplicate set of plates from the residuum in the spot plate.
(c) Dab small amounts of sample with the same spatula used above onto the
metal test surfaces on both the top and bottom microviscosity plates. The plates
should be warm so the sample can flow and wet the entire test surface. Do not
touch the sample or plate with your fingers!
(d) Move the coated plates to a flat, cool surface. Position shims, invert, and
locate top metal plate. (See Sketch A in Fig. 5.)
(e) Place a spacer over top metal plate that clears the tops of the side spacers.
(A glass microviscosity plate works fine.)
(/) Transfer entire assembly on a flat surface such as J/g-in.-thick, 3-in. by 6-in.
aluminum plate to 275 F (135 C) oven. Leave in oven for 15 min; then place a
500-g weight, which has been preheated in the oven, on top of the spacer, and
leave for another 15 min. (See Sketch B, Fig. 5.)
(g) Remove assembly. Replace hot 500-g weight with a cold weight. Place
assembly on a flat metal surface, check shim drag to be sure of seating, and allow
to cool in the air for 1 h.
(A) As soon as the plates have cooled to room temperature, recheck shim drag
by sliding them back and forth. If the shims move freely with no drag, the plates
were not properly seated against the shims. If this is the case, prepare another plate
immediately. After the plates have stood for 1 h in air (with the weight still in
place), remove the weight and spacer and measure the thickness of the plates (at
each end) containing the shims and sample with the metric micrometer. Subtract
the thickness at each end from the corresponding thickness of the bare plates. If
the difference is more than a few microns from the shim thickness, prepare another
plate immediately. If the thickness is close, place the plates and shims in the 77 F
(25 C) water bath. The asphalt film thickness used in calculating the viscosity is the
difference between the measurements taken at Step (a) and Step (h). Do not assume
thickness is equal to shim thickness.
5. After a period of 5 min in the water bath, remove the shims, place the plates
in the microviscometer, and proceed with the measurement as follows:
(a) Apply a load of 1000 g. Allow a movement of exactly four large divisions (high
speed) on the Varian recorder. This represents approximately 0.20-mm movement
on the plates. By using the same movement on each sample, a consistent shear
history on the sample is established.
(b) Remove 1000-g weight. Apply a 200-g weight. Allow the measurement to
proceed until a steady rate of shear is obtained—usually at least two large divisions
across the chart. If the slope with the 1000-g load was >25 deg, use the slow chart
speed with the 200-g loading.
(c) Calculate viscosities at the two load levels (1000 g and 200 g) by using the
angle of the recorded line and the average thickness as measured by the difference
between the bare plates and the plates containing the specimen. If the viscosities on
duplicate plates differ by more than 10 percent, prepare another duplicate set of
plates from the same residuum in the spot plate and repeat the viscosity measure-
ment.
SAMPLE
METAL
MICROVISCOSITY
SIDE SPACER PLATES
AND SHIM
TOP SPACER
SAMPLE
METAL
SIDE SPACER MICROVISCOSITY
AND SHIM PLATES
FIG. 5—Sample preparation with metal microviscosity plates.
FIG. 6—RTF circulating oven.

APPENDIX II
Description of Fitting Procedure for RMF-C Viscosity Versus
Recovered Viscosity at Several Locations
The data for the several sites were fitted with the following equation:
y = In (recovered viscosity) = aL + b • In (RMF-C viscosity)
where QL is an initial constant (level effect) fitted for each site, and b is a common
slope fitted across all the data.
The fitting was done with a standard least-squares regression program on an
electronic computer, with the aL values being obtained as the coefficients of a group
of variables (one for each test site) having the values 1 or 0, depending on whether
or not theyth data point came from that site. Thus, .y, = aiXJ + a^Xij + atX3j- • •
+ a»Xtj + b In (RMF-C viscosity), with XL, = 1 if y, is from site L; all other
Xa = 0 for this./'.
The results obtained in the regression analysis are summarized next:
site ah e>Ll
Martinez 3.0831 21.8

Los Altos 1.1157 3.0
Ludlow 1.8122 6.1
Olivehurst 1.2694 3.6
Blythe 1.1995 3.3
Willits 0.2461 1.3
Benicia 1.0743 2.9
Folsom 1.1739 3.2
b = slope 0.611 ± 0.089

R = 0 98547
Standard Error of Fit = 0.207 in In (recovered viscosities)
= 21 percent of recovered viscosities
5
The quantities e"L are the fitted recovered viscosities corresponding to a value of 1 for
RMF-C viscosities.
References
[1] Lewis, R. H. and Welborn, J. Y., Proceedings, Association of Asphalt Paving Tech-
• nologists, Vol. 12, 1940, p. 14.
[2] Schmidt, R. J. and Santucci, L. E., Proceedings, Association of Asphalt Paving
[3] Lewis, R. H. and Welborn, J. Y„ Public Roads. Vol. 22, No. 2, 1941.
[4] Lewis, R. H. and Halstead, W. J., Public Roads, Vol. 24, No. 8, 1946.
[5] Pauls, J. T. and Welborn, J. Y., Proceedings. Association of Asphalt Paving Tech-
[<5] Hveem, F. N., Zube, Ernest, and Skog, John, Proceedings, Association of Asphalt
p. 387.
[8] Hveem, F. N., Zube, E., and Skog, J., Road and Paving Materials, ASTM STP 277,
American Society for Testing and Materials, 1959.
[9] Santucci, L. E. and Schmidt, R. J., Grading of Paving Asphalts by Viscosity at 140 F
Versus Penetration at 77 F, ASTM STP 424, American Society for Testing and
Materials, 1967.
[70] Santucci, L. E. and Schmidt, R. J., Highway Research Board Bulletin, No. 333, 1962,
pp. 1-9.
[11] Schmidt, R. J., Painter, L. J., Skog, J. B., and Puzinauskas, V. P., Proceedings, Asso-
ciation of Asphalt Paving Technologists, Vol. 37, 1968, p. 476.
[12] Santucci, L. E. and Schmidt, R. J., Proceedings, Association of Asphalt Paving
[13] Schmidt, R. J., Viscosity Testing of Asphalt and Experience with Viscosity Graded
Specifications, ASTM, STP 532, American Society for Testing and Materials, 1973,
pp. 52-63.
[14] Griffin, R. L., Miles, T. K., and Penther, C. J., Proceedings, Association of Asphalt
[15] Simpson, W. C , Griffin, R. L., and Miles, T. K., Road and Paving Materials, ASTM
STP 277, American Society for Testing and Materials, 1959.
[16] Beaton, J. L. and Sherman, G. B., Minutes of 9th Pacific Coast Users and Producers
Conference, 1971.
[17] Skog, J., Road and Paving Materials, ASTM STP 277, American Society for Testing
and Materials, 1959.
[18] Kemp, G. R. and Johnson, M. H., Viscosity Testing of Asphalt and Experience with
Viscosity Graded Specifications. ASTM STP 532, American Society for Testing and
Materials, 1973, pp. 100-121.
G. R. Kemp^
Asphalt Durability Tests and Their

Relationship to Field Hardening
REFERENCE: Kemp, G. R., "Asphalt Durability Tests and Their Relationship to

Field Hardening," Viscosity Testing of Asphalt and Experience with Viscosity
Graded Specifications, ASTM STP 532, American Society for Testing and Ma-
terials, 1973, pp. 100-122.
ABSTRACT: A total of 24 asphalts in 8 test sections are being evaluated relative

to their change in viscosity with time. The test sections are located in different
climatic areas in California. Statistical correlations are presented covering various
laboratory test methods for predicting asphalt durability with 30 and 50 months
of pavement service life. The laboratory test methods employed for predicting
asphalt durability involve the concept whereby the asphalt is weathered by heat
and air in a thinfilm.The amount of heat and thickness of the film varies, but the
end result is basically volatization and oxidation to cause hardening of the asphalt.
Additional correlations are presented using original voids and chemical procedures
involving the Rostler analysis and Heithaus procedure. This report discusses the
findings to date; additional pavement service life will be required before final
conclusions can be drawn.
KEY WORDS: asphalts, bitumens, pavements, specifications, durability, vis-

cosity, thin films
The Materials and Research Department of the California Division of

Highways has engaged since the 1930's in studying asphalt durability.
During that period many different approaches have been tried. Most of
these methods have utilized a concept whereby the asphalt is weathered
by heat and air in thin films. Chemical procedures using the Rostler
analysis and Heithaus procedure have also been studied. The problem is to
design laboratory test procedures that correlate well with actual field
weathering so that realistic specifications can be devised for controlling
the field weathering qualities of an asphalt. In order for the laboratory tests
to be of the greatest value they should correlate with field hardening even
though climatic conditions may be quite dilTerent. This report gives statis-
tical correlations of laboratory tests with field hardening of asphalt in eight
different field test sections located in different climatic areas in California.
' Engineering associate. Materials and Research Department, Division of Highways,

State of California, Sacramento, Calif. 95819.
Downloaded/printed by too

KEMP ON DURABILITY TESTS AND FIELD HARDENING 101
Field Test Sections

In 1963 a new asphalt specification was proposed by the Materials and
Research Department of the California Division of Highways which in-
cluded requirements to control "setting" and durability [I].2 Table 1 shows
TABLE 1—California tentative asphalt specifications.
Proposed Modified Modified

Test Method 1963 1965 1966
Flash Point, PMCT deg. F AASHO T-73 475 475 450

min
Penetration of Original Sample AASHO T-49 85-100
at77F
Stain Number of Original ASTM D 1328 10 10 10
Sample, max after 120 h,
140 F, 50 psi.
Viscosity, on original sample
140 F, min, poises AASHO T-202 2200
225 F, min, centistokes AASHO T-201 1800
325 F, max, centistokes AASHO T-201 200
Cohesiograph Reading, origi- Calif. 350 0.80
nal min in.
Gain 0 to 24 hr min in. 0.08
Rolling Thin-Film Test, 325 F, Calif. 346
75 min
Penetration, residue, 77 F, AASHO T-49 55
min
Ductility residue, 77 F, min, AASHO T-51 75 75 75
cm
Viscosity
140 F, poises AASHO T-202 4000 to 6000 4000 to 6000
275 F, centistokes AASHO T-201 425 to 800 425 to 800
Durability Test Calif. 347
Viscosity of residue after Calif. 348
durability test, megapoises
at77F
Shear rate, 0.05 s _I max 20 25 25
Shear rate, 0.001 s_1 max 60 60 60
Microductility of Residue Y? Calif. 349 10 10 10
cm/min min, mm
Solubility, CCL4, original AASHO T-44 99 99 99
sample, % min
Rolling Thin-Film Test, 375 F, Calif. 346
75 min
Penetration residue 77 F, AASHO T-49 45
min
Ductility residue 77 F, min, AASHO T-51 60
cm
this 1963 specification, also a revision made in 1965 to eliminate some

repetitious tests, and the present tentative specification which was revised
in 1966 by altering the flash requirement.
2
The italic numbers in brackets refer to the list of references appended to this paper.
Starting in 1964 with the Willits test section, a series of test sections were
placed in various climatic areas of the state. In these test sections, special
asphalts, which were manufactured to meet the proposed tentative asphalt
specifications, were placed and evaluated, along with a selected portion
of the job asphalt as a control. The test sections were located, besides the
Willits [2] location, at or near Folsom [3], Benicia [4], Martinez [5], Los
Altos [6], Ludlow [7], Olivehurst [8], and Blythe [9], California. Table 2
TABLE 2—Durability test sections.
Structural Section
Date Thickness and Type
Location Placed AsphaWs
Contract (% Climatic Seal, Surface Base
Road asphalt)* Area Type® (ID) ft. (special), ft (type), ft
Willits Aug. 1964 coastal S 5.5 (D) fog 0.33 AC 0 33(A)

64-1T13C4-P (5.6%) valley S4.5 (D) (0.17 AC) PMCTB
hot-summer S 3.5 (D) 0.33(B)
wet-winter C-85/100 (G) PMCTB
Folsom Aug. 1965 central S-TS (D) 0.05 0.33 AC 0.25 ACB
65-3T13C-07424 (5.5%) valley OG A C (0.17) AC) 0 . 5 0 AB
03-Sac, ED-50 hot-summer C-85/100 (C) 1 0 AS
wet-winter
Benicia Oct. 1965 coastal S-TS (D) 0.06 0.17 A C (A) 0.67 AB
65-10MAC-O490O4 (6.5%) bay area OG A C 0 33 A C ( B ) 0 92 AS
lO-Sol-21, 680 warm-summer C-85/100 (B) (0.17 AC A)
wet-winter
Martinez Sept. 1966 coastal S-TS (D) none 0.33 AC 0.25 ACB
04-120224 (5.0%) bay rear (0.17 AC) 0 67 AB
04-CC-4-4.4/9.8 warm-summer C-85/100 (B) 1.5 AS
wet-winter
Los Altos Dec. 1966 coastal S-TS (D) none 0.50 AC 0.42(A)
04-170374 (5.5%) bay area (0.17 AC) PMCTB
04-SC1-280-11.5/18.9 warm-summer C-85/100 (D) 0.42(B)
wet-winter PMCTB
Ludlow June 1967 high S-AC12 (E) 0.04 0.35 AC 0.50(B)
08-039334 (4 asphalt) desert S-AC12 (A) OG A C (0.17 AC) PMCTB
08-SBd-40- Jul.,Aug.'67 hot-summer 85/100 (A) (S-TS 0 . 8 0 AS
R28.4/R42.1 (S-TS) cool-winter C-85/100 (F) 0.35 AC)
(5.3%) S-TS (A)
S-TS (A)
LV
Olivehurst Sept. 1967 central S-TS (C) none 0.33 AC 0 . 6 7 AB
03-020964 valley (0.17 AC) 1.0 AS
03-Yub-70-0.1/R8.9 hot-summer S-TS (A)
(5.4%) wet-winter C-85/100 (C)
Blythe Oct. 1968 low C-200/300 (H) 0.05 0.30 AC 0.25 A C
11-037564 (C-2O0/3O0) desert OG A C 0 . 6 0 AB
ll-Riv-10-R133.5/ Nov. Dec. very hot S-TS (H) ( 0 . 3 0 AC
1968 summers surface and
Jan.,Feb.'69 warm-winter 85/100 (M) 0.25 AC base;1
(85/100) (85, 100,
0 . 3 0 surface
( 4 . 1 % base)
(4.5%
surface)
1
S = special.
C « control.
TS = tentative specification.
gives pertinent information as to contract road designation, date of place-

ment, percent of asphalt used, type of climate, asphalt identification, and
structural section. In each case the special asphalts were used in the same
manner as the control or job asphalt. Figure 12 shows the geographical
location of each test section.
"the tentative specification based on viscosity grading after the rolling
thin film (RTF) test was designed, in addition to providing uniformity, to
control "setting" and asphalt durability. Skog [70] in his study of setting
and durability showed a good correlation between viscosity results on
residues from American Association of State Highway Officials (AASHO)
thin-film (TF) procedure and the rolling thin film oven (RTFO) procedure.
Skog and Sherman [11] later showed that the RTF residue viscosities cor-
relate well with viscosities on recovered asphalt from original field mixes.
Table 3 shows the viscosities of the RTF residues of the asphalts used in
the test sections and how they compare with the viscosities of the recovered
asphalts from the original mixes used in the test sections. No serious
problems concerning "setting" occurred in the placement of the test sec-
tions as reported [2-9], and the test requirements (Table 1) indicate that
the tentative specification asphalts being studied for durability in the test
sections possessed adequate setting properties.
TABLE 3—Asphalt durability test sections—Comparison of viscosities of asphalts

recovered from original mixes and residue from rolling thin-film oven test.
Original Mix Rolling Thin Film

Asphalt Residue
Vacuum Kinematic Vacuum Kinematic

Capillary Viscosity, Capillary Viscosity,
Test Asphalt Viscometer, 275 F Viscometer, 275 F
Section Grade 140 F (centi- 140 F (centi-
Road (source) (poise) stokes) (poise) stokes)
Folsom 03-Sac, S-TS (D) 4319 496 4060 467

ED-50 C-85/100(C) 4568 510 3829 466
Benicia 10-Sol- S-TS (D) 5005 541 5035 521
21,680 C-85/100(B) 5911 679 7005 705
Martinez 04-CC- S-TS(D) 4184 470 3878 465
4-4.4/9.8 C-85/100(B) 2462 363 2257 330
Los Altos 04-SC1- S-TS (D) 3426 522 3930 454
280 11.5/18.9 C-85/100(D) 2957 467 3268 425
Ludlow 08-SBd- S-AC12 (E) 2014 254 1808 244
40-R28.4/R42.1 S-AC12 (A) 3732 507 2884 447
85/100 (A) 4909 534 3955 516
C-85/100(F) 2914 291 2397 261
S-TS (A) 5165 578 4346 528
S-TS (A) LV 3136 457 2634 418
Olivehurst 03- S-TS (C) 6063 562 4316 465
Yub-70-0.1/R8.9 S-TS (A) 6080 614 4487 531
85/100(0 7901 675 4603 537
° TS = tentative specification.
Laboratory Test Procedures

In early research on asphalt durability, the California Division of High-
ways developed the California infrared weathering machine, which is de-
scribed as to its design and operation in a paper by Skog [12].
Subsequent investigations utilized other procedures, some of which em-
ployed a direct chemical analysis. The procedures which have been employed
in the investigations involved in this report are described as follows:
Ottawa Sand Mix, 140 F (60 C), 400 ft, 1000 h

This procedure, which is described under the title of California infrared
weathering procedure, uses a mixture of Ottawa sand and 2 percent asphalt
which is mixed for 3 min in a specially designed mixer at 325 F (163 C).
The resulting mixture, which has been ascertained to have asphalt films on
the aggregate of approximately 5 to 7 y.m, is weathered in an uncompacted
state in small lids in an oven heated by infrared lights shining on the asphalt
mixture so that the surface and interior temperatures of the mix are kept
at 140 F (60 C), while air heated to 105 F (41 C) is blown across the surface
of the asphalt mix.
Weathering Plates, 210 F (99 C), 24 h

This procedure, described in California Test Method 347, is a modifica-
tion of the Shell Aging Index of Bituminous Materials test [75] and utilizes
a 20-jum-thick film of asphalt on a 50-mm square Pyrex glass plate which
is weathered in an oven for 24 h at 210 F (99 C).
Extended RTF, 325 F (163 C), 5 h
This procedure is performed the same as the ASTM Test for Effect of
Heat and Air on a Moving Film of Asphalt (ASTM D 2872-70) (rolling
thin-film oven test) except that the heating period is 5 h plus 15 min warmup
[14]. The asphalt is exposed to a temperature of 325 F (163 C) in a
changing thin film with a jet of heated air injected into the bottle for
approximately 1 s 15 times per min.
Field Density Briquette, 140 F (60 C), 90 days
This procedure utilizes briquettes 4 in. in diameter by approximately
2x/i in. high. These briquettes are prepared from the same aggregate as
used in the field with the same gradation, the same asphalt type, the same
percent asphalt, and mixed at the same temperature. Compaction of the
briquette specimen is such that the percent voids are the same as the in-
place field voids after construction. The prepared briquettes are then
weathered in the California infrared weathering oven at 140 F (60 C) and
air is blown across them at 105 F (41 C) for various periods of time. A
testing period of 90 days was used in this study.
Rostler Ratio (N + A7)/(P + A2)

This ratio is derived from the Rostler analysis [15] to define reactivity
of the malthene fraction and is determined by dividing the sum of the two
most reactive components of the malthene fraction by the two least reactive
components. The method may be outlined as follows: A 1-g sample of the
asphalt is dissolved in 50 ml of normal pentane, and the precipitated
asphaltenes are removed. This fraction is labeled A by Rostler. The solu-
tion is then treated with 85 percent sulfuric acid (H 2 S0 4 ) which separates a
fraction containing the most reactive portions of asphalt, and, according
to Rostler, contains all of the nitrogen bases of the asphalt. This fraction
is called N. H 2 S0 4 of 98 percent strength is then used to separate the next
most reactive fraction called "first acidaffins" and labeled Al. Next the
solution is treated with fuming H 2 S0 4 containing 30 percent sulfur trioxide
(S0 3 ). This treatment produces a fraction next in rank of reactivity, named
"second acidaffins" and labeled A2. The remaining least-reactive fraction,
the paraffins, is called P. These residual fractions are then used to calculate
the Rostler ratio.
Percent Original Voids

The percent voids was determined on the basis of a theoretical maximum
specific gravity determined from the fine aggregate, coarse aggregate, and
the asphalt used in the original mix. A ratio of the specific gravity of field
cores to the calculated theoretical maximum specific gravity subtracted
from 100 gives the percent voids (California Test Method 304). The signifi-
cance of percent original voids to asphalt weathering has been documented
by several researchers; it apparently indicates the susceptibility of a pave-
ment to oxidation.
Heithaus Value P
The Heithaus test [16] provides a method for studying the internal phase
relationship of an asphalt. The technique provides a measure of the pep-
tizing power of the malthenes Po, and the peptizability of the asphaltenes
Pa. When the two factors are combined the state of peptization P [P =
Po/(l — Pa)] can be determined. The higher the value of P, the more
stable the internal phase of the asphalt system.
Chevron Research RMFC Procedure

The rolling microfilm circulating (RMFC) procedure, which was devel-
oped by the Chevron Research Corporation, utilizes the apparatus of the
rolling thin-film test (ASTM D 2872-70) plus some additional apparatus.
Essentially an approximately 20-/*m film of asphalt is formed in a rolling
thin-film bottle by dissolving a specified amount of asphalt in benzene and
coating the inside of the bottle during the evaporation of the benzene by
rotating the bottle in the RTF oven for a specified period with the power
off till the benzene is evaporated. After complete evaporation, the bottle
is stoppered except for a small capillary tube through the stopper, and the
weathering period is then conducted for 48 h in the RTF oven at 210 F
(99 C) at the specified 15 rpm with no air being used. The weathered residue
is then removed and tested. The concept of the weathering is that there is
hardening through oxidation but that the evaporation present is contained
and resolvated into the system as is theorized to happen to the asphalt in
the enclosed void spaces in the pavement.
Laboratory Test Results

All of the tables presenting the test results of the various laboratory pro-
cedures to predict asphalt durability are related to the various test sections
in the field by name designation and road. The asphalts used in the various
test sections are identified by code letters to indicate their refinery source.
Table 4 shows the microviscosity results of the recovered asphalts from
the test sections for 30 and 50 months weathering periods, the extended
rolling thin-film procedure for 5 h, the weathering plates at 210 F (99 C)
for 24 h, 400 and 1000-h weathering periods of Ottawa sand mix in the
California infrared machine, and the weathering of briquettes fabricated
to field density for 90 days in the California infrared machine. Also shown
on this table are values of the Rostler ratio (N + A\)/(P + AT), the percent
voids of original cores from the pavements, and the Heithaus procedure.
The last column includes the results furnished by Chevron Research from
their rolling microfilm method using a modified rolling thin-film oven.
Analysis of Test Section Data

It was planned to core and test the various test sections at regular inter-
vals of about two years. As can be seen in Table 5, which shows data from
all the test sections for the different corings, the intervals between corings
varied from about 15 to 30 months.
In the Willits test section the special asphalt was manufactured in a very
small quantity and was made to conform to the 1963 tentative specifications.
Three different "setting" grades were made although their essential proper-
ties were the same. Figure 1 shows how the three special asphalts and the
control asphalt compare to each other after 62 months of field weathering.
The special asphalts used in this test section show superior durability
properties. In this section the control asphalt, although not as good as the
special asphalts, is still excellent in its weathering properties. Undoubtedly
the very low percentage of voids is a significant factor in the superior per-
formances by all of the asphalts in the Willits test section.
TABLE 4—Asphalt durability test sections—30 and 50-monthfieldweathering and

various laboratory procedures viscosity test data (surface course—microviscosity
at 77 F, 0.05s~l—all test results megapoise except Rostler ratio,
percent voids, and Heithaus P).
Various Laboratory Procedures Test Results
>>
Sand Mix
Sand Mix
'late
+ £
void;
Briquette 90 i
Field Density
0001
Field ^ 53
00 h
Asphalt Results o. PS
•O M
Test Type0 — •8« S . « "1 3 c
J". S3^ K "^
o
201 F
140 F
Exten
WeatI
1 fc
Percei
Chevi
Ottaw
Origii
Heith
a+
RTF,
Section (source 30 50
Road grade) month month
o 3 as 5
1
Willits S(D)5.5 5.3 6.1 15 13 13 20 1.25 5.2 6.0 18
01-Men S(D)4.5 4.2 5.0 14 13 11 18 0.95 5.1 6.3 13
S(D)3.5 3.6 5.2 12 12 9 15 1.22 6.7 6.0 10
C(G)85/100 7.9 9.8 24 29 44 110 1.85 6.5 3.3 36
Folsom S(D)TS 23 25 21 26 37 108 1.98 7.0 3.5 29
03-Sac, C(C)85/10O 17 21 26 45 23 52 1.24 7.8 3.3 21
ED-50
Benicia S(D)TS 27 34 23 16 32 103 28 1.78 8.2 3.5 45
10-Sol-21 C(B)85/100 20 22 32 91 50 192 31 1.90 8.3 3.9 44
Martinez S(D)TS 72 197 22 29 43 100 1.83 10.3 3.5 37
04-CC-4 C(B)85/100 59 224 21 41 24 71 1.63 8.7 4.2 45
Los Altos S(D)TS 25 29 23 33 32 80 1.80 6.4 3.6 35
04-SC1-280 C(D)85/100 25 29 26 83 22 89 1.70 4.8 4.1 41
Ludlow S(E)AC12 22 13 12 11 28 17 1.50 4.0 5.8 9
08-SBd-40 S(A)AC12 36 25 27 27 100 26 1.70 5.1 3.8 23
(A)85/100 52 27 31 39 105 28 1.80 6.4 3.5 30
C(F)85/100 50 27 74 27 80 35 1.59 5.4 3.8 40
S(A)TS 36 30 25 52 135 30 1.50 6.8 3.6 39
S(A)LV TS 24 16 17 26 74 19 1.40 5.8 3.6 20
Olive- S(C)TS 27 27 34 32 76 31 1.37 7.0 3.4 28
hurst S(A)TS 26 28 24 50 125 35 1.47 6.6 3.5 27
03-Yub-70 C(C)85/100 25 25 34 28 78 31 1.22 9.8 3.5 35
Blythe S(H)TS 36 28 31 17 44 1.45 7.2 4.3 27
ll-Riv-10 C(H)200/300 5.4 2.5 6 3.8 12 1.50 4.4 4.4 2
C(M)85/100 135 28 44 48 132 6.9
• S = special asphalts.
C = control asphalts.
TS = Calif. Tentative Specification (4000 to 6000 poises at 140 F after RTF).
The special asphalts used in the Folsom, Benicia, Martinez, and Los Altos
test sections come from one refinery. All of these sections were laid within
18 months of each other as can be seen in Table 2. Figures 2 through 5
show how the special asphalts compared with the control asphalts. As can
be seen in Fig. 6, which compares the special asphalts from the different
sections, they are each weathering similarly except for the Martinez test
section which had the largest percentage of initial voids.
In the Ludlow test section, as shown in Fig. 7, special asphalts designed
to meet The Asphalt Institute AC 12 grading were compared with special
asphalts meeting the 1966 California tentative specification. In addition to
TABLE 5—Asphalt durability test sections—Recovered asphalt test results (surface course).
Tests
Microviscosity
Pave- Pene- Soft- Duc- Shear Micro-

Test ment tra- ening tility Mega poise Sus- duc-
Section Asphalt Age, tion Point 77 F pecti- tility
Road (source) Grade" month 77 F deg F cm 0.05 s"1 0.001 s-i bility m m
Willits special (D) 5.5 0 2.82 5.53 0.18 31

01-Men-101 VFS 22 4.93 11.0 0 18 22
62 6.59 17.6 0.23 17
4.5 0 2.62 4.42 0.13 35
FS 22 3.83 7.1 0.16 27
62 5.67 14.3 0.21 21
3.5 0 2.05 3.88 0.16 37
MS 22 3.2 7.7 0.22 24
62 6.43 17.0 0.21 20
control (G) 85/100 0 3.57 5.12 0.09 57
22 7.23 9.8 0.09 46
62 11.2 20.8 0.11 36
Folsom special (D) Calif. 0 54 126 100 + 4.22 6.38 0.11 63
03-Sac, ED-50 TS 29 25 137 100 + 22.8 32.6 0.08 36
52 23 140 100 + 24.7 40.9 0.12 31
control (C) 85/100 0 54 124 100 + 3.4 4.8 0.09 56
29 29 135 100 + 16.8 32.2 0.15 27
52 24 140 95 20.9 44.1 0.18 16
Benicia special (D) Calif. 0 50 127 100 + 4.9 6.6 0.08 69
10-Sol-21,680 TS 30 23 145 87 27.3 46.7 0.13 18
57 21 148 100 + 35.5 73.8 0.19 9
control (B) 85/100 0 53 128 100 + 4.6 7.3 0.12 53
30 26 142 94 20.3 39.7 0.17 14
57 24 146 100 + 22.4 49.8 0.19 16
Martinez special (D) Calif. 0 53 126 100 + 6.2 7.6 0.06 71
04-CC-4-4.4/9.8 TS 22 25 142 100 + 30.9 54.6 0.15 5
48 12 153 16 197.0 531.0 0.22 2
control (B) 85/100 0 54 122 100 + 4.4 5.1 0.04 72
22 26 136 100 + 23.3 30.4 0.06 65
48 10 150 60 224.0 416.0 0.15 5
Los Altos special (D) Calif. 0 54 124 100 + 2.34 3.73 0.12 65
04-SC1-280- TS 19 26 137 100 + 19.6 30.7 0.11 34
11.5/18.9 45 22 141 100 + 28.4 55.5 0.17 14
control (D) 85/100 0 54 122 100 + 2.19 2.64 0.05 61
19 27 135 100 + 16.9 22.6 0.08 41
45 21 139 100 + 27.7 50.3 0.15 26
Ludlow special (E) AC12 0 59 114 100 + 2.21 2.32 0.01 92
08-SBd-40- 20 25 131 100 + 15.5 15.5 0.00 127
R28.4/R42.1 40 21 136 150 + 28.0 34.7 0.05 53
special (A) AC12 0 68 122 100 + 2.62 3.86 0.10 58
20 23 147 62 26.1 72.0 0.26 7
40 15 148 67 44.7 130.0 0.28 5
standard 85/100 0 47 128 100 + 4.29 5.28 0.06 71
special (A) 20 18 146 22 41.0 98.8 0.20 12
40 18 149 22 59.8 156.0 0.23 7
control (F) 85/100 0 50 126 100 + 3.50 3.95 0.03 89
21 18 142 78 35.8 71.2 0.18 19
40 15 150 61 65.2 159.0 0.24 4
TABLE 5—Continued.
Tests
Microviscosity
Pave- Pene- Soft- Duc- Shear Micro-

Test ment tra- ening tility Megapoise Sus- duc-
Section Asphalt Age, tion Point 77 F pecti- tility
0
Road (source) Grade month 77 F deg F cm 0.05 s-> 0.001 s-i bility mm
special (A) Calif. 0 50 126 100 + 4.59 5.75 0.06 65

TS 19 22 144 73 25.9 61.2 0.21 10
38 19 149 20 40.7 106.0 0.25 5
special (A) Calif. 0 66 120 100 + 2.32 3.15 0.08 57
(low TS 18 31 136 100 + 14.5 23.6 0.12 19
viscosity) 38 23 146 58 31.2 82.5 0.23 7
Olivehurst special (C) Calif. 0 41 127 100 + 6.6 8.7 0.07 75
03-Yub-70,65 TS 15 29 135 100 + 15.7 21.7 0.13 43
33 23 145 68 30.0 87.8 0.24 8
special (A) Calif. 0 45 129 100 + 5.7 7.4 0.07 57
TS 15 30 137 100 + 15.1 28.1 0.15 22
33 23 144 61 29.7 74.5 0.23 6
control (C) Calif. 0 45 130 100 + 6.6 10.1 0.11 39
TS 15 31 143 51 23.6 80.0 0.31 7
33 25 145 45 24.8 70.7 0.26 7
Blythe 11-Riv- special (H) Calif. 6 40 126 100 + 6.1 7.72 0.06 76
R133.5/R149.5 TS 24 20 144 103 29.2 76.2 0.25 11
control (H) 200/300i 8 63 117 100 + 1.74 1.93 0.03 71
25 44 125 150 + 4.78 6.1 0.05 70
standard 85/100 4 29 136 100 + 11.6 11.6 0.00 91
special (M) 22 11 150 23 102.0 205.0 0.18 2
* TS = tentative specification,
VFS = very fast setting.
FS = fast setting.
MS = medium setting.
the regular 85/100 contract control asphalt, 85/100 asphalt from another
source meeting the standard specifications was placed as an additional test
section. As seen in Fig. 7, the various asphalts are weathering individually
somewhat differently, but the whole group of asphalts shows that the con-
ditions evident in this test section are affecting the group similarly.
In the Olivehurst and Blythe test sections, two additional sources of
special tentative-specification asphalt were employed. The Olivehurst sec-
tion compares two sources of tentative-specification asphalt, one of which
is the same as used in the Ludlow test section; while the special asphalt in
the Blythe test section is represented by yet another source. Figures 8 and 9
show how the different asphalts compare with each other in the Olivehurst
and Blythe test sections. The data for these three sections are not as con-
clusive, as the weathering period has been only about 30 months. This is
evident in the Blythe test section, as shown in Fig. 9, by the wide separation
1 10 VISCOSITY TESTING OF ASPHALT
40
SURFACE COURSE
30 —> SPECIAL (5.5) VERY FAST SET ASPHALT
—• SPECIAL ( 4 5) FAST SET ASPHALT
20- --o SPECIAL (3.5) MEDIUM SET ASPHALT
—• CONTROL 85/100 ASPHALT
2.'"
at in
£ in
— o
>- o
tuT
tn •
o £"
10 20 30 40 50
MONTHS FIELD WEATHERING
FIG. \~WUlits test section, RoadOl-Men-101.
between the weathering curves. Partly this can be attributed to the control
asphalt being a 200/300 material, but the special 85/100 control asphalt
is just as different from the tentative-specification asphalt as the 200/300
asphalt. Further weathering in this area should clarify the weathering
picture.
An analysis of all of the test sections shows that in each the control
asphalt is weathering at about the same rate as the special asphalt. Excep-
tions are the Willits and Blythe test sections where the control 85/100
10 20 30 40 50 60
FIG. 2—Folsom lest section, Road 03-Sac, ED-50.
CONTROL 85/100 ASPHALT

SPECIAL TENTATIVE
SPECIFICATION ASPHALT
0 10 20 30 40 50 60
FIG. Z—Benicia test section, Road W-Sol-21, 680.
paving asphalts are weathering at a faster rate. Another slight exception is

the special AC-12 asphalt, E, in the Ludlow test section which exhibits
much better ductlie properties than the other asphalts in this test section,
although its viscosity results do not indicate a much different weathering
pattern. With all the asphalts, special and control, weathering at about the
same rate in all the sections, there are some exceptions in which the rate
of weathering is greatly accelerated. In the Martinez test section, which
contains the same asphalts generally as the Folsom, Benicia, and Los Altos

— • SPECIAL TENTATIVE
10 20 30 40 50 60 70
FIG. 4—Martinez test section, Road 04-CC-4, 4.4/9.8 (surface course).
SURFACE COURSE

SPECIAL TENTATIVE
10 20 30 40 50 60 70
FIG. 5—Los Altos test section, Road 04-SCI-280, 11.5118.9.
test sections and has similar climatic conditions, the weathering has been
dramatically increased. The asphalts in the Martinez section have penetra-
tions of 10 and 12 for the control and special asphalts, where in the other
test sections the asphalts average penetrations of 22 and 23, respectively,
for the special and control asphalts for about the same weathering period.
Table 6 shows the results of visual observations for cracks and deflection
measurements. In the Martinez section we see that the road has extensive
block and chicken-wire cracking in both the special and control sections.
None of the other test sections exhibit any cracking at this time. In all the
i i i l S ;
100
80 /
60 /
"4>
£ QL 40 X _
_
a> v>
E m 20
— O
>- o
O ^ 10 - yy
/y
/^~ ' 1
O 1^- e Jy
_
> 6 /ft
7/fl ° o FOLSOM TEST SECTION -
• • BENICIA TEST SECTION
4 *7
I' " « MARTINEZ TEST SECTION -
/i a -o LOS ALTOS TEST SECTION
0
{ I > I ,
10 20 30 40 50 60 70
FIG. 6—Comparison of special asphalt D in various test sections {surface course).
KEM? ON DURABILITY TESTS AND FIELD HARDENING 113
TABLE 6—Asphalt durability test sections—deflection and visual inspection data.
Deflection Data
Visual Inspection
Wheel Track Data
Test
Section Asphalt Inner, Outer, Latest Cracks, Latest
Road (source) Grade0 in. in. Date etc. Date
Willits special (D) 5.5 VFS 0.006 0.005 4 69 some reflection cracking 7 71
Ol-Men-101 special (D) 4.5 FS 0.007 0.007 4 69 some reflection cracking 7 71
special (D) 3.5 MS 0.008 0.006 4 69 some reflection cracking 7 71
control (G) 85/100 0.008 0.006 4 69 some reflection cracking 7 71
Folsom special (D) TS 0.005 0.004 4 71 no cracking 5 71
03-Sac, ED-50 control (C) 85/100 0.004 0.004 4 71 no cracking 5 71
Benicia special (D) TS 0.013 0.013 2 71 no cracking 3 71
lO-Sol-21,680 control (B) 85/100 0.011 0.011 2 71 no cracking 3 71
Martinez special (D) TS 0.013 0.012 4 71 block and alligator 3 71
04-CC-4 control (B) 85/100 0.012 0.011 4 71 cracking both asphalts 3 71
Los Altos special (D) TS 0.001 0.001 3 71 no cracking ' 5 71
04-SC1-280 control (D) 85/100 0.001 0.001 3 71 no cracking 5 71
Ludlow special (E) AC-12 0.003 0.003 3 69 no cracking 1 71
08-SBd-40 special (A) AC-12 0.0014 0.0013 3 69 no cracking 1 71
(A) 85/100 0.0014 0.0013 3 69 no cracking 1 71
control (F) 85/100 0.0014 0.0013 3 69 no cracking 1 71
special (A) TS 0.001 0.0005 3 69 no cracking 1 71
special LV (A) TS 0.0015 0.0017 3 69 no cracking 1 71
Olivehurst special (C) TS 0.012 3 70 no cracking 5 71
03-Yub-70 special (A) TS 0.020 3 70 no cracking 5 71
control (C) 85/100 0.019 3 70 no cracking 5 71
Blythe special (H) TS no cracking 1 71
ll-Riv-10 control (H) 200/300 no cracking (extensive 1 71
bleeding)
(M) 85/100 no cracking 1 71
° TS = tentative specification.
VFS = very fast setting.
FS = fast setting.
MS = medium setting.
100 I I I i
80 '
60
,- ^ ^
.- ^—*G'
•
//
- •''•'/
.'^^^^
>'//
E in
—O -- / / / -
-!/'./ -° SPECIAL AC-12 ASPHALT "E"
TI / «—-• SPECIAL AC-12 ASPHALT " A " '_

°t t, -»
a
85/100 ASPHALT "A"
Q CONTROL 85/100 ASPHALT "F"
o I"-
(o ~
— o
w
7 - 0 SPECIAL TENTATIVE SPECIFICATION"
ASPHALT "A"
••» SPECIAL TENTATIVE SPECIFICATION
ASPHALT "A" (LOW VISCOSITY)
10 20 30 40 50 60
FIG. 7—Ludlow test section, Road 08-SBd-40, R 28.4/R 42.1 (surface course).
1 14 VISCOSITY TESTING OF ASPHALT
SURFACE COURSE
20
CONTROL 8VKX) ASPHALT 1

SPECIAL TENTATIVE
SPECIFICATION ASPHALT " v)
-i
-•> SPECIAL TENTATIVE
SPECIFICATION ASPHALT 'A"!
0 10 20 30 40 50 60 70
FIG. %—Olivehurst test section, Road 03- Yub-70, 65.
test sections the deflection values are low, even in the Martinez test section.
This indicates that the structural sections were well designed as there is no
instability, although there is extensive bleeding in various portions of the
control 200/300 asphalt of the Blythe test section. No bleeding is evident
in the tentative specification asphalt section or the special 85/100 asphalt
section, even though they used the same mix design and asphalt content as
the 200/300 asphalt. No bleeding is evident in any of the other test sections.
i i i i
100 ?
80 — t
// o o CONTROL 200/300 ASPHALT _

-\
60
• > SPECIAL TENTATIVE
/ SPECIFICATION ASPHALT
/ * -» 85/100 ASPHALT
/
/•
o 'd 20 /
V V)
E m
— o / /
10
ft
J
- /
/
6 - / -
4
- y^ -
2 - / -
10 20 30 40 50 60
FIG. 9—Blythe test section, Road ll-Riv-10, R 133.51R 149.5 (surface course).
KEMP ON DURABILITY TESTS AND FIELD HARDENING 1 15
Correlation of Laboratory and Field Test Results

The basic purpose of this research effort was to determine if laboratory
test procedures can reliably predict field asphalt durability. Since change in
consistency with aging is a recognized parameter of asphalt durability, the
viscosities obtained from field samples after 30 and 50 months were cor-
related with the viscosity results of various laboratory aging tests. The
values from the Rostler ratio, Heithaus procedure, and percent original
voids, of course, were not in viscosity terms. In all instances the viscosity
values are in megapoises and were determined with the sliding-plate micro-
viscometer at 0.05 s~1 shear rate.
It was a very difficult job to select the statistical methods that would
provide the most meaningful results. The field data are from eight different
test sections with different asphalts, asphalt contents, gradings, compaction,
climatological conditions, etc. To rate all the different asphalts under
identical conditions is not possible. This is not necessarily a bad situation
in that asphalt specifications should be written with the intent that the
asphalt will be used under a variety of conditions.
Two statistical programs were used in the correlation, with each having
advantages and disadvantages.
1. Least-Squares Regression Analysis—A linear transformation of the
power function (Y + AXb) and the exponential function (Y = Aebx) was
utilized to determine the better regression. The power function was used
when viscosities were plotted against viscosity, and the exponential func-
tion was used when field viscosity was plotted against a linear number
such as percent of voids, Rostler ratio, or Heithaus P. All the data were
correlated on the basis that all test sections together represent a single test
section. Table 7 lists the correlation coefficients and standard error of
estimate for the various test methods when comparing 30 and 50 months
of field weathering. Figure 10 presents a graphic comparison of the correla-
tion coefficients.
2. Covariance Analysis—This analysis is outlined by Dixon and Massey
[17]. In this system the individual correlations of the test sections are com-
pared based on a common slope. This analysis tends to negate individual
differences of the separate test sections such as voids, asphalt content, and
geographical location. The correlation coefficients are given in Table 8 and
graphically shown in Fig. 11.
Both statistical programs were performed on data from all eight test
sections at 30 months field service life and also on five test sections where
the service life exceeded 50 months. The five test sections at 50 months
were also analyzed at 30 months for comparison.
In comparing the correlation coefficients by the different statistical
methods, we see that none of the correlation values will sufficiently indicate
TABLE 7—Correlation coefficients (least-squares regression analysis) of field

viscosities at 30 and 50 months versus various laboratory procedures test data.
Field No. of Standard Error

Weather- Test of Estimate
Test Method ing Sec- Correlation
Designation Period tions Coefficient Laboratory Field
Field density briquette, 90 30 months 3 0.34 5.8 11.0
days
Percent original voids 8 0.39 1.52 28.6
Rostler ratio, (N + A,)/ 8 0.51 0.23 18.0
(P + A,)
Weathering plates, 210 F, 8 0.58 21.7 28.9
24 h
Extended RTF, 325 F, 5 h 8 0.59 6.3 27.5
Chevron R M F C , 48 h 8 0.64 10.5 15.8
Heithaus P 8 0.65 0.75 17.4
Ottawa sand mix, 140 F, 8 0.65 12.9 26.3
400h
Ottawa sand mix, 140 F, 8 0.72 43.8 26.6
1000 h
Weathering plates, 210 F, 30 months 5 0.57 27.0 23.8
24 h
Extended RTF, 325 F, 5 h 5 0.66 5.1 23.7
Rostler ratio, (N + A{)/ 5 0.67 0.26 21,4
(P + A2)
Ottawa sand mix, 140 F, 5 0.71 12.8 21.4
400 h
Ottawa sand mix, 140 F, 5 0.73 53.8 22.0
1000 h
Heithaus P 5 0.73 0.82 21.4
Chevron R M F C , 48 h 5 0.81 9.7 18.5
Weathering plates, 210 F, 50 months 5 0.46 28.2 84.0
24 h
Extended RTF-325 F, 5 h 5 0.50 5.6 84.7
Rostler ratio, (N + AJ/ 5 0.55 0.29 82.0
(P + A,)
Ottawa sand mix, 140 F, 5 0.57 14.1 81.9
400h
Ottawa sand mix, 140 F, 5 0.59 58.6 82.5
1000 h
Heithaus P 5 0.60 0.96 81.9
Chevron R M F C , 48 h 5 0.72 11.4 75.6
NOTE—Y = Aebx used for Heithaus P, percent voids, and Rostler ratio.
Y = AXh used in all other correlations.
that any of the laboratory test procedures could be depended upon to

predict adequately field weathering durability. One possible exception is
the correlation between the Chevron RMFC versus the field data for eight
test sections at 30 months by covariance analysis. In regard to the correla-
tion coefficients determined on the 50 months data, we see that the best
correlation is by the percent of original voids no matter what statistical
TABLE 8—Correlation coefficients (linear covariance analysis) offieldviscosities

at 30 and 50 months versus various laboratory procedures test data.
Field Weathering No. of Test Correlation

Test Method Designation Period Sections Coefficient
Weathering plates, 210 F, 24 h 30 months 8 no analysis

Percent original voids 8 0.39
Rostler ratio, (N + AJftP + A2) 8 0.51
Extended RTF, 325 F, 5 h 8 0.59
Heithaus P 8 0.64
Ottawa sand mix 140 F, 400 h 8 0.67
Ottawa sand mix 140 F, 1000 h 8 0.72
Field density briquette, 90 days 3 0.83
Chevron RMFC, 48 h 8 0.93
Chevron RMFC, 48 h 30 months 5 no analysis
Weathering plates, 210 F, 24 h 5 0.56
Rostler ratio, (N + A,)/(P + A2) 5 0.67
Ottawa sand mix, 140 F, 400 h 5 0.71
Heithaus P 5 0.73
Weathering plates, 210 F, 24 h 50 months 5 no analysis
Rostler ratio, (N + A,)/(P + A*) 5 0.55
Heithaus P 5 0.60
Chevron RMFC, 48 h 5 0.72
program is used. Therefore, to measure adequately the efficiency of the

test methods, a field comparison under relatively uniform void conditions
is desirable.
In analyzing the two statistical programs we see that the least-squares
regression analysis is encompassing all effects, including test-section vari-
ables such as grading, aggregates, voids, and climatic conditions. The co-
variance-analysis program offers a way to eliminate these test-section
variables and is probably the best program, except that each test section
should contain sufficient test asphalts to produce a reliable correlation
within each test section. In this study four of the five oldest test sections
contain only two asphalts, which makes the covariance analysis question-
able.
Findings To Date
1. All of the tentative specification asphalts graded on the basis of RTFO
residue viscosity had excellent "setting" properties regardless of supplier.
2. The superior qualities of the experimental asphalts used in the Willits
test section appear to be a combination of moderately low voids and good
asphalt durability properties.
s o I O A -ivNiaiao iN30«3d
««ie»-3dNH N0HA3H3
„d„ SnVHll3H
••"10001 - J o O H - W S 0
«JHOO»- i o O t l - ' N ' S ' O
!
V+d/'V+N 0I1VU d311S0M
l»i S - JoSZS - J l H Q3QN31X3
• ' 1 »Z - JoOIZ - S31V-|d 3NIM3H1V3M 13
s
3
•'I et-DJHH NOUA3H3

„d„ SHVH1I3H
< UJ -~?
•JMOOOI-JoOt-l (ISO
«•"( OOfr-doOH TI'S'O
SQIOA lVNIOiaO i.N3D«3d
'v+d/'v+N ouva H3nisow

«'<IS-doS2£-dlM Q30N31X3
«JM » 2 - J „ 0 I 2 - S 3 1 » 1 d 0NIM3H1V3M
«JM OOOI-doOH HSO

«J«00»-JoO»l HSO
„d„ S0VH1I3H €.•
"H 8V-DJHM u°"»10
>JH S - d o S Z £ - d l M
" M M-d,>0IZ-S3±Vld
"V + d / ' V + N 0I1VM
SaiOA
030N31X3
0NIM3HJ.V3M
M311S0M
"IVNISIdO %
S co
1!
(iuou»»g £ ) | .top Q6 - o m a A1ISN30 013l"d
_L _L c -2
o C
•3 i
(xqaV=<OB(qXV=A) SISA1VNV N0ISS3H93H SBHVflOS 1SV3~I)
iN3ioud300 Nouvn3ayoo
M
o a
i- E
— "1
| SOIOA 1VNIOIMO %
1 "1 a* -OJHd NOHA3HD
1 „d„ SnVHXI3H
1 ••"10001-dodW W'SO
1 •J<IO0»- d o 0W U S D
1 'V+d/'v+N OI1VH H3"US0M
I "I s-doszc J i « 1X3
* ••"1 » Z - d o 0 I Z d ' M
-1 „d„ S n » H H 3 H
•JUOOOI-JoOH W S O
v>
1 •ill 0 0 » - J o O f r l W S O
a.
1 J
SOIOA "IVNI9IU0 %
<>
/ 1 v+d/'v+N OIIVM asiisoH
>- 1 "US- doSZC i l U 1 X 3
< 1 «•"! » Z - J o O I Z d ' M
< • •JM 8fr - 3dWH N0«A3H3

o
z
| " < l 8 * O j m < NOMA3H3

| («U0|l3»s £ ) sXop 0 6 0IM8 0 i
1 «JM000l-do0»l WSO
1 IJH OOfr- JoO»l W S O
I „ d „ SPVH1I3H
1 !
« J M S - J o S 2 £ dlW 1 X 3
1 v+d/'v+N oiiva MS-USOU
L SOIOA I V N I O I d O %
• " 1 » Z - J o O I Z 'd'M
' 1 1 1 1 1 1 1 1
(SISA1VNV 30NVIHVA00 UV3Nn)

|
1N3I0IJ33O0 N0llV13da00
J,
Si
FIG. 12—Location of test sections.
3. In most of the test sections it appears that the special asphalts did
not weather significantly different than the control asphalts and that the
test-section conditions also dictated the weathering rates of the asphalts
involved.
4. The original void content of the pavement appears to be a definite
factor in long-term asphalt durability. The evidence was most indicative
at the 50 months weathering correlation.
5. All of the test sections were in good physical condition at 30+
months except for the Martinez test section, which is exhibiting extensive
block cracking in both the special and control-asphalt areas, and the con-
trol 200/300 asphalt portion of the Blythe project which is exhibiting
"bleeding."
6. The poorest correlation with field weathering, in general, was ob-

tained with the weathering plate test at 210 F (99 C) used in the tentative
specification. The best correlation appears to be with the Chevron RMFC
test at 48 h.
Conclusions
1. Grading asphalt on the basis of viscosity at 140 F (60 C) after a treat-
ment test that simulates field mixing (rolling thin-film test or thin film test)
will produce mixes that have better uniformity in "setting" during rolling.
This uniformity was consistent regardless of manufacturer.
2. Initial pavement air voids is an important factor in controlling early
asphalt hardening. Pavement air voids should be as low as possible without
sacrificing stability or causing pavement "bleeding."
3. More field weathering time is needed before final conclusions can be
made on the reliability of the various laboratory test methods for predict-
ing asphalt durability.
Acknowledgments
The author wishes to acknowledge the work of J. B. Skog in starting this
study to improve asphalt specifications. Acknowledgment is given C. A.
Frazier for supervising the statistical treatment. Special acknowledgment
is given N. H. Predoehl for his help in all phases of the work. Acknowledg-
ment is also given to the asphalt producers for their cooperation in furnish-
ing the tentative specification asphalts.
This work was done in cooperation with the U. S. Department of Trans-
portation, Federal Highway Administration. The opinions, findings, and
conclusions expressed in this report are those of the author and not neces-
sarily those of the Federal Highway Administration.
G. B. Sherman is the principal investigator, and M. H. Johnson and G. R.
Kemp are the co-investigators of the research project from which the infor-
mation contained in this report was obtained.
References
[]] Hveem, F. N., Zube, E., and Skog, J. B., Proceedings, Association of Asphalt Paving
Technologists, Appendix D, San Francisco, 1963, p. 318.
[2] Zube, E. and Skog, J. B., "Experimental Asphalt Test Section, Road 01-Men-l-E,
Wits.," Research Report 643185, State of California, Department of Public Works,
Division of Highways, Materials and Research Department, Feb. 1967.
[3] Zube, E., Skog, J. B., and Kemp, G., "Experimental Asphalt Test Section, Road
03-Sac, ED-50," Research Report 643229, State of California, Department of Public
Works, Division of Highways, Materials and Research Department, Aug. 1966.
lO-Sol-21, 680," Research Report 643229, State of California, Department of Public
Works, Division of Highways, Materials and Research Department, Aug. 1966.
04-CC-4-4.4/9.8," Research Report 643229, State of California, Department of Public
Works, Division of Highways, Materials and Research Department, Sept. 1967.
04-SC1-280-11.5/18.9," Research Report 643229, State of California, Department of
Public Works, Division of Highways, Materials and Research Department, Aug. 1967.
[7] Zube, E. and Skog, J. B., "Experimental Asphalt Test Section, Road 08-SBd-
R28.4/R42.1," Research Report 633134, State of California, Department of Public
Works, Division of Highways, Materials and Research Department, June 1968.
03-Yub-70-0.1/R8.9," Research Report 633134, State of California, Department of
Public Works, Division of Highways, Materials and Research Department, March
1968.
[9] Sherman, G. B., Kemp, G. R., and Predoehl, N. H., "Experimental Asphalt Dura-
bility Test Section, Road ll-Riv-10-R133.5/R149.5," Research Report 633134, State
of California, Department of Public Works, Division of Highways, Materials and
Research Department, Sept. 1971.
[10] Skog, J. B., "Setting and Durability Studies of Paving Grade Asphalts," Proceedings,
Association of Asphalt Paving Technologists, Denver, 1967.
[II] Skog, J. B. and Sherman, G. B., "Viscosity Grading for Asphalt Cements," Research
Report 633134, State of California, Department of Public Works, Division of High-
ways, Materials and Research Department, presented at 50th Annual Meeting of the
Highway Research Board, Jan. 1971.
[12] Skog, J. B., "The Operation, Control, and Application of the Infrared Weathering
Machine—California Design," Road and Paving Materials, ASTM STP 212, American
Society for Testing and Materials, 1957.
[13] Proposed Method of Test for: Aging Index of Bituminous Materials, 1964 Annual
Book of ASTM Standards, American Society for Testing and Materials, p. 665.
[14] Materials Manual, California Test Method 346-D, State of California, Department of
Public Works, Division of Highways.
[75] Rostler, F . S. and White, R. M., "Composition and Changes in Composition of
Highway Asphalts, 85-100 Penetration Grade," Proceedings, Association of Asphalt
Paving Technologists, New Orleans, 1962.
[16] Heithaus. J. J., "Measurement and Significance of Asphaltene Peptization," Sym-
posium on Fundamental Nature of Asphalt, sponsored by Division of Petroleum
Chemistry of the American Chemical Society, New York, Vol. 5, No. 4-A, Sept. 1960.
[17] Dixon, W. J. and Massey, F. J., Jr., Introduction to Statistical Analysis, 2nd ed.,
McGraw-Hill, Chapter 12.

Viscosity Testing of Asphalt and Experience With Viscosity Graded Specifications STP532-EB.29769

Uploaded by

Document Information

Original Title

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

Viscosity Testing of Asphalt and Experience With Viscosity Graded Specifications STP532-EB.29769

Uploaded by

Copyright:

Available Formats

Copyright by ASTM Int'l (all rights reserved); Tue Feb 20 02:01:08 EST 2018

ASTM SPECIAL TECHNICAL PUBLICATION 532

AMERICAN SOCIETY FOR TESTING AND MATERIALS

Printed in BaltinK>re, Md.

Printed in New Carlisle, Obio

Rapid Test Methods for the Determination of Bitumen

Asphalt Viscosity-Related Properties of In-Service Pavements in Pennsyl-

Copyright 1973 b y A S F M International www.astm.org

Rheology of Asphalt Cements at 60 C (140 F)

REFERENCE: Schweyer, H. E., Moore, J. P., and Ling, J. K. "Rhelogy of Asphalt

KEY WORDS: asphalts, viscosity, pressure, Florida capillary method, Cannon-

As theological studies of asphalt cements have become more sophisti-

Copyright 1973 b y A S T M International www.astm.org

using shift factors that showed flattening of the viscosity-temperature

FIG. 1—Temperature susceptibility of certain asphalts.

TABLE 1—Asphalts studied.

P25 Susceptibility Parameters

1. R70-50 Arkansas 89 3.454 2.1169

TABLE 2- -Calculated viscosity variations with small temperature changes at 60 C

No. 1 Arkansas -3.454 59.5 (139.1) 2.55

variation in testing. They also give a quantitative measure of how asphalts

Manning Zeitfuchs Pressure

FIG. 2—Cannon instruments.

FIG. 3—Transient mode apparatus arrangement.

FIG. 4—Rheometer assembly schematic.

the Cannon-Pressure instruments there is some indication of shear thinning,

Avg. Air Blown Canadian

" Arithmetic average of upstream and downstream pressures.

As a matter of interest, certain limited data for measurements at 40 C

TABLE 5—Comparative viscosity data at 40 C (104 F) in kilopoises.

1. R70-50 62.0 0.96 79 74 74

FIG. 5— Measurement mode effects on V60, viscosity at 60 C (140 F).

TABLE 6—Comparative Florida capillary data at 60 C (140 F) at constant stress

Pressure No. 6 Asphalt No. 10 Asphalt

10 1.34 1.80 4.95 4.06 0.972 2.46 1.93

" Average of absolute upstream and downstream pressures.

for the transient mode of the Cannon-Pressure method. Furthermore, it

TABLE 7—Effect of pressure on viscosity at 60 C (140 F).

Cannon Florida Capillary at 1 Os" 1

2. R70-51 1.86 1.85 4.04 3.61 23.9

[5] Welborn, J. Y. and Halstead, W. J., Proceedings, Association of Asphalt Paving

Study of Asphalt Viscosity Testing at 6 0 F

REFERENCE: Evans, J. V. and Gorman, J. W., "Study of Asphalt Viscosity

In recent years, asphalt technology has emphasized tests for measuring

Copyright 1973 b y A S T M International www.astm.org

organized by the Asphalt Institute has conducted a cooperative study in

TABLE 1—Laboratories participating and instruments used in study.

Cone- Falling Controlled

American Oil Co. x

Each laboratory made seven measurements on each asphalt; three meas-

TABLE 2—Nominal viscosities of asphalts tested.

A 3.64 1265 5.3

TABLE 3—Sample handling procedures and loading sequences.

The seven measurements on each asphalt were run in r a n d o m order at

Data and Analysis

Table 4 summarizes the viscosities at 60 F at a shear rate of 0.05 s _ 1 as

1. Viscometer Instrument - Cone Plate, Medium Cone

3. Date of Test 2/2/66

4. Sample Material Designation (Circle)

5. Sample Handling Procedure (Circle) ( Rigorous ) Normal

6. Loading Sequence (Circle) ( Increasing) Decreasing Normal

7. Time of placement of sample in 275 F oven 1Q;45

9. Time of placement of equipment in 140 F oven 11:15