You are on page 1of 226

Effects of Laboratory Mixing Methods and RAP

Materials on Performance of Hot Recycled
Asphalt Mixtures
By Viet Hung Nguyen
Thesis submitted to the University of Nottingham
for the degree of Doctor of Philosophy
July 2009
To Quynh Ngoc
i
Abstract
The primary work reported in this thesis is concerned mainly with the effects of different
mixing methods and RAP materials on homogeneity and mechanical properties of hot
recycled asphalt mixtures. The recycled asphalt mixture conforms to the requirement of BS
4987-1 (2005) for dense bitumen macadam size 10 mm (DBM 10 mm). The proportion of
RAP in the recycled mixture is 40%. RAP materials are artificially aged and processed in
the laboratory to prevent the variability of RAP gradation, bitumen content, and the origin.
Laboratory RAP is also used to assure that every single RAP particle is an agglomerate of
RAP aggregate and binder.
The mixing procedures include Black Rock (BR), Complete Blending (CB), the SHRP
procedure, and a newly developed field simulation method (FS). The primary difference
between these methods is the mixing mechanism. The BR case implies the situation in
which there is completely no interaction between RAP and virgin binder. On the contrary,
RAP and virgin binder are fully interacted in the CB case. The mixing procedures for BR
and CB cases conform to those for conventional asphalt mixtures. However, the bitumen
for BR case is pure virgin bitumen. In addition, the bitumen for CB is the blend between
RAP and virgin binder. The RAP/virgin binder proportion is 4/6. In the SHRP method,
RAP is preheated at 110
o
C for two hours before being mixed with virgin aggregate and
binder for 2 minutes at 130
o
C. In the FS method on the contrary, the mixing procedure
duplicates what occurs in the asphalt mixing plant. RAP is mixed with superheated virgin
aggregate (215
o
C) for different durations before this combination is blended with virgin
bitumen for 2 minutes at 130
o
C. The RAP/superheated virgin aggregate mixing duration
starts from short mixing time where RAP still exists at approximately original size and
gradually increases until the change in RAP lump size is insignificant. Depending on the
size of RAP used, RAP/superheated virgin aggregate mixing duration varies from 1 to 8
minutes.
The homogeneity of hot recycled asphalt mixture is examined by using virgin binder with a
different colour from that of RAP binder. The colour of virgin binder is obtained by mixing
clear binder (Shell Mexphalt C 160/220 Pen) with iron oxide pigment. The proportion of
pigment is 10% by weight of the binder making this binder red. The use of virgin binder
ii
with different colour from that of RAP binder helps to clearly differentiate the locations of
RAP and virgin materials. Surfaces of slices cut from compacted recycled specimens are
photographed by digital camera. The analysis of these surfaces in vertical order allows the
locations of RAP material to be qualitatively identified in a 3D manner.
Stiffness modulus values of samples for homogeneity assessment are also determined by
indirect tensile stiffness test. The stiffness test is carried out in four directions along the
circumference of each specimen with 45
o
angular increments. The experimental results
show that the stiffness measurement in four directions can indicate the heterogeneity of
recycled mixture. The variation in stiffness values in different measured directions will be
substantial for heterogeneous mixtures and minor in the case where recycled mixtures are
homogeneous. The results indicate there are mutual relations between mixing effort,
homogeneity, and stiffness values of recycled asphalt mixtures. The longer mixing time
will enhance the homogeneity and reduce the variation in stiffness values of recycled
mixture. In addition, as more RAP and virgin binder are incorporated, the stiffness values
of recycled mixture generally increase once the mixing time is extended.
As the clear binder is dyed red by 10% by weight of iron oxide, the proportion of the
pigment certainly alters the flow characteristic of binder. This might affect the mixing
process and rejuvenating effect between virgin and aged binder. Therefore, the effects of
mixing methods and RAP sizes on mechanical performance of hot recycled asphalt
mixtures are further investigated using normal straight run bitumen 160/220 Pen as virgin
binder. The assessment indicators include stiffness modulus, resistance to fatigue damage,
and resistance to permanent deformation.
The experimental results indicate that the conventional laboratory mixing method (SHRP)
tends to overestimate the mechanical properties of recycled asphalt mixture. The long RAP
preheating time that never exists in the industry coincidentally enhances the reaction
between RAP and virgin binder. The long RAP preheating time also slightly alters the
properties of RAP binder.
For the FS method, the increase in mixing duration significantly improves the homogeneity
level of recycled mixtures. The homogeneity level is also substantially affected by the size
of RAP material. For the same mixing effort, the mixtures comprised of small RAP are
iii
generally more homogeneous than those made from larger RAP. The more homogeneous
the mixture, the more interaction between RAP and virgin binder. Therefore, recycled
mixtures become stiffer and have better resistance to permanent deformation and fatigue
failure. A slightly linear increase in stiffness can result in an exponential increase in fatigue
life of the recycled mixture.
The mechanical properties including stiffness modulus, resistance to fatigue damage, and
resistance to permanent deformation of hot recycled asphalt mixtures are not similar to
those of the BR or CB mixtures, even at the favourable condition where RAP is preheated
for 2 hours at 110
o
C in the SHRP method and 8 minutes mixing duration in the FS method.
This implies that RAP does not act as Black Rock. In addition, the assumption that RAP
and virgin binder are fully blended also never exists in the recycled asphalt production
process.
Keywords: Reclaimed asphalt pavement (RAP), recycled, asphalt, mixtures, mechanical
properties, homogeneity
iv
Acknowledgements
This research project has been sponsored by the Government of Vietnam. The financial
support by Vietnamese government is gratefully acknowledged.
I owe my deepest gratitude to my supervisor, Professor Gordon Airey, for his invaluable
advice, encouragement, and supervision. This project would never have been completed
without his guidance and support.
I would like to show my sincere thanks to Dr Nicholas Thom, Dr Salah Zoorob, Dr James
Grenfell, and Professor Stephen Brown in Nottingham Transportation Engineering Centre
for their expert advice and support. Thanks also are due to Professor Cees van der Eijk in
the Methods and Data Institute for the discussion and advice on data analysis.
Acknowledgements must also be made to the Nottingham Transportation Engineering
Centre for providing necessary laboratory equipment, Dene limestone quarry in terms of
free aggregate supply, and Shell Bitumen for free of charge bitumen during the course of
the research.
I also would like to acknowledge the assistance and friendship of all current and past
lecturers, researchers, technicians, and administrative staff in the Nottingham
Transportation Engineering Centre, Department of Civil Engineering, and Faculty of
Engineering, especially those involved with the Pavement Research Group. Their
enthusiasm and encouragement have made my research enjoyable and memorable.
Special thanks also go to my host family, Mr Giles Harvey and his wife, Perdita, for
understanding and caring.
Finally, I am particularly grateful to my family, mom, dad, my brother, and especially my
wife, for their unconditional, patience, and continued support throughout the research
period.
v
Declaration
The work described in this thesis was carried out at the Nottingham Transportation
Engineering Centre, Department of Civil Engineering, The University of Nottingham
between April 2006 and April 2009. The thesis is the result of my own work, except the
work from others that has been specifically referenced. No part of this thesis has been, or
currently is, submitted for any degree, diploma, or other qualifications.
Viet Hung Nguyen
July 2009
vi
Table of Contents
Abstract ....................................................................................................................................i
Acknowledgements................................................................................................................iv
Declaration.............................................................................................................................. v
Table of Contents...................................................................................................................vi
List of Tables .........................................................................................................................ix
List of Figures........................................................................................................................xi
Glossary ...............................................................................................................................xiv
Glossary ...............................................................................................................................xiv
1 Introduction..................................................................................................................... 1
1.1 Background............................................................................................................. 1
1.2 Problem statement................................................................................................... 2
1.3 Research objectives................................................................................................. 4
1.4 Research methodology............................................................................................ 4
1.5 Scope of work ......................................................................................................... 5
2 Literature review............................................................................................................. 7
2.1 Asphalt mixture....................................................................................................... 7
2.1.1 Definition of asphalt mixture.......................................................................... 7
2.1.2 Classification of asphalt mixture .................................................................... 7
2.1.3 Properties of asphalt mixture .......................................................................... 8
2.2 Durability of asphalt mixture................................................................................ 11
2.2.1 Definition of durability ................................................................................. 11
2.2.2 Mechanism of ageing in asphalt mixture...................................................... 11
2.2.3 Factors affecting ageing mechanism............................................................. 14
2.2.4 Consequences of ageing in bituminous mixture ........................................... 17
2.2.5 Laboratory tests simulating field ageing....................................................... 19
2.3 Design methodology for recycling of bituminous pavements .............................. 21
2.3.1 Objectives and design procedure of pavement recycling ............................. 21
2.3.2 Methods for recycling bituminous pavements.............................................. 22
2.3.3 Selection of rejuvenators............................................................................... 23
2.3.4 Estimation of the consistency of the aged bitumen – modifier blend........... 25
2.4 Hot recycled mixture production .......................................................................... 31
2.4.1 Batch plant .................................................................................................... 31
2.4.2 Drum facility (Drum mixer).......................................................................... 34
2.4.3 RAP sizes used for production of recycled mixture ..................................... 37
2.5 Mixing mechanism................................................................................................ 38
2.5.1 Mechanical mixing........................................................................................ 38
2.5.2 Diffusion process or chemical mixing.......................................................... 47
2.6 Segregation and consequences.............................................................................. 62
2.6.1 Segregation ................................................................................................... 62
2.6.2 Consequences of segregation on the performance of asphalt mixture.......... 65
3 Laboratory RAP production.......................................................................................... 68
3.1 Introduction........................................................................................................... 68
3.2 Materials ............................................................................................................... 68
3.2.1 Bitumen......................................................................................................... 68
3.2.2 Aggregate...................................................................................................... 68
3.3 Procedure for RAP manufacture........................................................................... 70
3.4 Processing RAP .................................................................................................... 70
vii
3.5 Determine RAP properties.................................................................................... 72
3.5.1 RAP aggregate .............................................................................................. 72
3.5.2 RAP binder.................................................................................................... 72
4 Zero shear viscosity and the accuracy of viscosity mixing equations .......................... 74
4.1 Introduction........................................................................................................... 74
4.2 Experiments .......................................................................................................... 76
4.2.1 Materials ....................................................................................................... 76
4.2.2 Rheological testing........................................................................................ 77
4.3 Results and discussion .......................................................................................... 77
4.3.1 Zero shear viscosity (ZSV) ........................................................................... 77
4.3.2 Efficiency of viscosity mixing equations...................................................... 80
5 Effects of laboratory mixing methods on the homogeneity of hot recycled asphalt
mixture .................................................................................................................................. 86
5.1 Introduction........................................................................................................... 86
5.2 Development of laboratory mixing protocol......................................................... 86
5.2.1 Estimation of superheated temperature of virgin aggregate ......................... 87
5.2.2 Determine the superheated virgin aggregate/RAP mixing duration............. 88
5.2.3 Determine the mixing temperature ............................................................... 90
5.3 Method for segregation evaluation ....................................................................... 90
5.4 Specimens preparation.......................................................................................... 92
5.4.1 Materials ....................................................................................................... 92
5.4.2 Mixing procedure.......................................................................................... 92
5.4.3 Compacting procedure.................................................................................. 93
5.4.4 Machining specimens for segregation assessment........................................ 93
5.5 Results and analysis .............................................................................................. 93
5.5.1 Visual assessment for segregation ................................................................ 93
5.5.2 Mechanical assessment ............................................................................... 108
5.6 Summary............................................................................................................. 112
6 Effects of mixing procedures and RAP materials on stiffness distribution of hot
recycled asphalt mixtures.................................................................................................... 115
6.1 Introduction......................................................................................................... 115
6.2 Experiment design .............................................................................................. 115
6.2.1 Effects of mixing protocols and RAP sizes on stiffness of hot recycled
asphalt mixture............................................................................................................ 115
6.2.2 Effect of mixing equipment on stiffness distribution of hot recycled asphalt
mixtures ..................................................................................................................... 116
6.2.3 Effect of mixing protocols on RAP binder properties ................................ 118
6.3 Materials and specimens manufacture................................................................ 119
6.3.1 Material preparation and mixing procedure................................................ 119
6.3.2 Compaction................................................................................................. 122
6.3.3 Machining and storage of specimens.......................................................... 122
6.4 Assessment method............................................................................................. 122
6.5 Results and analysis ............................................................................................ 124
6.5.1 Air void contents......................................................................................... 124
6.5.2 Effects of mixing time on stiffness ............................................................. 127
6.5.3 Effects of RAP sizes on stiffness................................................................ 145
6.5.4 Effects of mixing methods on stiffness....................................................... 150
6.5.5 Effect of mixing equipment on stiffness..................................................... 154
6.5.6 Effects of mixing methods on RAP binder properties................................ 160
6.6 Summary............................................................................................................. 162
viii
7 Effects of mixing methods on fatigue life of hot recycled asphalt mixtures .............. 164
7.1 Introduction......................................................................................................... 164
7.2 Materials preparation and testing plan................................................................ 164
7.3 Results and analysis ............................................................................................ 166
7.3.1 Control mixtures ......................................................................................... 166
7.3.2 Recycled mixtures....................................................................................... 167
8 Effects of mixing methods on permanent deformation of hot recycled asphalt mixtures
..................................................................................................................................... 173
8.1 Introduction......................................................................................................... 173
8.2 Materials preparation and testing plan................................................................ 173
8.3 Results and discussion ........................................................................................ 175
9 Conclusions and recommendations for future research.............................................. 180
9.1 Conclusions......................................................................................................... 180
9.2 Recommendations for future research ................................................................ 183
References:.......................................................................................................................... 185
Appendix A......................................................................................................................... 193
Appendix B......................................................................................................................... 204
ix
List of Tables
Table 1: Mechanism of bitumen aging (Traxler, 1963)........................................................ 13
Table 2: Industrial supercritical fractions and commercial rejuvenating agent properties
(Chaffin et al., 1997)..................................................................................................... 24
Table 3: Ageing index (AI) after TFOT and PAV ageing of recycled blend with different
rejuvenators (Chaffin et al., 1997) ................................................................................ 25
Table 4: Viscosity at 60
o
C of aged bitumen and softening agents (Chaffin et al., 1995)..... 28
Table 5: Aged bitumen - softening agent Grunberg interaction parameter G
12
(Chaffin et
al., 1995) ....................................................................................................................... 29
Table 6: RAP preheating temperature required in Drum Mixer (Brock and Richmond, 2005)
....................................................................................................................................... 37
Table 7: State DOT specification requirements for the use of reclaimed asphalt pavement
(RAP) in hot asphalt paving mixtures (United States Department of Transportation,
2007) ............................................................................................................................. 38
Table 8: Viscosity at 60
o
C (Poise) of RAP binder (McDaniel et al., 2000) ......................... 41
Table 9: Critical temperature and performance grades of virgin and recovered RAP binders
....................................................................................................................................... 42
Table 10: Properties of RAP binder (Carpenter and Wolosick, 1980) ................................. 48
Table 11: Grading of RAP aggregate (Carpenter and Wolosick, 1980) ............................... 48
Table 12: Properties of RAP bitumen (Noureldin and Wood, 1987) ................................... 53
Table 13: Test results on reclaimed staged-extraction of RAP (Noureldin and Wood, 1987)
....................................................................................................................................... 53
Table 14: Test results on reclaimed, staged-extraction, no virgin aggregate (Noureldin and
Wood, 1987) ................................................................................................................. 54
Table 15:Tests results on reclaimed, staged-extraction, virgin aggregate used (Noureldin
and Wood, 1987)........................................................................................................... 55
Table 16: Properties of RAP aggregate (Huang et al., 2005) .............................................. 56
Table 17: Properties of Virgin Aggregate (Huang et al., 2005)............................................ 56
Table 18: Approximate absorption wave length of functional groups (Petersen, 1986) ...... 59
Table 19: Gradations, bitumen contents, and target air voids of laboratory-simulated
segregation mixtures (Gardiner and Brown, 2000)....................................................... 66
Table 20: Summary of the influence of segregation on mixture properties (Gardiner and
Brown, 2000) ................................................................................................................ 66
Table 21: Properties of bitumen 40/60 Pen........................................................................... 68
Table 22: Gradation of Dene limestone aggregate ............................................................... 69
Table 23: Properties of Dene limestone aggregate ............................................................... 70
Table 24: Gradation of processed RAP materials................................................................. 71
Table 25: Properties of RAP binder...................................................................................... 73
Table 26: Properties of bitumens .......................................................................................... 76
Table 27: Mix A - Regression analysis between experiment and predicted values using
different viscosity mixing equations at different temperatures..................................... 84
Table 28: Mix B - Regression analysis between experiment and predicted values using
different viscosity mixing equations at different temperatures..................................... 84
Table 29: G12 parameter of Mix A and B at different temperatures.................................... 85
Table 30: Stiffness versus different RAP/superheated virgin aggregate mixing duration.. 109
Table 31: Mean stiffness versus RAP/superheated virgin aggregate mixing duration....... 112
Table 32: Test plan to study the effects of mixing method on stiffness ............................. 116
Table 33: Test plan to study the effects of mixer on stiffness of recycled asphalt ............. 118
x
Table 34: Test plan to study the effects of mixing methods on binder properties.............. 118
Table 35: Properties of bitumen 160/220 Pen and 70/100 Pen........................................... 119
Table 36: Air void content of recycled mixture manufactured by different mixing method
..................................................................................................................................... 124
Table 37: Air void comparison by t-test ............................................................................. 126
Table 38: Summary of stiffness values (MPa) of recycled mixtures manufactured by
different mixing methods............................................................................................ 127
Table 39: Stiffness comparison by t-test............................................................................. 128
Table 40: Stiffness values of LR FS-2 specimens .............................................................. 134
Table 41: Stiffness values of LR FS-4 specimens .............................................................. 134
Table 42: Stiffness values of LR FS-6 specimens .............................................................. 135
Table 43: Stiffness values of LR FS-8 specimens .............................................................. 135
Table 44: Stiffness values of CB specimens....................................................................... 136
Table 45: Stiffness values of SR FS-2 specimens .............................................................. 149
Table 46: Stiffness values of LR FS-4 specimens .............................................................. 149
Table 47: Air void summary of recycled specimens manufactured by different method and
equipments .................................................................................................................. 154
Table 48: Stiffness (MPa) of recycled specimens manufactured by different methods and
equipment.................................................................................................................... 155
Table 49: Stiffness values of LR FS-2 specimens – Mixer B............................................. 157
Table 50: Stiffness values of LR FS-6 specimens – Mixer B............................................. 157
Table 51: Parameters of fatigue equation at 95% confidence of control and recycled asphalt
mixtures manufactured by different mixing methods................................................. 168
Table 52: Extrapolated fatigue life at 100 microstrain of recycled asphalt mixtures
manufactured by different mixing methods................................................................ 169
Table 53: Test plan to study the effects of different mixing methods on resistance to
permanent deformation............................................................................................... 173
Table 54: Permanent deformation data of control and recycled specimens manufactured by
different mixing methods............................................................................................ 176
Table 55: Rutting indicator data of control and recycled specimens manufactured by
different mixing methods............................................................................................ 179
xi
List of Figures
Figure 1: Comparison on gradation of gap-graded and continuous-graded asphalt mixture
(Read, 1996).................................................................................................................... 8
Figure 2: Strain response due to applied stress of Visco-Elasto-Plastic Constitutive model
(Perl et al., 1983)............................................................................................................. 9
Figure 3: Visco-elastic response to millions of load application (Read, 1996) .................... 10
Figure 4: Relationship between the temperature of the mixture and change in softening
point (Read and Whiteoak, 2003) ................................................................................. 15
Figure 5: The effect of void content on the hardening of bitumen on the road (Read and
Whiteoak, 2003)............................................................................................................ 16
Figure 6: Asphalt film thickness vs viscosity after short term ageing (Kandhal and
Chakraborty, 1996) ....................................................................................................... 17
Figure 7: Asphalt film thickness versus viscosity after long term ageing (Kandhal and
Chakraborty, 1996) ....................................................................................................... 17
Figure 8: Effect of ageing time on bitumen viscosity extracted from pavements (Kandhal
and Koehler, 1984)........................................................................................................ 18
Figure 9: Phase angle versus ageing levels (Daniel et al., 1998).......................................... 19
Figure 10:Fracture temperature versus ageing time (Kliewer et al., 1996) .......................... 19
Figure 11: Nomograph for predicting 60
o
C viscosity of recycled asphalt (Davidson et al.,
1977) ............................................................................................................................. 26
Figure 12: Batch Plant (MAPH-2, 2000).............................................................................. 32
Figure 13: Drum mixer with RAP centre inlet (Brock and Richmond, 2005)...................... 34
Figure 14: RAP in parallel drum mixer with isolated area (Brock and Richmond, 2005) ... 35
Figure 15: RAP in parallel drum facility with added continuous mixer (Brock and
Richmond, 2005)........................................................................................................... 35
Figure 16: Double Barrel Drum facility (Brock and Richmond, 2005)................................ 36
Figure 17: Ideal mixture with homogeneity (Harnby et al., 2001) ...................................... 39
Figure 18: Mixture with “self-loving” particles (Harnby et al., 2001) ................................. 40
Figure 19: RAP preheating time versus indirect tensile and unconfined compression
strength (Stephens et al., 2001)..................................................................................... 46
Figure 20: Arizona RAP binder complex modulus versus preheating time and temperature
(McDaniel et al., 2000) ................................................................................................. 46
Figure 21: Florida RAP binder versus preheating time and temperature (McDaniel et al.,
2000) ............................................................................................................................. 47
Figure 22: Effect of diffusion on resilient modulus (Carpenter and Wolosick, 1980) ......... 49
Figure 23: Schematic of modifier coating an aggregate particle during recycling process
(Carpenter and Wolosick, 1980) ................................................................................... 50
Figure 24: Diffusion model (Carpenter and Wolosick, 1980) .............................................. 51
Figure 25: Penetrations of outer and inner layers as function of time (Carpenter and
Wolosick, 1980)............................................................................................................ 52
Figure 26: Layers extraction process (Huang et al., 2005) ................................................... 57
Figure 27: Viscosity at 135
o
C of different micro-layers coated RAP particles (Huang et al.,
2005) ............................................................................................................................. 57
Figure 28: Schematic of FTIR – ART (Karlsson and Isacsson, 2003a) ............................... 58
Figure 29: Spectra obtained after 10, 50, 100, 500, and 925 minutes during diffusion of
rejuvenator into A-B180 at 60
o
C (Karlsson and Isacsson, 2003a) ............................... 59
Figure 30: Schematic of Fick’s Law diffusion model (Karlsson and Isacsson, 2003b) ....... 60
xii
Figure 31: Influence of bitumen type on diffusion coefficient (Karlsson and Isacsson,
2003b) ........................................................................................................................... 61
Figure 32: Influence of temperature on diffusion coefficient (Karlsson and Isacsson, 2003b)
....................................................................................................................................... 61
Figure 33: Influence of chemical composition of markers on diffusion coefficient (Karlsson
and Isacsson, 2003b)..................................................................................................... 62
Figure 34: Design gradation of RAP aggregate.................................................................... 69
Figure 35: Appearance of RAP materials ............................................................................. 71
Figure 36: Gradation of RAP aggregate before and after processing................................... 72
Figure 37: Complex viscosity of bitumen 160/220 Pen versus different frequency at 60
o
C75
Figure 38: Complex viscosity bitumen 100/150 Pen versus different frequency at 60
o
C.... 75
Figure 39: Extrapolated ZSV of Mix A blends at temperature of 25
o
C ............................... 79
Figure 40: Complex viscosity versus frequencies of blend (20 % RAP and 80% 160/220
Pen) at 70
o
C .................................................................................................................. 80
Figure 41: Experiment and predicted viscosity using different viscosity mixing equations of
Mix A (Blends of different proportion of aged binder and 160/220 pen) at 20
o
C........ 81
Figure 42: Viscosity difference and G12 versus temperatures ............................................. 85
Figure 43: RAP size after 2 minutes RAP/superheated virgin aggregate mixing time ........ 89
Figure 44: RAP size after 8 minutes RAP/superheated virgin aggregate mixing time ........ 89
Figure 45: Images taken by X-Ray scanner and normal digital camera of Shell Mexphalte C
dyed by 10 % iron oxide and RAP binder .................................................................... 91
Figure 46: LR mixture – 2 minutes mixing time .................................................................. 97
Figure 47: LR mixture – 4 minutes mixing time .................................................................. 98
Figure 48: LR mixture – 6 minutes mixing time .................................................................. 99
Figure 49: LR mixture – 8 minutes mixing time ................................................................ 100
Figure 50: SR mixture – 1 minute mixing time .................................................................. 101
Figure 51: SR mixture – 2 minutes mixing time................................................................. 102
Figure 52: SR mixture – 4 minutes mixing time................................................................. 103
Figure 53: SR mixture – 6 minutes mixing time................................................................. 104
Figure 54: SR mixture – 8 minutes mixing time................................................................. 105
Figure 55: Large version of Figure 54 g............................................................................. 106
Figure 56: Large version of Figure 54 h............................................................................. 107
Figure 57: Stiffness measurement scheme.......................................................................... 108
Figure 58: Stiffness modulus versus air void content......................................................... 110
Figure 59: Small particles tend to move downward to the bottom during mixing process 111
Figure 60: Relation between mixing effort, homogeneity and mechanical properties of
recycled asphalt........................................................................................................... 113
Figure 61: Schematic of Mixer A....................................................................................... 117
Figure 62: Schematic of Mixer B........................................................................................ 117
Figure 63: Master-curves of 70/100 Pen and recycled blend with 40% RAP and 60%
160/200 Pen bitumen .................................................................................................. 120
Figure 64: Coring and cutting scheme for compacted slabs ............................................... 123
Figure 65: Stiffness versus core location and measuring direction of LR FS-2 mixture.... 130
Figure 66: Stiffness versus core location and measuring direction of LR FS-4 mixture.... 131
Figure 67: Stiffness versus core location and measuring direction of LR FS-6 mixture.... 132
Figure 68: Stiffness versus core location and measuring direction of LR FS-8 mixture.... 133
Figure 69: Stiffness versus mixing time of LR FS mixtures .............................................. 137
Figure 70: Stiffness distribution of control mixtures.......................................................... 139
Figure 71: Probability plot of stiffness values – control mixtures...................................... 140
Figure 72: Probability plot of stiffness values – LR FS mixtures....................................... 140
xiii
Figure 73: Stiffness histogram of LR FS-2 mixture ........................................................... 141
Figure 74: Stiffness histogram of LR FS-2 mixture – Stiffness grouping analysis............ 142
Figure 75: Stiffness probability plot of LR FS-2 – Stiffness grouping analysis................. 142
Figure 76: Inter-quartile stiffness ranges of LR FS-2 group 1 and BR mixture................ 143
Figure 77: Inter-quartile stiffness ranges of LR FS mixtures ............................................. 143
Figure 78: Inter-quartile stiffness ranges of SR FS mixtures ............................................. 144
Figure 79: Stiffness versus core location and measuring direction of SR FS-2 mixture.... 146
Figure 80: Stiffness versus core location and measuring direction of SR FS-4 mixture.... 147
Figure 81: Stiffness range of LR and SR recycled mixtures manufactured by FS methods
..................................................................................................................................... 148
Figure 82: Inter-quartile stiffness ranges of control and LR mixtures manufactured by
different methods ........................................................................................................ 150
Figure 83: Inter-quartile stiffness ranges of SHRP mixtures and FS-2 mixtures ............... 151
Figure 84: Inter -quartile stiffness ranges of control and SR mixtures manufactured by
different methods ........................................................................................................ 153
Figure 85: Stiffness versus core location and measuring direction of LR FS-2 mixture –
Mixer B....................................................................................................................... 158
Figure 86: Stiffness versus core location and measuring direction of LR FS-6 mixture –
Mixer B....................................................................................................................... 159
Figure 87: Complex modulus versus log reduced frequency of RAP binder before and after
processed by different mixing methods ...................................................................... 161
Figure 88: Phase angle versus log reduced frequency of RAP binder before and after
processed by different mixing methods ...................................................................... 161
Figure 89: Schematic of ITFT test ...................................................................................... 165
Figure 90: Fatigue lines of control mixtures....................................................................... 169
Figure 91: Fatigue lines of LR mixtures manufactured by different methods.................... 170
Figure 92: Fatigue lines of SR mixtures manufactured by different methods.................... 170
Figure 93: Fatigue lines with boundaries of 95% confidence interval of LR FS-2 and LR
FS-8 mixtures.............................................................................................................. 171
Figure 94: Fatigue life at 30 µm strain and stiffness versus different mixing time of LR
mixtures....................................................................................................................... 171
Figure 95: Fatigue life at 30 µm strain and stiffness versus different mixing time of SR
mixtures....................................................................................................................... 172
Figure 96: Relation between RAP size and fatigue life of recycled mixtures manufactured
by different mixing methods....................................................................................... 172
Figure 97: Schematic of RLAT test to determine resistance to permanent deformation.... 174
Figure 98: Permanent deformation versus number of loading application of LR FS-2
specimens.................................................................................................................... 178
Figure 99: Inter-quartile rutting indicator ranges of control and recycled specimens
manufactured by different mixing methods................................................................ 178
xiv
Glossary
Actual blending: or “Actual Practice” is the case when RAP is preheated before being
blended with virgin aggregate and rejuvenator
Aggregate: inert materials, for instance, sand, gravel, crushed stone, and slag for asphalt
concrete production
Asphalt mixture: the combination of aggregate and bitumen
Batch Plant: or Batch facility, is equipment designed to produce hot asphalt concrete in
batch
Bitumen: the residual product of fractional distillation process of crude oil, used as binder
in asphalt mixture
Black Rock: the situation when RAP is inert in the recycled mixture
Counter flow drum mixer: the drum mixer with the direction of flame counter to the
movement direction of aggregate
Diffusion: is the process when rejuvenator covers, incorporates with RAP binder and
recovers the properties of RAP binder
Double Barrel: a trademark of Astec Industry for a unique counter flow drum mixer
Drum mixer: or Drum facility, is a combination of drum dryer and mixer for hot asphalt
mixture production
Mechanical mixing: the effort of using mechanical effect to produce homogeneous mixture
of different ingredients
Parallel drum mixer: drum mixer with the direction of the flame is similar to that of
aggregate movement
RAP aggregate: the aggregate extracted from RAP
RAP binder: the aged bitumen extracted from RAP
RAP: acronym for reclaimed asphalt pavement or recycled asphalt pavement, the material
is crushed, milled from deteriorated pavement for recycling
Recycled mixture: the asphalt mixture that used RAP material, virgin aggregate and
rejuvenator
Rejuvenator: the materials which include modifiers, reclaiming, recycling, modifying,
softening agents, recycling modifiers, rejuvenators, fluxing oils, extender oils, aromatic
oils, and virgin binder which will be used to recover the properties of RAP binder
Segregation: the non-homogeneity of the mixture, or the locally high concentration of
certain type of material in the whole mixture. Segregation can be divided into size
xv
segregation, and chemical segregation in term of recycling when the complete blending
status between RAP binder and rejuvenator has not been reached
Total blending: or “complete blending”, is the case when RAP binder is extracted from
RAP, being mixed with rejuvenator. The blend of RAP and rejuvenator is then mixed with
RAP and virgin aggregate to produce the recycled mixture
Virgin binder: bitumen not previously used
Virgin aggregate: aggregate not previously used
1
1 Introduction
1.1 Background
Recycling pavements has been used for many years as a rehabilitation technique in the
highway industry. The first recorded asphalt pavement recycling project was in 1915 (Epps
et al., 1980). Since that moment, there has been a wide range of recycling methods
regarding the equipment and procedures. There is also a variety of materials used as
rejuvenator, for instance, soft bitumen, bitumen fractions, and also commercial recycling
agents. Karlsson and Isacsson (2003d) summarised the methods for recycling asphalt
pavement as follows:
In-plant asphalt recycling
In this method, reclaimed asphalt pavement (RAP) is mixed with new materials in the
mixing plant. Depending on the processing temperatures, in-plant recycling is divided into
hot recycling (above 120
o
C), warm recycling (70-120
o
C) and cold recycling (below 70
o
C).
In-place or In-situ asphalt recycling
The difference of in-place from in-plant asphalt recycling is that all the work is carried out
in the field. In-place recycling is divided into four sub-categories (Karlsson and Isacsson,
2003d):
 Remixing is a process in which approximately 30-50 mm thick of the
deteriorated asphalt surface is milled and scarified. The reclaimed material is
mixed with new materials before being paved and compacted.
 Repaving process is almost the same as remixing except no fresh asphalt is
added to the reclaimed material. The deteriorated pavement is heated and
milled before being spread over again on the road surface. If rejuvenator is
used, the reclaimed material is mixed with rejuvenator before repaved. In
addition, a new thin asphalt layer is put on top and two layers are compacted
together.
 Cold in-place recycling is a process in which bitumen emulsion, foamed
bitumen, or extremely soft bitumen is used as rejuvenator. The application of
this method is favourably used in the climatic regions with temperature
below 10
o
C.
2
 Full depth reclamation is a process that allows the reconstruction of the
whole pavement structure using existing pavement materials.
In some aspects, pavement recycling philosophy has certain advantages compared to
conventional pavement construction. These benefits of recycling pavements can be
summarized as follows (Kandhal and Mallick, 1997):
- Reduce costs of new construction and rehabilitation projects
- Being consistent with environmental sustainable development in terms of
conservation of energy, mineral aggregates, and bitumen binder
- Preservation of road geometry
- Reduce the construction time delay
Due to the benefits brought to society by pavement recycling techniques, up to this moment
80% of the asphalt pavements removed each year from widening and resurfacing projects
are put back on roads, roadbeds, shoulders and embankments. In the US, almost all the
States allow reuse of RAP in the surface course with percentage of RAP from between 10
and 30% (United States Department of Transportation, 2007). The popularity of recycling
asphalt pavements has made RAP the most recycled material in the US in terms of both
percentage and tonnage. Mike Acott, president of the National Asphalt Pavement
Association, reported that approximately 73 million tons of RAP were reused every year
(NAPA, 2004).
1.2 Problem statement
The quality of an asphalt mixture is primarily affected by the design method, materials and
production process. The design step will determine the type and proportion of materials in
the mixture. The production process assures the mixture has the level of homogeneity as
required in the design step by using proper mixing methods such as mixing temperature and
duration. However, recycled mixture is different from conventional asphalt mixture as the
input materials include RAP, which is a combination of aged binder and aggregate. Hence,
the problems encountered in the recycling asphalt industry include not only those found in
conventional asphalt but also issues associated with RAP sizes, mixing methods and
diffusion mechanisms.
In the design process, the proportion of RAP binder, type and amount of rejuvenator are
selected based on the viscosity mixing equations. The fundamental philosophy of these
3
equations is that RAP binder and rejuvenator binders are completely blended. In addition,
the output and input of these viscosity mixing equations rely primarily on just the viscosity
values and proportions of each bitumen constituents. The other issue is whether the results
of these viscosity mixing equations are reliable, especially when two bitumen binders with
complicated chemical composition are mixed together.
Even when the results of these viscosity mixing equations are accurate, the other problem is
whether the complete blending between RAP binder and rejuvenator assumed in these
equations actually occurs in recycled asphalt mixture. In the laboratory, RAP binder is
extracted and recovered before being deliberately blended with rejuvenator without any
interventions. On the contrary, the mixing process between RAP binder and rejuvenator in
the asphalt mixing plants is affected by many factors such as the presence of aggregate,
filler, existence of RAP materials as agglomerates, mixing temperature, and efficiency of
the mixer.
The methods of preparing the recycled mixtures in the laboratory also indicate a
shortcoming as these methods could not represent the mixing mechanism in the field. In the
laboratory, RAP material is conventionally preheated for a long time before being mixed
with virgin aggregate and rejuvenator. The long preheating time might coincidently soften
RAP and enhance the mixing between RAP and virgin materials. On the contrary, RAP
material at ambient temperature is mixed with superheated virgin aggregate in the plant
mixer for a really short time. The question here is whether the short mixing duration in the
plant mixer can produce the recycled mixture with the level of homogeneity similar to that
of the laboratory procedure product. If the complete blending cannot occur in the industry
but only in the laboratory, the laboratory procedure has overestimated the properties of the
recycled mixture.
Size of RAP is also a problem as the bigger the size of RAP, the longer the time for the heat
to penetrate and break the RAP materials into separated pieces. Although there is a wide
range of RAP sizes handled in the asphalt pavement recycling industry, for instance, the
maximum RAP size is 50 mm, and even 75 mm is allowed in asphalt hot recycling process
(United States Department of Transportation, 2007), up to this moment, there is a lack of
research to investigate the effect of RAP sizes on the properties of recycled mixtures.
4
1.3 Research objectives
The objectives of this research are as following:
- To better understand the hot recycling asphalt technique.
- To develop a protocol to prepare the hot recycled mixture in the laboratory that
duplicates the production mechanism in the industrial asphalt mixing plant.
- To investigate the effects of mixing methods and RAP materials on homogeneity of
hot recycled asphalt mixtures.
- To investigate the effect of mixing methods and RAP materials on mechanical
properties of recycled mixture including stiffness modulus, resistance to permanent
deformation, and resistance to fatigue damage.
- To correlate the homogeneity and mechanical properties of hot recycled asphalt
mixture.
1.4 Research methodology
To study the effects of mixing process on the properties including homogeneity and
mechanical performance, recycled asphalt mixtures are manufactured by different mixing
procedures. The mixing procedures include black rock (BR), complete blending (CB), the
SHRP procedure, and a newly developed field simulation method (FS). The primary
difference between these methods is the mixing mechanism.
The BR case implies the situation in which there is no interaction between RAP and virgin
binder. On the contrary, RAP and virgin binder are fully interacted in CB case. The mixing
procedures for BR and CB methods conform to those of conventional asphalt mixtures.
However, the binder for BR case is pure virgin bitumen. On the contrary, the binder for CB
case is the blend between RAP and virgin binder. The RAP/virgin binder proportion is 4/6.
In the SHRP method, RAP is preheated at 110
o
C for two hours before being mixed with
virgin aggregate and binder. In FS method on the contrary, the mixing procedure duplicates
what occurs in the asphalt mixing plant. RAP is mixed with superheated virgin aggregate
(215
o
C) for different durations before this combination is blended with virgin bitumen. The
RAP/superheated virgin aggregate mixing duration starts from short mixing time where
RAP still exists at approximately original size and gradually increases until the change in
RAP lump size is insignificant.
5
In addition, two RAP sizes are used to study the effects of RAP materials on properties of
hot recycled asphalt mixture. RAP material is artificially aged and processed in the
laboratory. This is to prevent the variability of RAP aggregate gradation, binder content and
origin. In addition, the use of artificial RAP is used to assure that every single RAP particle
is an agglomerate of RAP aggregate and binder.
The homogeneous level of hot recycled asphalt mixture is examined by positioning the
locations of RAP and virgin materials. The use of virgin binder with a different colour from
that of RAP binder helps to clearly differentiate the locations of RAP and virgin materials.
Surfaces of slices cut from compacted recycled specimens are photographed by digital
camera. The analysis of these surfaces in vertical order allows qualitative identifying of the
locations of RAP material in a 3D manner. The virgin binder is obtained by mixing clear
binder (Shell Mexphalt C 160/220 Pen) with iron oxide pigment. The proportion of pigment
is 10% by weight of the binder making this binder red. Stiffness modulus values of samples
for homogeneity assessment are also determined by indirect tensile stiffness test. This is
aimed to correlate the homogeneous level and mechanical properties of recycled hot asphalt
mixtures.
As the clear binder is dyed red by 10% by weight of iron ioxide, the proportion of the
pigment alters the flow characteristic of the binder. This might affect the mixing process
and rejuvenation between virgin and aged binder. Therefore, the effects of mixing methods
and RAP sizes on mechanical performance of hot recycled asphalt mixtures are further
investigated using normal straight run bitumen 160/220 Pen as virgin binder. The
assessment indicators include stiffness modulus, resistance to fatigue damage, and
resistance to permanent deformation.
1.5 Scope of work
The scope of work in this thesis includes:
Chapter 1: Introduction
The content of this chapter briefly demonstrates current problems of asphalt
recycling technique that leads to the research objectives.
Chapter 2: Literature review
This chapter contains up-to-date knowledge on hot asphalt recycling
6
techniques and related issues.
Chapter 3: Laboratory RAP production
This chapter presents the purposes and procedure for laboratory RAP
production.
Chapter 4: Zero shear viscosity and the accuracy of viscosity mixing equations
This chapter contains the evaluation of different viscosity mixing equations
using zero shear viscosity.
Chapter 5: Effect of laboratory mixing methods on homogeneity of recycled asphalt
mixtures
This chapter presents the effects of different RAP/superheated virgin
aggregate mixing durations and RAP sizes on homogeneity of hot recycled
asphalt mixture. The homogeneity of recycled mixtures is studied by using
virgin binder with different colour from that of RAP binder. The red colour
of virgin binder is obtained by mixing clear binder with iron oxide pigment.
The proportion of pigment is 10% by weight of the binder making this binder
red. The content of this chapter also includes the correlation between
homogeneity and stiffness distribution of hot recycled asphalt mixtures.
Chapter 6: Effect of mixing process and RAP materials on stiffness of recycled
asphalt mixtures
The proportion of the pigment certainly alters the flow characteristic of red
virgin binder (Chapter 5). This might affect the mixing process and
rejuvenation between virgin and aged binder. Therefore, the effects of mixing
methods and RAP sizes on mechanical performance of hot recycled asphalt
mixtures are further investigated using normal straight run bitumen 160/220
Pen as virgin binder. The data and analysis on the effect of mixing method
and RAP materials on stiffness modulus, resistance to fatigue damage and
resistance to permanent deformation of hot recycled asphalt mixtures are
presented in Chapters 6,7, and 8.
Chapter 7: Effects of mixing methods and RAP materials on fatigue life of recycled
asphalt mixtures
Chapter 8: Effect of mixing methods on permanent deformation resistance of hot
recycled asphalt mixtures
Chapter 9: Conclusions and Recommendations
7
2 Literature review
2.1 Asphalt mixture
2.1.1 Definition of asphalt mixture
Asphalt or bituminous mixture is a combination of bitumen and mineral aggregate. This
kind of mixture is normally used for construction of highway pavement layers, parking
areas, and pedestrian streets. Once asphalt mixture is compacted to required air void
content, mineral aggregate with different size fractions plays a role as skeleton to provide
the strength. Bitumen, a residual fraction of fractional distillation process of crude oil, will
act as an adhesive, to bond aggregate particles together and improve the performance of the
mixture (Read and Whiteoak, 2003).
2.1.2 Classification of asphalt mixture
Depending on the proportion and particle size distribution, asphalt mixtures are divided into
two main categories, gap-graded and continuous-graded mixtures. In gap-graded, the
particle size distribution is not continuous (Figure 1). Normally, there is one single size of
coarse aggregate. The high void content due to single size of coarse aggregate will be filled
with sand, filler and bitumen. The structural strength of gap-graded mixtures is built by the
mortar of sand, bitumen and filler. Depending on the required aggregate gradation and
binder content, gap-graded mixture is classified into mastic asphalt (BS-EN:13108-6,
2006), hot rolled asphalt (BS-EN:13108-4, 2006), stone mastic asphalt (BS-EN:13108-5,
2006), and porous asphalt (BS-EN:13108-7, 2006).
Different from gap-graded, there normally exist many different aggregate size fractions in
continuous-graded mixtures with the hypothesis that the smaller particles will fill up the
voids generated by the bigger particles (Roberts et al., 1991). Hence, the strength of
aggregate skeleton is based on the interlock among aggregate particles. This makes the
continuous-graded mixture better at deformation resistance than gap-graded one. However,
as the required amount of binder content in gap-graded mixture is higher due to using fine
aggregate as sand and filler, this type of mixture has better fatigue resistance (Read, 1996).
8
Figure 1: Comparison on gradation of gap-graded and continuous-graded asphalt mixture (Read, 1996)
2.1.3 Properties of asphalt mixture
Stiffness modulus
Stiffness is the resistance to deformation under applied stress conditions. As asphalt
mixture is a visco-elastic material, the stiffness of asphalt mixture normally includes elastic
and viscous components. The proportions of each component rely primarily on the
temperature and the loading time. Under low temperature and short loading time, the
asphalt mixture will behave elastically. On the contrary, the relation between stress and
strain will follow viscous manner under high temperature and long loading time (Read and
Whiteoak, 2003).
The stiffness of asphalt mixture can be determined by the following equation:


= E (1)
Where:
E : stiffness modulus
 : applied stress
9
 : strain caused by applied stress
Permanent deformation
Permanent deformation is the phenomenon that unrecoverable strain is accumulated after
the load is released in each loading cycle. Figure 2 illustrates the strain response to the
applied load. The strain starts increasing when the load is applied. Once the load is released,
the elastic component of the strain will recover instantaneously. There is also a component
called visco-elastic strain which will recover with time. However, the permanent
deformation, due to plastic characteristic of asphalt mixture, cannot recover (Perl et al.,
1983). Although this viscous and plastic deformation is really small after each loading
cycle, the accumulation will become large after millions of loads (Figure 3). This will cause
rutting phenomenon in the pavement structure.
Figure 2: Strain response due to applied stress of Visco-Elasto-Plastic Constitutive model (Perl et al.,
1983)
10
Figure 3: Visco-elastic response to millions of load application (Read, 1996)
Fatigue characteristic of asphalt mixture
Fatigue can be defined as: “The phenomenon of fracture under repeated or fluctuating stress
having a maximum value generally less than the tensile strength of the materials” (Pell,
1988). However, tensile stress induced in the pavement is not only due to traffic loading but
also the effect of surrounding environment, for instance, the fluctuation of surrounding
temperature. Read (1996) also defined the fatigue as: “Fatigue in bituminous pavement is
the phenomenon of cracking. It consists of two main phases, crack initiation and crack
propagation, and is caused by tensile strains generated in the pavement by not only traffic
loading but also temperature variations and construction practices”. The empirical data
shows that the tensile range from 30 to 200 microstrain is the condition that fatigue damage
might possibly occur (Brown, 2000).
The relationship defining the fatigue life of bituminous mixture based on crack initiation is
as follows (Brown, 2000):
m
f
f
c N
|
|
.
|

\
|
=

1
(2)
Where
11
f
N : number of applications of load to initiate a fatigue crack
t
 : initial value of tensile strain
m c, : factors depending on the composition and properties of the mixture
2.2 Durability of asphalt mixture
2.2.1 Definition of durability
Asphalt pavement in general has to carry the traffic under certain climatic conditions. In
order to satisfy the performance demand, a pavement must have the ability to withstand any
damage during the whole service life. Therefore, the durability of asphalt pavement is
defined as follows (as cited in Scholz, 1995):
“Durability as it applies to bituminous paving mixtures is defined as the ability of
the materials comprising the mixture to resist the effects of water, aging and
temperature variations, in the context of a given amount of traffic loading, without
significant deterioration for an extended period”.
2.2.2 Mechanism of ageing in asphalt mixture
As asphalt mixture is a combination of bituminous binder and a skeleton of mineral
aggregate, the ageing mechanism of asphalt mixture is understood as that of the bitumen.
Traxler (1963) studied the causes of ageing or hardening phenomenon in asphalt binder and
concluded there were 15 factors which might cause ageing in bitumen (Table 1). However,
some of the causes were just listed but not verified by experimental data. Petersen (1984)
suggested the fundamental factors caused hardening in asphalt materials, which were:
1. Loss of oily components of bitumen by volatility or absorption by mineral
aggregate.
2. Changes in chemical composition as chemical molecules of bitumen react with
oxygen (oxidation).
3. Molecular structuring causing thixotropic effects (steric hardening).
Oxidation is the phenomenon when chemical molecules in bitumen are oxidised by oxygen
in the atmosphere and form polar groups containing oxygen, for instance, hydroxyl,
carbonyl and carboxylic groups (Read and Whiteoak, 2003). The polar molecules own
unevenly distributed electrical charges and tend to interact with the others. Depending on
12
the strength of the bond, these polar molecules will from a network and comprise a wide
range of molecular types and sizes (John, 1993).
Rostler and White (1962) studied the compositional change of 85/100 penetration grade
bitumen. Bitumen composition was divided into asphaltene (A), nitrogen base (N), first
accidaffin (A
1
), second accidaffin (A
2
), and paraffins (P) fractions. Each fraction had its
particular function. Nitrogen base has the function of peptizing inert asphaltene. The
accidaffin group will keep the peptized asphaltene solvated. This solution will be gelled by
paraffins. Due to ageing, the proportion of nitrogen base and first accidaffin change greatly,
first accidaffin changes to nitrogen base, and nitrogen base turns to asphaltene. The
resulting proportion of each fraction in bitumen would cause incompatibility (or sysneresis)
and substantially affect the durability of bitumen.
Noureldin (1995) also studied the ageing effect of bitumen and reported that bitumen is a
combination of asphaltene, resin, and oil. The viscosity of bitumen is mainly attributed to
the asphaltene component. During the ageing process, the oil will convert to resin, and the
resin will turn to asphaltene. The increasing proportion of asphaltene plus the fact that the
maltene phase necessary to disperse asphaltene is insufficient will increase the viscosity of
bitumen.
Ageing is also due to the loss of volatiles in bitumen. Oily proportion of bitumen primarily
volatizes due to high temperature. In addition, the loss of volatiles is also attributed to long
term exposure of asphalt to the environment. In addition, depending on the mineralogy of
aggregate, oily component is also absorbed by porosity when bitumen is in contact with
aggregate. However, the hardening due to these phenomenon is not considerable compared
to ageing by oxidation (Read and Whiteoak, 2003).
13
Table 1: Mechanism of bitumen aging (Traxler, 1963)
14
2.2.3 Factors affecting ageing mechanism
Chemical composition of bitumen binder
White et al. (1970) studied the effects of chemical composition on durability of bitumen.
Four bitumen binders from different origins, California, Venezuela, Arkansas, and Alberta
were fractioned into five basic components, asphaltene (A), nitrogen base (N), first
accidaffin (A
1
), second accidaffin (A
2
), and paraffins (P). Bitumen blends were produced
by blending different components in various proportions. Each blend was subjected to
syneresis analysis. The homogeneity of each blend was studied by microscope. The
compatibility of each blend was also evaluated by filter paper test under ultraviolet light.
The aim of filter paper test was to identify the separation of oil phase. White et al. (1970)
claimed that compatibility of bitumen relied primarily in the proportion of Nitrogen bases
and Paraffins, expressed as ratio N/P, or syneresis parameter. This ratio should be greater
than 1 for the bitumen to be free of syneresis.
Durability of each blend was also studied by Pellet abrasion test before and after aging. The
pellet was made by compressing a mixture of Ottawa sand and bitumen. The mixture was
produced by mixing Ottawa sand and bitumen at 160
o
C for 6 minutes. The bitumen content
of the mixture was 2% by weight of the sand. During ageing process, the mixture was
conditioned under infrared light of 7 days at 60
o
C. The test was carried out by shaking 2
gram pellet in a square bottle. The loss in milligrams after 500 revolutions of the pellet was
recorded. The test results showed that durable bitumen must have the compositional
parameter, expressed as (N+A
1
)/(P+A
2
), above 0.4.
Temperature
Temperature seriously affects the ageing or hardening rate of bitumen. The higher the
temperature bitumen is exposed to, the more bitumen ages. Especially in the condition of
above 100
o
C, Read and Whiteoak (2003) stated that the oxidation rate increases twofold
for each increment of 10
o
C. Temperature extremely affects the properties of bitumen. Data
in Figure 4 illustrates that for 30s mixing time, a raise of 5.5
o
C in mixing temperature will
elevate the softening point by 1
o
C.
15
Figure 4: Relationship between the temperature of the mixture and change in softening point (Read and
Whiteoak, 2003)
Air void content in total mixture
The air void contents after compaction presents highway engineers with a paradox.
Although low air void content reduces the ageing rate, it might increase rutting
phenomenon. Vice versa, if the air void content is increased, the asphalt binder coating
aggregate can be oxidized faster. The stiff oxidized bitumen will easily rupture. Water can
get in and destroy the bond between bitumen film and aggregate and reduce the tensile
strength of the mixture. Most highway agencies prefer the range of air voids after
compaction from 3 to 5% (Roberts et al., 1991). However, if the compaction quality is not
well controlled, the high air void content will cause faster ageing speed. Figure 5 shows the
ageing of pavement made of bitumen that has penetration of 100 dmm at 25
o
C. The
penetration after mixing is 70 dmm. If the air void content is less than 5%, the penetration
of bitumen after five years in service is almost the same as the initial value. On the
contrary, if the void content is higher than 9%, the pavement is extremely aged as the
penetration reduces from 70 to 20 (Read and Whiteoak, 2003)
16
Figure 5: The effect of void content on the hardening of bitumen on the road (Read and Whiteoak,
2003)
Bitumen film thickness
During the mixing process, the bitumen is exposed to extremely high temperature in
condition of very thin layer approximately 5 to 15µm thick. Bitumen will be extremely
aged due to oxidation and loss of volatiles. The loss of bitumen penetration during mixing
is approximately 30% (Read and Whiteoak, 2003). Another study by Kandhal and
Chakraborty (1996) also showed that the thicker the bitumen film, the less viscosity of
bitumen increase after short and long term ageing (Figures 6 and 7).
17
Figure 6: Asphalt film thickness vs viscosity after short term ageing (Kandhal and Chakraborty, 1996)
Figure 7: Asphalt film thickness versus viscosity after long term ageing (Kandhal and Chakraborty,
1996)
2.2.4 Consequences of ageing in bituminous mixture
The compositional changes in bitumen due to ageing will result in the increase of bitumen
viscosity. Kandhal and Koehler (1984) carried out a study on durability of dense-grade
18
pavements using different types and sources of bitumen in Pennsylvania, US. The data
measured from 1961 to 1976 (Figure 8) shows that the longer time the pavement is exposed
to environment, the higher the pavement viscosity.
Figure 8: Effect of ageing time on bitumen viscosity extracted from pavements (Kandhal and Koehler,
1984)
Ageing will also change the visco-elastic properties of bitumen. Research by Daniel et al.
(1998) shows that the longer the ageing time, the lower the phase angle of bitumen (Figure
9). In fact, the decrease of loss modulus plus the increase of viscosity of bitumen due to
ageing will result in the pavement being more prone to cracking at low temperature.
Kliewer et al. (1996) studied the effect of ageing on fracture temperature. The experimental
results (Figure 10) demonstrated that the increase of ageing time would result in the
increase of fracture temperature. In fact, the more aged the pavement, the more prone to
cracking.
19
Figure 9: Phase angle versus ageing levels (Daniel et al., 1998)
Figure 10:Fracture temperature versus ageing time (Kliewer et al., 1996)
2.2.5 Laboratory tests simulating field ageing
Plancher et al. (1976), studied the effect of lime on durability of bitumen by comparing the
resilient modulus of cylindrical specimens with diameter of 4 cm and thickness of 2.5 cm
before and after ageing procedure. Both normal and lime treated bitumen were used in this
research. The results showed that mixtures using lime treated bitumen had better resistance
to ageing. The ageing procedure included:
- pressing the mixture of sand and bitumen in specified dimension mould at 150
o
C
under pressure of 27.6 MPa.
- conditioning in the oven 150
o
C for 1 hour.
20
- cooling at 25
o
C for 72 hours then testing the resilient modulus for un-aged
specimen.
- storing at 150
o
C for 5 hours.
- cooling at 25
o
C for 72 hours then testing the resilient modulus for aged specimen.
Hugo and Kenedy (1985) implemented the accelerated ageing procedure in both dry and
moist conditions. Both types required the slab with thickness of 4 cm to be conditioned
under temperature of 100
o
C for 4 to 7 days. An open reservoir is put in the oven in order to
maintain the relative humidity higher than 80% in moisturised condition. After ageing, the
bitumen was recovered and subjected to Shell plate viscosity testing at different
temperatures. The ageing effect was obtained by comparing the viscosity of recovered
bitumen with that of original bitumen. Although it is advantageous to study the ageing in
moisturised condition, the affect of different void content and temperature on hardening is
not mentioned in this research.
These mentioned scholars studied the ageing of bitumen in compacted mixtures. However,
the ageing during production process has been ignored. Von Quintas in 1988 simulated the
hardening of bitumen during production, or short term ageing, by conditioning the loose
mixture at 135
o
C in a force draft oven for 8, 16, 24, and 36 hours. The laboratory data,
although scattered, generally showed approximately the same ageing level as in the field.
Not only short term, long term ageing was also studied by conditioning specimens for 2
days at 60
o
C followed by 3 days at 107
o
C (Airey, 2003).
The most comprehensive study on ageing of aggregate mixture is the project A-003A by
Bell et al. (1994a) under the Strategic Highway Research Program (SHRP). In this research,
many factors affecting the ageing mechanism were considered, for instance, air void
content, temperature, and production. Short term ageing, presenting the hardening of
bitumen during production process, is simulated by conditioning loose mixture at 135 or
163
o
C during 1, 6, and 15 hours (STOA). The aged mixture is then compacted to 4 and 8
percent air void after STOA. For long term ageing, the compacted specimens are
conditioned in force draft oven for 2 or 7 days at constant temperature of 107
o
C. Before
LTOA, compacted specimens are preconditioned for 2 days at 40 or 60
o
C to assure the
stability of specimens. Characteristics before and after ageing are determined by resilient
modulus and indirect tensile tests. The results from the tests showed that in some mixtures,
21
the resilient modulus is 4 times higher than the initial value after STOA and 6 times for
LTOA. Hence, the accelerated ageing procedure is recommended as follows:
- STOA: loose mixture is conditioned at 135
o
C for 4 hours.
- LTOA: compacted mixture is conditioned at 85
o
C for 5 days or 100
o
C for 2 days.
The recommended accelerated ageing procedures were then verified by field practice. After
this field validation, Bell et al. (1994b) concluded that:
- STOA: is the simulation of bitumen hardening in pavement during production and
construction plus less than two year performance.
- LTOA (8 days at 85
o
C): is the simulation of bitumen hardening in pavement for 18
years performance in the condition of wet-no-freeze climate, or 9 years in dry-freeze
climate.
Airey (2003) suggested the condition of 4 days at 85
o
C to simulate LTOA in pavement for
15 years performance in the condition of wet-no-freeze climate, or 7 years in dry-freeze
climate. However, in AASHTO PP2 (1994), it is demonstrated that this condition is equal
to LTOA during 5 days at 85
o
C.
2.3 Design methodology for recycling of bituminous pavements
2.3.1 Objectives and design procedure of pavement recycling
There are different objectives depending on the organizations in charge or related to the
pavement recycling job. For instance, the contractors want to recycle the pavements as the
costs of construction are cheaper due to the reuse of existing materials (RAP). The
government is also interested due to the saved money from construction hence the problem
of constrained budgets for highway maintenance can be solved. In term of highway
engineers, the overall objective of recycling pavement is, to restore the properties of
existing deteriorated pavements materials, to such a level that can satisfy the service
requirements.
The bitumen, due to the chemical changes during service life of asphalt pavement, cannot
be used without any modifications. Any materials that can alter the properties of RAP
binder are defined as modifiers. This wide definition includes reclaiming, recycling,
modifying, softening agents, recycling modifiers, rejuvenators, fluxing oils, extender oils,
and aromatic oils (Karlsson and Isacsson, 2003d). The purpose of asphalt pavement
recycling (Epps et al., 1980) is to choose proper modifiers to:
22
- Restore the consistency of aged bitumen.
- Restore the chemical composition of aged bitumen for durability.
- Provide sufficient binder to coat the new aggregate added and satisfy the stability
requirement of the mixture.
To obtain the overall objective of recycling bituminous pavements, (Davidson et al., 1977),
recommended the design procedure for asphalt pavement recycling includes:
- Determine the properties of existing pavement (RAP) including bitumen content,
consistency of bitumen, bitumen demand for aggregate, aggregate gradation.
- Select the reclaiming agent. The reclaim agent must reduce the viscosity of aged
binder to the desired level and improve the durability.
- Analyse and use data for design.
2.3.2 Methods for recycling bituminous pavements
The concept of recycling or reusing the existing pavement material in new construction or
rehabilitation project has lasted for many years. The first recycling pavement project
recorded was in 1915 (Epps et al., 1980). Since that moment, there has been a wide variety
of pavement recycling techniques. These approaches are normally classified based on the
materials to be recycled for instance, bituminous or portland cement pavement, the
structural layers which will benefit from recycling as surface, base or sub base, and the
procedure as well as equipment for recycling purpose.
Generally, recycling techniques are categorized into two main types, in-plant and in-situ
recycling. Regarding in-plant recycling, the reclaimed pavements are transported to asphalt
production plant, ripped, milled, and crushed into required sizes before being blended with
virgin bitumen and aggregate in the mixing plant. Depending on the mixing temperatures
required of the recycled mixture, in-plant recycling is further divided into cold, warm, and
hot recycling (Karlsson and Isacsson, 2003d). Apart from the other in-plant recycling
methods, hot recycling proves to be the most advantageous due to the ability to correct
most of the pavement surface defects, deformation and cracking. In addition, hot recycled
mixture with 10 to 30% of RAP can have the same performance compared to virgin mixture
(Kandhal and Mallick, 1997).
Different from in-plant recycling, in-situ methods are processed on site. In-situ recycling is
also divided into hot in-situ recycling which includes remixing and repaving and cold in-
23
situ recycling or full depth reclamation. In remixing method, existing pavement is milled,
mixed with new materials before being laid and compacted. The procedure in repaving
method is almost the same except no new materials are added. However, a new layer is laid
on top of the recycled pavement to increase structural strength of the pavements. In full
depth reclamation, all the surface and a part of base course is milled, scarified, and mixed
with bitumen emulsion, foamed bitumen or soft bitumen to produce a stabilized base. A
new surface is then laid and compacted on the new recycled base (Kandhal and Mallick,
1997).
2.3.3 Selection of rejuvenators
Davison et al. (1977) studied the recycling aspect with a wide range of 6 aged binders and
12 reclaiming agents. The percentages of aged bitumen for each combination were in turn
0, 5, 25, 50%. In addition, the effect of asphaltene origin was also investigated by using one
reclaiming agent with different asphaltenes. The consistencies as well as the chemical
fractions of recycled mixture before and after ageing were compared. It was suggested that
the compositional parameter of reclaiming agent should be in the range of 0.4 to 1,
preferably 0.4 to 0.8, and the syneresis parameter (Section 2.2.3) should be higher than 1 to
assure the durability improvement of aged bitumen binders. Dunning and Mendenhall
(1978) suggested that the flash point of modifier should be enough to produce the blend
with flash point of 205
o
C. In addition, the viscosity at 60
o
C of modifier should be in range
of 90 to 300 cP.
The origin where bitumen came from also affects the durability of recycled materials. In the
study by White et al. (1970), four bitumens from different sources, California, Venezuela,
Arkansas, and Alberta, were used. All bitumen binders were fractionated into five basic
fractions. The different asphatene fractions with the same proportion were in turn mixed
with different maltene phases. When the asphaltene and maltene from California and
Alberta were cross blended, there were segregations with all the blends with asphaltene
coming from Alberta. The other blends with California asphaltene showed no sign of
heterogeneity or syneresis.
White et al. (1970) also concluded that the properties of bitumen were governed
predominantly by the composition of the maltene phase. This is actually the ability to
peptize asphaltenes by mean of the maltenes phase. The molecular weight of asphaltenes
24
also affected the viscosity of the blend. His experimental data indicated that the blends with
higher molecular weight asphaltenes had higher viscosity than blends with lower molecular
weight asphaltene.
The durability of recycled bitumen was also studied by Chaffin et al. (1997) by assessing
the potential of some rejuvenators for asphalt pavement recycling. In this study, industrial
supercritical fractions (ISCF-A, ISCF-B, ISCF-C), commercial recycling agents (CRA-A,
CRA-B, CRA-C) as well as bitumen fractions were appraised. Bitumen fractions are
fractions F3 extracted from SHRP bitumen (YBF, AAF, ABM) in a supercritical pilot plant
at Texas A&M University at a temperature of 221
o
C and pressure of 49.3 Bar (Bullin et al.,
1995). The SHRP bitumens come from different crude sources. YBF is the popular AC-20
and there is no clear origin recorded for this bitumen. Meanwhile, AAF and ABM in turn
originated from West Texas Sour and California Valley (Mortazavi and Moulthrop, 1993).
The aged bitumen was artificially produced from SHRP bitumen ABF in an air bubbling
apparatus (denoted as ABF-AB1). The properties of recycling agents are illustrated in
(Table 2).
Viscosity Saturate Asphaltene Aromatic
(dPa.s) (wt%) (wt%) (wt%)
ISCF A 17.6 20.4 0.3 79.3
ISCF B 58.0 30.8 0.7 68.5
ISCF C 434.0 11.4 3.4 85.2
CRA A 2.4 8.7 0.7 90.6
CRA B 1.2 12.4 0.9 86.7
CRA C 1.0 28.0 0.5 71.5
Table 2: Industrial supercritical fractions and commercial rejuvenating agent properties (Chaffin et al.,
1997)
The blends were produced by mixing aged ABF-AB1 with different rejuvenating agents.
The amount of aged bitumen was determined by ASTM 4887 (2003) so the final viscosities
of the blends were approximately the same as that value of SHRP AAF-1 (approximately
2000 P at 60
o
C). All the blends as well as AAF-1 were subjected to TFOT and PAV ageing
to determine the ageing index. The aging index was the ratio between viscosities before and
after artificial aging by TFOT and PAV method. The results (Table 3) indicated that the
blends with bitumen fractions had lowest ageing index, followed by the blends with
25
industrial supercritical fractions and commercial rejuvenating agents. The SHRP AAF-1
had the highest ageing index.
Composition Viscosity TOFT PAV
Asphalt/agent (dPa.s) AI AI
SHRP AAF-1 N/A 1890 2.80 12.42
AAF-AB1/ISCF A 72/28 1900 1.68 4.21
AAF-AB1/ISCF B 61/39 2140 N/A N/A
AAF-AB1/ISCF C 43/57 2080 1.67 3.89
AAF-AB1/CRA A 81/19 1840 1.85 4.30
AAF-AB1/CRA B 83/17 1850 1.70 4.46
AAF-AB1/CRA C 83/17 1900 1.96 5.53
AAF-AB1/YBF F3 61/39 2000 1.50 3.00
AAF-AB1/AAF F3 67/33 2090 1.67 3.85
AAF-AB1/ABM F3 44/56 1670 1.59 2.93
Table 3: Ageing index (AI) after TFOT and PAV ageing of recycled blend with different rejuvenators
(Chaffin et al., 1997)
2.3.4 Estimation of the consistency of the aged bitumen – modifier
blend
A crucial phase of the design procedure for recycling asphalt pavement is to predict the
viscosity of recycled bitumen binder. The design viscosity of the recycled binder is usually
obtained by blending different proportions of aged binder and modifier until the desired
viscosity is obtained (Epps et al., 1980). This process is time consuming due to many
blending trials and complicated chemical constituents of bitumen. Bitumen is a compound
of many different chemical substances, and consistency of each is far different from the
others.
Davidson et al. (1977) built a viscosity blending chart to simplify the viscosity estimation
process (Figure 11). The construction of this nomograph was based on the trial blending
data of 6 aged binders and 12 reclaiming agents. Figure 11 is an illustration of how to use
the viscosity blending chart. The first step is to draw a line that connects the viscosity
values of aged and virgin binder (points 3 and 5). To determine the proportion of virgin
binder in the blend, the next step is to draw a horizontal line at desired viscosity value of
the blend on Y-axis. Final step is to draw the vertical line from the point that the first two
lines cross each other (point 2). The point that this vertical line crosses the X-axis is the
proportion of virgin binder. The viscosity blending chart can also be used to determine the
viscosity of recycled blend or for the selection of rejuvenators.
26
Figure 11: Nomograph for predicting 60
o
C viscosity of recycled asphalt (Davidson et al., 1977)
Viscosity of bitumen blend between aged and virgin binders can also be predicted using
viscosity mixing equations. In these equations, aged and virgin bitumen binders are
considered as liquids and the viscosity of the bitumen blend can be estimated approximately
by mixing theory for binary liquids (Chaffin et al., 1995). Arrhenius (1887) suggested the
following equation. This equation was then developed to ASTM D4887 (2003), one of the
most popular tools to predict the viscosity of recycled binder (Karlsson and Isacsson,
2003d).
( ) ( ) ( )
2 2 1 1
ln ln ln    x x
mix
+ = (3)
Where:
mix
 : viscosity of the mixture of two binary liquids
1
 ,
2
 : viscosity of both liquids
1
x ,
2
x : volume percentages of both liquids
The Arrhenius equation, due to its simplicity, could not express the interaction between two
liquids in the blend. However, the interaction between two liquids, whether in chemical or
physical form, may alter the characteristic of the blend. This is due to the complicated
chemical composition of bitumen itself and especially in case of recycled binder which is a
mixture of at least two bituminous materials coming from different crude oil sources. The
27
difference in origin might lead to the fact that although two rejuvenators have the same
viscosity, after mixing with aged binder, the viscosities of recycled blends are still different
(White et al., 1970).
Grunberg and Nissan (1949) also introduced an equation to estimate the viscosity of binary
liquids. One advantage of this viscosity mixing rule over Arrhenius equation is the
interaction between two liquids in the blend. The viscosity of binary liquid mixture is
expressed as follows:
( ) ( ) ( )
12 2 1 2 2 1 1
ln ln ln G x x x x
mix
+ + =    (4)
Where:
12
G : is a characteristic of the system of two binary liquids or interaction parameter
Epps et al. (1980) also introduced an equation to estimate the viscosity of a mixture
between two liquids. The viscosity of binary liquid is based on the log log relationship with
the log log viscosity of each individual liquid and the mass percentage in the whole
mixture. This equation is expressed as follows:
( ) ( ) ( ) ( ) ( ) ( )
2 2 1 1
ln ln ln ln ln ln    x x
mix
+ = (5)
Actually, the Arrhenius equation is a special case of Grunberg and Nissan viscosity mixing
rule where
12
G is equal to zero. A study by Irving (1977) concluded that generally,
Grunberg and Nissan equation was the best equation to estimate the viscosity of binary
liquids. Irving (1977) also claimed that if constant parameter
12
G was used universally, the
predicted viscosity could be less than 30% difference from actual value.
Chaffin et al. (1995) carried out a study to verify the effectiveness of different viscosity
mixing equations. In this research, three SHRP bitumens, AAA-1, AAF-1, and ABM-1
were used to produce aged bitumen by using pressurised oxygen vessel and air bubbled
reaction apparatus. A wide range of softening agents was used, for instance, low viscosity
bitumen binders, commercial recycling agent, and supercritical fractions. Low viscosity
bitumen included AC-3, AC-5 from Diamond Shamrock (DS) in Duma, Texas, AC-5 from
Shell in Deerpark, Texas, and AAV, ABH from SHRP. Commercial recycling agents were
Sun Hydrolene 125, Witco Cyclogen, Exxon Nuso 95 and Mobil Mobisol 120. Supercritical
fractions were extracted from AC-20 (YBF), AAA-1, ABM-1, and AAF-1. Viscosities of
aged bitumens, softening agents are illustrated in Table 4. The consistencies of aged
28
bitumens and softening agents were measured by rheological apparatus at 60
o
C and
frequency of 1.6Hz.
Aged bitumens were blended with low viscosity bitumen, softening agents including
commercial agents and supercritical fractions at increments of 20%. Viscosity values
estimated by different viscosity mixing equations were then compared to experiment data.
Using Grunberg and Nissan equation, interaction parameter
12
G for each pair of aged
bitumen/softening agent was calculated. The results of
12
G for the whole experiment (Table
5) showed that the interaction parameters
12
G of each pair of aged bitumen/softening agent
varied considerably. The finding of Chaffin et al. (1995) is also in agreement with Irving
(1977). In fact, using a constant value of
12
G would result in considerable errors in
viscosity estimation.
Table 4: Viscosity at 60
o
C of aged bitumen and softening agents (Chaffin et al., 1995)
29
Table 5: Aged bitumen - softening agent Grunberg interaction parameter G
12
(Chaffin et al., 1995)
The variation of
12
G was attributed to the viscosity difference between softening agents and
aged bitumen. The larger the difference in viscosity between aged bitumen and rejuvenator,
the greater the absolute value of interaction parameter
12
G . To eliminate the viscosity
effects, Chaffin et al. (1995) introduced the dimensionless log viscosity (DLV). Using
DLV, Grunberd and Nissan equation is mathematically transformed as follows:
( )
( ) ( ) ( )
2
2
1 2
12
2
1 2
12
1 2
1
/ ln / ln
1
/ ln
/ ln
x
G
x
G
DLV
m
|
|
.
|

\
| ÷
+
|
|
.
|

\
|
+ = =
     
 
(6)
Where:
30
2 1
, ,   
m
Viscosity of mixture, rejuvenator and aged binder
2
x
Proportion of aged binder in the mixture
Based on experimental data, Chaffin et al. (1995) came up with the following equation to
estimate the viscosity of bitumen blend:
2
73 . 0 26 . 0 01 . 0
as as
x x DLV + + = (7)
Where:
as
x : the proportion of aged binder in the blend
In Grunberg and Nissan method, the viscosity of the blend has a second order polynomial
relationship with the proportion of aged binder. Meanwhile, the viscosity mixing equation
by Chaffin et al. (1995) expressed the second polynomial relationship between DLV and
proportion of aged binder. Although the ways to find the solutions are mathematically
different, both methods have the same philosophy, to use one constant interaction parameter
universally. This will lead to the fact that the method performs well with some types of
blend and not with others. Chaffin et al. (1995) compared the predicted viscosity values
using different method to the experimental data. The result indicated DLV method
performed well when supercritical and commercial recycling agents were used. On the
other hand, the deviation is considerably larger than those obtained by the other mixing
rules when soft asphalt was used as recycling agent. In this case, Arrhenius or ASTM 4887
is the best to estimate viscosity of the blend.
Mixing rule is also expressed in BS EN 13108 part 1 (2006). In this document, the
penetration and softening point of the mixture is estimated by those of individual bitumen
comprising the mixture and the mass proportions. The penetration of the mixture is
estimated by the following equation:
( ) ( ) ( )
2 1
log log log pen b pen a pen
mix
+ = (8)
Where:
mix
pen : calculated penetration of the mixture
1
pen ,
2
pen : penetrations of both binders comprising mixture
a , b : mass proportions of each bitumen in the mixture ( 1 = + b a )
The softening point of the mixture is estimated by the following equation:
31
2 1
bT aT T
mix
+ = (9)
Where:
mix
T : calculated softening point of the mixture.
1
T ,
2
T : softening point of each individual binder in the mixture.
2.4 Hot recycled mixture production
The conventional plants for producing hot asphalt mixture are batch plant and drum plant.
With existence of RAP, some modified features should be made to the conventional mixing
plant to incorporate RAP materials to produce the hot recycled asphalt mixtures (Roberts et
al., 1991). The way RAP is processed is different with different modified mixing plants.
2.4.1 Batch plant
The schematic of a batch mixing plant is illustrated in Figure 12. For conventional asphalt
mixture, each component has its own function. The burner is in charge of drying aggregate
to remove the moisture content and preheat aggregate to the temperature required for
mixing. After drying process, the heated aggregate is conveyed to the screening deck to
determine the quantities necessary and stored in the hot bin. Bitumen is also preheated and
weighed before sprayed and mixed with heated aggregate in the pugmill.
However, there are problems using conventional method to produce hot recycled asphalt
mixture. The materials for recycled mix are not just virgin aggregate and bitumen but also
RAP. If the RAP material is dried the same way as the virgin aggregate, there will be
environmental problems as the RAP bitumen is burned directly by the flame in the burner
causing blue smoke (Hughes, 1978). Hence, the conventional method is modified to prevent
the “blue smoke” phenomenon. RAP is stored separately before being introduced to virgin
aggregate in the hot bin. Virgin aggregate is superheated and the excessive thermal energy
will be used to heat up RAP material from ambient to required mixing temperature and
remove the RAP moisture content. After storing in hot bin, the combination of RAP and
virgin aggregate will be mixed with virgin bitumen to produce the recycled mixture.
32
Figure 12: Batch Plant (MAPH-2, 2000)
There are different methods of introducing RAP to the superheated virgin aggregate (Brock
and Richmond, 2005):
33
- Method 1: RAP is mixed with superheated virgin aggregate before being screened,
weighed, and stored.
- Method 2: RAP is weighed and screened before introduced and stored with
superheated virgin aggregate in hot bin
- Method 3: RAP is dried separately and then mixed with heated virgin aggregate and
bitumen.
In the first method; it is really difficult to control the gradation of the recycled mixture as
the size of RAP and the size of RAP aggregate are not similar, sometimes considerably
different. Two RAP particles have the same size, for example, ½ inch (12.5 mm). However,
one might be made of only one ½ inch (12.5 mm) aggregate; the other might be combined
of many 1 mm particles. In addition, the bottom deck screen, especially sizes smaller than
6.4 mm will be blind due to the RAP binder filling up those tiny holes of the screens. In
addition, bigger RAP size cannot be used in this process as the bigger size, for instance,
size 2 inches (50 mm), cannot pass the screen decks.
In the second method, as the RAP is introduced to superheated virgin aggregate and stored
in the hot bin, the blend is not well mixed. Hence, the heat from superheated virgin
aggregate is not well transferred to the RAP materials to remove all the moisture content
and heat up RAP.
The only problem with the third method is the production rate and the cost. RAP materials
do not have the same characteristics as virgin aggregate as RAP is a combination of aged
bitumen and aggregate. With bitumen, if the drying temperature is higher than 100
o
C, it
will increase the ageing speed of RAP bitumen (Shell bitumen handbook). On the other
hand, if low drying temperature is employed, the time for heating will be longer and might
seriously affect the production rate as well as the cost of asphalt production.
As RAP is mixed with virgin aggregate at extremely high temperatures, the properties of
RAP binder might change. The ageing process might also be affected by the steam
generated due to RAP moisture content during the mixing process in the pugmill.
34
2.4.2 Drum facility (Drum mixer)
The schematic of drum mixer for producing hot asphalt mixture is demonstrated in Figure
13. The procedure in the drum facility is quite different from batch plant. In batch plant, the
process of drying and heating aggregate is separate from the mixing aggregate with
bitumen. Drying and heating aggregate are processed in the burner. The mixing between
aggregate is implemented in the pugmill. However, with conventional drum facility,
aggregate is heated and mixed with bitumen in the same drum. The drying and heating time
of aggregate is twice as long as the mixing time between aggregate and bitumen. Mixing
time of aggregate with bitumen is normally 30 seconds (Read and Whiteoak, 2003).
Bitumen is introduced and mixed with heated aggregate at the two third point of the overall
time.
Figure 13: Drum mixer with RAP centre inlet (Brock and Richmond, 2005)
There are two main types of drum facility, parallel and counter drum mixer. Both types are
different regarding the way that materials are introduced into the mixer. In parallel mixer,
materials go in the same direction with the flame. Vice versa, the movement of materials is
opposed to the direction of the flame in counter drum mixer (Brock and Richmond, 2005).
In order to intake RAP materials, the conventional drum facility must be modified for hot
recycled asphalt production. Depending on what type of drum mixer, the modifications are
different. The first version of parallel drum mixer has a centre RAP inlet (Figure 13). In this
facility, RAP is mixed with superheated virgin aggregate before being mixed with virgin
35
binder. The design of kicker flight in the middle of the drum mixer is aimed to form a dam
of virgin aggregate so that the direct exposure of RAP to the flame is prevented.
Figure 14: RAP in parallel drum mixer with isolated area (Brock and Richmond, 2005)
In parallel type, there is also a drum mixer with separated mixing area (Figure 14). In this
facility, RAP is introduced with superheated virgin aggregate before entering isolated
mixing area where the blend is mixed with virgin binder. Sometimes, the mixer is designed
separately and called added continuous mixer (Figure 15).
Figure 15: RAP in parallel drum facility with added continuous mixer (Brock and Richmond, 2005)
36
The latest type of drum facility is called double barrel mixer. In this facility, the shell of the
drum is used as a shaft of the coater (Figure 16). As drying and mixing compartment are
separated by the shell of the drum, the exposure of RAP to direct flame of the burner is
absolutely prevented.
Figure 16: Double Barrel Drum facility (Brock and Richmond, 2005)
Except for the parallel drum facility with isolated area in which, the water steam is removed
considerably from mixing area, RAP material, virgin aggregate and virgin bitumen are
mixed in the condition of high temperature and steam. An example of preheating
temperature for RAP in drum mixer is illustrated in Table 6. Hence, the characteristics of
RAP binder might be changed, for instance, the ageing after mixing with superheated
aggregate due to the direct exposure to high temperature and steam condition. An advantage
of drum mixer is big sizes of RAP can be processed as the RAP is incorporated directly into
the mixer.
Generally, in both types of mixing plant, RAP material at ambient temperature is
introduced with superheated virgin aggregate. The heat transferred from superheated virgin
aggregate will help to remove the RAP moisture content. In addition, the energy RAP
absorbs from superheated virgin aggregate also increases RAP temperature and weakens
the bitumen bonds between RAP aggregate particles. Under the mechanical mixing, RAP
materials are separated and mixed with virgin aggregate. The combination of RAP and
37
virgin aggregate is then mixed with virgin binder. The total duration for a production cycle
in batch plant is about 60 seconds (Read and Whiteoak, 2003). In drum mixer, the
production cycle is approximately the same except for double barrel mixer at about 90
seconds (MAPH-2, 2000).
RAP RAP Moisture Superheat Temperature Required (
o
C)
Content (%) Content (%) 116
o
C Mix 127
o
C Mix 138
o
C Mix 149
o
C Mix
10
0 132 144 156 168
1 134 147 159 171
2 137 149 162 174
3 140 152 164 177
4 143 155 167 179
5 146 158 170 182
20
0 144 158 172 186
1 151 164 178 192
2 157 171 184 198
3 163 177 191 204
4 169 183 197 211
5 175 189 203 217
30
0 162 178 166 209
1 173 188 315 219
2 183 199 214 230
3 194 209 225 241
4 204 220 236 251
5 215 231 246 262
40
0 186 203 221 239
1 218 219 237 256
2 234 235 253 272
3 250 251 269 288
4 266 267 286 304
5 282 283 302 320
50
0 216 238 260 282
1 240 262 284 309
2 264 287 309 331
3 289 311 333 356
4 313 336 358 380
5 338 360 382 404
Table 6: RAP preheating temperature required in Drum Mixer (Brock and Richmond, 2005)
2.4.3 RAP sizes used for production of recycled mixture
There is a variety of RAP sizes used for production of RAP in recycled mixtures regarding
the equipment used for recycling and the percentages of RAP in the mixture. Table 7
illustrates the maximum sizes of RAP allowed for batch and drum mixing plants in relation
38
to RAP proportion of some States in America. The maximum size of RAP used is normally
less than 2 inches (50 mm). However, there are some States, for instance, Arkansas, and
Minnesota which use up to 3 inches (75 mm) RAP sizes (United States Department of
Transportation, 2007).
State
Max. RAP % - Batch Plants Max. RAP % - Drum Plants Top Size
for RAP
Base Binder Surface Base Binder Surface
Alabama 40 40 15 50 50 15 2 in
Alaska - - - - - - -
Arizona 30 30 30 30 30 30 1.5 in
Arkansas 70 70 70 70 70 70 3 in
California 50 50 50 50 50 50 2 in
Colorado 15 15 15 15 15 15 1.5
Connecticut 40 40 40 40 40 40 2 in
Delaware 35 35 25 50 50 30 2 in
Florida 60 50 None 60 50 None Specs
Georgia 25 25 25 40 40 40 2 in
Hawaii 30 None None 40 None None 1.5 in
Idaho Open Open Open Open Open Open 2 in
Illinois 50 25 15 50 25 15 Specs
Indiana 50 50 20 50 50 20 2 in
Iowa Open Open Open Open Open Open 1.5 in
Kansas 50 50 50 50 50 50 2 in
Kentucky 30 30 30 30 30 30 Specs
Louisiana 30 30 None 30 30 None 2 in
Maine 40 40 None 40 40 None 1 in
Maryland Open Open Limit Open Open Limit Specs
Massachusetts 20 20 10 40 40 10 0.75 in
Michigan 50 50 50 50 50 50 Specs
Minnesota 59 50 30 50 50 30 3 in
Mississippi 30 30 15 30 30 15 2 in
Missouri 50 50 50 50 50 50 1.5 in
Montana 50 50 10 50 50 10 2 in
Nebraska Not Used Not Used Not Used Open Open Open 2 in
Nevada 50 50 15 50 50 15 1.5 in
New Hampshire 35 35 15 50 50 15 Specs
New Jersey 25 25 10 25 25 10 2 in
Table 7: State DOT specification requirements for the use of reclaimed asphalt pavement (RAP) in hot
asphalt paving mixtures (United States Department of Transportation, 2007)
2.5 Mixing mechanism
2.5.1 Mechanical mixing
There are many aspects that scientists have to cope with in the mixing industry, for
instance, mixing of solid particles, liquids, gases and liquids, and cohesive powders.
Generally, the aim of a mixing procedure is to reduce the scale and intensity of segregation,
39
for example, size segregation, making the most homogeneity of mixing ingredients (Harnby
et al., 2001). The ideal mixture with homogeneity is illustrated in Figure 17. In this
research, just the aspect of mixing cohesive powders is considered as asphalt mixtures (due
to bitumen having a liquid state at high mixing temperature causing interparticulate forces
among aggregate particles) possess relatively comparable characteristics to those of
cohesive powders.
Figure 17: Ideal mixture with homogeneity (Harnby et al., 2001)
In order to obtain the most homogeneity of the mixture, the external forces, for instance,
rotational, vibrating effects must be satisfied to break any bonds that exist between particles
and relocate these particles in the mixture. Otherwise, segregation will occur and cause
adverse effects to the quality of the mixture. There are many factors affecting the mixing
process that lead to segregation, for instance, type of particles, interparticulate forces, type
of mixing machine, and mixing time.
In term of interparticulate forces among particles, there are electrostatic bonding, van der
Waal’s forces, and liquid bridge bonding. If these particulate forces are not deactivated
during mixing process, there will be movements of small free flowing agglomerates of
aggregate particles (Figure 18). Particle types also attribute to the magnitude of liquid
bridge bonding. For instance, if the surfaces of particles are not smooth, the surface
roughness will reduce the contact areas hence reduce the bonding forces. In addition,
different sizes of particle also cause segregation as these particles will have different
velocities during mixing process (Harnby et al., 2001).
40
Figure 18: Mixture with “self-loving” particles (Harnby et al., 2001)
Production of conventional asphalt mixture comprises weighing, heating aggregate, and
blending hot virgin aggregate with bitumen binder. During the whole process, all the
materials are rolled centrifugally in a mixing vessel. The heating time of aggregate is aimed
to dry the material. In addition, heating and the rotational effects on aggregate create
aggregate movement, breaking up all the bonds between aggregate particles, due to
moisture, electrostatic charging, van der Waal’s forces. This is to make a homogeneous
blend of different particle sizes of aggregate. In addition, the drying process also separates
all the particles, even filler, before mixing with bitumen so binder can cover each particle
surface to improve the bonding between particles in the final mixture.
However, as the size of most aggregate particles is larger than 75µm, the flow of every
single aggregate particle is considered free due to the fact that electrostatic charging and
van der Waal’s forces among particles bigger than 75µm are inconsiderable and can be
ignored (Harnby et al., 2001). In addition, as the cohesion among particles due to existence
of water is not considerable compared to that of bitumen, hence just the bitumen contributes
to the bond among aggregate particles. If the mixing machine and procedure cannot
overcome these issues, for instance, breaking the bond between aggregate particles,
relocating all the particles so the mixture can reach the homogeneous status, segregation
will occur.
The mixing mechanism of recycled asphalt mixture is quite different from that of
conventional asphalt mixture. Both types of asphalt mixtures are finally, the combinations
41
of the same type of materials, bitumen binder and aggregate. However, conventional
asphalt mixture is a combination of virgin aggregate and bitumen binder while in recycled
mixture, the aggregate is not just virgin aggregate but also RAP, a blend of aged bitumen
and aggregate as well. In fact, this ultimate difference between two types of mixture makes
the production, in fact, mixing mechanism quite different.
The production of recycled asphalt includes the following steps:
- superheating virgin aggregate
- blending with RAP
- mixing the blend of virgin aggregate and RAP with virgin bitumen
“Black rock” concept
The ultimate characteristic distinguishing RAP (reclaimed asphalt pavement) from virgin
materials is the content of aged bitumen. This proportion of aged bitumen makes RAP not
conform to any requirements of normal design methods for virgin mixture, for instance,
Hveem, Marshall, or Superpave. Although the RAP aggregate covered by aged binder is
considered stiff, inert as black rock at ambient temperature, the level that a proportion of
aged bitumen might be incorporated with virgin binder during the mixing process and
through service life of pavement has still been ambiguously identified.
Research has been carried out to investigate whether RAP material acts as black rocks once
accommodated in the recycled mixture (McDaniel et al., 2000). In this study, three types of
RAP, low stiffness RAP from Florida (FL), medium stiffness RAP from Connecticut (CT),
and high stiffened RAP from Arizona (AZ) were used. All the RAP materials were
processed to less than ½ inch (12.5 mm) size. Viscosity of RAP binder is showed in Table
8. Two soft virgin binders, PG 52-43 and PG 64-22 were used as rejuvenators (Table 9).
The proportions of RAP in the recycled mixture were 10, and 40%.
RAP Source Viscosity at 60
o
C, Poise
FL 23760
CT 65191
AZ 124975
Table 8: Viscosity at 60
o
C (Poise) of RAP binder (McDaniel et al., 2000)
42
Virgin Recovered RAP Binders
Binders (Unaged)
Aging Property PG 52-34 PG 64-22 FL CT AZ
Original High Temp. Stiffness 53.9 67.8 82.2 82.4 89
RTFO High Temp. Stiffness 54.6 66.6 75.4 75.8 85.3
PAV
Intermediate Temp. Stiffness 11.5 21.7 19.3 25.1 33.8
BBR S -23.7 -18.1 -15.9 -15.1 -5.6
BBR m-value -25.9 -16.2 -16.4 -14.4 -7.1
PG
Actual (Critical Temperature) PG 53-33 PG 66-26 PG 82-25 PG 82-24 PG 89-15
MP1 (Performance Grade) PG 52-28 PG 64-22 PG 82-22 PG 82-22 PG 88-10
Table 9: Critical temperature and performance grades of virgin and recovered RAP binders
Test samples are designed to simulate three possible cases with different levels of blending,
total blending, actual blending, and black rock. In the total blending, the aged bitumen is
extracted and recovered from the RAP before being mixed with RAP and virgin aggregate.
This total blending simulates what occurs during the design process. In case of actual
practice, RAP and virgin aggregate are mixed directly before being blended with virgin
bitumen. For the black rock case, just RAP aggregate is extracted and mixed up with virgin
aggregate and binder.
To prepare for the samples, all the virgin aggregate, and RAP aggregate in case of black
rock condition, are heated overnight at 150
o
C. RAP material is also preheated at 110
o
C for
the duration of 2 hours before mixing. Preheated temperatures of virgin binder rely on the
performance grade: 155-160
o
C for PG 64-22 and 135-140
o
C for PG 52-34. After mixing,
the loose mixture is held in the oven at 135
o
C for 4 hours successfully to simulate the short
term ageing. In terms of long term ageing, the compacted sample is stored in condition of
85
o
C for five days continuously. All the specimens are compacted by Superpave gyratory
compactor at different temperatures, 143-148
o
C for PG 64-22, and 122-130
o
C for PG 58-
34.
Superpave tests including frequency sweep (SF), simple shear (SS), repeated shear at
constant height (RSCH), indirect tensile creep (ITC) and indirect tensile strength (ITS) are
employed to identify the differences in properties of three blending situations, total
blending, actual blending, and black rock, with both 10 and 40% of RAP. These tests cover
a wide range of asphalt properties, fatigue, rutting, and low temperature cracking.
43
Frequency sweep at constant height is used to determine the complex shear modulus and
phase angle of asphalt mixture. In this test, repeated shear load is applied on the test
specimen to produce a horizontal strain of 0.005%. Axial stress is also applied to keep the
specimen height constant. The results at frequencies of 0.01, 10 Hz and temperatures of 20,
40
o
C show that stiffness of the samples increases in accordance with the increase of RAP
proportion for total and actual blending cases. On the contrary with the black rock situation,
stiffness of samples is the same with 10 and 40% of RAP.
In simple shear at constant height test, the shear load is increased at the rate of 70kPa/sec
until it reaches the specified shear load relevant with test temperatures of 4, 20, and 40
o
C
(Procedure D of AASHTO TP7-94). The load is held constant for 10 seconds before being
released at a rate of 25kPa/sec. Maximum deformation results from the test indicates that
with the same RAP, the maximum deformation increases if the softer virgin binder is
applied. This is similar with increasing test temperature.
To evaluate rutting phenomenon of asphalt mixture, the repeated shear at constant height
test is employed. Plastic shear strain of mixture is determined under given loading mode
and temperature (Procedure C, AASHTO TP7-94). During the test, the stress controlled
shear load is applied to specimen until the number of loading cycles, each consisting of 0.1
sec loading time and 0.6 sec for rest period, reaches 5000 or the permanent strain exceeds
5%. Test temperature is 58
o
C for the PG 64-22 and 52
o
C for the PG 52-34 binder. The test
result also indicates that with 10% of RAP, the recycled mixture with softer virgin binder
has higher plastic shear strain. With the same virgin binder, the plastic shear strain is not
influenced substantially with different sources of RAP and blending situation. However,
plastic shear deformation of black rock case is considerably higher than those of total
blending and actual practice with 40% RAP content.
Indirect tensile test (AASHTO TP9) is used to analyze thermal susceptibility and low
temperature cracking of asphalt mixture. The relation of load magnitude, deformation, and
loading time was studied at three temperatures 0, -10, and -20
o
C. Horizontal and vertical
deformations of the specimen due to the static compression load applied across the
diametric plane of specimen are recorded over a period of time (240 seconds) to calculate
the creep compliance. On the other hand, in terms of temperature cracking, specimen is
tested at -10
o
C by applying the load with strain rate of 12.5 mm/minute until fracture
44
occurs. The test result shows that with 10% RAP contents, although there is not substantial
variation, the stiffness of the actual practice has tendency of being between black rock and
total blending cases.
Data from all the tests showed that with 10% RAP, among 66 combinations, there were 36
cases in which the test results of actual blending, total practice and black rock were
approximately the same, 9 cases that actual practice and total blending were almost similar,
and only 6 cases in which actual blending resembled the black rock. The overall trend
indicated that the there is no considerable differences between actual blending, total
practice and black rock cases.
The test results with 40% RAP showed inconsistent trend. There were 21 cases where the
test results showed no considerable variations between total blending and actual practice.
There are 12 cases where the results from total blending, actual practice, and black rock
were different. In fact, 10 out of 12 cases occurred with PG 64-22. In addition, there were 3
cases where actual blending resembled the black rock.
Due to the results from the tests, where 6 cases actual blending resembled the black rock
with 10% RAP, 3 cases with 40% of RAP, McDaniel et al. (2000) concluded that RAP did
not work as “Black Rock” or inert component in the recycled mixture. Actually, RAP
binders interacted with virgin binder and the mixture generally had approximately the same
properties as that of complete blending case. However, the laboratory mixing procedure
used in this research is quite different from that of the mixing plant. The long RAP
preheating time might enhance the interaction between RAP and virgin binder. McDaniel et
al. (2000) also carried out a study to compare the laboratory mixing procedure with that of
an industrial mixer. The test results showed that the laboratory mixture processed the same
characters as those of a mixture mixed by real mixing plant. However, the proportion of
RAP was only 15%. It has been argued whether the complete blending obtained in the
laboratory can actually occur in an industrial mixer, especially in the case where high
proportion and larger size of RAP is used.
Laboratory procedures for recycled mixture preparation
In general, the procedures for preparing the recycled mixture include heating the RAP to a
specified temperature for a certain period, and mixing with preheated rejuvenator before
45
compaction. For instance, Carpenter and Wolosick (1980) heated the RAP to 116
o
C before
mixing with rejuvenator. Noureldin and Wood (1987), heated the RAP at 116
o
C for 30
minutes before mixing with preheated rejuvenator (AC 2.5, AE-150, and Mobilsol-30) at
82
o
C. The loose mixture was then conditioned at 60
o
C for 15 hours. McDaniel el al. (2000),
heated the RAP to 110
o
C for 2 hours mixing with rejuvenator at required mixing
temperature. The requirement of heating RAP is to make the RAP workable so that it can
be mixed with rejuvenator (McDaniel and Anderson, 2001). In fact, the purpose of
preheating RAP is to soften the RAP binder to break RAP into separate pieces so the
rejuvenator can cover the RAP for diffusion process. Preheating time was also used for
preventing the effect of RAP moisture content on the properties of the recycled mixture
(Stephens et al., 2001).
Effect of RAP preheating duration
Stephens et al. (2001) studied the effect of RAP preheating time on the strength of recycled
mixture with the hypothesis that if RAP acted as black rock, the effect of preheating on the
recycled mixture would be insignificant. Samples were prepared with the same 15% RAP
content except with preheating times from 0 to 540 minutes. There were also samples with
the same aggregate gradation and virgin binder for comparison. All the samples were
subjected to indirect tensile and unconfined compression tests. The test results indicated
that the longer the RAP preheating time, the higher the strength of the recycled mixtures
(Figure 19).
Stephens et al. (2001) concluded that the increase of indirect tensile and unconfined
compression strength was attributed to the long RAP preheating time, the lump of RAP was
totally heated through and broken down during mixing for complete blending. This was not
a firm conclusion as the increase of indirect tensile and unconfined compression strength
might also be accredited to the RAP being hardened due to exposure at high temperature
during the long preheating time.
However, McDaniel et al. (2000) investigated the effect of preheating time on the
properties of RAP binder. Two types of RAP taken from Arizona and Florida were
subjected to different preheating times and temperatures. The properties of original RAP
binders were compared to those of the RAP after different preheating time. The test results
(Figures 20 and 21) indicated that the complex modulus of RAP after two hours preheated
46
at 110 and 150
o
C did not change considerably. This supports the conclusion of Stephen et
al. (2001) during the first two hours of preheating, the increase in indirect tensile and
unconfined compression strength of the mixture is attributed to complete blending with
virgin binder as the RAP lump is heated through and completely separated.
Figure 19: RAP preheating time versus indirect tensile and unconfined compression strength (Stephens
et al., 2001)
Figure 20: Arizona RAP binder complex modulus versus preheating time and temperature (McDaniel
et al., 2000)
47
Figure 21: Florida RAP binder versus preheating time and temperature (McDaniel et al., 2000)
2.5.2 Diffusion process or chemical mixing
The performance of recycled asphalt mixtures is not only affected by the quality of
mechanical mixing but also the interaction between RAP and virgin binder. The mechanical
mixing only supports the fact that virgin binder can coat the particles covered by RAP
bitumen. Meanwhile, the requirement of the recycling process is for the virgin binder to be
well mixed and penetrate or diffuse into RAP bitumen so as to reduce the viscosity and
bring back its expected requirements. However, the mechanism of diffusion is not well
understood even though asphalt recycling is not a new aspect in the industry.
Diffusion mechanism
There have been some studies carried out in the past to investigate the diffusion mechanism
of virgin binder into RAP bitumen. Carpenter and Wolosick (1980) undertook research to
evaluate the effects of modifiers on the performance of recycled asphalt mixture. The
reclaimed asphalt was taken from a city street in Champaign, Illinois in June, 1976 and
crushed until passing 12.5 mm sieve. RAP bitumen was then extracted and tested. The
properties of RAP bitumen and the grading of RAP aggregate are presented in Tables 10
and 11.
48
Properties Value
Viscosity at 60
o
C (Pa.s)
Penetration at 25
o
C (dmm)
Penetration at 4
o
C (dmm)
Softening point (
o
C)
Asphalt content (%)
Specific gravity (g/cm
3
)
4490
26
22
63
5.3
1.198
Table 10: Properties of RAP binder (Carpenter and Wolosick, 1980)
Sieves Size Percent Passing
12.5 mm
9.5 mm
6.3 mm
4.75 mm
2.00 mm
850 µm
425 µm
150 µm
75 µm
100
81
78
68
58
34
23
13
9
Table 11: Grading of RAP aggregate (Carpenter and Wolosick, 1980)
Carpenter and Wolosick used 100% of RAP for this research and Paxole 1009, viscosity at
60
o
C of 234 mm
2
/s (0.23 Pa.s) and specific gravity of 1.028, was employed as rejuvenator.
The amount of rejuvenator was determined after trial blending with different proportions of
rejuvenator over RAP binder. The result of the blending trial is a graph showing the relation
between viscosity of aged bitumen, rejuvenator and % of rejuvenator. In order to get the
recycled binder with target viscosity of 100 Pa.s, the amount of rejuvenator is 20% of the
weight of RAP bitumen.
Resilient modulus, creep compliance, and permanent deformation tests were implemented
in this research. In order to study the influence of diffusion mechanism on the properties
and performance of recycled mixture, two types of samples, rejuvenated and recycled, were
prepared. Both had the same material, bitumen and air void content but different methods of
preparation. With rejuvenated samples, RAP bitumen was extracted and blended with
49
rejuvenator before being mixed with RAP aggregate. Differently in recycled sample,
reclaimed materials, after being heated at 116
o
C, was mixed directly with rejuvenator to
simulate the mixing procedure in the asphalt industry. The samples then were tested at
predetermined time intervals. The difference between the test results of recycled and
rejuvenated mixture was studied to evaluate the influence of diffusion on the recycled
samples.
Figure 22: Effect of diffusion on resilient modulus (Carpenter and Wolosick, 1980)
The test results showed that the resilient modulus of the recycled mixture critically
decreased during the first two weeks before starting to increase again (Figure 22). The
phenomenon did not happen in the case of rejuvenated samples. In the rejuvenated samples,
the test result is almost the same with different testing times.
Carpenter and Wolosick (1980) simply developed the diffusion model of rejuvenator into
RAP bitumen. The diffusion process included the following steps:
1. the rejuvenator forms a low-viscosity layer covering the RAP particles
2. rejuvenator starts to penetrate into RAP bitumen, simultaneously softening the aged
bitumen and reducing the amount of rejuvenator
3. all the rejuvenator diffused into RAP bitumen, viscosity of the aged binder coated
aggregate decrease and that of outer layer (rejuvenator) increases.
4. The blend of rejuvenator and aged bitumen reaches equilibrium
In step 1, there is almost no interaction as rejuvenator just coats the RAP particles (Figure
23). Hence, the viscosity of rejuvenator and aged binder remain approximately the same.
Rejuvenator then starts to diffuse into aged binder. The proportion of rejuvenator that
50
covers the RAP particle will gradually decrease. The mutual interaction between
rejuvenator and aged binder results in the reducing viscosity of the outer layer of aged
binder. Simultaneously, the viscosity of the remaining rejuvenator increases (Steps 2 and
3). The process will progress until equilibrium status where all the rejuvenator diffuses into
aged binder and generates a homogeneous blend (Step 4).
Figure 23: Schematic of modifier coating an aggregate particle during recycling process (Carpenter and
Wolosick, 1980)
This simple diffusion mechanism can be used to interpret the critical duration in which the
resilient modulus decreases and permanent deformation increases. Figure 24 shows after
compaction, as the penetration has not started, the binding force between aggregate
particles is due to the high-viscosity RAP bitumen. Actually, the rejuvenator at this moment
has been located in the voids among particles. This situation leads to the fact that the
51
resilient modulus of recycled mixture is higher than that of rejuvenated sample. As the
diffusion or penetration of rejuvenator into RAP bitumen starts, the resilient modulus
decreases as the outer layer of RAP bitumen begins to be softened by rejuvenator. This
phenomenon is also due to the remaining low-viscosity rejuvenator starting to bind the
aggregate particles together. The resilient modulus keeps decreasing until no rejuvenator
remains. At this point, the viscosity of outer layer keeps increasing until the blend between
rejuvenator and RAP bitumen reaches equilibrium.
Figure 24: Diffusion model (Carpenter and Wolosick, 1980)
Staged-extraction process
Carpenter and Wolosick (1980) also carried out an extra experiment to verify the developed
diffusion model. In this experiment, the binder is divided into two layers, outer layer and
inner layer. If the diffusion of rejuvenator into RAP bitumen exists, the consistency of each
layer will be different in relation to different testing time intervals.
52
The mixture is prepared with the same procedure as that for the performance tests.
However, the mixture is left uncompacted. In order to divide the bitumen coat into two
layers, a sample of loose mixture is immersed in trichloroethelyne for 3 minutes. The
bitumen recovered by Abson method will represent the outer layer. The inner layer is
achieved by washing and recovering all the bitumen. The consistency of these recovered
bituminous materials will be tested. The whole procedure will be repeated at different
times. The results (Figure 25) show that simultaneously, the penetration of inner layers
increases and that of outer layers decreases until both penetrations are the same.
Figure 25: Penetrations of outer and inner layers as function of time (Carpenter and Wolosick, 1980)
The research carried out by Carpenter and Wolosick (1980) did not take into account the
influence of virgin aggregate during mixing process. This research dealt only with the
diffusion mechanism of the mixture of 100% RAP and rejuvenator. This might not
represent the phenomenon that occurs in recycled mixture containing virgin aggregate.
Depending on the efficiency of mixing process, there might be not only RAP particles
coating by rejuvenator but also virgin aggregate particles coating by rejuvenator or blend of
RAP binder and rejuvenator in recycle mixture containing virgin aggregate.
Noureldin and Wood (1987) also studied the diffusion of rejuvenator into RAP bitumen. In
this research, not only the mixture of RAP and rejuvenator but also the mixture of RAP,
rejuvenator and the addition of virgin aggregate were considered. The existence of virgin
aggregate in the mixture aimed to simulate the real situation in the recycling asphalt
industry. RAP material was milled from road US-52 in Indianapolis (Indiana). Properties of
RAP binder are showed in Table 12.
53
Three rejuvenators, AC 2.5, Mobisol 30, and AE 150, were used. The amount of
rejuvenators and percentage of RAP bitumen were estimated based on the Asphalt Institute
design recycled asphalt mixture method (Arrhenius viscosity mixing equation). The target
was that after being recycled by rejuvenator, the recycled bitumen must have the viscosity
at 60
o
C approximately similar to that of AC 20 (from 190 to 240 Pa.s). There were three
combinations of RAP bitumen and rejuvenator, 40% RAP bitumen/60% AC 2.5, 45% RAP
bitumen/55% AE 150, and 85 %RAP bitumen/15% Mobilsol 30.
Properties Value
Penetration at 25
o
C, (dmm)
Viscosity at 60
o
C (Pa.s)
Kinematic Viscosity at 135
o
C (cSt)
Softening Point (
o
C)
Bitumen content (%)
28
2089
726
60
6
Table 12: Properties of RAP bitumen (Noureldin and Wood, 1987)
Noureldin and Wood (1987) also used the staged extraction method. However, different
from Carpenter and Wolosick (1980), the bitumen coat was divided into four microlayers.
One advantage of this method is to show the non-uniform ageing pattern of RAP material.
After four microlayers extraction of RAP materials, the results (Table 13) indicated that
RAP bitumen from the outer two layers were seriously hardened. On the contrary, the two
inner layers close to the aggregate surface were slightly aged, the consistency were almost
the same as those of the original bitumen AC 20.
Solvent increment
(mL)
Binder
(% by weight)
Penetration
(dmm)
Viscosity at
60
o
C (Pa.s)
200
200
300
700
55.5
26.5
11.2
6.8
24
33
65
57
2400
1500
250
330
Table 13: Test results on reclaimed staged-extraction of RAP (Noureldin and Wood, 1987)
To prepare for the sample, RAP was heated at 115
o
C for 30 minutes and rejuvenator at
82
o
C before mixing together with virgin aggregate for 2 minutes. Virgin aggregate was also
54
heated at 115
o
C for 30 minutes. The loose mixture was then preserved at 60
o
C for 15 hours.
To obtain the bitumen of each microlayer, the sample of 1200g was in turn immersed in
200, 200, 300, and 700 mL of trichloroethylene for 5 minutes. Bitumen of each microlayer
was recovered by Abson method and its consistency was determined.
In the case of mixing only RAP with rejuvenators, results from staged-extraction tests
showed that all the rejuvenators could restore the consistency of the two outer microlayers.
However, the other two inner microlayers showed almost unchanged tendency (Table 14).
Due to the test being carried out just at one point in time, the result could not show the
changing tendency of each layers consistency.
Binder
Solvent
Increment
(mL)
Binder
(% by
Weight)
Penetration
at 25
o
C
(dmm)
Viscosity
at 60
o
C
(Pa.s)
60% AC 2.5
40% RAP binder
200
200
300
700
67.5
21.5
7
4
67
68
59
50
167.4
188.0
239.4
300.0
55% AE 150
45% RAP binder
200
200
300
700
69
16.5
8.5
6
75
70
62
49
168.3
201.0
229.0
302.0
15% Mobilsol 30
85% RAP binder
200
200
300
700
71
18
6
4
75
69
63
48
186.4
198.0
204.0
315.2
Table 14: Test results on reclaimed, staged-extraction, no virgin aggregate (Noureldin and Wood, 1987)
In the case where the recycled mixture was a combination of RAP material, rejuvenator,
and virgin aggregate, the amount of aggregate was estimated so the recycled binder
accounted for 6% by weight of the mixture. The grading of virgin aggregate was also
selected hence the gradation of the whole mixture satisfied the requirement of Indiana
specification. In accordance to 6% bitumen over total weight of the mixture, the amount of
aggregate added were in turn 60, 55, and 15% in relation with mixture using rejuvenator
55
AC 2.5, AE 150, and Mobilsol 30. The staged-extraction test results showed that only the
mixture using rejuvenator AE 150 had the same tendency as that of mixture using only
RAP and rejuvenators. The other two using rejuvenators AC 2.5 and Mobilsol 30 had
different trends (Table 15). The viscosities of inner layers were higher than those of outer
layers.
The shortcoming of this research was that the effect of time on the diffusion process was
not considered. In addition, the use of a single diffusion pattern could not fully describe the
diffusion mechanism that occurs in recycled asphalt mixture. The inconsistent viscosity
pattern of micro-layers compared to that of the mixture using only RAP and rejuvenator
indicated the segregation of the bitumen phase in recycled mixture. This substantiates the
fact that in recycled mixture, there exist RAP particles covered by rejuvenator, virgin
aggregate particle covered by rejuvenator, and RAP or virgin aggregate particles covered
by blend of aged binder and rejuvenator. Each situation has its own diffusion mechanism.
Binder
Solvent
Increment
(mL)
Binder
(% by
Weight)
Penetration
at 25
o
C
(dmm)
Viscosity
at 60
o
C
(Pa.s)
60% AC 2.5
40% RAP binder
200
200
300
700
72
19
5.5
3.5
60
51
52
130
210.0
289.2
247.0
80.9
55% AE 150
45% RAP binder
200
200
300
700
71
19
6
4
70
67
60
50
197.2
173.4
242.4
361.6
15% Mobilsol 30
85% RAP binder
200
200
300
700
74
17.5
5.5
3.5
73
80
90
100
204.9
166.4
126.0
124.0
Table 15:Tests results on reclaimed, staged-extraction, virgin aggregate used (Noureldin and Wood,
1987)
56
A staged-extraction process was also used by Huang et al. (2005) to investigate the effects
of rejuvenator on RAP material. In this research, Huang et al. (2005) used the combination
of RAP, rejuvenator, and virgin aggregate. However, just the fine RAP, containing No 4
(4.75mm) passing RAP particles was used. Limestone was used as virgin aggregate. The
gradation of RAP and virgin aggregate are showed in Tables 16 and 17. However, all the
particles passing size No 4 were removed before mixing with RAP. The mixing process
simulates the procedure in the practical industry, 20% RAP, virgin bitumen, and aggregate
were mixed together at 190
o
C. After mixing, the rejuvenated RAP was easily separated
from the whole mixture due to size difference.
Sieves Size % Pass
No.4
No.8
No.30
No.50
No. 100
No. 200
100
81
46
30
23.2
19.3
Table 16: Properties of RAP aggregate (Huang et al., 2005)
Sieves Size % Pass
37.5 mm
25.4 mm
19 mm
12.7 mm
9.5 mm
4.75 mm
100
97.6
77.7
35.3
14.3
1.9
Table 17: Properties of Virgin Aggregate (Huang et al., 2005)
The rejuvenated RAP was then recovered under staged-extraction process. Rejuvenated
RAP was soaked in solvent trichcloethylene for 3 minute. The binder was then recovered
by Abson method. The process is repeated three times for the first three microlayers. The
layer in contact with aggregate surface was obtained by washing the remaining rejuvenated
RAP with solvent. The recovered rejuvenated RAP bitumen is then subjected to rheological
testing to identify the difference between layers. The schematic of extracted layers is
illustrated in Figure 26.
57
Figure 26: Layers extraction process (Huang et al., 2005)
The test results for viscosities of bitumen from extracted layers at different temperatures
showed uniform tendency (Figure 27). The outer layer due to rejuvenation had lower
viscosities at 135
o
C than those of the inners layers. The layer in contact with surface of
aggregate was the stiffest. It was concluded that after mixing, about 40% of RAP binder
was blended with rejuvenator. By washing all the bitumen on virgin aggregate, Huang et al.
(2005) also demonstrated that during mixing process, approximately 6-6.8% of RAP binder
was transferred from RAP materials to virgin aggregate.
Figure 27: Viscosity at 135
o
C of different micro-layers coated RAP particles (Huang et al., 2005)
58
Investigating the diffusion mechanism using marker
Karlsson and Isacsson (2003) also carried out a study to examine the diffusion of bitumen
rejuvenator. In this study, the diffusion was not characterized by accessing the consistency
or rheology of rejuvenated binder microlayers. The diffusion mechanism was investigated
by measuring the variations in energy absorption capability of bitumen layers over a period
of time by FTIR-ATR (Fourier Transform Infrared Spectroscopy by Attenuated Total
Reflectance).
A Mattson Infinity 60 AR spectrophotometer was used in this research (Karlsson and
Isacsson, 2003a). This equipment includes a non-absorbing trapezoidal prism made of ZnSe
(Figure 28). Two frames, with different slot width are stuck together on the surface of the
prism. The thicknesses of each frame are in turn 200 and 500 µm in accordance with the
thickness of top and bottom layers. To prepare for the specimen, each bitumen layer is
scraped into the mould by two scrapers that fit into the frames.
Figure 28: Schematic of FTIR – ART (Karlsson and Isacsson, 2003a)
The philosophy of this method is to measure the movement of chemical analytes through
bitumen layers. As the top bitumen layer contains analyte substance, the movement of the
analyte also means the penetration of that bitumen layer into the bottom one. Each analyte
has particular absorption wavelength or wave number (Table 18). During 72 hours, the
changes in energy absorption are scanned with a resolution of 4 cm-1 and recorded every 1
or 5 minutes. The number of scans performed during each recorded interval is in turn 64 or
256 (Karlsson and Isacsson, 2003c). The absorption value at certain time will be
automatically calculated by WinFirst software using wavelength and relevant energy
absorbed. Figure 29 is an example of absorbance difference versus time in relation with
various wavelengths.
59
Table 18: Approximate absorption wave length of functional groups (Petersen, 1986)
Figure 29: Spectra obtained after 10, 50, 100, 500, and 925 minutes during diffusion of rejuvenator into
A-B180 at 60
o
C (Karlsson and Isacsson, 2003a)
Based on the calculated absorption results, the diffusion coefficient is estimated by
mathematical expression of Fick’s Law (Karlsson and Isacsson, 2003b). Fick’s model is
simply described in Figure 30. Initially, the concentration of bitumen is 0 and that of
rejuvenator is c
o
. Both layers have total thickness of L in which rejuvenator thickness
accounts for (1- α)L. The diffusion process is assumed to occur in constant pressure and
temperature.
2
2
.
x
c
D
t
c
c
c
=
c
c
(10)
c: concentration in terms of time
60
D: diffusion coefficient
t: time
x: position
Figure 30: Schematic of Fick’s Law diffusion model (Karlsson and Isacsson, 2003b)
To solve the equation of Fick’s Law with boundaries expressed in Figure 30, the result is
the mathematical relation between the concentration of rejuvenator at position x and time t
as follows:
( )
¿
·
=
(
¸
(

¸

÷
(
¸
(

¸

÷ ÷ =
1
0
0
2
. cos .
) sin(
.
2
). 1 ( ,
n
Dt
L
n
e
L
x n
n
n c
c t x c

  


(11)
(Karlsson and Isacsson, 2003b)
There is a wide variety of materials used by Karlsson and Isacsson (2003). The bitumen in
this research includes:
- Rejuvenator V115 from Nynas (R115) which is a heavy naphthetic petroleum
distillate
- B180 from Mexico having penetration of 180 (A-B180)
- B180 from Saudi Arabia, penetration of 180 (B-B180)
- B85 from Venezuela, penetration 85 (C-B85)
- B60 from Venezuela, penetration 60 (C-B60)
The study was carried out under conditions of different testing temperatures and thickness
of bitumen and rejuvenator layers. Results from the test indicated that the diffusion of
rejuvenator was influenced by many factors, for instance, temperature, type of bitumen,
rejuvenator, and the chemical composition.
Influence of type of bitumen on diffusion coefficient
Rejuvenator R115 is used together with three bitumens A-B180, B-B180 and C-B60. The
thickness of bitumen and rejuvenators are equally 500 µm. The experiment is carried out at
different temperatures. The results (Figure 31) indicate that the diffusion coefficients of
61
rejuvenator into bitumen A-B180 and B-B180 are almost the same and higher than B-C60.
This phenomenon could be attributed to the fact that bitumen B-C60 has lower penetration
than those of bitumens A-B180 and B-B180.
Figure 31: Influence of bitumen type on diffusion coefficient (Karlsson and Isacsson, 2003b)
Influence of temperature on diffusion coefficient
In this aspect, rejuvenator R115 is used together with A-B180. However, the test is repeated
with different temperatures and four combinations of bitumen-rejuvenator layer thickness;
200/200 µm, 200/500 µm, 500/200 µm, and 500/500 µm. The test results (Figure 32)
indicate a minor effect of layer thickness on the diffusion coefficient. However, below
90
o
C, the diffusion coefficient in relation with different layer thickness deviates slightly. In
general, the experimental results show the increasing tendency of diffusion coefficient if the
test temperature increases.
Figure 32: Influence of temperature on diffusion coefficient (Karlsson and Isacsson, 2003b)
62
Influences of chemical composition of rejuvenators on diffusion coefficient
To investigate the effect of composition on diffusion mechanism, a wide variety of markers
are used together with bitumen A-B180. The proportion of marker is 3% by weight of the
bitumen. Bitumen A-B180 mixed with different markers are in turn applied on top of pure
A-B180. The test results (Figure 33) demonstrate that generally, the diffusion coefficient
decreases if the molecular weight of marker increases. The same result has been reported by
(Qiu and Bousmina, 2003).
Figure 33: Influence of chemical composition of markers on diffusion coefficient (Karlsson and
Isacsson, 2003b)
2.6 Segregation and consequences
2.6.1 Segregation
Segregation is defined traditionally as non-homogeneity of the asphalt mixture. Actually,
segregation can be perceived as the concentration of coarse or fine materials in some areas
of the paved mat in conventional asphalt mixture (Brock et al., 2003). In asphalt recycling
aspect, there is not only the concentration of certain sizes of aggregate in one area but also
the concentration of different binders in some areas. This happens in the case where the
virgin bitumen or rejuvenator is not well distributed in the whole mixture or even when
these two bitumens are well distributed, the diffusion process between these two binders
cannot take place.
RAP material generally has different characteristics from that of pure virgin aggregate or
bitumen. At ambient temperatures, aged bitumen normally with high viscosity, will act like
63
a solid and RAP has the same characteristics as those of black rock. However, when the
temperature is high enough to turn aged bitumen into liquid state, RAP material will no
longer act like black rock but be the same as the blend of aggregate and liquid bitumen.
During the recycled asphalt production, after RAP at ambient temperature is blended with
superheated virgin aggregate, the energy transferred from virgin aggregate will heat up
RAP. The increase of RAP temperature will gradually change the aged bitumen from solid
to liquid state. Therefore, RAP material will change from a mixture of solid agglomerates
into a mixture of solid particles with different sizes and aged liquid bitumen. The mixing
process transforms from mixing among solid particles to mixing between solid particles and
liquid bitumen. Hence, while blending virgin aggregate with RAP, there exist not only inter
particulate forces as in virgin aggregate case but also the bridge force due to aged bitumen
among aggregate particles.
The aim of heating RAP is also to use thermal energy transferred from superheated
aggregate to soften the bond between aggregate particles as the higher the temperature, the
lower the viscosity of aged bitumen. Once the viscosity of RAP binder reaches the critical
point that the bond cannot hold two particles together, two particles will be separated and
relocated. This mechanism will repeat during the mixing process in order to get a
homogeneous blend between RAP and virgin aggregate. If the assumption is made that each
RAP particle is black, with white colour for virgin aggregate, the ideal homogeneous
mixture is illustrated in Figure 17.
However, this mechanism is affected by many factors. The first issue is that there are many
sizes of RAP available, for instance, 50mm, 30mm, and 20mm. In addition, each size of
RAP may not be made of one particle but probably comprise many smaller-sized particles.
Hence, the time for the heat transferred from virgin aggregate to soften and break RAP into
separate pieces for the relocation is quite different. According to heat transfer theory, the
larger the size the particle is, the longer the time for heat transfer (Cutnell and Johnson,
2004).
Energy transferred from superheated aggregate will heat up the RAP and soften the bond
between aggregate particles as the higher the temperature, the lower the viscosity of
bitumen binder. This is aimed to destroy the bond among aggregate to relocate the position
64
of each particle. If the blending time among RAP and virgin aggregate is not enough to
break RAP into separate pieces, certainly there will be some agglomerates of different
particle sizes of RAP. The same phenomenon also happens even when the duration and
temperature of mixing process are enough to make the whole RAP bitumen become liquid.
Depending on the sizes and proportion of aggregate in the mixture, there might still exist
agglomerates moving in the mixture, for instance, agglomerates of fillers and liquid
bitumen.
If the bitumen bond between aggregate is strong enough, the system will be dominated by
the free flowing of agglomerates of aggregate particles (Harnby et al., 2001). This mixture
is demonstrated in Figure 18. The black agglomerates comprising different aggregate
particles will be moving with virgin aggregate. If this situation exists, the virgin bitumen, or
rejuvenator cannot completely interact and recover the properties of aged bitumen. This
will lead to the fact that in the mixture, some particles are coated with aged bitumen which
is stiff, the others with soft rejuvenator.
The study by McDaniel el al. (2000) maintained that the actual practice and total blending
were almost the same. This conclusion could be applied for the case of 10% RAP as there
was a consistent trend of the results. However, there was apparently not enough support for
this conclusion with 40% of RAP in the recycled mixture (Huang et al., 2005). Even with
10% of RAP, the fact that total blending and actual practice were the same might not be
firmly proved as in this case, the data showed there were 36 cases that test results from total
blending, actual practice, and black rock were almost the same. The existence of small
amount of RAP, particular 10% in this research, did not affect substantially the properties
of the recycled mixture. If segregation occurs, there will be no considerable adversity to the
performance of the recycled mixture.
There would be segregation during the mixing process, especially with higher percentages
of RAP. For instance, with 40% RAP, there were 12 cases in which test results of total
blending, actual practice and, black rock cases were different (McDaniel et al., 2000). The
segregation was also identified in research by Noureldin and Wood (1987), where RAP was
mixed with 3 rejuvenators, AC 2.5, AE 150, Mobilso 30. In 2 out of 3 cases, after stage-
extractions and consistency tests, the inner layers had lower viscosities and higher
penetrations than those values of the outer ones. In the whole mixture, some aggregates are
65
covered by soft rejuvenator, some still covered by stiff RAP bitumen. RAP binder and
rejuvenator are not well mixed. This phenomenon is also reported in Huang et al (2005).
Huang et al. (2005) reported that, during mixing RAP and virgin binder, the RAP bitumen
transferred from RAP aggregate to virgin aggregate was 6 to 6.8%. After this combination
is mixed with rejuvenator, there would certainly be some aggregate covered by soft
rejuvenator first, and the others covered by stiff aged binder in the mixture. The segregation
consequently occurs due to the existence of bitumen with different consistency in the
mixture. This situation will also reduce the capability of diffusion process as in order to
diffuse efficiently; the rejuvenator must cover the RAP aggregates.
2.6.2 Consequences of segregation on the performance of asphalt
mixture
Segregation can cause adverse effects on the quality of the mixture as well as the
performance of the pavement during service life. Gardiner et al. (2000) stated that the
existence of segregation in the mixture can cause substantially:
- Decreased fatigue life in areas that have high concentration of coarse aggregate.
- Increased moisture damage due to high air voids caused by segregation.
- Increased rutting and raveling, especially with high volumes of traffic.
Gardiner and Brown (2000) implemented a study to assess the adverse effects of
segregation on the quality and performance of asphalt mixture. The data of segregation was
gathered from the field, using sections between 80 and 160 meters long. Based on the data
collected, the level of segregation compared to Job Mix Formula was classified as followed:
- Non-segregation: the percent passing any sieve differed less than 5%
- Low segregation: at least two sieves with a change more than 5%
- Medium segregation: at least two sieves with a change more than 10%
- High segregation: three sieves with a change more than 15%
The test samples then were prepared in the laboratory in accordance with the difference to
Job Mix Formula (Table 19). Parameters used to evaluate the effects of segregation to the
performance of mixture were permeability, resilient modulus, dynamic modulus, tensile
strength at dry and wet, and low temperature condition. The influence of segregation to
fatigue life was also evaluated. All the tests show the adverse effects of segregation to the
performance of mixture (Table 20).
66
Table 19: Gradations, bitumen contents, and target air voids of laboratory-simulated segregation
mixtures (Gardiner and Brown, 2000)
Table 20: Summary of the influence of segregation on mixture properties (Gardiner and Brown, 2000)
In the recycled mixture, beside those kinds of segregation normally occuring in
conventional mixtures, there also chemical segregation as the rejuvenator cannot diffuse
into RAP binder immediately. It might reduce the strength of the mixture or pavement after
67
being paved and compacted. Carpenter and Wolosick (1980) reported that the resilient
modulus of recycled mixture reduced during the first two weeks and then started to increase
to the equilibrium value due to the occurrence of diffusion process. There is also binder
segregation due to the rejuvenator not being well distributed in the mixture due to improper
mixing.
68
3 Laboratory RAP production
3.1 Introduction
One of the issues related to using RAP from the industry is the RAP variability. The
variability includes not only the gradation of RAP but also the RAP binder content and
origin. Each time RAP is acquired from the industry, the properties of RAP might be
different. Even if the amount of RAP necessary for the whole research is obtained, the
homogeneity of RAP is still not assured. The fact that RAP contains materials with
different or unknown origin might seriously affect the result of the research.
The purpose of laboratory RAP production is to eliminate the problematic variability of
RAP materials. Although the task is time consuming, laboratory RAP production helps to
control RAP aggregate gradation as well as RAP binder content and origin. In addition, one
of the objectives of this research is to study the effect of RAP sizes on mechanical
properties of recycled asphalt mixture. The laboratory RAP production also helps to assure
that every RAP piece is a lump, an agglomerate of RAP aggregate with different sizes
bound together by RAP binder.
3.2 Materials
3.2.1 Bitumen
The bitumen used for laboratory RAP production is 40/60 Pen bitumen supplied by Shell.
The properties of 40/60 Pen are shown in Table 21. The viscosity of 40/60 Pen bitumen is
zero shear viscosity extrapolated from DSR data using Cross model at 60
o
C (Section 4.3.1).
Penetration at 25
o
C (dmm) BS EN 1426 (2000) 50.6
Softening Point (
o
C) BS EN 1427 (2000) 56
Density (g/cm
3
) BS EN ISO 3838 (1996) 1.03
Viscosity at 60
o
C (Pa.s) DSR-ZSV 440
Table 21: Properties of bitumen 40/60 Pen
3.2.2 Aggregate
The aggregate for laboratory RAP production primarily comprises three nominal sizes, 10
mm, 6 mm and dust. The aggregate is supplied by Dene limestone quarry. Gradation of
each nominal size is in Table 22. The shaded areas present the aggregate sizes used in BS:
69
4987-1 (2005). After batching, the gradation of RAP aggregate conforms to the gradation
requirement for 10 mm DBM for close graded surface course (BS:4987-1, 2005) (Figure
34). The physical properties of each nominal size of aggregate are shown in Table 23.
Sieve Nominal size 10 Nominal size 6 Dust
Size
Pass
(%)
Retained
(%)
Pass
(%)
Retained
(%)
Pass
(%)
Retained
(%)
31.5 mm
20 mm
14 mm 100 0.00 100 0.00 100 0.00
10 mm 87.12 12.88 100 0.00 100 0.01
8 mm 46.25 40.87 99.47 0.53 100 0.02
6.3 mm 2.57 43.68 81.40 18.07 100 0.03
4 mm 1.79 0.78 16.91 64.48 98.53 1.47
2.8 mm 1.76 0.03 5.14 11.78 95.81 2.72
2.36 mm 5.14 0.00
2 mm 1.65 0.11 4.26 0.88 85.22 10.59
1 mm 1.60 0.05 4.19 0.07 62.08 23.14
0.5 mm 1.55 0.05 4.19 0.00 43.97 18.11
0.25 mm 1.32 0.23 4.12 0.07 31.10 12.87
0.125 mm 1.01 0.30 3.98 0.14 22.31 8.80
0.063 mm 0.61 0.40 3.76 0.21 17.69 4.61
Pan 0.61 3.76 17.69
Sum 100.00 100.00 100.00
Table 22: Gradation of Dene limestone aggregate
0
10
20
30
40
50
60
70
80
90
100
0.01 0.1 1 10 100
Sieve size (mm)
%

P
a
s
s
i
n
g
Upper Limit
Lower Limit
RAP aggregate gradation
Figure 34: Design gradation of RAP aggregate
70
Size 6mm Dust 10mm
Particle density
on an oven dried
basis g/cm
3
2.500 2.443 2.575
Particle density
on a saturated
surface-dried
basis g/cm
3
2.547 2.478 2.612
Apparent particle
density g/cm
3
2.622 2.533 2.673
Water absorption
%
1.87 1.46 1.42
Table 23: Properties of Dene limestone aggregate
3.3 Procedure for RAP manufacture
The bitumen content of RAP material is 5.2% by weight of the total mixture. Maximum
theoretical density of the mixture containing the designed aggregate gradation and 5.2%
bitumen, determined by BS EN 12697-5:2002, is 2453 kg/m
3
. The target air void content is
8%. Based on the target air void content and the maximum density of the mixture, the
amount of bitumen and each aggregate fractions are calculated thus after compaction, each
slab has the dimensions of 305 mm x 305 mm x 40 mm.
Procedure for manufacture of artificial aged slabs is as follows:
- Aggregate is heated overnight at 150
o
C
- Bitumen is heated at to 150
o
C for 3 hours
- Mixing for 2 minutes at 150
o
C
- Compact the loose mixture in the mould of 305 mm x 305 mm until the thickness of
40mm is reached.
- Remove the compacted specimens
- Condition in force draft oven at 85
o
C for 120 hours
3.4 Processing RAP
The artificial aged slabs are processed into two sizes, large RAP (denoted as LR) and small
RAP (SR). For large size of RAP, after conditioned at 100
o
C for 1 hour, artificially aged
slabs will be broken manually. The aim of the conditioning duration is to soften the RAP
and eliminate the degradation of RAP aggregate. During the breaking process, the size is
visually adjusted therefore the maximum dimension is less than 40 mm. Small RAP
71
material is obtained by crushing large RAP material using a jaw crusher. The gap between
two jaws of the crusher at static condition is set at 15 mm hence after being crushed, the
maximum dimension of small RAP material is about 20 mm. The gradation of RAP
materials are presented in Table 24. Figure 35 illustrates the appearance of both small and
large RAP.
Sieve Small RAP Large RAP
size
Pass
(%)
Retain
(%)
Pass
(%)
Retain
(%)
50 mm 100 0 100 0
37.5 mm 100 0 26.2 73.8
31.5 mm 100 0 0 26.2
20 mm 92.74 7.26
14 mm 47.82 44.92
10 mm 32.13 15.69
6.3 mm 13.6 18.53
4 mm 3.74 9.86
3.35 mm 2.23 1.51
2.36 mm 1.38 0.85
1.18 mm 0.38 1
Pan 0.38
Sum 100 100
Table 24: Gradation of processed RAP materials
Figure 35: Appearance of RAP materials
72
3.5 Determine RAP properties
3.5.1 RAP aggregate
Although the degradation of RAP is deliberately eliminated by 1 hour conditioning at
100
o
C, the RAP aggregate gradation might still alter during the crushing process. After
being extracted and recovered from small RAP material, RAP aggregate is subjected to
particle size distribution test (BS-EN:933-1, 1997). The result shows the gradation of RAP
aggregate after processing is almost the same as that of original (Figure 36).
0
10
20
30
40
50
60
70
80
90
100
0.01 0.1 1 10 100
Sieve size
%

P
a
s
s
i
n
g
Upper Limit
Lower Limit
RAP aggregate gradation
RAP aggregate gradation
after being crushed
Figure 36: Gradation of RAP aggregate before and after processing
3.5.2 RAP binder
RAP binder, after being extracted and recovered from RAP material (BS-EN:12697-4,
2005), is subjected to penetration test at 25
o
C, softening point and density test. Rhelogical
properties are also determined by dynamic shear rheometer (DSR) under strain-controlled
mode. The strain is 0.8% to assure visco-elastic properties of RAP binder. The test
temperatures range from 4 to 45
o
C when 8 mm plate is used with 2000µm thick specimen
and 20 to 80
o
C for 25mm plate and 1000µm thick specimen. The properties of RAP binder
are presented in Table 25. Viscosity of RAP binder is zero shear viscosity extrapolated
from DSR data using Cross model at 60
o
C.
73
Penetration at 25
o
C (dmm) BS EN 1426 (2000) 31
Softening Point (
o
C) BS EN 1427 (2000) 58
Density (g/cm
3
) BS EN ISO 3838 (1996) 1.03
Viscosity at 60
o
C (Pa.s) DSR-ZSV 1859
Table 25: Properties of RAP binder
74
4 Zero shear viscosity and the accuracy of viscosity
mixing equations
4.1 Introduction
To estimate the viscosity of the blend between aged and virgin binder is a vital part in the
recycled asphalt mixture design process. The estimation accuracy relies not only on the
efficiency of viscosity mixing rules but also on the viscosity of aged and virgin binders.
Inaccurate viscosity input might result in substantially erroneous prediction. The viscosity
of bituminous binder has been normally determined at an arbitrary temperature of 60
o
C.
Absolute viscosity can be conventionally determined by capillary method. However, the
limits of this approach are that it is time consuming and requires calibrations (Malkin and
Isayev, 2006).
Viscosity can be also determined by dynamic shear rheometer (DSR) at low frequency, for
instance, 0.05 or 0.1 rad/second. Chaffin et al. (1995) studied the efficiency of three mixing
rules, Arrhenius, Grunberg and Nissan, and Epps, on 47 bituminous materials including
straight run bitumen, bitumen fractions and commercial recycling agents. The dynamic
viscosity
*
 of all materials was measured at 60
o
C, angular frequency of 0.1 rad/second
under the geometry condition 25 mm plate and 500 µm gap. Chaffin et al (1995) claimed
that the
*
 at this specified frequency and temperature could be considered as a low
frequency limiting complex viscosity
*
o
 , independent of frequency or shear rate, and could
be used instead of absolute viscosity. However, the difference between
*
o
 and absolute
viscosity depends on the type of bitumen. Especially with stiff bitumen, the difference
might be considerable.
However, in practice, it is sometimes impossible to obtain the
*
o
 of stiff bitumen due to the
limit capability of the equipment (Chaffin et al., 1995). In addition, at high temperatures,
for instance at 60
o
C or higher, the viscosity values of soft bitumen are not stable under low
shear rates. Figures 37 and 38 illustrate the complex viscosities of bitumen 160/220 Pen and
100/150 Pen at 60
o
C versus different frequencies. The softer the bitumen, the more unstable
75
the complex viscosity values at low frequencies. If one value of low frequency is used to
determine
*
o
 , this value might not represent the viscosity of bitumen.
0.00E+00
2.00E+02
4.00E+02
6.00E+02
8.00E+02
1.00E+03
1.20E+03
1.0E-05 1.0E-04 1.0E-03 1.0E-02 1.0E-01 1.0E+00 1.0E+01
Frequency (Hz)
C
o
m
p
l
e
x

V
i
s
c
o
s
i
t
y

(
P
a
s
)
Figure 37: Complex viscosity of bitumen 160/220 Pen versus different frequency at 60
o
C
0.00E+00
5.00E+02
1.00E+03
1.50E+03
2.00E+03
2.50E+03
3.00E+03
3.50E+03
4.00E+03
4.50E+03
5.00E+03
1.0E-05 1.0E-04 1.0E-03 1.0E-02 1.0E-01 1.0E+00 1.0E+01
Frequency (Hz)
C
o
m
p
l
e
x

V
i
s
c
o
s
i
t
y

(
P
a
s
)
Figure 38: Complex viscosity bitumen 100/150 Pen versus different frequency at 60
o
C
The inaccurate viscosity determination might affect the efficiency of viscosity mixing
equations. To eliminate the above limitations of DSR in determining bitumen viscosity, it is
possible to increase the angular frequency. In addition, the temperature can also be reduced,
rather than using the arbitrary temperature 60
o
C, and using the Cross model to obtain the
76
low frequency limiting complex viscosity (ZSV) by mathematical extrapolation. The first
purpose of this experiment is to use zero shear viscosity to evaluate the efficiency of
different viscosity mixing equations.
In addition, Chaffin et al. (1995) claimed that the interaction parameter
12
G in Grunberg
and Nissan equation has relation with the difference between viscosities of rejuvenator and
aged binder. The bigger the viscosity difference, the higher the absolute value of
12
G . In
addition, Arrhenius is a special case of Grunberg and Nissan equation when
12
G is equal to
zero. This means the bigger the difference between viscosity of aged and virgin binder, the
more the viscosity predicted by Arrhenius mixing rule deviates from the Grunberg and
Nissan value. Based on experiment data, it is realized that the difference between
viscosities of rejuvenator and aged binder is dependent on the temperature. Therefore, the
purpose of this experiment is also to evaluate the effect of temperatures on efficiency of
Arrhenius and the other equations. If the hypothesis of Chaffin et al. (1995) is correct, the
adjustment in testing temperatures might improve the efficiency of Arrhenius equations.
4.2 Experiments
4.2.1 Materials
One aged bitumen and two rejuvenators are used in this study. The aged bitumen is
extracted and recovered from artificial RAP by fractionating column method (BS-
EN:12697-4, 2005). RAP was produced from 10 mm DBM (BS:4987-1, 2005) and binder
40/60 Pen with target air void content of 8%. The mixture is then conditioned at 85
o
C for
120 hours for LTOA (Airey, 2003). Two rejuvenators are soft bitumen 160/220 and
100/150 Pen. The properties of bitumens are in the Table 26.
Bitumen Penetration at 25
o
C (dmm) Specific Gravity
Aged binder
100/150 Pen
160/220 Pen
31
119
192
1.03 g/cm
3
1.02 g/cm
3
1.02 g/cm
3
Table 26: Properties of bitumens
There are two mixes in this study. Mix A represents blends of aged binder and 160/220 Pen
and Mix B represents those of aged binder and 100/150 Pen. For both Mix A and Mix B,
77
the increment of 20% aged binder produced a total of 8 pairs of aged binder/rejuvenator in
this project. Each blend is produced by pouring the determined proportions of liquid
bitumen into a glass tin by using digital balance with the accuracy of 0.001gram. The
weight of each blend is approximately 10 grams. All the blends are mixed manually by a
small paddle for 30 seconds at 160
o
C to assure the homogeneity.
4.2.2 Rheological testing
Rheological properties of each blend are determined using dynamic shear rheometer with
25 mm parallel plates and internal gap of 1000 µm. The complex viscosities of bitumen are
measured over temperature range from 20 to 80
o
C under strain-controlled mode. The strain
is chosen at 0.8% to assure bitumen working in the visco-elastic regime under frequency
range from 0.1 to 10 Hz.
4.3 Results and discussion
4.3.1 Zero shear viscosity (ZSV)
Zero shear viscosity or zero frequency complex viscosity is a theoretical concept. In
practice, it is impossible to obtain those values by current laboratory methods such as
rotational viscometry and rotational dynamic rheometry due to equipment limits. Hence, the
Cross model for pseudo-plastic materials is used for extrapolation of the zero shear rate
viscosity (Cross, 1965). This model describes the relationship between shear rate and
apparent viscosity by the following equation:
m
o
k
|
.
|

\
|
+
÷
+ =
-
·
·

 
 
1
(12)
Where:
 : viscosity
-
 : shear rate
o
 : zero shear viscosity
·
 : viscosity at infinite shear rate
k : material constant.
m: dimensionless material constant
78
As the rotational rheological device cannot cover sufficient range of shear rate, the
assumption was made that
·
>>  (Sybilski, 1996). Hence, the Cross model was
simplified as followed:
( )
m
o
k


+
=
1
(13)
Cross model (Equation 13) was also used for extrapolation of zero shear viscosity from the
dynamic oscillation data where
o
 was zero shear viscosity,
·
 was complex viscosity at
infinite frequency and  was oscillation angular frequency (Anderson et al., 2002).
Based on the rheological data, the theoretical zero shear viscosity is extrapolated by using
Cross model with the help of Matlab software. The results of extrapolation work are
presented in Table A-1 for blends of Mix A and Table A-2 for blends of Mix B with
different proportion of aged binder at different test temperatures (Appendix A). Figure 39 is
an example of the extrapolation work. The results show that the Cross model is fitted with
the experimental data as almost all the R square values of the regression analysis are
substantially high, except for high temperatures, for instance, higher than 60
o
C. At these
high temperatures, rejuvenators and those blends with low percentages of aged binder are
soft and the complex viscosity values are almost independent of angular frequencies. Figure
40 illustrates the dynamic viscosities versus frequency of a blend of Mix A (20% RAP and
80% 160/220 Pen) at 70
o
C. For these situations, a linear polynomial relation between
complex viscosity and frequency (Equation 14) is used for zero frequency viscosity
extrapolation instead of Cross model which requires the viscosity of material to be
dependent on frequency, or shear rate (Cross, 1965).
b a + =  
0
(14)
Where:
b a, : material constant
 : frequency
79
Figure 39: Extrapolated ZSV of Mix A blends at temperature of 25
o
C
80
1.0
10.0
100.0
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0
Frequency (Hz)
C
o
m
p
l
e
x

V
i
s
c
o
s
i
t
y

(
P
a
s
)
Figure 40: Complex viscosity versus frequencies of blend (20 % RAP and 80% 160/220 Pen) at 70
o
C
4.3.2 Efficiency of viscosity mixing equations
Four viscosity mixing equations are evaluated in this study, Grunberg and Nissan,
Arrhenius (ASTM 4887), Epps, and DLV (Section 2.3.4). For the Grunberg and Nissan
equation, parameter
12
G is determined by fitting this equation into the experiment data at
different testing temperatures using Matlab. For both Mixes A and B, the viscosity of each
blend is calculated by each of four mixing equations with the same proportion of RAP
binder, ZSVs of RAP binder and rejuvenator. The experiment data and predicted viscosity
using four mixing equations for Mix A and Mix B are illustrated in Tables A-3 and A-4
(Appendix A). Each pair of predicted and experimental data is compared by using
regression analysis with the help of Matlab software. Figure 41 illustrates the differences
between experimental and predicted values using four different viscosity mixing equations.
The difference between experimental and predicted data is evaluated by following
indicators, R square value, Residuals (R), and Root Mean Square Error (RMSE) (Tables 27
and 28).
The Residual represents the difference from the predicted to the experimental values and is
calculated by the following equation:
i i i
e p R ÷ = (15)
81
Where:
i
R : Residual of data point i
i
p : predicted value
i
e : experimental value
R is also presented by the percentage difference from the predicted and experimental value,
the closer the R value to zero, the better the fit.
0 10 20 30 40 50 60 70 80 90 100
10
6
10
7
Proportion of virgin binder (%)
Z
e
r
o

S
h
e
a
r

V
i
s
c
o
s
i
t
y

(
P
)
Experiment
ASTM 4887
G&N
Epps
DLV
Figure 41: Experiment and predicted viscosity using different viscosity mixing equations of Mix A
(Blends of different proportion of aged binder and 160/220 pen) at 20
o
C
The Root Mean Square Error (RMSE) is the standard deviation between predicted and
experimental data sets. The closer the RMSE to 0, the better fit. RMSE is determined by the
following equation:
( )
n
e p
RMSE
n
i
i i ¿
=
÷
=
1
2
(16)
Where:
82
n : the number of data points evaluated
i
p : predicted value
i
e : experiment or actual value
The results from regression analysis for blends of Mix A and Mix B are presented in Tables
27 and 28. The R square values of regression analysis between four equations and
experiment are almost the same. However in general, the RMSE generated by Grunberg
and Nissan equation are smallest at almost every test temperatures. The RMSE values
generated by DLV and Arrhenius equations (ASTM D4887) are in turn the first and second
largest. At high temperatures, for instance, higher than 55
o
C, these values from Epps and
Grunberg and Nissan equation are almost identical.
In term of Residual values, the viscosity predicted using DLV method are the most deviant
from actual values. Arrhenius equation, the most popular equation for estimating the
viscosity of bitumen mixture (Chaffin et al., 1995), generates the second highest residuals.
The results demonstrate that on the average, the predicted viscosities by Arrhenius equation
(ASTM D4887) are within 30% of the experimental data. This supports the finding of Irvin
(1977). DLV method could estimate the viscosity of the bitumen mixture within 50% of
the actual value. This also substantiates the fact that DLV method does not perform well
when soft bitumen is used as rejuvenator (Chaffin et al., 1995). Generally, Grunberg and
Nissan equation again generates the lowest values of residual. Maximum Residual values
generated by using Grunberg and Nissan and Epps equations are generally less than 20%.
At temperatures higher than 45
o
C, the residual generated by Grunberg and Nissan and Epps
equation are almost similar, approximately 10%. All viscosity mixing equations show the
tendency of improving accuracy when increasing the temperature (Tables 27 and 28).
Arrhenius equation (ASTM 4887) is a special case of Grunberg and Nissan when
12
G is
equal to zero. Table 29 presents the values of
12
G of Mix A and Mix B blends at different
temperatures. If the
12
G is negative, Arrhenius equation overestimates the viscosity of the
blend and vice versa. The higher the absolute value of
12
G , the greater the difference the
viscosity predicted by Arrhenius equation from the G&N value. Chaffin et al (1995)
83
claimed the variation is due to the difference between viscosity of aged binder and
rejuvenator. However, the results in this experiment demonstrate that the difference
between viscosity of aged binder and rejuvenator is not the only reason for the imprecise
viscosity estimation.
Figure 42 presents the difference between viscosity of aged binder and rejuvenator versus
12
G at different temperatures. The viscosity difference is expressed as the ratio of aged
binder viscosity over that of rejuvenator. The results indicate there is no relation between
viscosity difference and the
12
G parameter. Therefore, the deviation of viscosity predicted
by Arrhenius equation from actual values is not attributed to the viscosity difference. The
inaccurate estimation using Arrhenius viscosity mixing equations is probably caused by
interaction between aged and virgin binder.
As the bitumen has complicated chemical composition, the interaction occuring inside each
blend of different bitumen binders is different from the others. The interaction is also
different in mixtures comprised of the same two bitumen but different proportions (White et
al., 1970). Therefore, the fact that one constant value of interaction parameter
12
G is used
universally would result in substantial errors in viscosity estimation.
84
Temperature Grunbrerg and Nissan ASTM 4887 Epps DLV
o
C
R-
square RMSE
Max Res
(%)
R-
square RMSE
Max Res
(%)
R-
square RMSE
Max Res
(%)
R-
square RMSE
Max Res
(%)
20 0.996 6.5E+05 17.4 0.991 9.4E+05 14.82 0.983 1.4E+06 20.8 0.9126 3.1E+06 48.3
25 1.000 8.5E+04 9.8 0.999 1.3E+05 10.77 0.993 3.4E+05 20.8 0.9365 1.0E+06 53.1
30 0.990 1.7E+05 36.7 0.979 2.3E+05 21.07 0.963 3.0E+05 27.8 0.8936 5.6E+05 48.9
35 1.000 1.5E+04 7.2 0.999 1.9E+04 12.76 0.997 3.8E+04 10.1 0.9621 1.5E+05 47.8
40 0.999 1.1E+04 10.4 0.999 1.0E+04 13.82 0.991 2.5E+04 16.9 0.9484 6.7E+04 46.5
45 1.000 2.3E+03 8.5 0.997 5.7E+03 22.20 0.999 4.2E+03 7.5 0.971 2.0E+04 39.9
50 0.998 1.9E+03 15.2 0.995 2.9E+03 30.39 0.997 2.2E+03 8.8 0.9702 8.0E+03 33.5
55 1.000 3.0E+02 6.1 0.993 1.3E+03 26.07 1.000 3.0E+02 5.8 0.9768 2.5E+03 33.5
60 1.000 7.9E+01 3.4 0.991 5.8E+02 23.93 1.000 1.1E+02 4.4 0.9782 1.0E+03 31.2
65 0.999 9.4E+01 8.8 0.992 2.5E+02 26.79 0.999 1.0E+02 7.3 0.9736 5.1E+02 27.9
70 0.998 7.0E+01 11.8 0.990 1.3E+02 29.49 0.997 6.7E+01 10.9 0.9719 2.4E+02 25.3
75 1.000 1.5E+01 5.5 0.981 9.5E+01 31.66 0.999 2.7E+01 12.9 0.9847 9.4E+01 20.6
80 0.999 1.4E+01 11.2 0.991 3.2E+01 26.19 0.998 1.4E+01 9.6 0.964 6.2E+01 24.5
Table 27: Mix A - Regression analysis between experiment and predicted values using different viscosity mixing equations at different temperatures
Temperature Grunbrerg and Nissan ASTM 4887 Epps DLV
o
C
R-
square RMSE
Max Res
(%)
R-
square RMSE
Max Res
(%)
R-
square RMSE
Max Res
(%)
R-
square RMSE
Max Res
(%)
20 0.977 1.6E+06 21.7 0.971 1.6E+06 17.7 0.967 1.7E+06 20.2 0.886 3.48E+06 37.5
25 0.995 2.8E+05 16.3 0.985 4.1E+05 22.7 0.980 4.9E+05 25.2 0.856 1.30E+06 40.9
30 0.987 1.9E+05 14.7 0.951 3.3E+05 22.8 0.943 3.6E+05 25.8 0.819 6.96E+05 44.1
35 0.999 2.1E+04 10.3 0.998 2.9E+04 11.7 0.994 5.4E+04 16.2 0.933 1.93E+05 39.9
40 0.998 1.2E+04 8.4 0.997 1.5E+04 8.0 0.989 2.5E+04 13.5 0.923 7.91E+04 42.2
45 0.996 7.5E+03 16.4 0.996 7.1E+03 19.7 0.993 8.8E+03 10.8 0.948 2.68E+04 36.8
50 0.999 1.3E+03 8.7 0.993 3.4E+03 23.9 0.999 1.4E+03 10.8 0.971 7.53E+03 35.3
55 0.996 9.8E+02 11.1 0.996 9.0E+02 12.0 0.993 1.3E+03 11.6 0.94 3.93E+03 32.4
60 1.000 1.0E+02 6.2 0.991 5.5E+02 18.3 0.999 1.5E+02 6.5 0.973 1.09E+03 28.8
65 0.997 1.8E+02 11.5 0.994 2.2E+02 20.2 0.996 1.7E+02 8.5 0.957 6.24E+02 30.2
70 0.997 7.0E+01 9.0 0.996 8.4E+01 15.1 0.996 7.8E+01 12.5 0.951 3.03E+02 36.1
75 0.999 2.0E+01 4.5 0.990 6.8E+01 18.4 0.999 2.4E+01 7.1 0.972 1.23E+02 23.6
80 0.997 1.9E+01 7.1 0.996 4.9E+01 10.6 0.994 2.4E+01 8.0 0.943 8.19E+01 27.1
Table 28: Mix B - Regression analysis between experiment and predicted values using different viscosity mixing equations at different temperatures
85
Temperature Mix 1 Mix 2
20 0.2582 0.1719
25 0.1206 0.2160
30 0.3776 0.4509
35 -0.0916 0.0434
40 -0.0552 0.1172
45 -0.2128 -0.0507
50 -0.2242 -0.2477
55 -0.3121 -0.0141
60 -0.3276 -0.2512
65 -0.2724 -0.1357
70 -0.2624 -0.0995
75 -0.3950 -0.2253
80 -0.2260 -0.0574
Table 29: G12 parameter of Mix A and B at different temperatures
0
10
20
30
40
50
60
70
20 25 30 35 40 45 50 55 60 65 70 75 80
Temperature (Degree Celcius)
V
i
s
c
o
s
i
t
y

D
i
f
f
e
r
e
n
c
e
0.0
0.1
0.1
0.2
0.2
0.3
0.3
0.4
0.4
0.5
0.5
G
1
2
Mix A-V Mix B-V Mix A-G Mix B-G
Figure 42: Viscosity difference and G12 versus temperatures
86
5 Effects of laboratory mixing methods on the
homogeneity of hot recycled asphalt mixture
5.1 Introduction
This chapter contains the development of a new mixing protocol that simulates the mixing
mechanism between RAP and virgin materials in asphalt mixing plants. The effects of
different RAP/superheated virgin aggregate mixing durations and RAP sizes on
homogeneity of hot recycled asphalt mixture are considered. The homogeneity of recycled
mixtures is studied by using virgin binder with different colour from that of RAP binder.
The red colour of virgin binder is obtained by mixing clear binder with iron oxide pigment.
The proportion of pigment is 10% by weight of the binder making this binder red. The
content of this chapter also includes the correlation between homogeneity and stiffness
distribution of hot recycled asphalt mixtures.
5.2 Development of laboratory mixing protocol
It is essential that the laboratory mixing protocol should duplicate the mixing mechanism
occurring in the actual asphalt mixing plants. The laboratory mixing process should include
the following steps:
- Step 1: Virgin aggregate is superheated to predetermined temperature.
- Step 2: Superheated virgin aggregate is mixed with RAP material at ambient temperature
in the mixer maintained at mixing temperature.
- Step 3: the combination of superheated virgin aggregate and RAP material is blended
with virgin binder in the mixer maintained at mixing temperature.
Step 1 is to enable virgin aggregate enough thermal energy to heat up the RAP from
ambient to mixing temperature. The purpose of Step 2 is to use the heat from superheated
virgin aggregate to heat up and soften RAP materials, separating RAP lumps into single
particles covered by RAP binder. The heat source for separating RAP lumps also comes
from the mixer maintained at mixing temperature. During this process, RAP bitumen is also
transferred onto the surfaces of virgin aggregate particles. Then, the combination of RAP
and virgin aggregate is mixed with virgin binder in Step 3. The aim of this step is to assure
that virgin binder can incorporate and rejuvenate the RAP binder. In addition, this step is
also to ensure that the rejuvenated binder is well distributed all over the mixture and coats
87
every single aggregate particle. The mixing time for Step 3 is 2 minutes due to laboratory
experience with the mixer. For the manufacture of conventional asphalt mixture, this
duration is enough for virgin binder to coat all the aggregate particles.
The efficiency of the laboratory mixer is quite different from that of the industrial mixer for
asphalt production. In the actual asphalt mixing plant, besides mechanical mixing effects,
single virgin aggregate and RAP particles experience centrifugal force until this force is
smaller than gravity and materials start to fall downward. Once the heat transferred from
superheated virgin aggregate softens the RAP binder, the movement of each RAP particle
due to gravity also enhances the separating progress of RAP lumps.
In the laboratory mixer, on the contrary, the bulk of materials is moving circularly under
mechanical effect of mixing paddles. In fact, there is primarily horizontal and
inconsiderable vertical movement. Once the heat transferred from superheated virgin
aggregate weakens the bitumen bond, RAP particles are separated due to the mechanical
effect of mixing paddles and external friction among surface of particles. The breaking
progress is not enhanced by the vertical movement due to gravity of RAP materials.
In actual asphalt mixing plants, the continuous production process also assures the heat
conductivity between superheated virgin aggregate and RAP material being more efficient
than that in the laboratory. The laboratory production of recycled asphalt mixture requires
the lid of the mixer to be opened several times during manufacture process for material
intakes.
5.2.1 Estimation of superheated temperature of virgin aggregate
The amount of heat required to raise the RAP at ambient temperature to the mixing
temperature is equal to the amount of heat dispersed from virgin aggregate so the
temperature drops from superheated to mixing value. Based on the quantities of virgin
aggregate, RAP, specific heat of RAP and virgin aggregate, the superheated temperature
can be estimated.
The amount of heat (Cutnell and Johnson, 2004) required to raise the temperature of a mass
is:
( )
1 2
T T mc Q ÷ = (17)
88
Where:
m: the mass in quantity (kg)
c: specific heat (kJ/kg
o
C)
2 1
,T T : current and desired temperatures (
o
C)
The amount of heat required to raise temperature of RAP from ambient to mixing
temperature is:
( )
a m RAP RAP
T T c m Q ÷ =
1
(18)
The amount of heat dispersed from superheated virgin aggregate so the temperature reduces
from superheated to mixing temperature is:
( )
m s agg agg
T T c m Q ÷ =
2
(19)
Where
agg RAP
m m , : Quantities of RAP and virgin aggregate in the mixture (kg)
agg RAP
c c , : specific heat of RAP and virgin aggregate (kJ/kg
o
C)
a m s
T T T , , : superheated temperature (
o
C) for virgin aggregate, mixing temperature, ambient
temperature of RAP. The mixing temperature will be determined in the next section.
5.2.2 Determine the superheated virgin aggregate/RAP mixing duration
Due to the efficiency difference, the short mixing time in the real industrial mixer, for
instance, 60 second, might not be applicable in the laboratory mixer. To investigate the
effect of mixing on the properties of hot recycled mixture, the superheated virgin
aggregate/RAP mixing duration is adjusted based on the mixer used for the research.
During the mixing duration, thermal energy transferred from superheated virgin aggregate
and from the mixer will soften the bitumen bond and separate RAP lumps into single
particles. The adjustment method is to start from short mixing time and gradually increase
the mixing time until the change in RAP lumps size is insignificant. Consequently, the
RAP/virgin aggregate mixing duration is decided to be 2, 4, 6 and 8 minutes for large RAP
mixture and 1, 2, 4, 6, 8 minutes for small RAP mixture. Figure 43 illustrates that at 2
minutes mixing time, the RAP still remains approximately the original size. Figure 44 is for
8 minutes mixing time when the change in RAP size under mechanical mixing is negligible.
89
Figure 43: RAP size after 2 minutes RAP/superheated virgin aggregate mixing time
Figure 44: RAP size after 8 minutes RAP/superheated virgin aggregate mixing time
90
5.2.3 Determine the mixing temperature
The mixing temperature in Step 3 is determined based on the virgin binder used as
rejuvenator. The optimum mixing temperature is the temperature at which the viscosity of
virgin binder has the viscosity value of 0.2 Pa.s (Read and Whiteoak, 2003). The viscosity
of virgin binder is determined by Brookfield viscosity test at 120, 150, and 180
o
C. The
temperature at 0.2 Pa.s viscosity is interpolated by linear relationship between double log
viscosity and temperature (Heukelom, 1973). In this experiment, mixing temperature is
135
o
C
5.3 Method for segregation evaluation
The primary characteristic attributed to the difference between conventional asphalt and
recycled asphalt mixture is reclaimed material (RAP), a combination of aged binder and
aggregate. Hence, the segregation that occurs in recycled asphalt mixture includes not only
the different concentration of aggregate sizes and bitumen but also the distribution of RAP
material, new material and additive if applicable. Therefore, most measuring and
segregation detection techniques for conventional asphalt mixture are not applicable or not
efficient for recycled asphalt mixture. Unfortunately, the segregation level might
considerably affect the quality of recycled asphalt mixture and is determined mainly on
how well new and RAP materials are mixed (Tia et al., 1980).
The method using infrared scanner can only classify between the bitumen-rich and
aggregate-rich areas due to the difference in thermal properties (Gardiner and Brown,
2000). Similarly, the segregation detection method based on density, for instance, X-ray
scanner, can only identify the location of air void, aggregate and bitumen due to the
significant difference in density of these components. However, these kinds of equipment
cannot identify the position of RAP and new binder as the density of these binders are
approximately the same. Figure 45 illustrates the images taken by X-Ray photo machine
and normal digital camera of RAP binder and Shell clear binder (Mexphalte C 160/220 pen)
dyed by red iron oxide pigment. The proportion of the pigment is 10% by weight of the
clear binder. Although the density of iron oxide is quite different from those of clear and
RAP binder, 5.25 g/cm
3
to 1.03 g/cm
3
, the X-ray scanner image cannot show any difference
between RAP binder and clear binder dyed by iron oxide pigment. Meanwhile, RAP binder
91
is visually recognized as pure black and Mexphalte C dyed by 10% iron oxide has colour of
red (Figure 43).
Figure 45: Images taken by X-Ray scanner and normal digital camera of Shell Mexphalte C dyed by 10
% iron oxide and RAP binder
The dye chemistry method (Lee et al., 1983) can identify the location of recycling agent
due to the detection of chemical tracer. However, in undetected areas this method cannot
classify virgin aggregate and RAP lumps. The result of this method is only correct in cases
where the detected areas contain RAP binder and recycled agent, and the undetected areas
contain only aggregate. Unfortunately, the undetected areas contain not only aggregate but
also RAP material as RAP lumps.
To assess the effect of mixing process on homogeneity of recycled mixture, the RAP should
be clearly identified from virgin materials. In this research, RAP material is artificially
made in the laboratory. The colour of RAP binder will be purely black. On the contrary,
virgin binder, the blend of synthetic clear binder (Shell Mexphalt C 160/200 Pen) and 10%
of iron oxide by weight of the binder, will have the colour of red. The preparation of testing
specimens conforms to the procedures in Section 5.4 .
The hypothesis of using virgin binder with different colour is if RAP material is not well
mixed with virgin binder, the areas of black (RAP materials) and red (virgin material) are
clearly visualized. In addition, the phenomenon that RAP materials are not fully separated
92
into single aggregate particles covered by RAP binder but exist as RAP lumps in the
recycled mixture is also easily identified. After manufacture, surfaces of slices cut from
specimens are recorded by digital camera. The analysis of these surfaces based on vertical
order help to understand the distribution of RAP and virgin materials in the mixture in a 3D
manner.
Samples for visual segregation assessment are also subject to mechanical properties
evaluation. The purpose is to link the homogeneity level and the mechanical properties of
recycled asphalt mixture. Stiffness of each sample is measured by Indirect Tensile Stiffness
Modulus (ITSM) test at 20
o
C (BS-EN:12697-26, 2004). The stiffness data is then
statistically analyzed in conjunction with RAP (or virgin material) distribution pattern to
characterize the correlation between mixture homogeneity and mechanical properties.
5.4 Specimens preparation
5.4.1 Materials
The proportion of RAP is 40% in the recycled mixture. Aggregate gradation of recycled
mixture is designed the same as that of RAP (BS:4987-1, 2005). The bitumen content of
RAP and recycled mixture are the same, 5.2% by weight of total mixture. Virgin binder is
prepared by preheating synthesis clear binder (Shell Mexphalt C 160/220 Pen) to 135
o
C and
blending with iron oxide powder. The amount of pigment is 10% by weight of virgin binder
to make the color of virgin binder red.
5.4.2 Mixing procedure
The mixing process is implemented in the Mixer A (Section 6.2.2). The mixing procedure
is determined as follows:
 Step 1: Virgin aggregate is superheated to 215
o
C for 8 hours. Virgin binder is
preheated at 135
o
C for 2 hours
 Step 2: RAP, both large and small sizes, denoted as LR and SR, are blended with
superheated virgin aggregate at ambient temperature in the mixer maintained at
135
o
C for different durations. The mixing durations for large RAP (LR) are 2, 4, 6,
8 and 1, 2, 4, 6, 8 minutes for small RAP (SR).
93
 Step 3: the blend of RAP and virgin aggregate is mixed with virgin binder for 2
minutes. Before being poured into the mixer, virgin binder is well stirred manually
to prevent the settlement of iron oxide at the bottom of the container.
5.4.3 Compacting procedure
The loose mixture is compacted by gyratory compactor at 130
o
C until the targeted density
is obtained (BS-EN:12697-31, 2004). The targeted air void content of each specimen is 4%.
The amount of materials is estimated so the final cylindrical specimen has the diameter of
150 mm and 100 mm height.
5.4.4 Machining specimens for segregation assessment
After being de-moulded the next day, a cylinder of 100 mm diameter and 100 mm height is
cored from each compacted specimen. Two samples with approximately 40 mm thick,
samples 1 and 2, are cut from this core for stiffness evaluation (BS-EN:12697-26, 2004).
The surfaces of these samples are photographed by digital camera for visual segregation
assessment. After stiffness measurement, these samples are sliced up into slices
approximately 10 to 15mm thick. Surface image of each slice is also recorded by digital
camera for further visual analysis. This is aimed to investigate the relationship between the
heterogeneity and stiffness distribution of recycled mixtures with different mixing
durations.
5.5 Results and analysis
5.5.1 Visual assessment for segregation
Figures 46 to 49 illustrate the surface characteristic of specimens manufactured with large
RAP and different mixing times. Both surfaces of specimens for stiffness evaluation and
surface of slices cut from these specimens after stiffness evaluation are included. Figures
50 to 54 show the surfaces of slices from specimens prepared from small RAP with
different mixing times. The slices are in order from the bottom to top. The surfaces analysis
of slices in vertical order will give a clear picture of how RAP lumps are distributed in the
cylindrical specimen.
Due to the colour difference between that of virgin (red), RAP binder (black), and
aggregate, the locations of these components are easy to visualize. The advantage of using
virgin binder with different colour from that of RAP binder is to help identify the location
94
of RAP lumps which is impossible with the other methods. The area containing just black
binder and aggregate will be definitely RAP materials. This helps to better understand
whether virgin binder can incorporate and rejuvenate the properties of RAP binder or RAP
might work as inert lumps in the recycled mixture.
RAP materials exist as lumps, which is the agglomerate of single aggregate particles,
including filler stuck together by bitumen. Once RAP material is mixed with superheated
virgin aggregate, the thermal energy transferred from hot virgin aggregate will increase the
temperature of RAP materials. Due to the inherent properties of bitumen, the bonding
strength among RAP aggregate particles will be weakened. Under the mechanical effect of
the mixing paddles, RAP lumps are gradually separated into single pieces. In fact, the
factors that manipulate the efficiency of RAP separating process are, for instance, mixing
temperature and mechanical effect. In addition, due to the fact that the energy transfer
process occurs gradually, depending on size of material and contact surface (Cutnell and
Johnson, 2004), the efficiency of mixing process is also attributed to mixing duration.
The surface images show that RAP/superheated virgin aggregate mixing duration seriously
affects the homogeneity level of recycled mixtures, especially recycled mixture with large
RAP. For recycled mixture with large RAP and short mixing time, for instance 2 minutes,
the size of RAP areas on the surface of slices are large (Figure 46). This is because the
mixing duration (energy) is not enough to degrade RAP lumps into separate particles. RAP
lumps are almost inert and scattered in specimen at approximately their original size. In
addition, RAP lumps are not well distributed in the specimen. In some slices, for instance,
(Figures 46 b and c), the occurrence of RAP is intensive. This means the density of RAP
are extremely high. However, in the other slices, the presence of RAP is negligible (Figures
46 a and h). Actually, these areas contain primarily virgin materials.
The homogeneity of recycled mixture is clearly improved once the RAP/superheated virgin
aggregate mixing duration is extended. With large RAP recycled mixtures, Figures 46 to 49
show that when the mixing duration increases, the size of RAP areas in the surface of each
slice gradually becomes smaller. This phenomenon is as expected as the longer the mixing
time, the more thermal energy is transferred from superheated virgin aggregate. The effect
of mechanical mixing process plus the weakening of bitumen bond between RAP aggregate
particles slowly disintegrates the RAP lumps into smaller sizes. Therefore, the mixing
95
efficiency is enhanced as more active RAP particles are involved in the mixing process and
well distributed around the mixture.
During the mixing process, the separation of RAP lumps also offers the chance for RAP
binder to be rejuvenated by virgin binder. Visually, it is assumed that if the RAP binder is
rejuvenated by virgin binder or these two binders are integrated, the colour of virgin binder
will somehow be changed. The red colour of virgin binder will be darker due to the black
effect of RAP binder.
The surface images show the same tendency for both large and small RAP recycled
mixtures, the longer the RAP/superheated virgin aggregate mixing duration, the darker the
red color of the virgin binder. This indicates more integration between RAP and virgin
binder. In addition, the areas of pure black and red binder are also reduced. This is as
expected as the longer the mixing time, the more single pieces of RAP are separated from
RAP lumps. Hence, the total exposure surface of RAP particles for rejuvenating process
increases and more RAP binder is rejuvenated by virgin binder.
The rejuvenating process is enhanced not only by the total exposed surface of RAP particles
but also the temperature of RAP binder itself. This is because the diffusion of virgin into
RAP binder is significantly influenced by the temperature, the higher the temperatures, the
more efficient the diffusion process (Karlsson and Isacsson, 2003b). During the
RAP/superheated virgin aggregate mixing duration, the thermal energy transferred from
superheated virgin aggregate and the mixer increases the temperature of RAP binder and
progresses the rejuvenation between RAP and virgin binder.
The surface images also demonstrate that RAP/superheated virgin aggregate mixing
duration significantly affects the integration between RAP and virgin binder. RAP binder
cannot be diffused efficiently by virgin binder unless RAP binder is heated up enough or
activated using thermal energy from virgin aggregate and mixer during the mixing duration.
Figure 46 shows the surface images of recycled mixture made of large RAP, 2 minutes
mixing duration and Figure 50 for mixture composed of small RAP, 60 seconds mixing
duration. The colour of virgin binder almost remains unchanged as original red. In fact, the
mixing time or the thermal energy from superheated virgin aggregate is not enough to
96
activate the RAP binder for diffusion. RAP binder is inert and there is not considerable
integration between RAP and virgin binder.
The surface images show that the size of RAP seriously affects the homogeneity level of
recycled mixture. With the same RAP/superheated virgin aggregate mixing duration, for
instance 2 minutes, the areas of inert RAP lumps of samples composed of large RAP are
considerably larger and more intensive than those of small RAP (Figures 46 and 51). It is
reasonable as the larger the RAP size, the more thermal energy and mixing effort required
for separating RAP lumps. It is expect that when the mixing time is extended, all the RAP
lumps can be disintegrated as single pieces covered by RAP binder and well distributed
over recycled mixture. However, even when small RAP is used with 8 minutes
RAP/superheated virgin aggregate mixing duration, there is still a considerable portion of
RAP works as lumps (Figures 55 and 56). This validates the fact that “complete blending”
between RAP and virgin binder assumed in the design process never occurs during mixing.
Hence, the mechanical properties of recycled mixture might be not as consistent as
predicted. This circumstance might be worse under really short mixing time, about
maximum 90 seconds total, in the industrial mixer.
97
a b
c
d
e f
g h
Figure 46: LR mixture – 2 minutes mixing time
RAP lump
98
a b
c d
e f
g h
Figure 47: LR mixture – 4 minutes mixing time
99
a b
c d
e
f
g h
Figure 48: LR mixture – 6 minutes mixing time
100
a b
c d
e f
g h
Figure 49: LR mixture – 8 minutes mixing time
101
a b
c d
e f
g h
Figure 50: SR mixture – 1 minute mixing time
102
a b
c d
e f
g h
Figure 51: SR mixture – 2 minutes mixing time
103
a b
c d
e f
g h
Figure 52: SR mixture – 4 minutes mixing time
104
a b
c d
e f
g h
Figure 53: SR mixture – 6 minutes mixing time
105
a b
c d
e f
g h
Figure 54: SR mixture – 8 minutes mixing time
106
Figure 55: Large version of Figure 54 g
107
Figure 56: Large version of Figure 54 h
108
5.5.2 Mechanical assessment
Together with visual assessment for segregation, specimens composed of large RAP with
different mixing durations were also subjected to mechanical properties evaluation. The aim
is to link together the homogeneity level and mechanical performance of recycled asphalt
mixture. The stiffness of each specimen is measured by indirect tensile stiffness modulus
test (ITSM) (BS-EN:12697-26, 2004). The test is carried out at temperatures of 20
o
C, 124
milisecond risetime and 5 µm horizontal diametral displacement to ensure that the
specimen responds as an elastic material.
Conventionally, stiffness of each specimen is the mean value of stiffness measured at two
perpendicular directions unless the difference between these two values is greater than 10%
of the mean. However, the surfaces of specimens for stiffness evaluation demonstrate there
is considerable segregation in recycled mixture. In addition, it is also realized that stiffness
variation with different measured directions of most specimens are not satisfied by the
standard. This issue might be attributed to the heterogeneity of recycled mixture. Therefore,
the stiffness values of each specimen are measured by four directions along the
circumference (Figure 57) with the hypothesis that the specimens with higher level of
heterogeneity might have greater variations among stiffness at different directions.
Figure 57: Stiffness measurement scheme
1
2
3
4
Specimen
109
Mixing Air Stiffness in Different Directions Mean COV
Time Samples void (Mpa) Stiffness (%)
(mins) Content (MPa)
% 1 2 3 4
2 1 4.6 1813 2232 1778 2078 1975 11
2 8 1306 1197 1374 1844 1430 20
4 1 5.5 1452 2816 2004 2183 2114 27
2 6.5 2324 1959 1774 1491 1887 18
6 1 6.9 1690 1591 1408 1472 1540 8
2 5.8 2147 2106 2002 1850 2026 7
8 1 6 1977 2098 1999 1873 1987 5
2 5.5 1924 1950 1881 2175 1983 7
Table 30: Stiffness versus different RAP/superheated virgin aggregate mixing duration
Table 30 presents the results of stiffness values versus different mixing times. The data
demonstrates that mixing time significantly influences the stiffness variation. For short
mixing time, the differences among stiffness values obtained from the same specimen are
considerable. For sample 2 (2 minutes mixing time), the maximum stiffness value is 1844
MPa while the minimum is 1197 MPa. The coefficient of stiffness variation is 20%.
Similarly, for sample 1 (4 minutes mixing duration), maximum and minimum stiffness
values are in turn 2816 MPa, 1452 MPa, and the coefficient of variation is 27%. In
addition, the variations in stiffness values do occur not only in the same specimen but also
in different specimens with similar mixing time. For example, for 2 minutes mixing time,
mean stiffness of sample 1 is 1975 MPa compared to 1430 MPa of sample 2. Likewise, the
mean stiffness of sample 1 (6 minutes mixing time), is 1540 MPa, significantly different
from that of sample 2, 2026 MPa.
The reason for the stiffness variation might be due to the difference of air void content of
specimens. Stiffness decreases once the air void content increases (Tayebali et al., 1994).
Figure 58 also shows the tendency that stiffness generally decreases once the air void
content increases. However, the stiffness value difference is not only attributed to the
variation of air void content. For instance, sample 2 (2 minutes mixing time) has stiffness
value of 1844 MPa (measuring direction 4) which is considerably higher than regressed
110
stiffness of 8% air void content. Similarly, sample 1 (4 minutes mixing time) has stiffness
of 2816 MPa (measuring direction 2) which is considerably higher than expected with 5.5%
air void content. On the contrary, although the stiffness of sample 1 (2 minutes mixing
time) is expected to be high with 4.6% air void content, it is quite low with 1813 MPa
(measuring direction 1) and 1778 MPa (measuring direction 3).
y = -203.66x + 3110.1
R
2
= 0.7316
500
1000
1500
2000
2500
3000
4 5 6 7 8 9
Air Void Content (%)
S
t
i
f
f
n
e
s
s

M
o
d
u
l
u
s

(
M
P
a
)
2 minutes 4 minutes 6 minutes 8 minutes
Figure 58: Stiffness modulus versus air void content
The variation in stiffness values is attributable to the heterogeneity of recycled mixture. For
short mixing duration, RAP lumps are not fully disintegrated into single particles coating
by RAP binder. Therefore, the virgin binder that should be in contact and rejuvenate the
RAP binder could not be in the right place and fulfill its function. This results in some areas
with high concentration of RAP while the others contain primarily virgin materials.
This situation is exaggerated by the manufacturing process. There is a lack of vertical
movement of material in the mixing bowl due to the properties of the mixer. During mixing
process, RAP and virgin materials are not mixed together vertically. RAP lumps tend to
move up to the surface due to bigger sizes (Figure 59). This results in the situation that
when the loose mixture is transferred to the mould for compaction, RAP lumps tend to
move first and settle down at the bottom of the mould. Figures 46 and 47 show that
although came from the same compacted specimen, specimen 1 (at the bottom) tends to
Mixing duration
111
have considerably higher proportion of RAP than specimen 2 (on top). In addition, the
stiffness of specimen 1 is also considerably higher than that of specimen 2 (Table 30).
Figure 59: Small particles tend to move downward to the bottom during mixing process
It has been maintained that the uneven distribution of RAP materials results in the variation
of stiffness values. An effort has been made to investigate and quantify this relationship by
correlating:
 the stiffness: determining the stiffness of specimen at four measuring directions and
 RAP (or virgin materials) distribution pattern: by visual surface assessment of the
stiffness-measured specimen and slices machined from this specimen.
However, there is no clear correlation between the distribution of RAP (or virgin materials)
and stiffness values. This is because the material distribution patterns are not consistent
along the vertical direction of the specimen. Figures 46 a and d illustrate the surfaces at
both ends of specimen 1 (made from large RAP and 2 minutes mixing time) and b and c are
for the surfaces of the slices machined from this specimen. Unfortunately, the
characteristics of these surfaces are all different. For different mixing durations and RAP
112
materials, large and small, the phenomenon is the same. If the relation between stiffness
values and materials distribution patterns are evaluated based only on the surfaces at both
ends of specimen (Lee et al., 1983), the result will be certainly different.
The stiffness data demonstrates the tendency that the homogeneity of recycled mixture is
substantially enhanced once the RAP/superheated virgin aggregate duration is extended.
Not only the coefficient of stiffness variation in each specimen decreases (Table 30) but
also does the general coefficient of variation. Table 31 shows that the general coefficient of
variation significantly decreases from 22% to 5% once the mixing duration increases from
2 to 8 minutes.
Mixing time Mean Stiffness Standard Coefficient of
(minutes) (MPa) Deviation Variation (%)
2 1702 374 22
4 2000 449 22
6 1783 286 16
8 1984 105 5
Table 31: Mean stiffness versus RAP/superheated virgin aggregate mixing duration
5.6 Summary
The newly developed mixing method has duplicated the mixing mechanism that really
occurs in the industrial asphalt mixing plant. The mixing mechanism includes the following
steps:
 Virgin aggregate are superheated
 RAP material at ambient temperature is blended with superheated virgin aggregate.
The heat transferred from superheated virgin aggregate helps to soften RAP
agglomerate, weakening the bitumen binding among RAP aggregate particles.
Under mechanical mixing effort, these RAP lumps will be disintegrated and blended
with virgin aggregate.
 The RAP/superheated virgin aggregate blend will be mixed with virgin bitumen or
rejuvenator, the more the RAP material is disintegrated, the more likely and
effective the rejuvenating process.
113
Different mixing efforts (different RAP/superheated virgin aggregate mixing durations)
have depicted a clear picture of mixing mechanism between RAP and virgin materials. The
longer the mixing time, the more RAP and virgin materials are incorporated and hence the
homogeneity level of recycled mixture is enhanced. The visual assessment also
demonstrates that even when small RAP size is used and at considerable mixing duration,
RAP material, especially RAP binder, is not fully blended or rejuvenated by virgin binder.
This phenomenon is different from the assumption in the recycled mixture design process
where RAP and virgin binder are fully blended.
Figure 60: Relation between mixing effort, homogeneity and mechanical properties of recycled asphalt
The relation among mixing effort, homogeneity level, and mechanical properties of hot
recycled mixture is illustrated in Figure 60. Each factor will mutually influence the others.
Due to the fact that the RAP and virgin materials distribution do not follow any consistent
pattern, the relation between mixing effort and homogeneity level (1) could not be
quantified and neither could that between homogeneity level and mechanical properties (2).
The more mixing effort, the more homogeneity and less variation in mechanical properties.
There is no clear numerical parameter or values to characterize these relations. However,
the relation between mixing effort and mechanical properties (3) can be quantified. In this
preliminary experiment, there is a relation between mixing duration and stiffness values.
Hence, stiffness measurement proves to be potential tool to investigate the effect of mixing
effort on mechanical properties of hot recycled mixtures.
Mixing
Effort
Homogeneity
Level
Mechanical
Properties
1 3
2
114
The preliminary investigation of the effect of mixing process on mechanical properties of
hot recycled asphalt mixture shows the likely tendency that the more mixing, the less
variation in stiffness values. This might be due to the heterogeneity of recycled mixture.
Not enough mixing effort results in some places containing primarily RAP and the others
are dominated by virgin materials. The difference among RAP and virgin materials
properties, especially RAP and virgin binder, contributes to the considerable variation in
stiffness values. However, this preliminary statement needs to be further investigated and
verified as the number of testing samples is limited to 2 for each RAP/superheated virgin
aggregate mixing duration. In addition, due to visual assessment, the virgin binder is dyed
by iron oxide (10% by weight of virgin binder) which might affect the mixing characteristic
of virgin binder and the stiffness distribution.
115
6 Effects of mixing procedures and RAP materials on
stiffness distribution of hot recycled asphalt mixtures
6.1 Introduction
As the clear binder is dyed red by 10% by weight of iron oxide, the proportion of the
pigment certainly alters the flow characteristic of binder. This might affect the mixing
process and rejuvenation between virgin and aged binder. Therefore, this chapter also
investigates the effect of different mixing methods and RAP sizes on stiffness of recycled
asphalt mixtures. However, normal straight run bitumen 160/220 Pen is used as virgin
binder. Recycled asphalt mixtures are manufactured by different methods, black rock (BR),
complete blending (CB), the SHRP procedure, and the field simulation method (FS)
developed in Section 5.2. In FS method, RAP/superheated virgin aggregate mixing duration
varies from 2 to 8 minutes. Both large and small RAP materials are used in this experiment.
The stiffness distributions of recycled asphalt mixtures manufactured by different methods
are statistically compared to each other and those of BR and CB mixtures to investigate the
effect of mixing procedures and RAP materials.
6.2 Experiment design
6.2.1 Effects of mixing protocols and RAP sizes on stiffness of hot
recycled asphalt mixture
To investigate the effect of mixing protocols on stiffness distribution, recycled specimens
are manufactured by different methods. The mixing methods include the newly developed
method that duplicates the mixing mechanism in the industrial mixer (denoted as FS: field
simulation) with different RAP/superheated virgin aggregate mixing durations; and the
conventional method (from SHRP). The procedures are explained in Section 6.3.1.
Specimens are manufactured using the Mixer A (Section 6.2.2). In field simulation method
(FS), the RAP/superheated virgin aggregate mixing durations range from 2 to 8 minutes.
Both sizes of RAP, large and small are used (denoted as LR and SR). The stiffness of
recycled specimens are measured by indirect tensile stiffness test (ITSM) (BS-EN:12697-
26, 2004). The summary of experiments to investigate the effect of mixing protocols on
stiffness is shown in Table 32.
116
The experiment also includes the stiffness measurement of control mixtures that present the
“Black Rock” (BR) case where RAP binder is inert; and “Complete Blending” (CB)
assumed in the design process in which RAP binder is fully interacted with virgin binder or
rejuvenator. The aim is to compare the mechanical properties of control mixtures with
recycled mixtures manufactured by different protocols. The difference in mechanical
properties compared to those of control mixtures will demonstrate the effects of mixing
procedures and properties of RAP materials on the quality of hot recycled asphalt mixtures.
RAP/Virgin aggregate mixing duration (minutes)
LR FS 2 4 6 8
× × × ×
SHRP ×
RAP/Virgin aggregate mixing duration (minutes)
SR FS 2 4 6 8
× × × ×
SHRP ×
Table 32: Test plan to study the effects of mixing method on stiffness
6.2.2 Effect of mixing equipment on stiffness distribution of hot
recycled asphalt mixtures
Two mixers with different mechanical mixing effects, mixers A and B, are employed. The
characteristics of each mixer are as follows:
Mixer A
The schematic of mixer A is illustrated in Figure 61. There are two mixing paddles moving
with different orbits that help to drive and blend materials in the mixing bowl. The heat
supply of this mixer is maintained by heating the oil that moves between the external and
internal walls of mixing bowl. There is a thermocouple attached to the oil to control the
heat supply hence the mixer can be maintained at the required temperature.
Mixer B
The schematic of mixer B is presented in Figure 62. There are four mixing paddles used to
guide and blend materials in the mixing bowl. However, due to the rotating axis of mixer B
being not vertical (90
o
) but 60
o
compared to the ground plane, the movement of material
117
inside the mixing bowl consists of not only horizontal but also vertical direction. Mixer B
also allows reverse rotation. Also different from mixer A, the heat supplied for the mixer B
is controlled by the thermocouple that measures the air inside the mixer.
Figure 61: Schematic of Mixer A
Figure 62: Schematic of Mixer B
Primary Axis
Lid
Mixing Bowl
Rotating Orbit
Secondary Axis
Thermocouple
Mixing Paddle
Materials
Lid
Rotating Orbit
Thermocouple
Mixing Paddle
Materials
118
The summary of experiments to study the effect of mixers on stiffness distribution of hot
recycled mixtures is presented in Table 33.
Mixer RAP Mixing time (minutes)
Type Sizes 2 6
A LR × ×
SR × ×
B LR × ×
SR × ×
Table 33: Test plan to study the effects of mixer on stiffness of recycled asphalt
6.2.3 Effect of mixing protocols on RAP binder properties
Different mixing protocols will generate conditions that RAP materials are exposed at high
temperature for a certain period of time. This might alter the properties of RAP binder.
Depending on the exposure condition (Section 6.3.1), RAP binders are extracted and
recovered from RAP materials. The rheology properties of these conditioned RAP materials
are studied by Dynamic Shear Rheometer (DSR) and compared to that of original RAP
binder to quantify the effect of mixing procedure on properties of RAP. The summary of
experiments to investigate the effect of mixing protocol on properties of RAP binder is
shown in Table 34.
RAP/Virgin aggregate mixing duration (minutes)
LR FS 2 8
× ×
SHRP ×
RAP/Virgin aggregate mixing duration (minutes)
SR FS 2 8
× ×
SHRP ×
Table 34: Test plan to study the effects of mixing methods on binder properties
119
6.3 Materials and specimens manufacture
6.3.1 Material preparation and mixing procedure
Hot recycled mixtures
In this experiment, Shell bitumen 160/220 Pen is used as rejuvenator. The properties of
160/220 Pen bitumen are showed in Table 35. Although the origin of bitumen is not
allowed to be revealed due to Shell Global policy, all the bitumen binders used for the
whole research come from the same crude oil source. This is to eliminate the effect of
different crude oil sources on mechanical properties of recycled asphalt mixture. The
proportion of RAP in the recycled asphalt mixture is estimated using Grunberg and Nissan
viscosity mixing rule at 60
o
C. The amount of RAP in recycled mixture is 40% thus the
viscosity of recycled blend is approximately similar to that of 70/100 Pen bitumen (Table
35). Figure 63 illustrates the master curves of 70/100 Pen and recycled blend with
aged/virgin binder ratio of 4/6. The result shows that if aged and virgin binder is completely
mixed, the properties of recycled binder are almost similar to those of the 70/100 Pen
bitumen.
160/220 Pen 70/100 Pen
Penetration at 25
o
C (dmm) 192 83
Softening Point (
o
C) 37.4 47.2
Density (g/cm
3
) 1.021 1.028
Viscosity at 60
o
C (Pas) 64 192
Table 35: Properties of bitumen 160/220 Pen and 70/100 Pen
The materials for recycled mixtures conform to the requirements for surface course 10 mm
DBM (BS:4987-1, 2005). Both sizes of RAP, large and small, are used and denoted as LR
and SR. As bitumen content of RAP material is 5.2% and the aggregate gradation of RAP is
the same as that of virgin aggregate (Figure 34), amount of rejuvenator required is also
5.2% of the total weight of virgin materials.
Field Simulation method (FS)
 RAP materials, both large (LR) and small size (SR), are conditioned at room
temperature. The room temperature is maintained by thermal control system at
25
o
C.
120
 Virgin aggregate is superheated at 215
o
C for 8 hours.
 Rejuvenator is preheated at 135
o
C for 2 hours.
 The mixer temperature is maintained at 135
o
C.
 RAP material is mixed with superheated virgin aggregate for 2, 4, 6, 8 minutes.
 The combination of RAP and virgin aggregate is then blended with virgin binder for
2 minutes.
SHRP method
 RAP materials, both large (LR) and small size (SR), are conditioned at 110
o
C for 2
hours.
 Virgin aggregate is conditioned at 150
o
C for 8 hours.
 The mixer temperature is maintained at 135
o
C.
 Preheated RAP material is mixed with preheated virgin aggregate for 30 seconds in
the mixer maintained at 135
o
C.
 The combination of RAP and virgin aggregate is then blended with virgin binder for
2 minutes.
1.0E+00
1.0E+01
1.0E+02
1.0E+03
1.0E+04
1.0E+05
1.0E+06
1.0E+07
1.0E+08
1.0E+09
1.0E-07 1.0E-05 1.0E-03 1.0E-01 1.0E+01 1.0E+03 1.0E+05
Log Reduced Frequency
C
o
m
p
l
e
x

M
o
d
u
l
u
s

(
P
a
)
0
10
20
30
40
50
60
70
80
90
P
h
a
s
e

A
n
g
l
e

(
D
e
g
r
e
e
)
40% RAP recycled binder -
Complex Modulus
70/100 Pen - Complex Modulus
40% RAP Recycled binder - Phase
Angles
70/100 Pen - Phase Angles
Figure 63: Master-curves of 70/100 Pen and recycled blend with 40% RAP and 60% 160/200 Pen
bitumen
121
Control mixtures
“Black Rock” mixture (BR)
In Black rock case, there is an assumption that there is no interaction between RAP and
virgin binder. RAP binder is normally extracted and recovered from RAP material. After
batching, the combination of recovered RAP and virgin aggregate are then mixed with pure
virgin binder or rejuvenator (McDaniel et al., 2000). However, after binder recovery
process (BS-EN:12697-4, 2005), RAP binder still exists in recovered RAP aggregate, even
if RAP material is solvated in methylene chloride solution and soaked overnight. If
recovered RAP aggregate is used, the result is probably altered as rejuvenator might interact
with remaining RAP binder. Due to the fact that original RAP and virgin aggregate are the
same, the Black Rock mixture is manufactured from total virgin aggregate instead and
virgin binder 160/220 Pen. The bitumen content is 5.2% by weight of total mixture. The
preparation and mixing procedure for Black Rock mixture are as follows:
 Batched aggregate is preheated at 135
o
C for 8 hours.
 Virgin binder (160/220 Pen) is preheated at 135
o
C for 2 hours.
 Mixer is maintained at 135
o
C.
 Preheated aggregate and binder is mixed in the mixer for 2 minutes.
“Complete Blending” Mixture (CB)
In Complete Blending case, RAP binder is assumed to be fully blended with rejuvenator.
Therefore, RAP binder, after being extracted and recovered, is fully rejuvenated by virgin
binder before blending with batched aggregate. The same phenomenon occurs as for
recovery process, RAP binder is not fully extracted and recovered from RAP materials. If
RAP aggregate is used, the remaining RAP binder on recovered aggregate will deviate the
assumption of Complete Blending case. Hence, instead of batching recovered and virgin
aggregate, Complete Blending mixture in this case is made of pure virgin aggregate. The
bitumen content is 5.2% by weight of total mixture and the preparation of rejuvenated
binder for complete blending case is as follows:
 RAP binder is extracted and recovered from RAP materials
 RAP binder is blended with virgin binder (160/220 Pen) at 160
o
C by mechanical
mixer to produce homogeneous blend. The proportion of RAP/virgin binder is 4/6,
the same as the proportion of RAP material in hot recycled mixture.
The preparation and mixing procedure for Complete Blending (CB) mixture are as follows:
122
 Batched aggregate is conditioned at 150
o
C for 8 hours
 Rejuvenated binder is conditioned at 150
o
C for 2 hours
 Mixer is maintained at 150
o
C
 Aggregate and rejuvenated binder are mixed in the mixer for 2 minutes
The target of rejuvenated binder with 40% RAP binder is bitumen 70/100 Pen (Section
6.3.1). Hence, there is also a mixture made of virgin aggregate and bitumen 70/100 Pen,
denoted as CB-V. The bitumen content of this mixture is 5.2% by weight of total mixture.
This is to appraise the quality of recycled mixture compared to that made from virgin
materials. The material preparation and mixing procedure are the same as those for CB
mixture.
6.3.2 Compaction
The loose mixtures are compacted straight away after mixing by roller compacter (BS-
EN:12697-33, 2003). The internal dimensions of the mould are 305 mm x 305 mm. The
target height of compacted slab is 60 mm. For both recycled and control mixtures, the
amount of materials is determined based on maximum density, the bulk specific gravity at
target air void content 4%, and the volume of the slab after compaction.
6.3.3 Machining and storage of specimens
Compacted slabs are demoulded the next day and 5 specimens are cored from each slab.
The scheme of coring process is illustrated in Figure 64. To eliminate the surface defects
due to compaction, each specimen is cut at both ends to achieve the thickness of 40 mm.
The bulk specific density of each specimen is determined (BS-EN:12697-6, 2003) to
estimate the air void content (BS-EN:12697-8, 2003). All the specimens are then
conditioned at 20
o
C for 15 days before ITSM test. After stiffness determination, the
specimens are stored at 5
o
C for further testing.
6.4 Assessment method
The assessment method is primarily based on the comparison among stiffness data of
recycled mixtures manufactured by different mixing efforts. Due to RAP and virgin binder
properties being different, the hypothesis is that if RAP lumps are fully separated and well
blended with virgin material, stiffness values of different measuring directions will be
approximately the same. On the contrary, the heterogeneity of recycled mixture will result
123
in the considerable variance in stiffness values where mixing duration is not long enough.
In addition, as testing specimens are cored from roller-compacted slabs, the quality of
mixing is also revealed by the stiffness values between different samples.
Figure 64: Coring and cutting scheme for compacted slabs
The total number of specimens tested for each case, for instance 2 minutes
RAP/superheated virgin aggregate mixing duration, is ten. Stiffness is determined at 20
o
C,
5µm strain, 124 millisecond rise-time to ensure specimens are in the linear elastic regime
(BS-EN:12697-26, 2004). As the stiffness values of each sample are measured in four
directions at 45
o
C angular increments, 40 stiffness values are generated for each case. To
eliminate the effects of diffusion on mechanical properties, stiffness measurement is
implemented 15 days after the day of compaction.
The stiffness distribution is then analyzed by statistical tools, for instance, hypothesis test,
descriptive statistics, Anderson Darling statistic test. Stiffness distributions relevant to
different mixing efforts are compared to each other and also to those of “Black Rock” and
“Complete Blending” case to study the effect of mixing on mechanical properties of hot
recycled mixtures. Air void characteristic of testing specimens is also considered due to its
effect on stiffness values.
S1 S2
S3
S4 S5
305 mm
3
0
5

m
m
40 mm
100 mm
124
6.5 Results and analysis
6.5.1 Air void contents
Five cylindrical specimens are cored from roller-compacted slab. It is difficult to control the
air void content of each specimen. Although the target air void content is set at 4% volume,
there is some scatter in air void content data. Specimens at the centre of slab tend to have
lower air void content. The air void data is presented in Table 36. The standard deviation
(SD) or coefficient of variation (COV) indicates the scatter of air void content around the
mean of each set of specimens.
Mean (%) SD COV
FS-2 5.088 1.270 0.250
FS-4 5.626 0.692 0.123
LR FS-6 4.493 0.769 0.171
FS-8 5.263 0.903 0.172
SHRP 4.835 0.512 0.106
FS-2 5.218 0.460 0.088
FS-4 5.634 0.640 0.114
SR FS-6 6.086 1.645 0.270
FS-8 5.353 1.143 0.214
SHRP 3.481 0.377 0.108
CB 4.696 0.472 0.101
CB-V 4.313 0.917 0.213
BR 5.185 0.533 0.103
Table 36: Air void content of recycled mixture manufactured by different mixing method
The effect of air void content on stiffness of asphalt mixture has been investigated by
several researchers (Tayebali et al., 1994) (Read, 1996). Therefore, in order to compare
stiffness values of different asphalt mixtures, the air void contents of these materials must
be approximately the same. To compare whether the mean air void content of each data set
are identical, the t-test is used. The hypothesis of the test is if there is no significant
difference between the means of two data sets, the standard error of the difference in means
t must be smaller than the critical value t
critical
for the relevant confidence level. The
confidence level is 95% hence t
critical
must be in the range [-2.101; 2.101].
The summary of the statistical t-test is presented in Table 37. Pairs of air void content sets
that do not conform to the hypothesis of t-test are highlighted. The results show that almost
all mean air void contents are approximately similar to the others except control mixture
125
made of bitumen 70/100 Pen (CB-V) and recycled mixture made from small RAP by SHRP
mixing method (SR-SHRP). In general, the air void contents of these two sets of asphalt
mixtures are considerably lower than those of the others.
126
FS-8
FS-6 2.052
LR FS-4 -1.009 -3.462
FS-2 0.356 -1.266 1.176
SHRP 1.304 -1.169 2.906 0.584
FS-8 -0.195 -1.972 0.646 -0.49 -1.307
FS-6 -1.381 -2.701 -0.812 -1.514 -2.285 -1.15
SR FS-4 -1.059 -3.603 -0.027 -1.214 -3.082 -0.68 0.806
FS-2 -0.14 -2.577 1.551 -0.305 -1.76 0.345 1.598 1.666
SHRP 5.756 3.735 8.606 3.833 6.729 4.916 4.853 9.154 9.22
Control BR 0.234 -2.339 1.594 -0.224 -1.499 0.419 1.638 1.701 0.15 -8.25
Mixes CB 1.758 -0.711 3.508 0.913 0.629 1.677 2.554 3.723 2.5 -6.35 2.17
CB-V 2.333 0.476 3.613 1.563 1.571 2.242 2.963 3.732 2.79 -2.65 2.559 1.174
LR SR Control Mixes
FS-8 FS-6 FS-4 FS-2 SHRP FS-8 FS-6 FS-4 FS-2 SHRP BR CB CB-V
Table 37: Air void comparison by t-test
127
6.5.2 Effects of mixing time on stiffness
The summary of stiffness data is presented in Table 38. The mean stiffness comparison
among different sets of asphalt materials by t-test is shown in Table 39. The hypothesis is if
two means of stiffness are similar with confidence level of 95%, the calculated t value must
be in the range [-1.991; 1.991]. Each pair of stiffness sets that conforms to the hypothesis is
highlighted. The results show clearly that the mean stiffnesses of different sets of specimen
are quite different each other. For BR case, the difference is understandable as the viscosity
of bitumen 160/220 Pen is extremely low compared to that of the other bitumen. However,
except the BR case, the materials for the manufacture of the other recycled asphalt mixtures
are exactly the same. In fact, in these cases, if the RAP binders are fully extracted and
mixed with virgin bitumen (bitumen 160/220 Pen), these result blends will have similar
characteristics.
It has been argued that the stiffness difference among sets of recycled asphalt mixtures is
attributed to difference in air void characteristics. For instance, as a result of air void
analysis (Tables 36 and 37), the mean air void content of CB-V and SR-SHRP asphalt are
considerably lower than those of the other asphalt mixtures. However, the air void
comparison also demonstrates that apart from CB-V and SR-SHRP, the air void contents of
the other asphalt mixtures are statistically similar. Hence, if the stiffness values of these
asphalt mixtures are different, the reason must be due to different mixing characteristics.
Mean SD COV
FS-2 1262 483 0.383
FS-4 1412 314 0.222
LR FS-6 1610 197 0.122
FS-8 1720 66 0.038
SHRP 1614 102 0.063
FS-2 1732 240 0.139
FS-4 1733 77 0.044
SR FS-6 1808 96 0.053
FS-8 1866 139 0.075
SHRP 1774 47 0.026
CB 2294 144 0.063
CB-V 2409 210 0.087
BR 752 117 0.156
Table 38: Summary of stiffness values (MPa) of recycled mixtures manufactured by different mixing
methods
128
FS-8
FS-6 3.367
LR FS-4 6.685 3.379
FS-2 6.212 4.457 1.852
SHRP 5.532 -0.12 -3.874 -4.765
FS-8 -5.964 -6.704 -8.355 -7.756 -9.207
FS-6 -4.743 -5.709 -7.62 -7.274 -8.723 2.158
SR FS-4 -0.739 -3.661 -6.27 -6.345 -5.847 5.309 3.888
FS-2 -0.316 -2.705 -5.394 -5.902 -3.244 3.441 2.149 0.03
SHRP -4.162 -5.109 -7.205 -6.934 -8.966 3.971 2.035 -2.909 -1.277
Control BR 45.569 23.683 12.465 6.29 35.049 38.709 44.02 44.276 26.329 51.194
Mixes CB -22.973 -17.747 -16.158 -13.232 -24.389 -13.548 -17.785 -21.836 -14.264 -21.812 -52.615
CB-V 19.823 17.569 -16.705 -14.047 -21.553 -13.66 -16.483 -19.181 -14.654 -18.72 -43.641 -2.866
LR SR Control Mixes
FS-8 FS-6 FS-4 FS-2 SHRP FS-8 FS-6 FS-4 FS-2 SHRP BR CB CB-V
Table 39: Stiffness comparison by t-test
129
The most significant feature that separates recycled asphalt from virgin asphalt is the
recycled asphalt mixtures contain RAP materials, agglomerate lumps of RAP binder and
aggregate. In order to manufacture homogeneous recycled asphalt, RAP lumps should be
fully separated into single pieces and uniformly distributed in the whole mixture. Therefore,
virgin binder can rejuvenate and alter the properties of RAP binder. If this cannot occur,
then there will be some areas with high concentration of RAP and vice versa, other areas
are dominated by virgin material. Due to the fact that properties of RAP binder are entirely
different from those of virgin binder, there will be a substantial variation in stiffness values.
Figures 65 to 68 illustrate the relation among stiffness values, the location of specimens
cored from roller-compacted slabs, and the stiffness measured at different directions for
recycled asphalt mixtures manufactured from large RAP (LR) with different mixing
durations. The results show that mixing time between RAP and superheated virgin
aggregate significantly affects the homogeneity of recycled asphalt mixtures. With short
mixing duration, for instance 2 minutes, the degree of stiffness fluctuation is really high
(Figure 65). The stiffness values are not only different from specimen to specimen but also
different between measured directions in the same specimen. Tables 40 to 43 show the
stiffness values measured at different directions for specimens made of large RAP with
different mixing durations.
The data demonstrates that the homogeneity level of hot recycled asphalt mixture is
considerably improved once the RAP/superheating virgin aggregate mixing duration is
extended. Stiffness difference between specimens are considerably reduced and this
phenomenon is substantiated by the fact that the general coefficient of variation
significantly decreases from 38.3 to 3.8% (Table 38) once the mixing duration increases
from 2 to 8 minutes. In addition, the stiffness coefficient of variation in each specimen is
also reduced to less than 10% once the mixing duration is increased to 8 minutes (Table
43). The longer the mixing time, the closer the homogeneity of recycled mixture
approaches that of complete blending (CB) mixture (Table 44).
The mixing duration between RAP and superheated virgin aggregate is an important factor
that determines the quality of hot recycled mixture. During the mixing process, the heat
transferred from superheated aggregate will soften the RAP lumps. Under mechanical
effects of mixing paddles, RAP lumps are separated into single pieces covered by RAP
130
binder from circumference to the centre and distributed all over the mixture. If the mixing
time is not long enough, RAP lumps cannot be separated and hence, not well distributed in
the mixture. Consequently, this will result in the considerable variation in stiffness values.
1
2
3
4
S1
S2
S5
S4
S3
0
500
1000
1500
2000
2500
3000
S
t
i
f
f
n
e
s
s

(
M
P
a
)
Measured Directions
S
a
m
p
l
e
s
1
2
3
4
S7
S10
S9
S6
S8
0
500
1000
1500
2000
2500
3000
S
t
i
f
f
n
e
s
s

(
M
P
a
)
Measured Directions
S
a
m
p
l
e
s
Figure 65: Stiffness versus core location and measuring direction of LR FS-2 mixture
131
1
2
3
4
S1
S2
S3
S4
S5
0
500
1000
1500
2000
2500
3000
S
t
i
f
f
n
e
s
s

(
M
P
a
)
Measured Directions
S
a
m
p
l
e
s
1
2
3
4
S10
S6
S9
S7
S8
0
500
1000
1500
2000
2500
3000
S
t
i
f
f
n
e
s
s

(
M
P
a
)
Measured Directions
S
a
m
p
l
e
s
Figure 66: Stiffness versus core location and measuring direction of LR FS-4 mixture
132
1
2
3
4
S1
S4
S5
S3
S2
0
500
1000
1500
2000
2500
3000
S
t
i
f
f
n
e
s
s

(
M
P
a
)
Measured Directions
S
a
m
p
l
e
s
1
2
3
4
S8
S10
S7
S6
S9
0
500
1000
1500
2000
2500
3000
S
t
i
f
f
n
e
s
s

(
M
P
a
)
Measured Directions
S
a
m
p
l
e
s
Figure 67: Stiffness versus core location and measuring direction of LR FS-6 mixture
133
1
2
3
4
S1
S2
S3
S4
S5
0
500
1000
1500
2000
2500
3000
S
t
i
f
f
n
e
s
s

(
M
P
a
)
Measured Directions
S
a
m
p
l
e
s
1
2
3
4
S6
S7
S8
S9
S10
0
500
1000
1500
2000
2500
3000
S
t
i
f
f
n
e
s
s

(
M
P
a
)
Measured Directions
S
a
m
p
l
e
s
Figure 68: Stiffness versus core location and measuring direction of LR FS-8 mixture
134
Air
Void Stiffness in Different Directions Mean Standard COV
Samples Content (MPa) Stiffness Deviation (%)
(%) 1 2 3 4 (MPa)
S1 5.8 1499 1007 1047 731 1071 318 29.7
S2 4.1 722 920 753 768 791 88 11.1
S3 5.7 2049 2023 1792 1840 1926 129 6.7
S4 5.7 1205 1144 1168 1660 1294 245 18.9
S5 4.0 978 597 827 666 767 171 22.3
S6 4.2 1676 1908 1598 1768 1738 133 7.7
S7 7.9 823 603 691 756 718 94 13.1
S8 5.4 1928 1529 1723 1829 1752 171 9.8
S9 4.4 1665 1621 1583 1555 1606 48 3.0
S10 3.7 944 973 862 740 880 104 11.8
Table 40: Stiffness values of LR FS-2 specimens
Air
Void Stiffness in Different Directions Mean Standard COV
Samples Content (MPa) Stiffness Deviation (%)
(%) 1 2 3 4 (MPa)
S1 5.3 1063 881 872 1035 963 100 10.4
S2 7.3 1623 1530 1183 1068 1351 267 19.8
S3 5.5 1854 1703 1239 1285 1520 305 20.1
S4 4.9 1925 1517 2012 1405 1715 299 17.4
S5 6.1 1613 1620 1218 1387 1460 194 13.3
S6 5.4 1140 1062 952 1168 1081 97 9.0
S7 5.1 1692 1648 1298 1256 1474 228 15.5
S8 5.3 1914 1933 1503 1540 1723 233 13.5
S9 6.0 1543 1663 1250 1186 1411 229 16.2
S10 5.4 1150 1745 1114 1699 1427 341 23.9
Table 41: Stiffness values of LR FS-4 specimens
135
Air
Void Stiffness in Different Directions Mean Standard COV
Samples Content (MPa) Stiffness Deviation (%)
(%) 1 2 3 4 (MPa)
S1 4.6 1524 1328 1270 1232 1339 130 9.7
S2 3.6 1957 1974 1647 1911 1872 152 8.1
S3 4.1 1868 1863 1835 1839 1851 17 0.9
S4 6.1 1415 1447 1462 1480 1451 28 1.9
S5 4.0 1850 1701 1604 1700 1714 102 6.0
S6 4.6 1559 1584 1509 1486 1535 45 2.9
S7 4.3 1431 1505 1390 1406 1433 51 3.6
S8 4.2 1610 1512 1623 1568 1578 50 3.2
S9 3.9 1760 1888 1847 1779 1819 60 3.3
S10 5.5 1528 1606 1452 1461 1512 71 4.7
Table 42: Stiffness values of LR FS-6 specimens
Air
Void Stiffness in Different Directions Mean Standard COV
Samples Content (MPa) Stiffness Deviation (%)
(%) 1 2 3 4 (MPa)
S1 3.6 1727 1687 1697 1660 1693 28 1.7
S2 5.3 1666 1623 1629 1669 1647 24 1.5
S3 5.6 1768 1759 1744 1794 1766 21 1.2
S4 4.8 1795 1786 1853 1825 1815 30 1.7
S5 6.4 1751 1712 1657 1662 1696 45 2.7
S6 6.4 1806 1678 1744 1724 1738 53 3.0
S7 4.6 1747 1718 1690 1728 1721 24 1.4
S8 5.3 1780 1699 1685 1696 1715 44 2.6
S9 6.1 1637 1620 1651 1612 1630 17 1.0
S10 4.5 1823 1865 1721 1745 1789 67 3.7
Table 43: Stiffness values of LR FS-8 specimens
136
Air
Void Stiffness in Different Directions Mean Standard COV
Samples Content (MPa) Stiffness Deviation (%)
(%) 1 2 3 4 (MPa)
S1 4.3 2406 2415 2373 2386 2395 19 0.8
S2 5.1 2221 2255 2103 2186 2191 65 3.0
S3 5.5 2095 2103 2077 2084 2090 12 0.6
S4 4.7 2324 2233 2283 2317 2289 42 1.8
S5 4.7 2286 2202 2204 2282 2244 47 2.1
S6 5.0 2265 2358 2259 2269 2288 47 2.1
S7 4.6 2483 2470 2268 2488 2427 106 4.4
S8 4.9 2130 2119 2099 2076 2106 24 1.1
S9 4.2 2406 2356 2346 2259 2342 61 2.6
S10 4 2531 2562 2536 2496 2531 27 1.1
Table 44: Stiffness values of CB specimens
137
Longer mixing duration not only advances the distribution of RAP materials all over the
recycled mixture but also enhances the rejuvenation process between virgin and RAP
binder. If the rejuvenation can only occur once new virgin binder is in contact with RAP
binder, the more exposed area of RAP materials to virgin binder, the better this process.
When the mixing time is increased, the fact that more pieces are separated from RAP lumps
will enlarge the total RAP exposed area for rejuvenation. As more RAP binder is
incorporated with virgin binder, the stiffness values of recycled mixture generally increase.
1244.03
1412.26
1610.28
1720.83
2049
2012
1974
1865
597
872
1232
1612
0
500
1000
1500
2000
2500
0 2 4 6 8 10
Mixing Time (minutes)
S
t
i
f
f
n
e
s
s

(
M
P
a
)
Mean
Max
Min
Figure 69: Stiffness versus mixing time of LR FS mixtures
Figure 69 shows the relation between RAP/superheated virgin aggregate mixing durations
and stiffness of recycled mixtures made with large RAP (LR). The three continuous centre
lines present the mean stiffness and the boundaries of the mean with 95% confident level.
The top and the bottom dashed lines are for the maximum and minimum stiffness values
versus mixing durations. The data show that the mean stiffness increases considerably,
from 1200 MPa to 1700 MPa once the mixing durations are extended from 2 to 8 minutes.
The data also shows that even at very short mixing duration, for instance 2 minutes, the
means stiffness is far different from that of black rock (BR) case, 751 MPa.
138
The increase in stiffness value might be attributed to the incorporation or rejuvenation
between virgin and RAP binder. However, this is not the only reason as the increase of
stiffness values might be also due to the distribution of RAP materials. As the stiffness of
RAP is considerably higher than that of virgin binder, the specimens with higher
concentration of RAP will have higher stiffness. This situation is exaggerated in the case
where the mixing time is not sufficient to separate RAP lumps and these lumps are
randomly distributed over the mixture.
It is sometimes impossible to identify the causes, distribution or rejuvenation, credited to
the increase of stiffness values as the distribution of RAP over the mixture and
incorporation between RAP and virgin binder occurs simultaneously. However, it is
possible to recognize between distribution and rejuvenation, which could be dominantly the
reason for the increase of stiffness. If the increase in stiffness is attributed to the distribution
of RAP lumps, stiffness values are considerably different from specimen to specimen. In
addition, the stiffness values measured at different directions of each specimen are also
different. Most importantly if RAP material is merely distributed, the properties of the
mixture will be dominated by the characteristic of virgin material. On the contrary, the
properties of the mixture will be dominated by the properties of RAP binder where RAP
material predominates. When the mixing time is increased, as RAP lumps start to separate
and more RAP binder is rejuvenated by virgin binder, the properties of both RAP and
virgin binder have changed. Virgin binder stiffened by RAP bitumen will increase the
minimum value of stiffness. In addition, the incorporation between RAP and virgin binder
also reduces the variation among stiffness values.
Anderson-Darling test for standard distribution
To better understand the reason for the increase of stiffness values, stiffness data is further
investigated using the Anderson-Darling test (Anderson and Darling, 1952). This statistical
test is used to validate whether the stiffness values of recycled asphalt mixture match the
standard distribution with chosen confidence interval. The confidence level for the test is
95%. The theory of the test is to plot the empirical continuous distribution function (CDF)
of the data and compare it to that of the standard distribution. Anderson-Darling parameter
(AD) presents the difference between empirical data and theoretical normal distribution and
is expressed as following:
S N AD ÷ ÷ =
2
(20)
139
| | )) ( 1 log( ) ( log
) 1 2 (
1 i N i
N
i
Y F Y F
N
i
S
÷ +
÷ +
÷
=
¿
(21)
Where:
N: number of samples
F: assumed normal distribution function with estimated µ and δ
Y
i
: the sorted, standardized sample value
The hypothesis of the Anderson-Darling test is if the data set conforms to the standard
distribution, Anderson-Darling parameter has to be smaller than 0.787 and P value must be
higher than (1-Confidence level). The smaller the AD value is, the better the data fits into
normal distribution (Stephens, 1974).
Stiffness distribution of Control Asphalt Mixtures
The stiffness distributions of three control asphalt mixtures, BR, CB, and CB-V, are
illustrated in Figure 70. The results from the Anderson-Darling test demonstrate that
stiffness values of three control asphalt mixtures conform to the normal distribution (Figure
71). Coefficient of variation (COV) values are less than 15%. With 95% confidence
interval, the AD parameter ranges from 0.3 to 0.5. The mean stiffness values of CB and
CB-V are approximately the same, less than 10% difference. The result is as expected as
the rheological properties of recycled binder (CB) with 40% RAP binder is almost identical
to that of bitumen 70/100 Pen.
3000 2500 2000 1500 1000 500
35
30
25
20
15
10
5
0
St if f ness ( MPa)
F
r
e
q
u
e
n
c
y
2409 209.6 40
751.6 117.4 40
2297 143.8 40
Mean StDev N
CB- V
BR
CB
Variable
Hi st ogr am of CB-V, BR, CB
Normal
Figure 70: Stiffness distribution of control mixtures
140
3000 2500 2000 1500 1000 500
99
95
90
80
70
60
50
40
30
20
10
5
1
St if f ness ( MPa)
C
u
m
u
l
a
t
i
v
e
P
r
o
b
a
b
i
l
i
t
y
(
%
)
2409 209.6 40 0.329 0.506
751.6 117.4 40 0.363 0.425
2297 143.8 40 0.421 0.310
Mean StDev N AD P
CB- V
BR
CB
Variable
Pr obabi l i t y Pl ot of CB-V, BR, CB
Nor mal - 95% CI
Figure 71: Probability plot of stiffness values – control mixtures
3000 2500 2000 1500 1000 500 0
99
95
90
80
70
60
50
40
30
20
10
5
1
St if f ness ( MPa)
C
u
m
u
l
a
t
i
v
e
P
r
o
b
a
b
i
l
i
t
y
(
%
)
1262 482.8 40 1.546 < 0.005
1412 314.0 40 0.547 0.150
1610 197.0 40 0.698 0.063
1721 65.68 40 0.236 0.775
Mean StDev N AD P
LR FS- 2
LR FS- 4
LR FS- 6
LR FS- 8
Variable
Pr obabi l i t y Pl ot of LR FS-2, LR FS-4, LR FS-6, LR FS-8
Nor mal - 95% CI
Figure 72: Probability plot of stiffness values – LR FS mixtures
Figure 72 presents the stiffness cumulative probability of recycled asphalt mixtures
composed of large RAP manufactured with different RAP/superheated virgin aggregate
mixing durations. Results from the Anderson-Darling test show that only at 8 minutes
mixing time, the AD value is the same level as that of the control asphalt. At lower mixing
141
times, AD values are high, especially with 2 minutes mixing time, the stiffness distribution
does not conform to the standard distribution.
2100 1800 1500 1200 900 600
20
15
10
5
0
St if f ness ( MPa)
F
r
e
q
u
e
n
c
y
(
%
)
Figure 73: Stiffness histogram of LR FS-2 mixture
The stiffness histogram of recycled mixture composed of large RAP and 2 minute mixing
duration (LR FS-2) is shown in Figure 73. It is quite clear that rather than conforming to the
normal distribution, stiffness modulus values gather primarily into two main groups, G1
whose stiffness values range from 500 to 1200, and G2 from 1400 to 2200 MPa. If the
stiffness values of 2 minute mixing time are analyzed by these two groups, G1 and G2, the
distribution of stiffness values in each separated group conforms to the standard distribution
(Figure 74). The AD parameter for each group is almost the same level as that of control
mixtures (Figure 75).
If the comparison is based on only the mean stiffness, the properties of LR FS-2 are far
different from that of the BR case. In fact, this expresses a considerable interaction between
RAP and virgin binder hence, the mean stiffness of LR FS-2 is about 1262 MPa compared
to 751 MPa of the BR case. However, the stiffness distribution shows that group 1 of LR
FS-2 has 22 values varying from 500 MPa to 1200 MPa. In addition, the mean stiffness of
this group is 860 MPa, insignificantly different from that of the BR case. Figure 76 shows
the inter-quartile stiffness ranges of LR FS-2 group 1 and the BR case. This group
possesses the same characteristics as that of the BR mixture. The stiffness distribution
142
indicates that there is insubstantial interaction between RAP and virgin binder and the
properties of LR FS-2 is almost dominated by virgin binder, or RAP acts as “Black Rock”.
If the mixing duration is insufficient, RAP lumps are not heated up. Therefore, the bondage
between RAP aggregate pieces are neither weakened nor deactivated. Mechanical mixing
therefore only distributes RAP at approximately original size all over the mixture.
2100 1800 1500 1200 900 600
40
30
20
10
0
St if f ness ( MPa)
F
r
e
q
u
e
n
c
y
(
%
)
Figure 74: Stiffness histogram of LR FS-2 mixture – Stiffness grouping analysis
3000 2500 2000 1500 1000 500 0
99
95
90
80
70
60
50
40
30
20
10
5
1
St if f ness ( MPa)
C
u
m
u
l
a
t
i
v
e
P
r
o
b
a
b
i
l
i
t
y
(
%
)
1262 482.8 40 1.546 < 0.005
860.2 179.0 22 0.395 0.343
1753 175.5 18 0.173 0.914
Mean StDev N AD P
LR FS- 2
LR FS- 2 G1
LR FS- 2 G2
Variable
Pr obabi l i t y Pl ot of LR FS-2, LR FS-2 G1, LR FS-2 G2
Normal - 95% CI
Figure 75: Stiffness probability plot of LR FS-2 – Stiffness grouping analysis
143
BR LR FS- 2 G1
1400
1200
1000
800
600
400
200
0
S
t
i
f
f
n
e
s
s
(
M
P
a
)
Figure 76: Inter-quartile stiffness ranges of LR FS-2 group 1 and BR mixture
LR FS-8 LR FS-6 LR FS-4 LR FS-2
2250
2000
1750
1500
1250
1000
750
500
S
t
i
f
f
n
e
s
s
(
M
P
a
)
Figure 77: Inter-quartile stiffness ranges of LR FS mixtures
When the mixing time is increased, the RAP binder is heated up and becomes softer. RAP
lumps will start to disintegrate first from outside to centre once the bituminous bond
between RAP aggregate particles is prevailed by mechanical mixing effort. Under
mechanical effects, separated pieces will be moved around and mixed up with virgin
aggregate and binder. RAP material no longer acts as “Black Rock” but starts to integrate
with virgin binder. The more RAP pieces are disintegrated, the more interaction between
144
RAP and virgin binder. Due to the incorporation between RAP binder and rejuvenator, the
stiffness values generally increase. Figure 77 illustrates the inter-quartiles stiffness range
for recycled mixtures composed of large RAP (LR) by field simulation method (FS) with
different mixing durations. The bottom line presents the end of the first quartile data and the
centre line is for the median. The data shows the minimum stiffness value increases from
597 to 1612 MPa when the mixing time increases from 2 to 8 minutes. In addition, the data
also indicates the first quartile of stiffness data and the stiffness median increases
considerably once the RAP/superheated virgin aggregate mixing duration is extended.
SR FS- 8 SR FS-6 SR FS-4 SR FS- 2
2250
2000
1750
1500
1250
1000
750
500
S
t
i
f
f
n
e
s
s
(
M
P
a
)
Figure 78: Inter-quartile stiffness ranges of SR FS mixtures
While the mixing duration considerably affects the stiffness values of recycled mixture
composed of large RAP, it only slightly influences the stiffness values of small RAP
recycled mixtures. Figure 78 shows that when the mixing duration increases from 2 to 8
minutes, the mean stiffness values of SR recycled mixtures change unnoticeably. This is
because under the mechanical mixing efforts, the original size of large RAP is significantly
reduced. Therefore, the stiffness of LR recycled mixture increases due to more virgin
binder interacting with RAP binder. However, the original size of small RAP only changes
slightly due to small RAP being already crushed into smaller size before use. The
mechanical mixing effort could not disintegrate those RAP lumps composed of fine
aggregates and fillers or the coarse aggregate covered by filler mastic. However, the
increase in mixing time considerably increases the homogeneity of recycled mixture
145
composed of small RAP materials. Figure 78 shows that the inter-quartile stiffness range of
SR FS mixtures is considerably narrowed once mixing time is extended.
6.5.3 Effects of RAP sizes on stiffness
The size of RAP lumps substantially affects the homogeneity of the recycled mixtures. For
the same RAP/superheated virgin aggregate mixing duration, the recycled mixtures
composed of small RAP material have higher levels of homogeneity than the mixtures
composed of large RAP. Figures 79 and 80 illustrate the relation among stiffness values,
the location of specimens cored from roller-compacted slabs, and the stiffness measured at
different directions of recycled asphalt mixtures composed of small RAP (SR) with 2 and 4
minutes mixing duration. For the case of 2 minutes mixing time, the stiffness coefficient of
variation for SR mixture is 13.9 compared to 38.3% of LR mixture (Table 38). In addition,
the stiffness in Table 45 also shows that stiffness variation in each specimen of SR FS-2
mixture is also lower, maximum 10% compared to 30% of LR FS-2 mixture. For 4 minutes
mixing duration, the phenomenon is the same. Stiffness coefficient of variation for SR
mixture is 4.4% compared to 22.2% of LR mixture and the stiffness variation in each
specimen is a maximum of 6%compared to 24% of LR mixture (Tables 38 and 46).
Size of RAP affects not only the homogeneity level but also the interaction between RAP
and virgin binder. At 2 minutes mixing duration, the properties of recycled mixture
composed of large RAP are primarily dominated by virgin binder. In fact, RAP acts as
“Black Rock”. On the contrary, there is a considerable interaction between RAP and virgin
binder in recycled mixture composed of small RAP materials. The interaction is indicated
by the mean and the minimum stiffness values of SR mixtures being far different from
those of LR mixtures (Figure 81) for the same mixing duration. For instance, at 2 minutes
mixing duration, the minimum stiffness of SR mixture is 1497 MPa compared to 597 MPa
of LR mixture, and the mean stiffness is 1806 MPa compared to 1262 MPa.
The phenomenon is as expected as the interaction between RAP and virgin binder is
controlled by the total surface area of RAP materials. This total surface area is also the
exposed area of RAP material to virgin binder for rejuvenation. The larger the total exposed
area, the higher the probability that the virgin binder can coat and penetrate into RAP
binder. In addition, the total surface area is inversely related to the size of RAP materials.
As small RAP materials have larger total surface area, the recycled mixtures composed of
146
small RAP tend to have higher stiffness values due to more interaction between RAP and
virgin binder.
1
2
3
4
S4
S1
S2
S5
S3
0
500
1000
1500
2000
2500
3000
S
t
i
f
f
n
e
s
s

(
M
P
a
)
Measured Directions
S
a
m
p
l
e
s
1
2
3
4
S6
S7
S8
S9
S10
0
500
1000
1500
2000
2500
3000
S
t
i
f
f
n
e
s
s

(
M
P
a
)
Measured Directions
S
a
m
p
l
e
s
Figure 79: Stiffness versus core location and measuring direction of SR FS-2 mixture
147
1
2
3
4
S1
S2
S3
S4
S5
0
500
1000
1500
2000
2500
3000
S
t
i
f
f
n
e
s
s

(
M
P
a
)
Measured Directions
S
a
m
p
l
e
s
1
2
3
4
S6
S7
S8
S9
S10
0
500
1000
1500
2000
2500
3000
S
t
i
f
f
n
e
s
s

(
M
P
a
)
Measured Directions
S
a
m
p
l
e
s
Figure 80: Stiffness versus core location and measuring direction of SR FS-4 mixture
148
1412
1610
597
872
1232
1612
1244
1721
1866
1808
1733 1732
1437
1550
1582
1653
0
200
400
600
800
1000
1200
1400
1600
1800
2000
0 2 4 6 8 10
Mixing Time (minutes)
S
t
i
f
f
n
e
s
s

(
M
P
a
)
LR FS Mean
LR FS Min
SR FS Mean
SR FS Min
Figure 81: Stiffness range of LR and SR recycled mixtures manufactured by FS methods
149
Air
Void Stiffness in Different Directions Mean Standard COV
Samples Content (MPa) Stiffness Deviation (%)
(%) 1 2 3 4 (MPa)
S1 5.6 1709 1459 1437 1451 1514 130 8.6
S2 5.4 1871 1714 1660 1697 1736 93 5.4
S3 4.6 2161 2048 2005 2008 2056 73 3.6
S4 5.2 1953 1825 1854 1867 1875 55 2.9
S5 5.5 1843 1754 1714 1682 1748 70 4.0
S6 6.1 2052 1772 1793 1821 1860 130 7.0
S7 5.1 1550 1447 1498 1446 1485 50 3.4
S8 4.7 1646 1528 1518 1472 1541 74 4.8
S9 4.9 1525 1658 1504 1583 1568 69 4.4
S10 5.0 1946 1922 1942 1927 1934 12 0.6
Table 45: Stiffness values of SR FS-2 specimens
Air
Void Stiffness in Different Directions Mean Standard COV
Samples Content (MPa) Stiffness Deviation (%)
(%) 1 2 3 4 (MPa)
S1 6.2 1860 1702 1664 1698 1731 88 5.1
S2 5.7 1703 1681 1677 1667 1682 15 0.9
S3 5.4 1790 1674 1550 1661 1669 98 5.9
S4 5.2 1782 1766 1659 1673 1720 63 3.7
S5 6.1 1782 1619 1745 1632 1695 81 4.8
S6 6.0 1734 1712 1697 1601 1686 59 3.5
S7 6.6 1766 1706 1721 1698 1723 30 1.7
S8 4.3 1835 1779 1834 1889 1834 45 2.5
S9 5.5 1764 1714 1769 1803 1763 37 2.1
S10 5.3 1816 1831 1818 1831 1824 8 0.4
Table 46: Stiffness values of LR FS-4 specimens
150
6.5.4 Effects of mixing methods on stiffness
Mixing methods significantly affect stiffness distribution of hot recycled mixture. Figure 82
shows that stiffness of recycled mixtures manufactured by SHRP method is quite different
from that of mixtures produced by field simulation (FS) method, especially for the recycled
mixtures composed of large RAP. While stiffness of LR FS-2 mixture is relatively close to
that of the BR mixture, the stiffness of LR SHRP and LR FS-8 are comparatively close to
that of the CB mixture. In addition, in the SHRP method, the sizes of RAP contribute
almost no effects on stiffness of recycled mixture. Figure 83 shows that the stiffness values
of recycled mixture composed of large and small RAP by SHRP methods are
approximately the same. On the contrary, RAP sizes significantly affect stiffness values of
recycled mixtures manufactured by field simulation method (FS). This is due to the mixing
mechanisms as these approaches determine how RAP and virgin materials interact with
each other.
CB LR FS- 8 LR SHRP LR FS-2 BR
2500
2000
1500
1000
500
S
t
i
f
f
n
e
s
s
(
M
P
a
)
Figure 82: Inter-quartile stiffness ranges of control and LR mixtures manufactured by different
methods
In the SHRP method, RAP lumps are conditioned in a force draft oven at 110
o
C. The
conditioning temperature is considerably higher than the softening point, 58
o
C, of RAP
151
binder. Although the heat transfer process is significantly influenced by the size of material
(Cutnell and Johnson, 2004), under the conditioning duration of 2 hours, both small and
large sizes of RAP material are entirely heated. The heat that RAP materials absorb during
the conditioning duration will soften the RAP binder and weaken the bituminous bond
between RAP aggregate particles. Under the mechanical mixing effort, these RAP lumps
are disintegrated into smaller pieces and will be blended and rejuvenated by virgin binder.
This is why when using the SHRP method, mixtures composed of large and small RAP
have approximately the same stiffness value (Figure 83).
The mixing mechanism of the field simulation (FS) method, on the contrary, is quite
different from that of SHRP. In this approach, RAP materials at ambient temperature are
mixed with superheated virgin aggregate. RAP lumps exist as unbreakable agglomerates
under normal mechanical mixing at ambient temperature. However, under the heat
transferred from superheated virgin aggregate, RAP binder will be softened and the bonds
among RAP aggregate particles will be deactivated. The superheated temperature of virgin
aggregate is 215
o
C. Although this temperature is extremely higher than the softening point
of RAP binder, a certain duration or critical duration is required so that the heat can be
transferred from superheated virgin aggregate to completely heat up RAP materials.
SR FS-2 LR FS-2 SR SHRP LR SHRP
2000
1500
1000
500
0
S
t
i
f
f
n
e
s
s
(
M
P
a
)
Figure 83: Inter-quartile stiffness ranges of SHRP mixtures and FS-2 mixtures
152
As the mechanical mixing efforts remains constant during the mixing process, mixing
duration in FS method, or the amount of heat transferred from virgin aggregate, is a critical
factor that determines the quality of hot recycled mixture. Due to the heat transferred from
virgin aggregate, the bitumen bond between RAP aggregates will be gradually weakened
until being overcome by mechanical mixing power. At this critical point, RAP lumps start
to disintegrate. Due to the fact that heat transfer process is significantly influenced by the
sizes of material (Cutnell and Johnson, 2004), the bigger the size of RAP lumps, the longer
the critical mixing duration is required.
In the case that the mixing duration is not sufficient, RAP materials will be distributed all
over the mixture at approximately original size. The incorporation between RAP and virgin
binder in this case will depend primarily on RAP size. Due to the total surface area, the
smaller the size of RAP, the more interaction between RAP and virgin binder occurs. If the
large size of RAP is used, an inconsiderable proportion of RAP will interact with virgin
binder and RAP materials will act as black rock. The longer the mixing time, the more RAP
aggregate pieces are separated, and the more RAP binder will interact with virgin binder.
Hence, when the mixing time is extended to 8 minutes, the stiffness of LR FS-8 approaches
that of CB mixture. If the mixing duration is the same, recycled mixtures composed of
small RAP will have better incorporations between RAP and virgin binder than large RAP
mixtures.
The experimental data also indicates that SHRP method tends to overestimate the
mechanical properties of hot recycled mixtures. The RAP extensive preheating duration in
the SHRP method coincidently advances the incorporation between RAP and virgin
materials, especially when large sizes of RAP materials are used. In addition, such a long
preheating duration is by no means practical in the real industry due to the length constraint
of the mixer and economical issues (Lee et al., 1983). Hence, field simulation (FS) method
would better exemplify what occurs in the real industrial mixers.
The data in Table 38 show that stiffness of SHRP mixture is approximately similar to that
of FS-4 or FS-6 mixtures. This is because in FS method, there is a critical period of time
which is required to transfer the heat from superheated virgin aggregate to soften RAP
materials. If the RAP/superheated virgin aggregate mixing duration is shorter than the
required duration, the interaction between RAP and virgin binder of FS mixture will be less
153
than that of the SHRP mixture. Therefore, stiffness of FS mixture is certainly lower than
that of SHRP mixture.
It is interesting that both methods, SHRP and FS, could not produce the recycled mixtures
that possess the stiffness values approximately the same as that of the complete blending
case. Figure 84 and Table 38 show that the mean stiffness of SHRP mixture, even when
small RAP is used, is far different from that of complete blending mixture, 1773 MPa
compared to 2294 MPa. By FS method, even at 8 minutes mixing time which is never
practical in the industry, the mean stiffness is 1865 MPa. This finding is quite different to
the result of McDaniel et al. (2000) where the properties of actual blending mixture (similar
to SHRP method in this research) is the same as that of complete blending.
CB SR FS-8 SR SHRP BR
2500
2000
1500
1000
500
S
t
i
f
f
n
e
s
s
(
M
P
a
)
Figure 84: Inter -quartile stiffness ranges of control and SR mixtures manufactured by different
methods
Theoretically, the complete blending situation will exist once all the RAP binder is
activated and rejuvenated by virgin binder. If RAP lumps are entirely disintegrated into
single pieces of aggregates, including also fine aggregate particles and filler, coated by
RAP binder; and under mechanical mixing, virgin binder could cover every single piece of
RAP binder-coated aggregate, the complete blending might occur. In this case, the stiffness
154
of recycled mixture will be approximately the same as that of complete blending mixture.
However, this situation is hardly happens in reality. This is because mechanical mixing can
disintegrate coarse aggregate from the RAP lumps. However, this process does not separate
fine aggregates from RAP lumps, or especially, filler from filler mastic. Figures 55 and 56
show that after 8 minutes mixing time by the FS method, although the original size of RAP
is reduced, RAP materials still exist as agglomerates. Consequently, there will be a
proportion of RAP binder that is not activated and rejuvenated by virgin binder.
6.5.5 Effect of mixing equipment on stiffness
The heat supply in mixer A is controlled by the thermocouple attached on the mixing bowl.
On the contrary, the temperature in mixer B is controlled by the temperature of the air
inside the mixing apartment. Due to the difference in mixing operation, the temperatures of
the loose recycled mixtures after mixing are also different. During the manufacture process,
although the temperatures are set the same, the loose recycled mixtures manufactured by
Mixer B always have considerably higher temperature than those produced by Mixer A.
Therefore, a small experiment was carried out with the aim to eliminate the effects of
differences in mixing temperature on the mechanical properties of recycled mixture. In this
experiment, temperature is recorded by external thermocouple attached directly to the loose
mixture during the mixing process. The results indicate that the setting temperature of
Mixer B should be about 20
o
C lower than that of Mixer A in order to produce the loose
mixture with the same temperature.
LR FS-2 LR FS-6 SR FS-2 SR FS-6
Mean COV (%) Mean COV (%) Mean COV (%) Mean COV (%)
Mixer A 5.09 25.0 4.49 17.1 5.21 8.6 6.08 27.1
Mixer B 5.15 20.8 3.47 19.9 4.75 11.4 3.46 14.2
Table 47: Air void summary of recycled specimens manufactured by different method and equipments
Table 47 presents the summary of air void content and coefficient of variation of recycled
mixtures manufactured by both Mixers A and B. The result shows that the air void contents
are approximately similar except LR FS-6 and SR FS-6 mixtures produced by Mixer B
which have slightly lower volumes of air void content.
The summary of stiffness values are shown in Table 48. The results show that for 2 minutes
mixing time, the mean stiffness values of large RAP mixtures manufactured by both Mixer
155
A and B are slightly different. In addition, the stiffness values distribution has almost the
same pattern as that of mixture manufactured by Mixer A. Table 49 and Figure 85 illustrate
the relation among stiffness values, the location of specimens cored from roller-compacted
slabs, and the stiffness values measured in different directions of LR FS-2 mixtures
produced by Mixer B. Besides the general variation coefficient of stiffness being 29.2%, the
stiffness values measured in different directions for the same specimens also vary
substantially. In addition, this mixture has 20 stiffness values lower than 1200 MPa. The
mean stiffness of this group, 1026 MPa, is quite close to 751 MPa of Black Rock (BR)
mixture, indicates that there is inconsiderable interaction between RAP and virgin binder.
Method Mean SD COV (%) Max Min Median
LR FS-2 1262 483 38.3 2049 597 1156
Mixer A LR FS-6 1610 197 12.2 1974 1232 1576
(MA) SR FS-2 1732 204 11.8 2161 1437 1714
SR FS-6 1808 96 5.3 2002 1582 1794
LR FS-2 1342 392 29.2 2048 752 1219
Mixer B LR FS-6 1690 77 4.6 1839 1533 1679
(MB) SR FS-2 1771 74 4.2 1896 1631 1779
SR FS-6 1825 63 3.5 1947 1702 1827
Table 48: Stiffness (MPa) of recycled specimens manufactured by different methods and equipment
When the mixing time is extended from 2 to 6 minutes, the interaction between RAP and
virgin binder of large RAP mixture manufactured by Mixer B is far better than that of
Mixer A. Although there is insignificant difference in the mean stiffness, 1610 MPa by
mixer A and 1690 MPa by mixer B, the minimum stiffness value of Mixer B mixture is
1533 MPa compared to 1232 MPa of mixture manufactured by Mixer A. The minimum
stiffness value of Mixer B mixture is almost the same as the median stiffness of Mixer A
mixture, 1576 MPa (Table 48). In addition, the homogeneity of Mixer B mixture is also
much better. Not only the general stiffness coefficient of variation but also the stiffness
variations in each specimen (Table 50 and Figure 86) are lower than those of Mixer A
mixture. This is because Mixer B is more efficient in than Mixer A. The tilt axis of Mixer B
allows the material in the mixing apartment to move not only horizontally but also
156
vertically. In addition, Mixer B has a function that can reverse the mixing direction also
enhances the homogeneity of the recycled mixture.
The data indicates that when the mixing time is not sufficient to deactivate all the bitumen
bonds in RAP lumps, the effect of mechanical mixing can only distribute the RAP material
all over the mixture. However, when the mixing duration is adequate (reaches critical
point), mechanical mixing will separate RAP lumps and enhance the interaction between
RAP and virgin binder. The more efficient the mechanical mixing, the higher the
homogeneity level of recycled asphalt mixtures which are manufactured. The length of
critical duration depends primarily on RAP size. The bigger the size of RAP, the longer the
critical duration.
157
Air
Void Stiffness in Different Directions Mean Standard COV
Samples Content (MPa) Stiffness Deviation (%)
(%) 1 2 3 4 (MPa)
S1 6 996 934 1422 1415 1192 263 22.1
S2 6.3 813 752 1044 969 895 135 15.1
S3 4.7 1703 1082 1448 1994 1557 387 24.9
S4 6.3 1018 1622 1349 1261 1313 249 19.0
S5 5.5 1294 1034 1042 1245 1154 135 11.7
S6 5.5 1179 1070 1115 1043 1102 59 5.4
S7 5.2 1093 1078 1194 1022 1097 72 6.6
S8 5.2 1002 1044 1967 1996 1502 554 36.9
S9 3.0 2028 1934 1473 1271 1677 363 21.6
S10 3.8 2048 1931 2005 1752 1934 131 6.8
Table 49: Stiffness values of LR FS-2 specimens – Mixer B
Air
Void Stiffness in Different Directions Mean Standard COV
Samples Content (MPa) Stiffness Deviation (%)
(%) 1 2 3 4 (MPa)
S1 3.8 1832 1738 1739 1576 1721 107 6.2
S2 4.9 1566 1563 1533 1570 1558 17 1.1
S3 3.7 1727 1669 1655 1654 1676 35 2.1
S4 3.7 1744 1639 1662 1603 1662 60 3.6
S5 3.4 1733 1679 1648 1660 1680 36 2.1
S6 3.6 1714 1636 1634 1787 1693 73 4.3
S7 3.4 1772 1748 1634 1651 1701 69 4.1
S8 2.8 1647 1760 1678 1687 1693 48 2.8
S9 2.3 1839 1787 1754 1815 1799 37 2.1
S10 3.1 1762 1742 1730 1630 1716 59 3.4
Table 50: Stiffness values of LR FS-6 specimens – Mixer B
158
1
2
3
4
S1
S2
S3
S4
S5
0
500
1000
1500
2000
2500
3000
S
t
i
f
f
n
e
s
s

(
M
P
a
)
Measured Direction
S
a
m
p
l
e
s
1
2
3
4
S6
S7
S8
S9
S10
0
500
1000
1500
2000
2500
3000
S
t
i
f
f
n
e
s
s

(
M
P
a
)
Measured Direction
S
a
m
p
l
e
s
Figure 85: Stiffness versus core location and measuring direction of LR FS-2 mixture – Mixer B
159
1
2
3
4
S1
S2
S3
S4
S5
0
500
1000
1500
2000
2500
3000
S
t
i
f
f
n
e
s
s

(
M
P
a
)
Measured Direction
S
a
m
p
l
e
s
1
2
3
4
S6
S7
S8
S9
S10
0
500
1000
1500
2000
2500
3000
S
t
i
f
f
n
e
s
s

(
M
P
a
)
Measured Direction
S
a
m
p
l
e
s
Figure 86: Stiffness versus core location and measuring direction of LR FS-6 mixture – Mixer B
160
6.5.6 Effects of mixing methods on RAP binder properties
The increase of stiffness may also be attributed to the aging of RAP binder. This is because
during the mixing process, RAP binder might be aged more due to the exposure to high
temperature from superheated aggregate as well as the mixer for a long period of time. An
experiment has been carried out to investigate if there is any alteration to RAP binder
properties during the mixing process. In this experiment, RAP lumps including both small
(SR) and large sizes (LR), are mixed with superheated virgin aggregate for different mixing
durations. The procedures are the same as those in the manufacture process of field
simulation method (FS). For the SHRP method, RAP lumps are conditioned in the force
draft oven at 110
o
C for 2 hours.
After being extracted and recovered, rheological properties of processed RAP binders are
studied by Dynamic Shear Rheometer (DSR). The thickness of testing specimen is 1000
µm for 25 mm plate and 2000 µm for 8 mm plate. Testing temperatures range from 5 to
45
o
C for 8 mm plate and from 20 to 80
o
C for 25 mm plate. Rheological testing is carried
out under 0.8% strain to ensure bitumen responds in linear visco-elastic region. The test
frequencies range from 0.1 to 10 Hz. Master-curves of these processed RAP binders are
constructed from rheological data and compared to that of original RAP binder. Figures 87
and 88 show the complex modulus and phase angle versus log reduced frequency of
original RAP and processed RAP binders. The data indicates there is no significant
alteration to RAP binder after RAP lumps are mixed with superheated virgin aggregate by
FS method. For the SHRP method, the RAP binder extracted from small RAP lumps shows
noticeably more ageing compared to original RAP binder.
161
1.0E+00
1.0E+01
1.0E+02
1.0E+03
1.0E+04
1.0E+05
1.0E+06
1.0E+07
1.0E+08
-6 -5 -4 -3 -2 -1 0 1 2 3 4
Log Reduced Frequency (Hz)
C
o
m
p
l
e
x

M
o
d
u
l
u
s

(
P
a
)
LR FS-2
LR FS-8
SR FS-2
SR FS-8
RAP
SR SHRP
LR SHRP
Figure 87: Complex modulus versus log reduced frequency of RAP binder before and after processed
by different mixing methods
20
30
40
50
60
70
80
90
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5
Log Reduced Frequency (Hz)
P
h
a
s
e

A
n
g
l
e
LR FS-2
LR FS-8
SR FS-2
SR FS-8
RAP
SR SHRP
LR SHRP
Figure 88: Phase angle versus log reduced frequency of RAP binder before and after processed by
different mixing methods
162
6.6 Summary
Stiffness values of cylindrical specimens measured in different orientations indirectly
express the heterogeneity of recycled mixture. The variation in stiffness values at different
measured directions will be substantial for a heterogeneous mixture and minor in the case
where the recycled mixture is homogeneous.
The mixing methods considerably affect the reaction between RAP and virgin binders as
the mixing mechanisms determine how RAP and virgin binder are blended together. In the
SHRP method, the mixing condition favourably enhances the interaction between RAP and
virgin binders. The long preheating condition at a temperature considerably higher than the
softening point of the RAP binder coincidently deactivates the bitumen bond between RAP
aggregate particles. Under mechanical mixing, the RAP lumps are separated thereby
increasing the contact areas between RAP and virgin binders.
On the contrary in the FS method, RAP is mixed with superheated virgin aggregate. The
thermal energy transferred from virgin aggregate will help to increase the RAP temperature
and weaken the bitumen bond between RAP aggregates. As the thermal transfer process is
time dependent, the point that the bitumen bond is overcome by mechanical mixing effort is
defined as a critical point. At this critical point, RAP lumps under mechanical mixing will
start to disintegrate into separate pieces of aggregate coated by RAP binder. If the
RAP/superheated virgin aggregate duration does not exceed the critical duration, the
mechanical mixing effort only distributes the RAP all over the mixture at its approximately
original size. In this case, the incorporation between RAP and virgin binder depends
primarily on the original size of RAP. The smaller the original size of RAP materials, the
better the interaction between RAP and virgin binder.
The increase in mixing duration significantly improves the homogeneity level of recycled
mixture. Not only the stiffness variation among different specimens but also the variation of
stiffness values measured in different directions for the same specimen significantly
reduces. The homogeneity level is also substantially affected by the sizes of RAP material.
For the same mixing effort, the recycled mixtures composed of small RAP are generally
more homogeneous than those made from large RAP. The more homogeneous the recycled
mixture, the more interaction between RAP and virgin binder. Therefore, the stiffness of
163
recycled mixture generally increases once the RAP/superheated virgin aggregate mixing
duration is extended.
Although the increase of mixing duration has positive effects on the homogeneity, the
complete blending between RAP and virgin binder assumed in the design process would
never exist in the production of recycled asphalt mixtures. Qualitatively, even at favourable
conditions of considerably long mixing times compared to those in the real asphalt mixing
plant, there still exist a considerable proportion of RAP as lumps. Quantitatively, as RAP
binder is not completely blended with virgin binder, the stiffness values of recycled asphalt
mixtures are lower than those of Complete Blending mixture.
164
7 Effects of mixing methods on fatigue life of hot
recycled asphalt mixtures
7.1 Introduction
This chapter investigates the effect of different mixing methods and RAP sizes on fatigue
life of recycled mixtures. The materials used in this chapter are the same as those in
Chapter 6. Fatigue life of recycled mixtures manufactured by different mixing methods are
determined by indirect tensile fatigue test (BS-EN:12697-24, 2004). The fatigue lives of
recycled mixtures are compared based on the parameters of the fatigue equations and the
number of loading cycles to fatigue failure at 100 microsrain.
7.2 Materials preparation and testing plan
The materials for this experiment are those specimens used for ITSM (Section 6.3.1). After
stiffness measurement, these specimens are subjected to indirect tensile fatigue test (ITFT)
to determine the number of loading cycles to failure. The indirect tensile fatigue test is
carried out under stress-controlled mode by Nottingham Asphalt Tester (NAT) at 20
o
C. The
rise time is 124 milisecond with the cyclic load pulse of 0.67 Hz. The failure criteria is the
point where total vertical deformation of specimen reaches 9 mm (BS-EN:12697-24, 2004).
The schematic of the indirect tensile fatigue test is illustrated in Figure 89.
The number of specimens tested for each set of material, for instance, LR FS-2, is 10. The
target stress levels are selected to obtain a wide range of fatigue lives. The maximum is at
least ten times greater than the minimum fatigue life (DD-ABF, 1996). Normally, for this
experiment, the target stress levels vary from 100 to 400 kPa. However for Black Rock case
with lower stiffness modulus, the range of stress level is from 50 to 200 kPa.
Before fatigue testing, stiffness under the same stress level that specimen will experience in
the fatigue test is determined by indirect tensile stiffness test (DD-213, 1993).
Conventionally, the stiffness is the average of two stiffness values at two perpendicular
directions of specimen. However, due to the stiffness modulus values of recycled specimen
varying considerably with different measuring directions, stiffness under certain stress level
is measured at the direction that specimen has the lowest stiffness modulus. This is also the
165
direction for the fatigue test due to the assumption that failure will occur initially at the
weakest direction.
Figure 89: Schematic of ITFT test
The initial maximum tensile strains at the centre of specimen are plotted against the
relevant numbers of loading cycles to failure on logarithmic scales. The initial maximum
tensile strain at the centre of specimen is calculated as follows (DD-ABF, 1996):
( )
1000
3 1
max ,
max ,
×
+ ×
=
m
x
x
S
 
 (22)
Where:
max , x

is the maximum tensile stress at the centre of specimen (kPa)
 is Poisson’s ratio (assumed to be 0.35)
m
S is the indirect tensile stiffness modulus at
max , x
 (MPa)
Based on the testing data of 10 specimens for each set of material, linear regression analysis
using Least Squares method is applied to obtain the best-fitted equation for fatigue life. The
166
empirical relationship that is used for regression analysis is expressed as follows (Pell,
1973):
( )
2
1
K
i f
K N  = (23)
Where:
f
N
Number of load applications to failure at particular level of initial strain
i

Initial tensile strain
2 1
, K K
Material Coefficients
The fatigue lives of recycled asphalt mixtures are then compared to each other and to those
of control mixtures to study the effects of different mixing protocols and RAP materials on
fatigue life of hot recycled asphalt mixtures.
7.3 Results and analysis
The parameters of fatigue equations of control mixtures and recycled mixtures
manufactured by different methods are summarized in Table 51. The R square values
indicate that the empirical relationship (Equation 23) between number of loading cycles to
failure and initial maximum strain at the centre of specimen fits the experimental data. As
fatigue failure normally occurs between 30-200 microstrain range (Read, 1996), the
numbers of loading cycles to failure at 100 microstrain of different mixtures are also
extrapolated for comparison purposes (Table 52).
7.3.1 Control mixtures
Figure 90 illustrates the regressed fatigue lines of three control mixtures, BR, CB and CB-
V. The results show that regressed fatigue lines of CB and CB-V mixtures are almost
statistically the same. This is because the rheological properties of recycled bitumen for CB
mixture is almost the same as that of 70/100 Pen binder (Section 6.3.1). However, due to
the fact that 160/220 Pen is considerably softer than fully blended recycled binder and
70/100 Pen bitumen, CB and CB-V have longer fatigue life than that of BR mixture (Table
52). The results confirm the finding of Cooper and Pell (1974) that under stress-controlled
mode, the mixture with the stiffer bitumen will provide longer fatigue life. In addition, the
fatigue life of BR mixture is more stress dependent than that of CB and CB-V mixtures.
The K2 values of CB and CB-V are approximately similar and extremely lower than that of
the BR mixture.
167
7.3.2 Recycled mixtures
Figures 91 and 92 present the regressed fatigue lines of LR and SR recycled mixtures
manufactured by different mixing protocols. The results show that for both LR and SR
mixtures, the fatigue lines are slightly improved once the mixing time is extended.
Although there are not considerable differences between these fatigue lines, the boundaries
with 95% confidence of these regressed lines are extremely different, especially with LR
mixture. Figure 93 shows that the boundary with 95% confidence of fatigue line of LR
recycled mixture is significantly narrower once the mixing time increases from 2 to 8
minutes. This is due to the homogeneous level of recycled asphalt mixture. Actually, the
variation in stiffness values is considerably reduced when the mixing duration is increased
(Section 6.5.2). The results indicate that to increase mixing effort also means to improve the
reliability of fatigue line of recycled asphalt mixture.
As more RAP and virgin binder can be interacted, the stiffness of recycled mixture
generally increases once the mixing time is extended. This results in the substantial increase
of fatigue life. Figure 94 illustrates the correlation between stiffness and fatigue life at 100
microstrain versus mixing times of recycled mixtures composed of large RAP material. The
data clearly shows that when the mixing time increases, while the stiffness increases
linearly, the fatigue life increases exponentially. Even for recycled mixture composed of
small RAP material, while the stiffness increases slightly, there is a noticeable increase in
fatigue life, approximately five times once the mixing time is extended from 2 to 8 minutes
(Figure 95). The phenomena will be more pronounced if the fatigue lives of the recycled
asphalt mixtures are extrapolated at 30 microstrain. However, the confidence with such a
large extrapolation will be certainly low. The data indicates that a small increase in stiffness
value can be associated with a tremendous extension of the fatigue life of recycled asphalt
mixture.
In addition, the data also demonstrates that the fatigue line of recycled mixture is gradually
less dependent on the initial maximum tensile strain or stress level. For recycled mixture
composed of large RAP material, the mean K2 values reduced from -2.081 to -3.899 when
the mixing time increases from 2 to 8 minutes (Table 51). In fact, the dependence of fatigue
life on stress level gradually transforms from that of black rock to complete blending when
the mixing time increases.
168
The size of RAP also significantly affects the number of loading cycles to fatigue failure of
hot recycled mixtures. This is because to produce a hot recycled asphalt mixture with a
certain level of homogeneity, the bigger the size of RAP material, the more mixing effort is
required. In addition, the homogeneity and stiffness of recycled mixture are mutually
correlated. In fact, with the same mixing efforts, recycled mixtures composed of large RAP
generally have lower stiffness than that of small RAP mixtures (Section 6.5.3). The results
from fatigue testing also demonstrate that with the same mixing effort, recycled mixtures
composed of large RAP have much lower fatigue life than that of small RAP mixtures
(Figure 96).
Mixing methods also have considerable effects on fatigue life of recycled mixture as the
mixing mechanisms determine how RAP and virgin materials are blended with each other.
The data from fatigue test is in an agreement with stiffness data. The fatigue lives of SHRP
mixtures are approximately the same as those of mixtures manufactured by FS method with
4 to 6 minutes mixing time (Table 52). In addition, the fatigue lives of recycled asphalt
mixtures are substantially lower than those of the complete blending (CB) case. Even when
small RAP material is used with 8 minute mixing time, the fatigue life of SR FS-8 mixture
is just about half of the CB mixture. This supports the statement in Section 5.5.1 that the
complete blending between RAP and virgin binder assumed in the design would never
occur in the industry.
R K1 K2
Square Min Mean Max Min Mean Max
LR FS-2 0.73 8.28E+05 2.92E+08 1.03E+11 -3.097 -2.081 -1.066
LR FS-4 0.85 1.94E+07 2.35E+09 2.79E+11 -3.275 -2.441 -1.067
LR FS-6 0.95 4.69E+09 1.29E+11 3.47E+12 -3.698 -3.117 -2.536
LR FS-8 0.99 1.20E+12 1.07E+13 6.31E+13 -4.278 -3.899 -3.519
LR SHRP 0.91 2.65E+08 1.71E+10 1.10E+12 -3.468 -2.753 -2.039
SR FS-2 0.96 7.00E+09 1.23E+11 2.18E+12 -3.634 -3.132 -2.631
SR FS-4 0.94 2.79E+09 1.21E+11 5.26E+12 -3.757 -3.101 -2.446
SR FS-6 0.98 6.65E+10 5.85E+11 5.15E+12 -3.694 -3.314 -2.934
SR FS-8 0.98 6.78E+11 6.18E+12 5.65E+13 -4.088 -3.699 -3.311
SR SHRP 0.95 5.81E+10 2.49E+12 1.07E+14 -4.268 -3.619 -2.971
CB-V 0.94 2.58E+11 2.45E+13 2.34E+15 -4.705 -3.907 -3.108
CB 0.97 8.71E+11 2.04E+13 4.75E+14 -4.462 -3.896 -3.331
BR 0.96 3.73E+06 1.94E+07 1.01E+08 -1.892 -1.610 -1.328
Table 51: Parameters of fatigue equation at 95% confidence of control and recycled asphalt mixtures
manufactured by different mixing methods
169
Fatigue life at
100 microstrain
LR FS-2 20038
LR FS-4 30807
LR FS-6 74896
LR FS-8 170696
LR SHRP 53345
SR FS-2 67056
SR FS-4 75659
SR FS-6 137838
SR FS-8 247131
SR SHRP 143872
CB-V 377413
CB 327959
BR 11713
Table 52: Extrapolated fatigue life at 100 microstrain of recycled asphalt mixtures manufactured by
different mixing methods
10
100
1000
10000
1.0E+02 1.0E+03 1.0E+04 1.0E+05
Number of Loading Cycles to Failure
M
i
c
r
o
s
t
r
a
i
n
s
CB-V BR CB
Figure 90: Fatigue lines of control mixtures
170
10
100
1000
10000
1.0E+02 1.0E+03 1.0E+04 1.0E+05
Number of Loading Cycles to Failure
M
i
c
r
o
s
t
r
a
i
n
s
LR FS-2 LR FS-4 LR FS-6 LR FS-8 LR SHRP
Figure 91: Fatigue lines of LR mixtures manufactured by different methods
10
100
1000
10000
1.0E+02 1.0E+03 1.0E+04 1.0E+05
Number of Loading Cycles to Failure
M
i
c
r
o
s
t
r
a
i
n
s
SR FS-2 SR FS-4 SR FS-6 SR FS-8 SR SHRP
Figure 92: Fatigue lines of SR mixtures manufactured by different methods
171
10
100
1000
10000
1.0E+02 1.0E+03 1.0E+04 1.0E+05
Number of Loading Cycles to Failure
M
i
c
r
o
s
t
r
a
i
n
s
LR FS-2 LR FS-8
Figure 93: Fatigue lines with boundaries of 95% confidence interval of LR FS-2 and LR FS-8 mixtures
0.0E+00
2.0E+04
4.0E+04
6.0E+04
8.0E+04
1.0E+05
1.2E+05
1.4E+05
1.6E+05
1.8E+05
2.0E+05
LR FS-2 LR FS-4 LR FS-6 LR FS-8
F
a
t
i
g
u
e

L
i
f
e
1000
1100
1200
1300
1400
1500
1600
1700
1800
1900
2000
S
t
i
f
f
n
e
s
s

(
M
P
a
)
Fatigue
Stif f ness
Figure 94: Fatigue life at 100 microstrain and stiffness versus different mixing time of LR mixtures
172
0.0E+00
2.0E+04
4.0E+04
6.0E+04
8.0E+04
1.0E+05
1.2E+05
1.4E+05
1.6E+05
1.8E+05
2.0E+05
2.2E+05
2.4E+05
2.6E+05
SR FS-2 SR FS-4 SR FS-6 SR FS-8
F
a
t
i
g
u
e

L
i
f
e
1000
1100
1200
1300
1400
1500
1600
1700
1800
1900
2000
S
t
i
f
f
n
e
s
s

(
M
P
a
)
Fatigue
Stif f ness
Figure 95: Fatigue life at 100 microstrain and stiffness versus different mixing time of SR mixtures
0.0E+00
2.0E+04
4.0E+04
6.0E+04
8.0E+04
1.0E+05
1.2E+05
1.4E+05
1.6E+05
1.8E+05
2.0E+05
2.2E+05
2.4E+05
2.6E+05
LR SR
F
a
t
i
g
u
e

L
i
f
e
FS-2
FS-4
FS-6
FS-8
SHRP
Figure 96: Relation between RAP sizes and fatigue life at 100 microstrain of recycled mixtures
manufactured by different mixing methods
173
8 Effects of mixing methods on permanent deformation
of hot recycled asphalt mixtures
8.1 Introduction
This chapter studies the effect of different mixing methods on resistance to permanent
deformation of recycled mixtures. The material preparation is the same as that in Chapter 6.
Only large RAP material is used in this experiment. The resistance to permanent
deformation of recycled mixtures is determined by repeated loading axial test . To eliminate
the effect of air void content, the resistances to permanent deformation of recycled mixtures
are compared based on rutting characteristic after densification stage.
8.2 Materials preparation and testing plan
The materials for determining the resistance to permanent deformation of hot recycled
mixtures are summarized in Table 53. Recycled mixtures are manufactured by two different
methods, field simulation (FS) and the SHRP method. Only large size of RAP is used in
this experiment. The proportion of RAP is also 40%. In FS method, RAP/superheated
virgin aggregate varies from 2 to 8 minutes. There are also two control mixtures presenting
the “Black Rock” and “Complete Blending” cases. The manufactures of theses mixtures are
similar to the procedures in Section 6.3.1.
RAP/Virgin aggregate mixing duration (minutes)
LR FS 2 4 6 8
× × × ×
SHRP ×
Table 53: Test plan to study the effects of different mixing methods on resistance to permanent
deformation
The resistance to permanent deformation is determined following the procedure in DD
226:1996. The test is conducted at 40
o
C and under repeated axial dynamic load conditions.
Each load pulse comprises of 1 second for load application and 1 second for the rest period.
The magnitude of stress level is 100 kPa. The number of load applications that the
specimen will experience is 3600. Similar to stiffness test, permanent deformation
determination is implemented 15 days after the day of compaction. After manufacture,
specimens are stored in a cabinet at a temperature of 20
o
C. Before testing, specimen is
174
conditioned at testing temperature for at least 8 hours. In addition, to minimize the friction
between surfaces of specimen and testing plates, surfaces of specimen are coated evenly
and thinly with silicone grease and graphite flakes. Figure 97 illustrates the schematic of
RLAT test to determine the resistance of mixture to permanent deformation.
During the test, accumulated vertical deformations are recorded after each load application.
Permanent deformation is then plotted against number of loading applications that the
specimen experiences. Permanent deformation patterns of recycled mixtures manufactured
by different mixing methods are compared to each other to study the effects of mixing
methods on deformation resistance. The deformations of recycled mixtures are also
compared to those of control asphalt mixtures to study how different the properties of
recycled mixtures are to those assumed in the design process.
Figure 97: Schematic of RLAT test to determine resistance to permanent deformation
175
8.3 Results and discussion
The permanent deformation behavior of asphalt mixture under creep test conditions is
normally divided into three stages (Bernasconi and Piatti, 1978):
 Primary stage: in this stage, the volume of specimen under the load
decreases. This will consequently cause an increase in the density. The
primary stage is also called densification in the other literature (Bahuguna et
al., 2006). During this stage, although the accumulation of vertical strain
increases rapidly, the permanent deformation per each loading cycle
decreases.
 Secondary stage: the secondary stage starts when the permanent deformation
per loading cycle reaches approximately constant value. In this stage, the
volume decreases due to the load will be equal to the volume increases in the
adjacent areas. This stage represents the shear deformation and is considered
to be the primary deformation behavior of asphalt mixture (Sousa et al.,
1991).
 Tertiary stage: the vertical strain per loading cycle increases rapidly again to
failure.
However, under the condition of repeated load axial test with a total of 3600 loading cycles,
the results just represent the densification and a part of the secondary stage. Figure 98 is an
example of the permanent deformation versus the number of load application of LR FS-2
mixture. The data shows that the primary stage, or densification, occurs during the first
2000 cycles. During this stage, accumulated permanent deformation increases rapidly. On
the contrary, during the secondary stage (after 2000
th
load application), the accumulated
vertical deformation has a linear relationship with the number of loading cycles.
The effect of air void content on permanent deformation has been found in previous
literature (Sousa et al., 1991). Air void content plays a very important role in the primary
stage or densification. In this experiment, generally under 3600 loading applications, the
test data shows that deformation caused by densification accounts for a considerable part. If
the permanent deformations at the 3600
th
load application are used for the purpose of
comparison, the result might be tremendously dependent on the air void content. In
addition, the air void contents of testing specimens are always hard to control even in
176
laboratory manufacture. The permanent deformations and air void contents data is
summarized in Table 54.
Sample Air Void Deformation
Number (%) (%)
1 4.12 3.20
2 4.73 3.05
LR FS-2 3 5.18 2.64
4 6.77 5.18
5 6.40 3.53
1 5.38 4.19
2 5.50 3.86
LR FS-4 3 5.54 3.58
4 5.50 3.20
5 6.07 3.19
1 5.38 4.00
2 5.83 3.12
LR FS-6 3 4.40 3.18
4 3.91 3.61
5 4.93 3.91
1 5.46 3.28
2 5.58 3.29
LR FS-8 3 5.42 3.24
4 5.83 2.82
5 6.36 3.00
1 7.58 2.74
2 4.20 2.81
LR SHRP 3 4.89 3.42
4 4.77 4.03
5 5.14 3.83
1 3.26 2.67
2 4.32 2.74
CB 3 4.12 2.29
4 3.30 3.51
5 3.59 3.50
1 5.79 4.34
2 6.24 4.27
BR 3 6.20 4.54
4 5.38 4.77
5 6.93 4.85
Table 54: Permanent deformation data of control and recycled specimens manufactured by different
mixing methods
However, air void content will have insignificant effects on secondary stage as the
deformation in this stage occurs without volume change. Therefore, the coefficient of the
linear relationship between permanent deformation and number of load application in
secondary stage can be used as a rutting indicator. The higher the value of the rutting
177
indicator, the lower the resistance to permanent deformation. For hot recycled asphalt
mixtures, comparisons between rutting indicators might reveal the effects of mixing efforts
on the resistance to permanent deformation. As all the mixture variables are deliberately
controlled the same, the hypothesis is under mixing effort, rutting indicator might be altered
in relation with the incorporation between RAP and virgin binder. Actually, rutting
indicator will increase unless under mixing effort, integration between RAP and virgin
binder is enhanced and vice versa.
The rutting indicators of control and recycled mixtures manufactured by different mixing
methods are shown in Figure 99 and Table 55. Rutting indicator is the coefficient of the
equation expressing the linear relationship between accumulated permanent deformation
and number of load cycles after 2000
th
load applications. The results indicate that
RAP/superheated virgin aggregate mixing duration significantly affects the resistance to
permanent deformation of recycled mixtures. The longer the mixing time, the less the
susceptibility of recycled mixtures to permanent deformation under repeated loading
conditions. For 2 minutes mixing duration, there is a proportion of the rutting indicators
that are close to that of the BR case. On the contrary, rutting indicators of recycled mixture
mixed for 8 minutes are close to those of the CB case (Figure 99). This is because the
longer the mixing time, the more integration between RAP and virgin binder results in the
increase of mixture stiffness. In addition, as recycled mixture become more homogeneous,
rutting indicator variation also decreases. The coefficient of variation substantially
decreases from 49%to 10% once the mixing time is increased from 2 to 8 minutes.
The data also indicates that mixing method significantly affects the resistance to permanent
deformation of recycled mixtures. Recycled mixture manufactured by the SHRP method
generally has better resistance to permanent deformation than those of FS mixtures (Figure
99). However, even in favourable conditions that never exist in the asphalt production
industry, for instance, 2 hour preheating RAP at 110
o
C of SHRP or 8 minutes mixing time
in field simulation (FS) method, the resistance to permanent deformation of recycled
mixture never reaches that of the complete blending (CB) mixture.
178
0
1
2
3
4
5
6
0 500 1000 1500 2000 2500 3000 3500 4000
Number of Loading Circles
P
e
r
m
a
n
e
n
t

D
e
f
o
r
m
a
t
i
o
n

(
%
)
Figure 98: Permanent deformation versus number of loading application of LR FS-2 specimens
CB LR SHRP LR FS-8 LR FS-6 LR FS-4 LR FS-2 BR
0.0005
0.0004
0.0003
0.0002
0.0001
R
u
t
t
i
n
g
I
n
d
i
c
a
t
o
r
Figure 99: Inter-quartile rutting indicator ranges of control and recycled specimens manufactured by
different mixing methods
179
Rutting Indicator (s
-1
) Mean SD COV (%)
LR FS-2 5.18E-04 2.78E-04 1.78E-04 2.87E-04 1.71E-04 2.86E-04 1.40E-04 49.08
LR FS-4 3.06E-04 3.20E-04 2.93E-04 2.40E-04 2.83E-04 2.88E-04 3.05E-05 10.58
LR FS-6 2.61E-04 2.03E-04 1.61E-04 2.74E-04 2.67E-04 2.33E-04 4.91E-05 21.06
LR FS-8 1.84E-04 2.12E-04 2.20E-04 2.10E-04 2.46E-04 2.15E-04 2.21E-05 10.32
LR SHRP 2.39E-04 2.13E-04 2.66E-04 1.62E-04 1.62E-04 2.40E-05 5.48E-06 22.83
CB 2.21E-04 2.13E-04 1.60E-04 1.53E-04 1.40E-04 1.77E-04 3.70E-05 20.83
BR 4.16E-04 3.64E-04 3.79E-04 4.86E-04 4.65E-04 4.22E-04 5.29E-05 12.54
Table 55: Rutting indicator data of control and recycled specimens manufactured by different mixing methods
180
9 Conclusions and recommendations for future research
9.1 Conclusions
The principal conclusions which can be drawn from the literature review include:
 The philosophy of available viscosity mixing equations assumes that aged binder
and virgin binder are completely blended. The question is whether this assumption
actually occurs in the recycled asphalt production process. Otherwise, the
mechanical properties of recycled asphalt mixture would be deviated from expected.
 The result estimated by available viscosity mixing equations is normally not really
close to the actual value. These viscosity mixing equations can provide an
approximate value. Trial experiments with increment proportions of RAP or virgin
binder in the blend should be used to obtain the accurate value.
 The current laboratory mixing methods indicate shortcomings as these methods
could not depict the mixing mechanism that occurs in the asphalt mixing plant.
Laboratory procedure allows RAP to be preheated for a long duration at high
temperature before mixing with virgin material. On the contrary, RAP at ambient
temperature is mixed with superheated virgin aggregate for short time, maximum 90
seconds in the asphalt mixing plant. The extensive RAP preheating time at high
temperature might coincidentally enhance the interaction between RAP and virgin
binder. The question is whether the asphalt mixing plant could generate the recycled
mixture with the same homogeneity as that manufactured in the laboratory.
 RAP material used in laboratory is normally processed to less than ½ or 1 inch.
However, there is a wide range of RAP sizes that have been used in the highway
industry. The sizes of RAP used in practice are sometimes considerably bigger than
that used in the laboratory.
 The diffusion pattern is supposed to be the same for the whole recycled mixture.
The diffusion model assumes that single RAP aggregate particles coated by RAP
binder are covered by virgin binder. Virgin binder starts to diffuse into RAP binder.
The diffusion is long term and affected by many factors, for instance, temperature,
181
bitumen thickness, and chemical composition of bitumen. The diffusion model does
not take into account the effect of RAP lump and existence of virgin aggregate.
The principal conclusions which can be drawn from the experimental work presented
in this thesis include:
 Grunberg and Nissan equation proves to be the most efficient rule for predicting the
viscosity of bitumen blends. Using Grunberg and Nissan equation, if interaction
parameter G
12
is properly determined for each specific bitumen blend, the viscosity
could be predicted within 10% of the actual values. The other viscosity mixing rules
using one universal interaction parameter G
12
generally generate considerably high
residual errors. For instance, the predicted viscosities by Arrhenius equation (ASTM
D4887) are within approximately 30% of the actual values and 50% for DLV
method. Therefore, the interaction parameter G
12
should be determined for each
specific bitumen blend. The fact that one constant value of G
12
is used universally
would result in substantial errors in viscosity estimation.
 There are reciprocal relationships between mixing effort, homogeneity, and
mechanical properties of recycled mixtures. The mutual relationships between
mixing effort and homogeneity, or homogeneity and mechanical properties, could
not be evaluated in a quantitative manner. On the contrary, the relationship between
mixing effort and mechanical properties could be quantified.
 The laboratory mixing method conventionally used to prepare recycled asphalt
specimens tends to overestimate the mechanical properties of recycled asphalt
mixtures. The long RAP preheating time that never exists in the industry
coincidentally enhances the incorporation between RAP and virgin binder. The long
RAP preheating time also slightly alters the properties of RAP binder. Using this
method, the effect of RAP size is negligible.
 The newly developed laboratory mixing method provides a better means of
describing the mixing mechanism between RAP and virgin material in the industrial
asphalt mixer. The mixing mechanism is expressed as follows:
 Virgin aggregate is superheated to required temperature
182
 RAP material is heated up and softened by the thermal energy transferred
from superheated virgin aggregate. In this step, RAP materials start to
disintegrate and are distributed all over the mixture under mechanical mixing
 RAP/virgin aggregate blend is mixed with virgin binder. In this step, besides
the fact that RAP disintegration and distribution still progress, there is
incorporation or rejuvenation between RAP and virgin binder.
 The newly developed method allows the effect of mixing time on homogeneity level
to be investigated. This method also allows the effects of RAP sizes on quality of
recycled mixture to be studied. In addition, the use of colour binder helps to position
accurately the location of RAP binder, virgin bitumen and aggregate. The surface
analysis of slices from top to bottom clearly depicts 3D distribution of RAP
materials in recycled mixture.
 Measuring stiffness values of a cylindrical specimen in different orientations
indirectly expresses the heterogeneity of recycled mixtures. The variation in
stiffness values at different measured directions will be substantial for
heterogeneous mixtures and relatively minor in the case of recycled mixtures that
are homogeneous.
 The RAP superheated virgin aggregate mixing duration should be longer than a
critical duration in which all the bitumen bonds in RAP lumps are deactivated.
Hence, mechanical mixing will separate RAP lumps and enhance the interaction or
rejuvenation between RAP and virgin binder. In this situation, the more efficient the
mechanical mixing, the higher level of homogeneous recycled asphalt mixtures are
manufactured. Otherwise, the effect of mechanical mixing can only distribute the
RAP material all over the mixture. This situation is exaggerated if large RAP sizes
are used. As an inconsiderable proportion of RAP binder can be rejuvenated by
virgin binder, the recycled asphalt mixture will possess the properties of the Black
Rock mixture.
 The increase in mixing duration significantly improves the homogeneity level of
recycled mixture. The homogeneity level is also substantially affected by the sizes
of RAP material. For the same mixing effort, the mixtures composed of small RAP
183
are generally more homogeneous than those made from large RAP. The more
homogeneous the recycled mixture, the more interaction between RAP and virgin
binder. Therefore, recycled mixtures become stiffer and thus have better resistance
to permanent deformation and fatigue failure. A slightly linear increase in stiffness
can result in an exponential increase in fatigue life of the recycled mixture.
 Although increase of mixing duration has positive effects on the homogeneity, the
complete blending between RAP and virgin binder assumed in the design process
would never exist in the production of recycled asphalt mixtures. Qualitatively, even
at favourable conditions of extremely long mixing time compared to that in the real
asphalt mixing plant, a considerable proportion of RAP still exists as lumps.
Quantitatively, as RAP binder is not completely blended with virgin binder, the
stiffness values, resistance to permanent deformation, and fatigue life of recycled
asphalt mixture are considerably poorer than those of the Complete Blending
mixture. Therefore, the design methodology for recycled asphalt mixture tends to
overestimate the performance of hot recycled asphalt mixture.
9.2 Recommendations for future research
The mechanical mixing characteristic substantially affects the homogeneity and mechanical
properties of hot recycled asphalt mixture. However, the mechanical mixing characteristic
of a real asphalt mixer is quite different from that found in the laboratory. The movement of
material in the laboratory mixer is primarily horizontal. On the contrary, in asphalt mixing
plants, this includes not only horizontal but also vertical movement. The mixing efficiency
of a real industry mixer might be totally different from that of a laboratory mixer.
Therefore, research should be carried to validate the finding of the laboratory work.
The mixing condition in laboratory work is also different from the field. The mixing
process in this research occurs under conditions of no moisture content. However, in the
industry, RAP materials with different moisture contents are mixed with superheated virgin
aggregate. The mixing process between RAP and virgin aggregate might occur under
extremely hot and steamy conditions. This might alter the properties of RAP and virgin
binder. Therefore, different mixing conditions should be considered in future studies. In
addition, considering the whole service life, recycled asphalt pavement has to work in
different climatic conditions, for instance under the effects of frost and moisture. Therefore,
184
low temperature cracking resistance and long term durability of recycled asphalt mixture
should be also considered in future research.
The material used in this research is primarily Dense Bitumen Macadam (DBM). The
gradation of DBM conforms to continuous-graded theory. Therefore, there will be a
substantial proportion of RAP materials existing as agglomerate after crushing. These RAP
lumps will restrain the interaction between RAP and virgin binder. However, the situation
might be different with material using gap-graded aggregate. Future research should take
into account the effect of aggregate grading on the incorporation between RAP and virgin
binder.
As the complete blending between RAP and virgin material assumed in the design process
would never exist in the recycled asphalt production process, a new design method that
involves the partial blending between RAP and virgin binder should be considered in future
research.
Experiments in this thesis were carried out primarily at a macro scale level. However, the
micro scale level has not been taken into account. How virgin binder diffuses into and
recovers the properties of aged binder are not covered by this thesis. Once virgin binder is
in contact with RAP binder, the rejuvenation or diffusion progresses with time. The
rejuvenation might be influenced by many factors, for instance, temperature, or chemical
composition of bitumen binder, the presence of filler or proportion of filler in the filler
mastic. As a result, the effect of diffusion process on mechanical properties of recycled
asphalt mixture should be considered in future research.
185
References:
AIREY, G. (2003) State of the Art Report on Aging Test Methods for Bituminous
Pavement Materials. The International Journal of Pavement Engineering, Vol 4, p:
165-176, September 2003.
ANDERSON, A. A., HIR, Y. M. L., P., P. J. & MARTIN, D. (2002) Zero Shear Viscosity
of Asphalt Binders. Transportation Research Record, Vol 1810, p: 54-62,
Transportation Research Board, National Research Council, Washington D.C.
ANDERSON, T. W. & DARLING, D. A. (1952) Asymptonic Theory of Certain "Goodness
of Fit" Criteria Based on Stochastic Processes. The Annals of Mathematical
Statistics, 23, pp. 193-212.
ASTM-D4887-99 (2003) Standard Practice for Preparation of Viscosity Blends for Hot
Recycled Bituminous Materials. American Society of Testing Materials.
BAHUGUNA, S., PANOSKALTSIS, V. & PAPOULIA, K. D. (2006) Identification and
Modelling of Permanent Deformations of Asphalt Concrete. Journal of Engineering
Mechanics, 123, 231-239.
BELL, C. A., ABWAHAB, Y., CRISTI, M. E. & SOSNOVSKE, D. (1994a) Selections for
Laboratory Aging Procedure for Asphalt-Aggregate Mixture, SHRP A-338. Oregon
State University, Covallis OR 97311, USA.
BELL, C. A., WIEDER, A. J. & FELLIN, M. J. (1994b) Laboratory Aging of Asphalt
Mixtures: Field Validation. Oregon State University, Covallis OR 97311, USA.
BERNASCONI, G. & PIATTI, G. (1978) Creep of Engineering Materials and Structures.
Applied Science Publishers Ltd, London.
BROCK, D. J., MAY, J. D. & RENEGAR, G. (2003) Segregation: Causes and Cures. USA,
ASTEC Inc.
BROCK, D. J. & RICHMOND, J. L. (2005) Milling and Recycling. ASTEC Inc, 4010
Jerome Avenue, Chattanooga, TN 37407, USA.
BROWN, E. R., COLLINS, R. & BROWNFIELD, J. R. (1989) Investigation of
Segregation of Asphalt Mixtures in The State of Georgia. Transportation Research
Record 1217, p: 1-8, Transportation Research Board, National Research Council,
Washington D.C.
186
BROWN, S. F. (2000) Bituminous Pavement: Materials, Design and Evaluation,
Residential Course at The University of Nottingham, Section J, Bituminous
Materials, Stiffness and Fatigue Cracking.
BS-EN1427 (2000) Softening Point of Bitumen - Ring and Ball Method. British Standard
Institution.
BS-EN:933-1 (1997) Test for geometrical properties of aggregates - Part 1: Determination
of particle sizes distribution - Sieving method. British Standard Institution.
BS-EN:1426 (2000) Penetration of Bituminous Materials. British Standard Institution.
BS-EN:3838 (1996) Determination of Density Using Capillary - Stoppered Pyknometers.
British Standard Institution.
BS-EN:12697-4 (2005) BS-EN 12697-4: Bituminous Mixture - Test Method for Hot Mix
Asphalt. Part 4: Bitumen Recovery: Fractionating Column. British Institution.
BS-EN:12697-5 (2002) BS-EN 12697-5: Bituminous Mixture - Test Method for Hot Mix
Asphalt. Part 5: Determination of the maximum density. British Institution.
BS-EN:12697-6 (2003) BS-EN 12697-6: Bituminous Mixture - Test Method for Hot Mix
Asphalt. Part 6: Determination of bulk density of bituminous specimens. British
Institution.
BS-EN:12697-8 (2003) BS-EN 12697-8: Bituminous Mixture - Test Method for Hot Mix
Asphalt. Part 8: Determination of void characteristic of bituminous specimens.
British Institution.
BS-EN:12697-24 (2004) BS EN 12697-24: Bituminous Mixture-Test Methods for Hot Mix
Asphalt. Part 24: Resistance to fatigue. British Institution.
BS-EN:12697-26 (2004) BS-EN:12697-26: Bituminous Mixture Test Methods for Hot Mix
Asphalt. Part 26: Stiffness. British Institution.
BS-EN:12697-31 (2004) BS-EN 12697-31: Bituminous Mixture - Test Method for Hot Mix
Asphalt. Part 31:Specimens prepared by gyratory compactor. British Institution.
BS-EN:12697-33 (2003) BS-EN 12697-33: Bituminous Mixture - Test Method for Hot Mix
Asphalt. Part 33:Specimens Prepared by Roller Compactor. British Institution.
BS-EN:13108-1 (2006) BS EN 13108-1: Bituminous Mixture - Material Specification -
Part 1: Asphalt Concrete. British Standard Institution.
BS-EN:13108-4 (2006) Bituminous mixtures. Material Specifications - Hot Rolled Asphalt.
British Standard Institution.
BS-EN:13108-5 (2006) Bituminous mixtures. Material Specifications - Stone Mastic
Asphalt. British Standard Institution.
187
BS-EN:13108-6 (2006) Bituminous mixtures. Material Specifications - Mastic Asphalt.
British Standard Institution.
BS-EN:13108-7 (2006) Bituminous mixtures. Material Specifications - Porous Asphalt.
British Standard Institution.
BS:4987-1 (2005) BS 4987-1 (2005):Coated macadam (asphalt concrete) for roads and
other paved areas. Specification for constituent materials and for mixtures.
BULLIN, J. A., GLOVER, C. J., DAVIDSON, R. R., CHAFFIN, J. M. & LIN, M.-S.
(1995) Development of Superior Asphalt Recycling Agents. USA, University of
Texas A&M and Department of Chemical Engineering, Texas Transportation
Institute.
CARPENTER, S. H. & WOLOSICK, J. R. (1980) Modifier Influence in the
Characterization of Hot-Mix Recycled Material. Transportation Research Record
777, p: 15-22, Transportation Research Board, National Research Council,
Washington D.C.
CHAFFIN, J. M., DAVIDSON, R. R., GLOVER, C. J. & A., B. J. (1995) Viscosity Mixing
Rules for Asphalt Recycling. Transportation Research Record 1501, p:75-87,
Transportation Research Board, National Research Council, Washington D.C.
CHAFFIN, J. M., LIU, M., DAVIDSON, R. R., GLOVER, C. J. & BULLIN, J. A. (1997)
Supercritical Fractions as Asphalt Recycling Agents and Preliminary Aging Studies
on Recycled Asphalts. Journal of Industrial & Engineering Chemistry Research,
Vol 36, p: 656-666.
COOPER, K. E. & PELL, P. S. (1974) The Effects of Mix Variables on the Fatigue
Strength of Bituminous Materials. Report 633. Crowthorne, Berkshire, TRRL
Laboratory.
CROSS, M. M. (1965) Rheology of Non-Newtonian Fluids: A New Flow Equation for
Pseudoplastic Systems. Journal of Colloid Science, 20, 417-437.
CUTNELL, J. D. & JOHNSON, K. W. (2004) Physics, Wiley, John & Sons, Incorporated,
USA.
DANIEL, J. S., KIM, Y. R. & LEE, H.-J. (1998) Effects of Aging on Viscoelastic
Properties of Asphalt-Aggregate Mixtures. Transportation Research Record, Vol
1630, p: 21-27, Transportation Research Board, National Research Council,
Washington D.C.
188
DAVIDSON, D. D., CANESSA, W. & ESCOBAR, S. J. (1977) Practical Aspects of
Reconstituting Deteriorated Bituminous Pavements. Special Technical Publication,
No 662, p: 17-43 American Society of Testing Materials, USA.
DD-213 (1993) Method for determination of the indirect tensile stiffness modulus of
bituminous mixtures. British Standard Institution.
DD-ABF (1996) Method for the determination of the fatigue characteristics of bituminous
mixtures using indirect tensile fatigue. British Standard Institution.
DUNNING, R. L. & MENDENHALL, R. L. (1978) Design of Recycled Asphalt
Pavements and Selection of Modifiers. Special Techincal Publication, No 625, p:
35-46, American Society of Testing Materials, USA.
EPPS, J. A., LITTLE, D. N., HOLMGREEN, R. J. & TERREL, R. L. (1980) Guidelines for
Recycling Pavement Materials. NHCRP 224. National Highway Cooperative
Research Program, USA.
GARDINER, M. S. & BROWN, E. R. (2000) Segregation in Hot-Mix Asphalt Pavements.
Washington D.C. US, National Cooperative Highway Research Program, National
Academy Press.
GRUNBERG, L. & NISSAN, A. H. (1949) Mixing Law for Viscosity. Nature (London),
164, p: 799-800.
HARNBY, N., EDWARDS, M. F. & NIENOW, A. W. (2001) Mixing In The Process
Industries, Great Britain, Reed Education and Professional Publishing Ltd.
HEUKELOM, W. (1973) An Improved Method of Characterizing Asphaltic Bitumens with
the Aid of Their Mechanical Properties. Proc Association of Asphalt Paving
Technologist, Vol 42, pp 62-98.
HUANG, B., LI, G., VUKOSAVLJEVIC, D., SHU, X. & EGAN, B. K. (2005) Laboratory
Investigation of Mixing Hot-Mix Asphalt with Reclaimed Asphalt Pavement.
Transportation Research Record, Vol 1929, Transportation Research Board,
National Research Council, Washington D.C.
HUGHES, C. S. (1978) Problems and Solutions for A Hot Mix Recycling Plant.
Proceedings, Association of Asphalt Paving Technologist, Vol 30, P:644-664.
HUGO, F. & KENNEDY, T. W. (1985) Surface Cracking of Asphalt Mixture in Southern
Africa. Proceeding of Asphalt Association of Paving Technologist, Vol 54, p: 454-
501. AAPT, USA.
IRVING, J. B. (1977) Viscosities of Binary Liquid Mixtures: The Effectiveness of Mixture
Equations. National Engineering Laboratory, UK Department of Industry.
189
JOHN, D. R. (1993) Understanding How the Origin and Composition of Paving-Grade
Asphalt Cement Affect Their Performance. Strategic Highway Research Program,
The University of Texas Austin, Texas, USA.
KANDHAL, P. S. & CHAKRABORTY, S. (1996) Evaluation of Voids in the Mineral
Aggregate for HMA Paving Mixture. National Centre for Asphalt Technology,
Auburn University, Alabama, USA.
KANDHAL, P. S. & KOEHLER, W. C. (1984) Significant Studies in Asphalt Durability:
Pennsylvania Experience. Transportation Research Record 999, p:41-49,
Transportation Research Board, National Research Council, Washington D.C.
KANDHAL, P. S. & MALLICK, R. B. (1997) Pavement Recycling Guidelines for State
and Local Government. National Centre for Asphalt Technology, Auburn
University, AL 36849, USA.
KARLSSON, R. & ISACSSON, U. (2003a) Application of FTIR-ATR to Characterization
of Bitumen Rejuvenator Diffusion. Journal of Materials in Civil Engineering, Vol
15, No2, ASCE, U.S.
KARLSSON, R. & ISACSSON, U. (2003b) Investigations on Bitumen Rejuvenator
Diffusion and Structural Stability. Proceeding. Lenxington, Kentucky, Asphalt
Association Pavement Technologist.
KARLSSON, R. & ISACSSON, U. (2003c) Laboratory Studies of Diffusion in Bitumen
Using Marker. Journal of Materials Sciences, Vol 38, No 13, pp: 2835-2844, 38, p:
2835-2844.
KARLSSON, R. & ISACSSON, U. (2003d) Material-Related Aspects of Asphalt
Recycling - State-of-the-Art. Journal of Materials in Civil Engineering, ASCE, U.S.,
Vol 18, No 1, February 25.
KLIEWER, J. E., ZENG, H. & VINSON, T. S. (1996) Aging and Low-Temperature
Cracking of Asphalt Concrete Mixture. Journal of Cold Regions Engineering, Vol
10, No3, p: 134-164, September 1996, ASCE.
LEE, T.-C., TERREL, R. L. & MAHONEY, J. P. (1983) Test for Efficiency of Mixing of
Recycled Asphalt Paving Mixtures. Transportation Research Record, National
Research Council, Washington D.C. Vol 911, pp: 51-59.
MALKIN, A. Y. & ISAYEV, A. I. (2006) Rheology: Concepts, Methods, and Application,
Toronto, Chemtec Publishing.
MAPH-2 (2000) Hot mix asphalt paving handbook, Washington DC, United States of
America, Department of Transportation, Federal Aviation Administration.
190
MCDANIEL, R. S. & ANDERSON, R. M. (2001) Recommended Use of Reclaim Asphalt
Pavement in the Superpave Mix Design Method: Technician's Manual. National
Cooperative Highway Research Program, Transportation Research Board, National
Research Council, USA.
MCDANIEL, R. S., SOLEYMANI, H., ANDERSON, R. M., TURNER, P. & PETERSON,
R. (2000) Recommended Use of Reclaim Asphalt Pavement in the Superpave Mix
Design Method. National Cooperative Highway Research Program, Transportation
Research Board, National Research Council, USA.
MORTAZAVI, M. & MOULTHROP, J. S. (1993) The SHRP Materials Reference Library.
SHRP-A-646. Washington DC, SHRP, National Research Council.
NAPA (2004) Asphalt Pavement is the Surprise Leader in the Recycling of Various
Materials. http://www.hotmix.org/recycling.php.
NOURELDIN, A. S. (1995) Oxidation Of Asphalt Binders And Its Effect On Molercular
Sizes Distribution. Special Technical Publication, No 1241, American Society of
Testing Materials, Philadelphia, UAS.
NOURELDIN, A. S. & WOOD, L. E. (1987) Rejuvenator Diffusion in Binder Film for
Hot-Mix Recycled Asphalt Pavement. Transportation Research Record 1115, p: 51-
61, Transportation Research Board, National Research Council, Washington D.C.
PELL, P. S. (1973) Characterisation of Fatigue Behaviour. Symposium on Structural
Design of Asphalt Concrete Pavements to Prevent Fatigue Cracking. Nottingham,
UK.
PELL, P. S. (1988) Bituminous Pavement: Materials, Design and Evaluation, Residential
Course at The University of Nottingham, Section J, Bituminous Materials, Fatigue
Cracking, p: J1-J13.
PERL, M., UZAN, J. & SIDES, A. (1983) Visco-Elasto-Plastic Constitutive Law for a
Bituminous Mixture Under Repeated Loading. Transportation Research Record.
Vol 911, Transportation Research Board, National Research Council, Washington
D.C.
PETERSEN, J. C. (1984) Chemical Composition of Asphalt As Related to Asphalt
Durability: State of The Art. Transportation Research Record 999, p:13-30,
Transportation Research Board, National Research Council, Washington D.C.
PETERSEN, J. C. (1986) Quantitative Functional Group Analysis of Asphalt Using
Differential Infrared Spectrometry and Selective Chemical Reaction - Theory and
191
Application. Transportation Research Record 1096, p:1-11, Transportation Research
Board, National Research Council, Washington D.C.
PLANCHER, H., GREEN, E. L. & PETERSEN, J. C. (1976) Reduction of Oxydative
Hardening of Asphalt by Treatments with Hydrated Lime: A Mechanistic Study.
Proceeding of Asphalt Association of Paving Technologist, Vol 45, p: 1-24. AAPT,
USA.
QIU, H. & BOUSMINA, M. (2003) Molecular Weight Polydispersity Effects on Diffusion
at Polymer/Polymer Interfaces. Canada Research on Polymer Physics and
Nanomaterials, Department of Chemical Engineering, CRASP (Polytechnique,
Montreal), Laval University, Sainte-Foy, QC G1K 7P4, Canada.
READ, J. & WHITEOAK, D. (2003) Shell Bitumen Handbook, Thomas Telford.
READ, J. M. (1996) Fatigue Cracking of Bituminous Paving Mixtures - PhD Thesis.
Nottingham Centre for Pavement Engineering, School of Civil Engineering. The
University of Nottingham.
ROBERTS, F. L., KANDHAL, P. S., BROWN, E. R., LEE, D.-Y. & KENNEDY, T. W.
(1991) Hot Mix Asphalt Materials, Mixture Design and Construction, Lanham,
Maryland, US, NAPA Education Foundation.
ROSTLER, F. S. & WHITE, R. M. (1962) Composition and Changes In Composition of
Highway Asphalts, 85-100 Penetration Grade. Proceedings, Association of Asphalt
Paving Technologist, Vol 30, P:359-377. AAPT, USA.
SCHOLZ, T. V. (1995) Durability of Bituminous Paving Mixtures - PhD Thesis.
Nottingham Pavement Engineering Center. Nottingham, School of Civil
Engineering, Nottingham University.
SOUSA, J. B., CRAUS, J. & MONISMITH, C. L. (1991) Summary Report on Permanent
Deformation in Asphalt Concrete. Strategic Highway Research Program, National
Research Council, Washington, D.C. 1991.
STEPHENS, J. E., MAHONEY, J. & DIPPOLD, C. (2001) Determination of the PG Binder
Grade to Use in a RAP mix. University of Connecticut and the Connecticut
Department of Transportation.
STEPHENS, M. A. (1974) EDF Statistics for Goodness of Fit and Some Comparisons.
Journal of American Statistical Association, 69, pp. 730-737.
SYBILSKI, D. (1996) Zero-Shear Viscosity of Bituminous Binder and Its Relation to
Bituminous Mixture's Rutting Resistance. Transportation Research Record 1535,
192
p:15-21, Transportation Research Board, National Research Council, Washington
D.C.
TAYEBALI, A. A., TSAI, B.-W. & MONISMITH, C. L. (1994) Stiffness of Asphalt-
Aggregate Mixes. SHRP A-338, National Research Council, Washington DC, US.
TIA, M., WOOD, L. E. & HANCHER, D. E. (1980) The Effect of Added Virgin Binders
on the Properties of Cold-Mix Recycled Asphalt Paving Mixtures. Proc Association
of Asphalt Paving Technologist, Vol 49, pp: 199-218.
TRAXLER, R. N. (1961) Relation Between Asphalt Composition and Hardening by
Volatilization and Oxidation. Proceedings, Association of Asphalt Paving
Technologist, Vol 30, P:359-377. AAPT, USA.
TRAXLER, R. N. (1963) Durability of Asphalt Cement. Proceedings, Association of
Asphalt Paving Technologist, Vol 32, p: 44-58. AAPT, USA.
UNITED STATES DEPARTMENT OF TRANSPORTATION, F. H. A. (2007) Reclaimed
Asphalt Pavement - User Guidelines.
http://www.tfhrc.gov/hnr20/recycle/waste/rap131.htm.
WHITE, R. M., MITTEN, W. R. & SKOG, J. B. (1970) Fraction Components of Asphalt -
Compatibility and Interchangeability of Fractions Produced From Different Aphalts.
Proceeding of Asphalt Association of Paving Technologist, Vol 39, p: 492-531.
AAPT, USA.
193
Appendix A
Zero shear viscosity extrapolation data
194
Table A-1: Cross model parameters of Mix A (blends of different proportion of RAP binder and
160/220 Pen) at different temperatures
Temperature (
o
C) RAP % ·
 (Pa.s)
o
 (Pa.s)
k m R square
0 8.3E-06 6.9E+04 0.668 0.440 0.998
19.2 7.4E-04 1.5E+05 1.173 0.446 0.993
20 41.9 1.7E-07 3.3E+05 1.516 0.573 0.992
59.8 4.0E-07 6.2E+05 2.036 0.639 0.988
80 4.4E-08 1.5E+06 3.950 0.717 0.993
100 8.3E-06 2.8E+06 5.891 0.821 1.000
0 4.4E-08 2.5E+04 0.601 0.332 0.997
19.2 2.3E-08 5.1E+04 0.762 0.400 0.997
25 41.9 5.8E-09 1.2E+05 1.145 0.444 0.997
59.8 3.7E-12 2.6E+05 1.909 0.470 0.960
80 2.8E-06 5.2E+05 2.158 0.598 0.986
100 3.0E-05 1.1E+06 3.236 0.715 0.992
0 1.4E-07 9.9E+03 0.492 0.302 1.000
19.2 1.7E-08 1.9E+04 0.691 0.326 1.000
30 41.9 1.0E-04 4.4E+04 0.922 0.392 0.995
59.8 2.2E-08 9.4E+04 1.405 0.403 0.999
80 2.5E-08 2.6E+05 2.458 0.440 0.992
100 3.6E-11 4.4E+05 2.564 0.542 0.994
0 1.7E-06 3.3E+03 0.301 0.316 1.000
19.2 9.7E-05 6.9E+03 0.453 0.325 1.000
35 41.9 1.3E-06 1.8E+04 0.904 0.310 1.000
59.8 1.0E-04 3.5E+04 1.080 0.367 0.989
80 1.4E-07 8.9E+04 1.639 0.398 0.997
100 1.4E-07 2.0E+05 2.535 0.435 0.997
0 1.0E-04 1.3E+03 0.135 0.416 0.998
19.2 9.6E-05 2.5E+03 0.224 0.405 0.999
40 41.9 1.0E-04 6.3E+03 0.556 0.328 1.000
59.8 1.8E-06 1.4E+04 0.919 0.322 1.000
80 8.2E-10 3.6E+04 1.439 0.346 0.999
100 3.8E-07 7.7E+04 1.943 0.380 0.997
0 1.0E-04 5.4E+02 0.077 0.490 0.993
19.2 1.0E-04 1.1E+03 0.198 0.358 0.999
45 41.9 9.8E-05 2.4E+03 0.367 0.339 0.999
59.8 2.0E-08 5.1E+03 0.618 0.318 1.000
80 1.0E+04 1.3E+04 1.017 0.334 1.000
100 4.2E-09 3.1E+04 1.653 0.336 1.000
0 1.0E-04 2.6E+02 0.051 0.522 0.992
19.2 1.0E-03 4.7E+02 0.073 0.526 0.997
50 41.9 9.8E-05 1.0E+03 0.207 0.395 0.999
59.8 3.9E-09 2.0E+03 0.390 0.339 0.999
80 8.2E-06 5.4E+03 0.838 0.306 1.000
100 2.1E-07 1.2E+04 1.233 0.320 1.000
0 1.3E-07 1.2E+02 0.051 0.404 0.911
19.2 1.8E-06 2.2E+02 0.052 0.534 0.976
55 41.9 1.0E-04 4.5E+02 0.133 0.430 1.000
59.8 9.8E-05 8.3E+02 0.230 0.392 1.000
80 2.1E-06 2.0E+03 0.368 0.390 1.000
100 3.7E-07 4.4E+03 0.724 0.343 1.000
0 9.9E-05 6.4E+01 0.003 1.402 0.145
19.2 1.0E-04 1.1E+02 0.009 1.072 0.950
60 41.9 5.2E-07 2.1E+02 0.102 0.403 0.992
195
59.8 1.0E-04 3.9E+02 0.133 0.461 0.999
80 6.6E-06 8.5E+02 0.228 0.451 0.993
100 2.6E-07 1.9E+03 0.502 0.385 1.000
0 NA 3.5E+01 NA NA 0.196
19.2 NA 6.1E+01 NA NA 0.622
65 41.9 8.3E-07 1.0E+02 0.039 0.608 0.997
59.8 1.0E-05 1.9E+02 0.070 0.567 0.983
80 1.0E-04 4.2E+02 0.175 0.434 0.995
100 1.7E-07 8.5E+02 0.390 0.337 0.999
0 NA 2.1E+01 NA NA 0.366
19.2 NA 3.4E+01 NA NA 0.175
70 41.9 9.0E-08 5.5E+01 0.024 0.668 0.928
59.8 1.0E-04 9.5E+01 0.019 0.910 0.891
80 1.4E-05 2.1E+02 0.143 0.402 0.998
100 3.6E-05 4.0E+02 0.268 0.357 0.999
0 NA 1.3E+01 NA NA 0.294
19.2 NA 2.0E+01 NA NA 0.125
75 41.9 NA 3.2E+01 NA NA 0.298
59.8 4.8E-08 5.3E+01 0.008 1.124 0.862
80 9.5E-08 1.1E+02 0.078 0.507 0.991
100 5.0E-09 2.1E+02 0.295 0.284 0.999
0 NA 7.9E+00 NA NA 0.446
19.2 NA 1.2E+01 NA NA 0.605
80 41.9 NA 1.8E+01 NA NA 0.778
59.8 1.0E-05 3.1E+01 0.007 1.166 0.722
80 4.7E-09 5.8E+01 0.049 0.524 0.798
100 9.7E-06 1.0E+02 0.133 0.421 0.997
196
Table A-2: Cross model parameters of Mix B (blends of different % RAP binder and 100/150 Pen) at
different temperatures
Temperature (
o
C) RAP % ·
 (Pa.s)
o
 (Pa.s)
k m R square
0 3.5E-06 2.3E+05 1.435 0.537 0.997
19.7 1.1E-08 3.6E+05 1.720 0.585 0.994
20 41.9 9.1E-09 5.9E+05 2.001 0.678 0.994
60.1 1.8E-05 9.4E+05 2.782 0.724 0.995
80.9 2.4E-07 2.1E+06 5.161 0.731 0.993
100 2.0E-08 2.8E+06 5.891 0.821 1.000
0 4.4E-08 8.7E+04 1.136 0.438 0.998
19.7 1.4E-08 1.8E+05 1.967 0.423 0.992
25 41.9 8.0E-09 2.6E+05 1.797 0.525 0.997
60.1 3.1E-06 4.0E+05 2.204 0.571 0.994
80.9 1.1E-08 7.5E+05 2.791 0.620 0.991
100 5.4E-07 1.1E+06 3.235 0.715 0.993
0 4.9E-08 3.1E+04 0.883 0.361 0.999
19.7 1.0E-04 5.8E+04 1.087 0.421 0.997
30 41.9 2.0E-09 1.0E+05 1.546 0.418 0.999
60.1 7.0E-09 1.7E+05 2.038 0.429 0.995
80.9 7.7E-09 3.4E+05 2.664 0.477 0.996
100 3.1E-10 4.4E+05 2.564 0.542 0.994
0 1.0E-04 1.1E+04 0.622 0.341 1.000
19.7 1.5E-07 2.1E+04 0.914 0.348 0.999
35 41.9 1.0E-04 3.7E+04 1.099 0.387 0.992
60.1 1.0E-04 6.0E+04 1.382 0.405 0.995
80.9 2.2E-08 1.2E+05 1.816 0.424 1.000
100 1.4E-07 2.0E+05 2.535 0.435 0.997
0 8.7E-08 3.6E+03 0.339 0.368 1.000
19.7 7.3E-09 7.1E+03 0.555 0.350 1.000
40 41.9 2.2E-06 1.3E+04 0.893 0.329 1.000
60.1 1.6E-08 2.2E+04 1.170 0.338 1.000
80.9 1.0E-04 4.7E+04 1.415 0.388 0.994
100 3.8E-07 7.7E+04 1.943 0.380 0.997
0 9.8E-05 1.4E+03 0.206 0.403 0.999
19.7 3.1E-07 2.7E+03 0.392 0.333 0.999
45 41.9 9.3E-07 4.7E+03 0.556 0.340 1.000
60.1 1.0E-04 7.6E+03 0.649 0.372 1.000
80.9 1.5E-07 1.8E+04 1.125 0.338 1.000
100 4.2E-09 3.1E+04 1.653 0.336 1.000
0 7.0E-06 5.7E+02 0.123 0.454 0.998
19.7 9.8E-05 1.0E+03 0.225 0.383 1.000
50 41.9 1.0E-04 1.9E+03 0.335 0.365 0.999
60.1 1.0E-04 2.9E+03 0.430 0.373 1.000
80.9 9.8E-05 6.3E+03 0.649 0.368 1.000
100 2.1E-07 1.2E+04 1.233 0.320 1.000
0 9.8E-05 2.6E+02 0.060 0.538 0.990
19.7 9.9E-05 4.5E+02 0.128 0.440 0.999
55 41.9 9.9E-05 7.7E+02 0.193 0.419 1.000
60.1 2.5E-06 1.3E+03 0.354 0.344 0.999
80.9 8.1E-07 2.7E+03 0.492 0.352 1.000
anh 100 3.7E-07 4.4E+03 0.724 0.343 1.000
0 9.8E-05 1.2E+02 0.021 0.778 0.994
19.7 1.0E-04 2.1E+02 0.071 0.505 0.994
60 41.9 1.0E-04 3.4E+02 0.065 0.653 0.981
197
60.1 2.6E-08 5.3E+02 0.178 0.415 0.999
80.9 1.0E-06 1.0E+03 0.454 0.589 0.980
100 1.2E-07 1.9E+03 0.502 0.349 1.000
0 1.2E-09 6.6E+01 0.020 0.636 0.862
19.7 2.3E-14 1.0E+02 0.036 0.581 0.968
65 41.9 1.0E-04 1.8E+02 0.085 0.472 0.991
60.1 1.0E-04 2.5E+02 0.096 0.513 0.992
80.9 9.9E-05 5.2E+02 0.205 0.404 0.998
100 1.7E-07 8.5E+02 0.390 0.337 0.999
0 9.5E-05 3.6E+01 0.010 0.670 0.701
19.7 4.2E-06 5.6E+01 0.045 0.489 0.964
70 41.9 6.2E-09 1.0E+02 0.149 0.268 0.966
60.1 2.3E-07 1.3E+02 0.067 0.532 0.974
80.9 1.0E-04 2.5E+02 0.101 0.052 0.997
100 3.6E-05 4.0E+02 0.268 0.357 0.999
0 NA 2.1E+01 NA NA 0.267
19.7 NA 3.0E+01 NA NA 0.804
75 41.9 9.5E-07 4.9E+01 0.008 1.103 0.927
60.1 7.7E-06 7.1E+01 0.049 0.532 0.970
80.9 1.0E-04 1.3E+02 0.067 0.556 0.983
100 2.9E-09 2.1E+02 0.295 0.284 0.998
0 NA 1.3E+01 NA NA 0.182
19.7 NA 1.9E+01 NA NA 0.363
80 41.9 NA 2.9E+01 NA NA 0.316
60.1 7.1E-05 4.0E+01 0.014 0.935 0.870
80.9 1.0E-04 7.0E+01 0.025 0.798 0.987
100 9.7E-05 1.0E+02 0.133 0.421 0.997
198
Table A-3: Mix A – experiment and predicted viscosity (P) using different viscosity mixing equations at different temperatures
Temperature Rap Experiment ASTM Predicted G&N Predicted Epps Predicted DLV Predicted
%
Viscosity
(P)
Residue
(%)
Viscosity
(P)
Residue
(%)
Viscosity
(P)
Residue
(%)
Viscosity
(P)
Residue
(%)
0 6.89E+05 6.89E+05 0.0 6.89E+05 0.0 6.89E+05 0.0 7.15E+05 3.8
19.2 1.54E+06 1.40E+06 8.9 1.53E+06 0.1 1.31E+06 14.8 9.50E+05 38.2
20 41.9 3.33E+06 3.23E+06 2.4 3.74E+06 12.8 2.90E+06 12.4 1.72E+06 48.3
59.8 6.15E+06 6.26E+06 1.8 7.22E+06 17.4 5.62E+06 8.7 3.33E+06 45.9
80 1.55E+07 1.32E+07 14.8 1.45E+07 6.3 1.23E+07 20.8 8.64E+06 44.2
100 2.76E+07 2.76E+07 0.0 2.76E+07 0.0 2.76E+07 0.0 2.76E+07 0.0
0 2.50E+05 2.50E+05 0.0 2.50E+05 0.0 2.50E+05 0.0 2.60E+05 3.8
19.2 5.07E+05 5.13E+05 1.3 5.36E+05 5.7 4.76E+05 6.0 3.46E+05 31.7
25 41.9 1.18E+06 1.20E+06 2.6 1.29E+06 9.8 1.07E+06 8.9 6.31E+05 46.1
59.8 2.63E+06 2.35E+06 10.8 2.51E+06 4.6 2.09E+06 20.8 1.24E+06 53.1
80 5.19E+06 5.01E+06 3.4 5.24E+06 1.0 4.62E+06 10.9 3.26E+06 37.2
100 1.06E+07 1.06E+07 0.0 1.06E+07 0.0 1.06E+07 0.0 1.06E+07 0.0
0 9.91E+04 9.91E+04 0.0 9.91E+04 0.0 9.91E+04 0.0 1.03E+05 3.9
19.2 1.94E+05 2.05E+05 5.6 2.35E+05 20.8 1.89E+05 2.6 1.38E+05 29.1
30 41.9 4.40E+05 4.84E+05 10.6 5.98E+05 36.7 4.25E+05 2.9 2.53E+05 42.3
59.8 9.38E+05 9.54E+05 1.7 1.18E+06 25.3 8.37E+05 10.8 4.98E+05 46.9
80 2.60E+06 2.05E+06 21.1 2.35E+06 9.3 1.88E+06 27.8 1.33E+06 48.9
100 4.37E+06 4.37E+06 0.0 4.37E+06 0.0 4.37E+06 0.0 4.37E+06 0.0
0 3.33E+04 3.33E+04 0.0 3.33E+04 0.0 3.33E+04 0.0 3.47E+04 4.2
19.2 6.91E+04 7.34E+04 6.2 7.10E+04 2.8 6.62E+04 4.1 4.76E+04 31.1
35 41.9 1.77E+05 1.87E+05 6.0 1.78E+05 0.7 1.59E+05 10.1 9.21E+04 47.8
59.8 3.47E+05 3.91E+05 12.8 3.72E+05 7.2 3.31E+05 4.5 1.93E+05 44.4
80 8.87E+05 8.99E+05 1.3 8.69E+05 2.0 8.02E+05 9.5 5.60E+05 36.9
100 2.05E+06 2.05E+06 0.0 2.05E+06 0.0 2.05E+06 0.0 2.05E+06 0.0
0 1.27E+04 1.27E+04 0.0 1.27E+04 0.0 1.27E+04 0.0 1.32E+04 4.2
19.2 2.52E+04 2.79E+04 10.9 2.74E+04 8.7 2.50E+04 0.7 1.81E+04 28.0
40 41.9 6.27E+04 7.10E+04 13.8 6.88E+04 10.3 5.94E+04 4.8 3.50E+04 43.9
59.8 1.37E+05 1.48E+05 8.2 1.44E+05 4.9 1.24E+05 9.6 7.32E+04 46.5
80 3.61E+05 3.40E+05 6.0 3.33E+05 7.9 3.00E+05 16.8 2.12E+05 41.3
199
100 7.72E+05 7.72E+05 0.0 7.72E+05 0.0 7.72E+05 0.0 7.72E+05 0.0
0 5.39E+03 5.39E+03 0.0 5.39E+03 0.0 5.39E+03 0.0 5.62E+03 4.1
19.2 1.10E+04 1.18E+04 6.9 1.09E+04 0.9 1.05E+04 4.9 7.68E+03 30.3
45 41.9 2.44E+04 2.96E+04 22.2 2.63E+04 8.4 2.45E+04 1.1 1.47E+04 39.3
59.8 5.09E+04 6.14E+04 20.6 5.45E+04 7.2 5.07E+04 0.4 3.06E+04 39.9
80 1.32E+05 1.40E+05 5.4 1.29E+05 2.6 1.22E+05 7.5 8.75E+04 33.9
100 3.15E+05 3.15E+05 0.0 3.15E+05 0.0 3.15E+05 0.0 3.15E+05 0.0
0 2.56E+03 2.56E+03 0.0 2.56E+03 0.0 2.56E+03 0.0 2.66E+03 3.9
19.2 4.67E+03 5.37E+03 15.0 4.96E+03 6.2 4.79E+03 2.6 3.58E+03 23.4
50 41.9 1.00E+04 1.29E+04 29.6 1.14E+04 14.3 1.07E+04 7.6 6.65E+03 33.3
59.8 1.98E+04 2.58E+04 30.4 2.28E+04 15.2 2.14E+04 8.1 1.33E+04 32.8
80 5.43E+04 5.63E+04 3.7 5.18E+04 4.6 4.95E+04 8.8 3.61E+04 33.5
100 1.22E+05 1.22E+05 0.0 1.22E+05 0.0 1.22E+05 0.0 1.22E+05 0.0
0 1.24E+03 1.24E+03 0.0 1.24E+03 0.0 1.24E+03 0.0 1.29E+03 3.6
19.2 2.21E+03 2.46E+03 11.7 2.20E+03 0.1 2.21E+03 0.3 1.69E+03 23.2
55 41.9 4.52E+03 5.53E+03 22.8 4.65E+03 3.1 4.65E+03 3.2 3.00E+03 33.5
59.8 8.31E+03 1.05E+04 26.1 8.82E+03 6.1 8.79E+03 5.8 5.68E+03 31.6
80 1.96E+04 2.15E+04 9.9 1.92E+04 2.0 1.91E+04 2.5 1.43E+04 27.0
100 4.39E+04 4.39E+04 0.0 4.39E+04 0.0 4.39E+04 0.0 4.39E+04 0.0
0 6.40E+02 6.40E+02 0.0 6.40E+02 0.0 6.40E+02 0.0 6.62E+02 3.4
19.2 1.12E+03 1.22E+03 8.8 1.09E+03 3.2 1.10E+03 2.0 8.58E+02 23.7
60 41.9 2.15E+03 2.63E+03 22.9 2.19E+03 2.3 2.22E+03 3.7 1.47E+03 31.2
59.8 3.87E+03 4.80E+03 23.9 4.00E+03 3.4 4.04E+03 4.4 2.69E+03 30.4
80 8.50E+03 9.48E+03 11.5 8.40E+03 1.2 8.43E+03 0.9 6.44E+03 24.3
100 1.86E+04 1.86E+04 0.0 1.86E+04 0.0 1.86E+04 0.0 1.86E+04 0.0
0 3.47E+02 3.47E+02 0.0 3.47E+02 0.0 3.47E+02 0.0 3.58E+02 3.2
19.2 6.13E+02 6.41E+02 4.6 5.81E+02 5.1 5.78E+02 5.6 4.58E+02 25.2
65 41.9 1.05E+03 1.32E+03 26.8 1.14E+03 8.8 1.12E+03 7.3 7.64E+02 26.8
59.8 1.88E+03 2.34E+03 24.6 2.02E+03 7.2 1.98E+03 5.3 1.36E+03 27.9
80 4.17E+03 4.47E+03 7.2 4.04E+03 3.0 3.98E+03 4.5 3.10E+03 25.7
100 8.46E+03 8.46E+03 0.0 8.46E+03 0.0 8.46E+03 0.0 8.46E+03 0.0
0 2.08E+02 2.08E+02 0.0 2.08E+02 0.0 2.08E+02 0.0 2.14E+02 3.0
19.2 3.40E+02 3.66E+02 7.7 3.34E+02 2.0 3.33E+02 2.2 2.69E+02 21.0
200
70 41.9 5.54E+02 7.15E+02 29.5 6.17E+02 11.8 6.12E+02 10.9 4.31E+02 22.0
59.8 9.53E+02 1.21E+03 27.2 1.05E+03 10.0 1.04E+03 8.7 7.31E+02 23.3
80 2.10E+03 2.20E+03 4.8 1.99E+03 4.9 1.97E+03 5.9 1.57E+03 25.3
100 3.96E+03 3.96E+03 0.0 3.96E+03 0.0 3.96E+03 0.0 3.96E+03 0.0
0 1.28E+02 1.28E+02 0.0 1.28E+02 0.0 1.28E+02 0.0 1.31E+02 2.9
19.2 2.00E+02 2.19E+02 9.8 1.90E+02 4.7 1.99E+02 0.2 1.63E+02 18.3
75 41.9 3.16E+02 4.15E+02 31.7 3.32E+02 5.5 3.55E+02 12.9 2.56E+02 18.8
59.8 5.33E+02 6.86E+02 28.6 5.51E+02 3.4 5.87E+02 10.0 4.24E+02 20.6
80 1.07E+03 1.21E+03 13.5 1.05E+03 1.9 1.09E+03 1.9 8.77E+02 17.8
100 2.13E+03 2.13E+03 0.0 2.13E+03 0.0 2.13E+03 0.0 2.13E+03 0.0
0 7.90E+01 7.90E+01 0.0 7.90E+01 0.0 7.90E+01 0.0 8.11E+01 2.6
19.2 1.17E+02 1.29E+02 10.8 1.19E+02 2.2 1.18E+02 1.6 9.86E+01 15.3
80 41.9 1.83E+02 2.30E+02 26.2 2.03E+02 11.2 2.00E+02 9.6 1.49E+02 18.7
59.8 3.11E+02 3.64E+02 17.0 3.21E+02 3.2 3.16E+02 1.4 2.35E+02 24.5
80 5.82E+02 6.10E+02 4.8 5.61E+02 3.6 5.53E+02 5.0 4.55E+02 21.8
100 1.02E+03 1.02E+03 0.0 1.02E+03 0.0 1.02E+03 0.0 1.02E+03 0.0
201
Table A-4: Mix B – experiment and predicted viscosity (P) using different viscosity mixing equations at different temperatures
Temperature Rap Experiment ASTM Predicted G&N Predicted Epps Predicted DLV Predicted
%
Viscosity
(P)
Residue
(%)
Viscosity
(P)
Residue
(%)
Viscosity
(P)
Residue
(%)
Viscosity
(P)
Residue
(%)
0 2.29E+06 2.29E+06 0.0 2.29E+06 0.0 2.29E+06 0.0 2.35E+06 2.5
19.7 3.64E+06 3.74E+06 2.9 3.99E+06 9.5 3.63E+06 0.2 2.87E+06 21.3
20 41.9 5.89E+06 6.50E+06 10.5 7.16E+06 21.7 6.20E+06 5.4 4.24E+06 27.9
60.1 9.35E+06 1.02E+07 9.4 1.12E+07 20.2 9.76E+06 4.3 6.68E+06 28.6
80.9 2.09E+07 1.72E+07 17.7 1.82E+07 12.5 1.66E+07 20.2 1.30E+07 37.5
100 2.76E+07 2.76E+07 0.0 2.76E+07 0.0 2.76E+07 0.0 2.76E+07 0.0
0 8.67E+05 8.67E+05 0.0 8.67E+05 0.0 8.67E+05 0.0 8.89E+05 2.5
19.7 1.84E+06 1.42E+06 22.7 1.54E+06 16.4 1.37E+06 25.2 1.08E+06 40.9
25 41.9 2.65E+06 2.47E+06 6.6 2.79E+06 5.5 2.35E+06 11.2 1.61E+06 39.2
60.1 4.03E+06 3.90E+06 3.1 4.40E+06 9.2 3.71E+06 7.9 2.54E+06 36.8
80.9 7.46E+06 6.57E+06 11.9 7.10E+06 4.9 6.36E+06 14.8 4.98E+06 33.3
100 1.06E+07 1.06E+07 0.0 1.06E+07 0.0 1.06E+07 0.0 1.06E+07 0.0
0 3.07E+05 3.07E+05 0.0 3.07E+05 0.0 3.07E+05 0.0 3.15E+05 2.7
19.7 5.77E+05 5.18E+05 10.3 6.10E+05 5.7 4.98E+05 13.7 3.89E+05 32.5
30 41.9 1.05E+06 9.34E+05 11.0 1.20E+06 14.6 8.78E+05 16.3 5.92E+05 43.6
60.1 1.72E+06 1.51E+06 12.0 1.94E+06 12.9 1.42E+06 17.2 9.62E+05 44.1
80.9 3.40E+06 2.63E+06 22.7 3.09E+06 9.3 2.53E+06 25.7 1.96E+06 42.4
100 4.37E+06 4.37E+06 0.0 4.37E+06 0.0 4.37E+06 0.0 4.37E+06 0.0
0 1.05E+05 1.05E+05 0.0 1.05E+05 0.0 1.05E+05 0.0 1.08E+05 3.0
19.7 2.14E+05 1.89E+05 11.7 1.92E+05 10.3 1.79E+05 16.2 1.37E+05 35.8
35 41.9 3.65E+05 3.65E+05 0.0 3.74E+05 2.5 3.36E+05 7.8 2.19E+05 39.9
60.1 5.96E+05 6.27E+05 5.2 6.42E+05 7.7 5.78E+05 3.1 3.77E+05 36.7
80.9 1.22E+06 1.16E+06 4.6 1.18E+06 3.1 1.10E+06 9.6 8.36E+05 31.3
100 2.05E+06 2.05E+06 0.0 2.05E+06 0.0 2.05E+06 0.0 2.05E+06 0.0
0 3.64E+04 3.64E+04 0.0 3.64E+04 0.0 3.64E+04 0.0 3.75E+04 3.1
19.7 7.14E+04 6.64E+04 6.9 6.93E+04 2.9 6.25E+04 12.4 4.78E+04 33.0
40 41.9 1.34E+05 1.31E+05 2.2 1.40E+05 4.4 1.19E+05 11.0 7.74E+04 42.2
60.1 2.24E+05 2.28E+05 1.7 2.43E+05 8.5 2.08E+05 7.4 1.35E+05 39.7
80.9 4.68E+05 4.31E+05 7.9 4.49E+05 4.0 4.05E+05 13.4 3.07E+05 34.4
202
100 7.72E+05 7.72E+05 0.0 7.72E+05 0.0 7.72E+05 0.0 7.72E+05 0.0
0 1.39E+04 1.39E+04 0.0 1.39E+04 0.0 1.39E+04 0.0 1.43E+04 3.2
19.7 2.69E+04 2.56E+04 4.5 2.52E+04 6.3 2.40E+04 10.8 1.83E+04 31.7
45 41.9 4.74E+04 5.13E+04 8.1 4.98E+04 5.1 4.61E+04 2.8 3.00E+04 36.8
60.1 7.56E+04 9.05E+04 19.8 8.80E+04 16.5 8.13E+04 7.7 5.31E+04 29.8
80.9 1.81E+05 1.73E+05 4.0 1.70E+05 5.7 1.62E+05 10.5 1.23E+05 32.1
100 3.15E+05 3.15E+05 0.0 3.15E+05 0.0 3.15E+05 0.0 3.15E+05 0.0
0 5.70E+03 5.70E+03 0.0 5.70E+03 0.0 5.70E+03 0.0 5.88E+03 3.1
19.7 1.04E+04 1.04E+04 0.1 9.52E+03 8.7 9.71E+03 6.9 7.50E+03 28.1
50 41.9 1.88E+04 2.06E+04 9.5 1.79E+04 4.7 1.84E+04 2.1 1.22E+04 35.3
60.1 2.90E+04 3.59E+04 24.0 3.13E+04 8.2 3.21E+04 10.9 2.13E+04 26.6
80.9 6.34E+04 6.79E+04 7.2 6.22E+04 1.9 6.31E+04 0.5 4.84E+04 23.7
100 1.22E+05 1.22E+05 0.0 1.22E+05 0.0 1.22E+05 0.0 1.22E+05 0.0
0 2.59E+03 2.59E+03 0.0 2.59E+03 0.0 2.59E+03 0.0 2.67E+03 2.9
19.7 4.45E+03 4.53E+03 1.7 4.50E+03 1.2 4.23E+03 4.9 3.34E+03 25.0
55 41.9 7.72E+03 8.48E+03 9.9 8.42E+03 9.1 7.64E+03 1.0 5.21E+03 32.4
60.1 1.27E+04 1.42E+04 12.1 1.41E+04 11.2 1.28E+04 0.9 8.75E+03 30.9
80.9 2.70E+04 2.56E+04 5.2 2.55E+04 5.6 2.39E+04 11.5 1.87E+04 30.7
100 4.39E+04 4.39E+04 0.0 4.39E+04 0.0 4.39E+04 0.0 4.39E+04 0.0
0 1.24E+03 1.24E+03 0.0 1.24E+03 0.0 1.24E+03 0.0 1.27E+03 2.7
19.7 2.06E+03 2.11E+03 2.8 1.93E+03 6.2 1.98E+03 3.9 1.58E+03 23.2
60 41.9 3.40E+03 3.86E+03 13.4 3.35E+03 1.5 3.47E+03 2.1 2.42E+03 28.8
60.1 5.33E+03 6.31E+03 18.4 5.49E+03 3.1 5.68E+03 6.6 3.97E+03 25.5
80.9 1.02E+04 1.11E+04 8.7 1.01E+04 0.6 1.03E+04 1.4 8.21E+03 19.5
100 1.86E+04 1.86E+04 0.0 1.86E+04 0.0 1.86E+04 0.0 1.86E+04 0.0
0 6.59E+02 6.59E+02 0.0 6.59E+02 0.0 6.59E+02 0.0 6.76E+02 2.6
19.7 1.01E+03 1.09E+03 7.9 1.04E+03 2.7 1.02E+03 1.1 8.28E+02 18.0
65 41.9 1.78E+03 1.92E+03 8.2 1.78E+03 0.3 1.73E+03 2.3 1.24E+03 30.2
60.1 2.54E+03 3.06E+03 20.2 2.83E+03 11.6 2.76E+03 8.5 1.97E+03 22.3
80.9 5.21E+03 5.20E+03 0.3 4.95E+03 5.0 4.86E+03 6.8 3.91E+03 24.9
100 8.46E+03 8.46E+03 0.0 8.46E+03 0.0 8.46E+03 0.0 8.46E+03 0.0
0 3.63E+02 3.63E+02 0.0 3.63E+02 0.0 3.63E+02 0.0 3.72E+02 2.4
19.7 5.61E+02 5.81E+02 3.7 5.61E+02 0.0 5.46E+02 2.5 4.50E+02 19.8
203
70 41.9 1.03E+03 9.88E+02 3.6 9.35E+02 8.8 8.96E+02 12.6 6.55E+02 36.1
60.1 1.33E+03 1.53E+03 15.2 1.44E+03 9.0 1.38E+03 4.4 1.01E+03 23.5
80.9 2.47E+03 2.51E+03 1.6 2.42E+03 2.0 2.35E+03 4.8 1.93E+03 22.1
100 3.96E+03 3.96E+03 0.0 3.96E+03 0.0 3.96E+03 0.0 3.96E+03 0.0
0 2.11E+02 2.11E+02 0.0 2.11E+02 0.0 2.11E+02 0.0 2.15E+02 2.3
19.7 3.02E+02 3.32E+02 10.0 3.06E+02 1.3 3.12E+02 3.3 2.59E+02 14.2
75 41.9 4.88E+02 5.55E+02 13.7 4.89E+02 0.2 5.02E+02 2.9 3.73E+02 23.6
60.1 7.13E+02 8.45E+02 18.5 7.46E+02 4.6 7.64E+02 7.1 5.69E+02 20.2
80.9 1.29E+03 1.37E+03 5.9 1.26E+03 2.2 1.28E+03 0.9 1.06E+03 18.0
100 2.13E+03 2.13E+03 0.0 2.13E+03 0.0 2.13E+03 0.0 2.13E+03 0.0
0 1.28E+02 1.28E+02 0.0 1.28E+02 0.0 1.28E+02 0.0 1.31E+02 2.1
19.7 1.88E+02 1.92E+02 2.3 1.88E+02 0.2 1.82E+02 3.3 1.54E+02 18.1
80 41.9 2.93E+02 3.05E+02 4.1 2.95E+02 0.8 2.79E+02 4.7 2.13E+02 27.1
60.1 4.02E+02 4.44E+02 10.7 4.31E+02 7.2 4.06E+02 1.2 3.12E+02 22.4
80.9 7.00E+02 6.84E+02 2.3 6.70E+02 4.3 6.45E+02 7.9 5.43E+02 22.4
100 1.02E+03 1.02E+03 0.0 1.02E+03 0.0 1.02E+03 0.0 1.02E+03 0.0
204
Appendix B
ITFT data
205
Table B-1: ITFT data of LR FS-2 mixture at 20
o
C
Sample Width (mm) | (mm) E (MPa) o (kPa) c (µm) N
S1 40 98 1096 100 187 33537
S2 40 98 854 300 720 402
S3 40 98 1849 400 443 622
S4 37 98 1309 250 392 670
S5 39 98 937 200 438 636
S6 42 98 1709 100 120 7054
S7 40 98 899 150 342 1960
S8 41 98 1515 100 135 7436
S9 42 98 1636 250 313 1356
S10 40 98 811 200 506 1018
Table B-2: ITFT data of LR FS-4 mixture at 20
o
C
Sample Width (mm) | (mm) E (MPa) o (kPa) c (µm) N
S1 41 98 1052 100 195 9351
S2 39 98 1165 200 352 1678
S3 40 98 1376 300 447 971
S4 40 98 1523 350 471 610
S5 41 98 1346 150 228 4090
S6 42 98 1301 100 158 7516
S7 41 98 1390 250 369 1579
S8 42 98 1768 300 348 602
S9 43 98 1250 250 410 1075
S10 38 98 1299 200 316 2357
Table B-3: ITFT data of LR FS-6 mixture at 20
o
C
Sample Width (mm) | (mm) E (MPa) o (kPa) c (µm) N
S1 41 98 1178 200 348 1544
S2 42 98 1823 300 337 1166
S3 42 98 1698 350 423 688
S4 41 98 1273 250 403 1080
S5 41 98 1638 200 250 3922
S6 41 98 1360 150 226 5303
S7 40 98 1195 150 257 5128
S8 40 98 1208 250 424 1098
S9 41 98 1832 200 224 7192
S10 40 98 1050 100 195 8748
Table B-4: ITFT data of LR FS-8 mixture at 20
o
C
Sample Width (mm) | (mm) E (MPa) o (kPa) c (µm) N
S1 39 98 1649 150 186 20224
S2 41 98 1526 150 202 9354
S3 41 98 1516 350 473 399
S4 40 98 1691 300 364 946
S5 42 98 1435 250 357 1249
S6 40 98 1403 250 365 1131
S7 41 98 1637 200 250 3978
S8 40 98 1619 300 380 956
S9 40 98 1422 200 288 2714
S10 40 98 1623 350 442 592
206
Table B-5: ITFT data of SR FS-2 mixture at 20
o
C
Sample Width (mm) | (mm) E (MPa) o (kPa) c (µm) N
S1 41 98 1413 200 290 2436
S2 41 98 1605 350 447 523
S3 40 98 1729 350 415 691
S4 40 98 1669 300 368 1046
S5 40 98 1532 250 335 1649
S6 40 98 1667 200 246 2950
S7 40 98 1587 150 194 8990
S8 40 98 1788 250 287 2437
S9 39 98 1665 150 185 10847
S10 40 98 1515 300 406 1261
Table B-6: ITFT data of SR FS-4 mixture at 20
o
C
Sample Width (mm) | (mm) E (MPa) o (kPa) c (µm) N
S1 40 98 1616 250 317 2205
S2 41 98 1506 200 272 3377
S3 40 98 1565 300 393 1282
S4 41 98 1493 350 481 719
S5 41 98 1637 150 188 12325
S6 40 98 1448 150 212 8857
S7 41 98 1507 200 272 1884
S8 42 98 1581 300 389 1255
S9 41 98 1593 250 322 1984
S10 40 98 1606 350 447 565
Table B-7: ITFT data of SR FS-6 mixture at 20
o
C
Sample Width (mm) | (mm) E (MPa) o (kPa) c (µm) N
S1 40 98 1544 300 398 1327
S2 39 98 1751 350 410 1072
S3 39 98 1566 250 327 2629
S4 40 98 1656 200 248 5617
S5 40 98 1681 150 183 18661
S6 39 98 1619 250 317 3471
S7 40 98 1683 300 365 2476
S8 40 98 1605 350 447 953
S9 40 98 1379 150 223 9802
S10 40 98 1705 200 240 7952
Table B-8: ITFT data of SR FS-8 mixture at 20
o
C
Sample Width (mm) | (mm) E (MPa) o (kPa) c (µm) N
S1 38 98 1560 250 329 2445
S2 39 98 1609 200 255 6589
S3 40 98 1743 350 412 1250
S4 42 98 1514 150 203 18133
S5 40 98 1497 150 205 19370
S6 39 98 1704 250 301 4404
S7 41 98 1740 300 353 2437
S8 41 98 1734 300 355 2815
S9 39 98 1727 200 237 9840
S10 40 98 1808 350 397 1549
207
Table B-9: ITFT data of LR SHRP mixture at 20
o
C
Sample Width (mm) | (mm) E (MPa) o (kPa) c (µm) N
S1 40 98 1535 350 467 709
S2 40 98 1528 300 402 1116
S3 40 98 1338 150 230 7945
S4 39 98 1379 200 297 2275
S5 40 98 1366 250 375 2001
S6 40 98 1356 250 378 1788
S7 41 98 1590 300 387 971
S8 41 98 1579 150 195 5966
S9 41 98 1470 350 488 578
S10 40 98 1376 200 298 2782
Table B-10: ITFT data of SR SHRP mixture at 20
o
C
Sample Width (mm) | (mm) E (MPa) o (kPa) c (µm) N
S1 41 98 1450 300 424 1621
S2 41 98 1584 350 453 800
S3 40 98 1675 250 306 5587
S4 42 98 1634 200 251 8259
S5 39 98 1640 150 188 20869
S6 40 98 1539 150 200 22343
S7 40 98 1453 200 282 5940
S8 40 98 1481 250 346 2218
S9 40 98 1385 300 444 1218
S10 39 98 1437 350 499 913
Table B-11: ITFT data of BR mixture at 20
o
C
Sample Width (mm) | (mm) E (MPa) o (kPa) c (µm) N
S1 39 98 744 100 276 1963
S2 39 98 655 250 782 414
S3 40 98 791 50 130 6885
S4 41 98 754 50 136 7062
S5 39 98 764 150 402 1612
S6 37 98 853 100 240 2672
S7 40 98 878 200 467 978
S8 39 98 818 200 501 930
S9 40 98 815 250 629 431
S10 41 98 919 150 335 2510
Table B-12: ITFT data of CB mixture at 20
o
C
Sample Width (mm) | (mm) E (MPa) o (kPa) c (µm) N
S1 39 98 2185 350 328 2622
S2 39 98 1805 300 341 2842
S3 39 98 1918 250 267 9778
S4 39 98 2001 200 205 14064
S5 40 98 1708 150 180 37012
S6 39 98 1941 300 317 3460
S7 38 98 1738 150 177 33720
S8 39 98 1822 200 225 16648
S9 39 98 1850 250 277 6324
S10 38 98 1942 350 369 2024
208
Table B-13: ITFT data of CB-V mixture at 20
o
C
Sample Width (mm) | (mm) E (MPa) o (kPa) c (µm) N
S1 40 98 1840 350 390 2267
S2 39 98 1873 400 438 1201
S3 40 98 1757 350 408 1774
S4 42 98 2025 300 304 4891
S5 40 98 2047 300 300 2936
S6 41 98 2251 250 228 12190
S7 40 98 2306 400 356 2497
S8 40 98 1754 200 234 20100
S9 40 98 2210 250 232 12239
S10 41 98 1984 200 207 28312
209

To Quynh Ngoc

Abstract
The primary work reported in this thesis is concerned mainly with the effects of different mixing methods and RAP materials on homogeneity and mechanical properties of hot recycled asphalt mixtures. The recycled asphalt mixture conforms to the requirement of BS 4987-1 (2005) for dense bitumen macadam size 10 mm (DBM 10 mm). The proportion of RAP in the recycled mixture is 40%. RAP materials are artificially aged and processed in the laboratory to prevent the variability of RAP gradation, bitumen content, and the origin. Laboratory RAP is also used to assure that every single RAP particle is an agglomerate of RAP aggregate and binder. The mixing procedures include Black Rock (BR), Complete Blending (CB), the SHRP procedure, and a newly developed field simulation method (FS). The primary difference between these methods is the mixing mechanism. The BR case implies the situation in which there is completely no interaction between RAP and virgin binder. On the contrary, RAP and virgin binder are fully interacted in the CB case. The mixing procedures for BR and CB cases conform to those for conventional asphalt mixtures. However, the bitumen for BR case is pure virgin bitumen. In addition, the bitumen for CB is the blend between RAP and virgin binder. The RAP/virgin binder proportion is 4/6. In the SHRP method, RAP is preheated at 110oC for two hours before being mixed with virgin aggregate and binder for 2 minutes at 130oC. In the FS method on the contrary, the mixing procedure duplicates what occurs in the asphalt mixing plant. RAP is mixed with superheated virgin aggregate (215oC) for different durations before this combination is blended with virgin bitumen for 2 minutes at 130oC. The RAP/superheated virgin aggregate mixing duration starts from short mixing time where RAP still exists at approximately original size and gradually increases until the change in RAP lump size is insignificant. Depending on the size of RAP used, RAP/superheated virgin aggregate mixing duration varies from 1 to 8 minutes. The homogeneity of hot recycled asphalt mixture is examined by using virgin binder with a different colour from that of RAP binder. The colour of virgin binder is obtained by mixing clear binder (Shell Mexphalt C 160/220 Pen) with iron oxide pigment. The proportion of pigment is 10% by weight of the binder making this binder red. The use of virgin binder

i

with different colour from that of RAP binder helps to clearly differentiate the locations of RAP and virgin materials. Surfaces of slices cut from compacted recycled specimens are photographed by digital camera. The analysis of these surfaces in vertical order allows the locations of RAP material to be qualitatively identified in a 3D manner. Stiffness modulus values of samples for homogeneity assessment are also determined by indirect tensile stiffness test. The stiffness test is carried out in four directions along the circumference of each specimen with 45o angular increments. The experimental results show that the stiffness measurement in four directions can indicate the heterogeneity of recycled mixture. The variation in stiffness values in different measured directions will be substantial for heterogeneous mixtures and minor in the case where recycled mixtures are homogeneous. The results indicate there are mutual relations between mixing effort, homogeneity, and stiffness values of recycled asphalt mixtures. The longer mixing time will enhance the homogeneity and reduce the variation in stiffness values of recycled mixture. In addition, as more RAP and virgin binder are incorporated, the stiffness values of recycled mixture generally increase once the mixing time is extended. As the clear binder is dyed red by 10% by weight of iron oxide, the proportion of the pigment certainly alters the flow characteristic of binder. This might affect the mixing process and rejuvenating effect between virgin and aged binder. Therefore, the effects of mixing methods and RAP sizes on mechanical performance of hot recycled asphalt mixtures are further investigated using normal straight run bitumen 160/220 Pen as virgin binder. The assessment indicators include stiffness modulus, resistance to fatigue damage, and resistance to permanent deformation. The experimental results indicate that the conventional laboratory mixing method (SHRP) tends to overestimate the mechanical properties of recycled asphalt mixture. The long RAP preheating time that never exists in the industry coincidentally enhances the reaction between RAP and virgin binder. The long RAP preheating time also slightly alters the properties of RAP binder. For the FS method, the increase in mixing duration significantly improves the homogeneity level of recycled mixtures. The homogeneity level is also substantially affected by the size of RAP material. For the same mixing effort, the mixtures comprised of small RAP are ii

This implies that RAP does not act as Black Rock.generally more homogeneous than those made from larger RAP. A slightly linear increase in stiffness can result in an exponential increase in fatigue life of the recycled mixture. resistance to fatigue damage. and resistance to permanent deformation of hot recycled asphalt mixtures are not similar to those of the BR or CB mixtures. asphalt. In addition. mixtures. Keywords: Reclaimed asphalt pavement (RAP). homogeneity iii . mechanical properties. recycled mixtures become stiffer and have better resistance to permanent deformation and fatigue failure. even at the favourable condition where RAP is preheated for 2 hours at 110oC in the SHRP method and 8 minutes mixing duration in the FS method. Therefore. The more homogeneous the mixture. The mechanical properties including stiffness modulus. the assumption that RAP and virgin binder are fully blended also never exists in the recycled asphalt production process. recycled. the more interaction between RAP and virgin binder.

Finally. Dene limestone quarry in terms of free aggregate supply. encouragement. Acknowledgements must also be made to the Nottingham Transportation Engineering Centre for providing necessary laboratory equipment. This project would never have been completed without his guidance and support. Professor Gordon Airey. researchers. and supervision. patience. Dr James Grenfell. The financial support by Vietnamese government is gratefully acknowledged. and Professor Stephen Brown in Nottingham Transportation Engineering Centre for their expert advice and support. for their unconditional. Perdita. mom. and Shell Bitumen for free of charge bitumen during the course of the research. for understanding and caring. I owe my deepest gratitude to my supervisor. I am particularly grateful to my family. technicians. and especially my wife. especially those involved with the Pavement Research Group. Dr Salah Zoorob. Mr Giles Harvey and his wife. Department of Civil Engineering.Acknowledgements This research project has been sponsored by the Government of Vietnam. Their enthusiasm and encouragement have made my research enjoyable and memorable. iv . for his invaluable advice. and administrative staff in the Nottingham Transportation Engineering Centre. I also would like to acknowledge the assistance and friendship of all current and past lecturers. dad. and Faculty of Engineering. Special thanks also go to my host family. Thanks also are due to Professor Cees van der Eijk in the Methods and Data Institute for the discussion and advice on data analysis. and continued support throughout the research period. I would like to show my sincere thanks to Dr Nicholas Thom. my brother.

Declaration The work described in this thesis was carried out at the Nottingham Transportation Engineering Centre. No part of this thesis has been. Department of Civil Engineering. The thesis is the result of my own work. submitted for any degree. diploma. The University of Nottingham between April 2006 and April 2009. or other qualifications. Viet Hung Nguyen July 2009 v . except the work from others that has been specifically referenced. or currently is.

..................................8 2...............................11 2........................1 Asphalt mixture.....................4 Hot recycled mixture production ...............................................2..............................................................................................................68 3.......................................................................................................1.......................................................................................................................70 vi ..............7 2..........................................2 Mechanism of ageing in asphalt mixture ...............................................................68 3..6 Segregation and consequences...ix List of Figures ................3........................................................................................................3 Procedure for RAP manufacture...........................................62 2.2 Methods for recycling bituminous pavements....................................23 2................................................4 Processing RAP ...........2 Diffusion process or chemical mixing .............................................................................................1.2.........................................25 2.................2...................................................................................iv Declaration.5 2 Literature review...........6................................................................3............................................3..............62 2............................4 1................................................................................5...............4 Research methodology...................47 2..7 2................................................................................xiv 1 Introduction...............4 Consequences of ageing in bituminous mixture ..........11 2......................................17 2.................3.1 Introduction....................................................................................................................................................2 Durability of asphalt mixture ..................................................................................................5 Scope of work ..............2 Aggregate.............................................................................................................................................3 Selection of rejuvenators...........................................................3 RAP sizes used for production of recycled mixture .................................................................................................................................................................................1 Mechanical mixing.............................................31 2..................................i Acknowledgements..xiv Glossary ..........................................21 2.....................2 Problem statement.......................................22 2......................................................................................1 1................................................................................................Table of Contents Abstract ......................................................4..................................2 Materials ......................1 Objectives and design procedure of pavement recycling .....................................................................................................6..................................................................................................3 Properties of asphalt mixture ...............................................11 2...................................7 2.................................................................2 Drum facility (Drum mixer)...............................................................4 Estimation of the consistency of the aged bitumen – modifier blend............38 2.......................70 3.........................................................3 Design methodology for recycling of bituminous pavements ........................................................................................................1 Definition of durability ....2................4 1...........68 3...............1 Bitumen.....19 2...............4..2 1........................................1 Background .......................2...................................................................5....................1 Batch plant ...........2 Consequences of segregation on the performance of asphalt mixture.........vi List of Tables ...xi Glossary .......................................................................14 2..38 2....2....1 1.....................37 2...........5 Mixing mechanism........1 Definition of asphalt mixture ..........................65 3 Laboratory RAP production.................5 Laboratory tests simulating field ageing........................................................................................1..................31 2........4..............68 3................34 2.............68 3.....................................v Table of Contents..................21 2.....................................1 Segregation ...............................................3 Research objectives.............3 Factors affecting ageing mechanism.........2 Classification of asphalt mixture .....................2...............................7 2....................

............................5 Results and analysis ...............................................88 5....................................122 6..................................................................................................77 4....4.............................................................................3 Effects of RAP sizes on stiffness ....................................119 6..............................74 4..........................5.............................6 Effects of mixing methods on RAP binder properties ........................5 Determine RAP properties...................................................................................1 Material preparation and mixing procedure...................................................................................................................6 Summary ................4......................................................................................................................4....1 Air void contents.......................................................115 6...................................5............................................92 5.72 3.........................................3......................154 6.............86 5...................2 Mechanical assessment .....................2 Development of laboratory mixing protocol.............................................................................................................................1 Estimation of superheated temperature of virgin aggregate ............77 4..160 6..........................................................................2...............5...................................5 Results and analysis .......162 vii ...74 4..............................1 Zero shear viscosity (ZSV) ................................76 4.......116 6...................................2 Determine the superheated virgin aggregate/RAP mixing duration ..............................3 Compacting procedure ...............4 Assessment method.......................5 Effect of mixing equipment on stiffness......................115 6.........2 Rheological testing......3........90 5.......................................................1 Introduction.2 Efficiency of viscosity mixing equations.........................................................................................3 Effect of mixing protocols on RAP binder properties ......................4 Machining specimens for segregation assessment.6 Summary ...............2........................................3..........122 6............................1 Visual assessment for segregation ...............................2................5.................................................1 Effects of mixing protocols and RAP sizes on stiffness of hot recycled asphalt mixture...........................................................................3 Machining and storage of specimens.................................................72 4 Zero shear viscosity and the accuracy of viscosity mixing equations ..1 RAP aggregate ..5.........2......150 6....................................115 6.........................................72 3...............1 Materials ....................................................1 Materials ....................5...........................4...1 Introduction.....................................3.........................145 6.......................................2 Effect of mixing equipment on stiffness distribution of hot recycled asphalt mixtures ....................................5.............90 5.5............................................4 Specimens preparation ........86 5........93 5....................................................115 6..3............................................119 6...................2 RAP binder..............1 Introduction............................................................................93 5....................3.................2 Mixing procedure...........122 6..92 5.......112 6 Effects of mixing procedures and RAP materials on stiffness distribution of hot recycled asphalt mixtures.......................................2 Effects of mixing time on stiffness ...................................................................................................3 Results and discussion ........................3 Materials and specimens manufacture .77 4..............................................2 Experiment design ...........................2 Compaction .................................80 5 Effects of laboratory mixing methods on the homogeneity of hot recycled asphalt mixture .............................................2.............................................86 5........76 4.124 6................93 5...................108 5.......................................2................................................................................................................................93 5....3 Method for segregation evaluation ....87 5.............................................127 6...................................................................................................................................................................2 Experiments ......................118 6.................................................5........................................................2.5..........124 6.......92 5................2......................3 Determine the mixing temperature ...........4 Effects of mixing methods on stiffness.........................

...........................................................................3 Results and discussion ...................................................166 7...................................................................................................................................................3......1 Introduction.......................................166 7..................................................................................................................................................193 Appendix B ....................................7 Effects of mixing methods on fatigue life of hot recycled asphalt mixtures .173 8......................................................1 Introduction..180 9....1 Conclusions...................................................................164 7.......183 References:...........2 Recycled mixtures...........175 9 Conclusions and recommendations for future research ..............................164 7........................................................................180 9.....................164 7...................................1 Control mixtures ...........................................................................................................................................................173 8..........185 Appendix A..............................................................................................................3 Results and analysis ...............................................................2 Recommendations for future research .....................2 Materials preparation and testing plan..........................................................................................................................3..............................167 8 Effects of mixing methods on permanent deformation of hot recycled asphalt mixtures ...........................204 viii .................................2 Materials preparation and testing plan.....................173 8..........................

.......................................Regression analysis between experiment and predicted values using different viscosity mixing equations at different temperatures..24 Table 3: Ageing index (AI) after TFOT and PAV ageing of recycled blend with different rejuvenators (Chaffin et al.............................70 Table 24: Gradation of processed RAP materials. 1987) .......... 1995)..............................................37 Table 7: State DOT specification requirements for the use of reclaimed asphalt pavement (RAP) in hot asphalt paving mixtures (United States Department of Transportation.........55 Table 16: Properties of RAP aggregate (Huang et al...109 Table 31: Mean stiffness versus RAP/superheated virgin aggregate mixing duration............................. 1995) ..........................................................68 Table 22: Gradation of Dene limestone aggregate ..........................84 Table 28: Mix B ........ 1980) ............. 2005)...................... 1997) ..............................42 Table 10: Properties of RAP binder (Carpenter and Wolosick...........................................69 Table 23: Properties of Dene limestone aggregate ...........................................................................................29 Table 6: RAP preheating temperature required in Drum Mixer (Brock and Richmond............................................ and target air voids of laboratory-simulated segregation mixtures (Gardiner and Brown.......softening agent Grunberg interaction parameter G12 (Chaffin et al...................................................................................................................................112 Table 32: Test plan to study the effects of mixing method on stiffness .....................................66 Table 20: Summary of the influence of segregation on mixture properties (Gardiner and Brown......Regression analysis between experiment and predicted values using different viscosity mixing equations at different temperatures.....76 Table 27: Mix A .....................56 Table 17: Properties of Virgin Aggregate (Huang et al....84 Table 29: G12 parameter of Mix A and B at different temperatures...................... staged-extraction..118 ix ............. 2000).................................................................... virgin aggregate used (Noureldin and Wood......... 1987) .........................85 Table 30: Stiffness versus different RAP/superheated virgin aggregate mixing duration.................. 1987) .....53 Table 13: Test results on reclaimed staged-extraction of RAP (Noureldin and Wood................... staged-extraction......................................................................................................................................................66 Table 21: Properties of bitumen 40/60 Pen............................................................................. 1987)............................. no virgin aggregate (Noureldin and Wood.......... bitumen contents..............38 Table 8: Viscosity at 60oC (Poise) of RAP binder (McDaniel et al..........................73 Table 26: Properties of bitumens ....................................................................... 1986) .............. 2007) .......................25 Table 4: Viscosity at 60oC of aged bitumen and softening agents (Chaffin et al.................................................. 2005) .48 Table 12: Properties of RAP bitumen (Noureldin and Wood..................................................... 2000) .....71 Table 25: Properties of RAP binder......................................................................................................................................... 2000) .............................. 1997)...............................................116 Table 33: Test plan to study the effects of mixer on stiffness of recycled asphalt .......................53 Table 14: Test results on reclaimed................................56 Table 18: Approximate absorption wave length of functional groups (Petersen....................54 Table 15:Tests results on reclaimed............................................13 Table 2: Industrial supercritical fractions and commercial rejuvenating agent properties (Chaffin et al...................................................................................List of Tables Table 1: Mechanism of bitumen aging (Traxler............................................................................................41 Table 9: Critical temperature and performance grades of virgin and recovered RAP binders ...................48 Table 11: Grading of RAP aggregate (Carpenter and Wolosick..........................28 Table 5: Aged bitumen ........................................................................... 2005) . 1963)..................................59 Table 19: Gradations............................................... 1980) .........

...........................................................................................................149 Table 46: Stiffness values of LR FS-4 specimens ..................................................136 Table 45: Stiffness values of SR FS-2 specimens .............................................................179 x ................................................118 Table 35: Properties of bitumen 160/220 Pen and 70/100 Pen.....................................126 Table 38: Summary of stiffness values (MPa) of recycled mixtures manufactured by different mixing methods............................................................................................................134 Table 42: Stiffness values of LR FS-6 specimens ................................................................................................................................119 Table 36: Air void content of recycled mixture manufactured by different mixing method ..........................................................127 Table 39: Stiffness comparison by t-test..........................154 Table 48: Stiffness (MPa) of recycled specimens manufactured by different methods and equipment............................................................................................................................................................................169 Table 53: Test plan to study the effects of different mixing methods on resistance to permanent deformation .....................................................................................155 Table 49: Stiffness values of LR FS-2 specimens – Mixer B...........................149 Table 47: Air void summary of recycled specimens manufactured by different method and equipments ......................135 Table 44: Stiffness values of CB specimens..................................................................................................................128 Table 40: Stiffness values of LR FS-2 specimens ....................................................................................................................................................134 Table 41: Stiffness values of LR FS-4 specimens .........Table 34: Test plan to study the effects of mixing methods on binder properties.157 Table 51: Parameters of fatigue equation at 95% confidence of control and recycled asphalt mixtures manufactured by different mixing methods..124 Table 37: Air void comparison by t-test ........135 Table 43: Stiffness values of LR FS-8 specimens ...........................................................................................157 Table 50: Stiffness values of LR FS-6 specimens – Mixer B.......................................................168 Table 52: Extrapolated fatigue life at 100 microstrain of recycled asphalt mixtures manufactured by different mixing methods....................................................................................................................................173 Table 54: Permanent deformation data of control and recycled specimens manufactured by different mixing methods...............176 Table 55: Rutting indicator data of control and recycled specimens manufactured by different mixing methods........................................................................................

.....................................16 Figure 6: Asphalt film thickness vs viscosity after short term ageing (Kandhal and Chakraborty. 2003)..18 Figure 9: Phase angle versus ageing levels (Daniel et al............................................. 50.... 1996) ... 2001) .....................35 Figure 15: RAP in parallel drum facility with added continuous mixer (Brock and Richmond.....58 Figure 29: Spectra obtained after 10.......................................................................................................... 2000) ................ 2003a) ................15 Figure 5: The effect of void content on the hardening of bitumen on the road (Read and Whiteoak.................................................................................... 1996) .......................................................................................... 1998).. 2005).......59 Figure 30: Schematic of Fick’s Law diffusion model (Karlsson and Isacsson.. 1996) ............47 Figure 22: Effect of diffusion on resilient modulus (Carpenter and Wolosick. 1996) ..26 Figure 12: Batch Plant (MAPH-2.......................................................................................... 2005) ................. 2000)........................................ and 925 minutes during diffusion of rejuvenator into A-B180 at 60oC (Karlsson and Isacsson..................................................................................... 2001) ..............39 Figure 18: Mixture with “self-loving” particles (Harnby et al.............................................................................. 1984)...................50 Figure 24: Diffusion model (Carpenter and Wolosick......57 Figure 27: Viscosity at 135oC of different micro-layers coated RAP particles (Huang et al....................................... 2005) .............................................32 Figure 13: Drum mixer with RAP centre inlet (Brock and Richmond...........................36 Figure 17: Ideal mixture with homogeneity (Harnby et al.................................................................. 2003b) .34 Figure 14: RAP in parallel drum mixer with isolated area (Brock and Richmond. 100............. 500...................................................................46 Figure 20: Arizona RAP binder complex modulus versus preheating time and temperature (McDaniel et al......51 Figure 25: Penetrations of outer and inner layers as function of time (Carpenter and Wolosick................................................. 2005).................................................60 xi ........................49 Figure 23: Schematic of modifier coating an aggregate particle during recycling process (Carpenter and Wolosick...... 1980) .......10 Figure 4: Relationship between the temperature of the mixture and change in softening point (Read and Whiteoak..........17 Figure 8: Effect of ageing time on bitumen viscosity extracted from pavements (Kandhal and Koehler..... 1996)........................ 1980) ......................................................................................35 Figure 16: Double Barrel Drum facility (Brock and Richmond....................................................................................List of Figures Figure 1: Comparison on gradation of gap-graded and continuous-graded asphalt mixture (Read........................... 2001)....................................... 1980) .. 2005) ........................................................... 2003) ...8 Figure 2: Strain response due to applied stress of Visco-Elasto-Plastic Constitutive model (Perl et al......... 2003a) ..........40 Figure 19: RAP preheating time versus indirect tensile and unconfined compression strength (Stephens et al.................................46 Figure 21: Florida RAP binder versus preheating time and temperature (McDaniel et al...............19 Figure 11: Nomograph for predicting 60oC viscosity of recycled asphalt (Davidson et al..............................................17 Figure 7: Asphalt film thickness versus viscosity after long term ageing (Kandhal and Chakraborty............57 Figure 28: Schematic of FTIR – ART (Karlsson and Isacsson.........9 Figure 3: Visco-elastic response to millions of load application (Read................................................................................ 1983)......................................................................................................................................................................................................52 Figure 26: Layers extraction process (Huang et al............................................. 1980)............. 2000) ...............................19 Figure 10:Fracture temperature versus ageing time (Kliewer et al........ 2005)............... 1977) ..........................

............................................................................................................105 Figure 55: Large version of Figure 54 g .79 Figure 40: Complex viscosity versus frequencies of blend (20 % RAP and 80% 160/220 Pen) at 70oC .....................97 Figure 47: LR mixture – 4 minutes mixing time ......................62 Figure 34: Design gradation of RAP aggregate...................................113 Figure 61: Schematic of Mixer A ...............75 Figure 39: Extrapolated ZSV of Mix A blends at temperature of 25oC ..............................107 Figure 57: Stiffness measurement scheme.....102 Figure 52: SR mixture – 4 minutes mixing time..............................................................................................................................................69 Figure 35: Appearance of RAP materials ................................. homogeneity and mechanical properties of recycled asphalt............85 Figure 43: RAP size after 2 minutes RAP/superheated virgin aggregate mixing time ..139 Figure 71: Probability plot of stiffness values – control mixtures..........................................................................................61 Figure 33: Influence of chemical composition of markers on diffusion coefficient (Karlsson and Isacsson.............................120 Figure 64: Coring and cutting scheme for compacted slabs ..........130 Figure 66: Stiffness versus core location and measuring direction of LR FS-4 mixture...................................................103 Figure 53: SR mixture – 6 minutes mixing time.........................................................................................................123 Figure 65: Stiffness versus core location and measuring direction of LR FS-2 mixture...........................................................................................................................................91 Figure 46: LR mixture – 2 minutes mixing time .........................................................................................................................................................................................104 Figure 54: SR mixture – 8 minutes mixing time.......................................................................................140 Figure 72: Probability plot of stiffness values – LR FS mixtures.........................................................................................................71 Figure 36: Gradation of RAP aggregate before and after processing................................................................140 xii .......................................................................... 2003b) ..............................................................................................................................81 Figure 42: Viscosity difference and G12 versus temperatures ................. 2003b) ...........133 Figure 69: Stiffness versus mixing time of LR FS mixtures ...............137 Figure 70: Stiffness distribution of control mixtures.....117 Figure 63: Master-curves of 70/100 Pen and recycled blend with 40% RAP and 60% 160/200 Pen bitumen ...........................................89 Figure 45: Images taken by X-Ray scanner and normal digital camera of Shell Mexphalte C dyed by 10 % iron oxide and RAP binder ..................................................110 Figure 59: Small particles tend to move downward to the bottom during mixing process 111 Figure 60: Relation between mixing effort................98 Figure 48: LR mixture – 6 minutes mixing time ......................117 Figure 62: Schematic of Mixer B................................................................................106 Figure 56: Large version of Figure 54 h ................................72 Figure 37: Complex viscosity of bitumen 160/220 Pen versus different frequency at 60oC 75 Figure 38: Complex viscosity bitumen 100/150 Pen versus different frequency at 60oC ..................................................131 Figure 67: Stiffness versus core location and measuring direction of LR FS-6 mixture..............101 Figure 51: SR mixture – 2 minutes mixing time.................................. 2003b)............................................................................................................................Figure 31: Influence of bitumen type on diffusion coefficient (Karlsson and Isacsson...........132 Figure 68: Stiffness versus core location and measuring direction of LR FS-8 mixture..80 Figure 41: Experiment and predicted viscosity using different viscosity mixing equations of Mix A (Blends of different proportion of aged binder and 160/220 pen) at 20oC...............................................................................................108 Figure 58: Stiffness modulus versus air void content....................................................................................100 Figure 50: SR mixture – 1 minute mixing time ........................................................................99 Figure 49: LR mixture – 8 minutes mixing time ......................89 Figure 44: RAP size after 8 minutes RAP/superheated virgin aggregate mixing time .......................61 Figure 32: Influence of temperature on diffusion coefficient (Karlsson and Isacsson..............

...................159 Figure 87: Complex modulus versus log reduced frequency of RAP binder before and after processed by different mixing methods ......................................................................................................................................................................................................174 Figure 98: Permanent deformation versus number of loading application of LR FS-2 specimens.......................................................................................................................................................178 Figure 99: Inter-quartile rutting indicator ranges of control and recycled specimens manufactured by different mixing methods...141 Figure 74: Stiffness histogram of LR FS-2 mixture – Stiffness grouping analysis .....................................................................................142 Figure 76: Inter-quartile stiffness ranges of LR FS-2 group 1 and BR mixture .148 Figure 82: Inter-quartile stiffness ranges of control and LR mixtures manufactured by different methods ............................................171 Figure 95: Fatigue life at 30 m strain and stiffness versus different mixing time of SR mixtures...146 Figure 80: Stiffness versus core location and measuring direction of SR FS-4 mixture.............158 Figure 86: Stiffness versus core location and measuring direction of LR FS-6 mixture – Mixer B ....................................................................................................................................170 Figure 93: Fatigue lines with boundaries of 95% confidence interval of LR FS-2 and LR FS-8 mixtures........Figure 73: Stiffness histogram of LR FS-2 mixture ..............................................................................................................................................................................................................................................................................................................................143 Figure 77: Inter-quartile stiffness ranges of LR FS mixtures .....................................................................................................172 Figure 96: Relation between RAP size and fatigue life of recycled mixtures manufactured by different mixing methods............................................172 Figure 97: Schematic of RLAT test to determine resistance to permanent deformation...........................144 Figure 79: Stiffness versus core location and measuring direction of SR FS-2 mixture......................178 xiii .......151 Figure 84: Inter -quartile stiffness ranges of control and SR mixtures manufactured by different methods ...................................................................................................143 Figure 78: Inter-quartile stiffness ranges of SR FS mixtures ......................................................................................................................................................142 Figure 75: Stiffness probability plot of LR FS-2 – Stiffness grouping analysis.............153 Figure 85: Stiffness versus core location and measuring direction of LR FS-2 mixture – Mixer B ................................................170 Figure 92: Fatigue lines of SR mixtures manufactured by different methods.........................................................169 Figure 91: Fatigue lines of LR mixtures manufactured by different methods........................................................................165 Figure 90: Fatigue lines of control mixtures...161 Figure 89: Schematic of ITFT test ..........................147 Figure 81: Stiffness range of LR and SR recycled mixtures manufactured by FS methods ................................................................150 Figure 83: Inter-quartile stiffness ranges of SHRP mixtures and FS-2 mixtures ............171 Figure 94: Fatigue life at 30 m strain and stiffness versus different mixing time of LR mixtures.............................161 Figure 88: Phase angle versus log reduced frequency of RAP binder before and after processed by different mixing methods ....

and virgin binder which will be used to recover the properties of RAP binder Segregation: the non-homogeneity of the mixture. the material is crushed. and slag for asphalt concrete production Asphalt mixture: the combination of aggregate and bitumen Batch Plant: or Batch facility. crushed stone. gravel. used as binder in asphalt mixture Black Rock: the situation when RAP is inert in the recycled mixture Counter flow drum mixer: the drum mixer with the direction of flame counter to the movement direction of aggregate Diffusion: is the process when rejuvenator covers. for instance. recycling modifiers.Glossary Actual blending: or “Actual Practice” is the case when RAP is preheated before being blended with virgin aggregate and rejuvenator Aggregate: inert materials. softening agents. aromatic oils. reclaiming. recycling. or the locally high concentration of certain type of material in the whole mixture. Segregation can be divided into size xiv . fluxing oils. incorporates with RAP binder and recovers the properties of RAP binder Double Barrel: a trademark of Astec Industry for a unique counter flow drum mixer Drum mixer: or Drum facility. rejuvenators. virgin aggregate and rejuvenator Rejuvenator: the materials which include modifiers. sand. milled from deteriorated pavement for recycling Recycled mixture: the asphalt mixture that used RAP material. is a combination of drum dryer and mixer for hot asphalt mixture production Mechanical mixing: the effort of using mechanical effect to produce homogeneous mixture of different ingredients Parallel drum mixer: drum mixer with the direction of the flame is similar to that of aggregate movement RAP aggregate: the aggregate extracted from RAP RAP binder: the aged bitumen extracted from RAP RAP: acronym for reclaimed asphalt pavement or recycled asphalt pavement. extender oils. is equipment designed to produce hot asphalt concrete in batch Bitumen: the residual product of fractional distillation process of crude oil. modifying.

and chemical segregation in term of recycling when the complete blending status between RAP binder and rejuvenator has not been reached Total blending: or “complete blending”. is the case when RAP binder is extracted from RAP. The blend of RAP and rejuvenator is then mixed with RAP and virgin aggregate to produce the recycled mixture Virgin binder: bitumen not previously used Virgin aggregate: aggregate not previously used xv .segregation. being mixed with rejuvenator.

In-place or In-situ asphalt recycling The difference of in-place from in-plant asphalt recycling is that all the work is carried out in the field.1 Introduction 1. there has been a wide range of recycling methods regarding the equipment and procedures. The application of this method is favourably used in the climatic regions with temperature below 10oC. or extremely soft bitumen is used as rejuvenator. a new thin asphalt layer is put on top and two layers are compacted together. Karlsson and Isacsson (2003d) summarised the methods for recycling asphalt pavement as follows: In-plant asphalt recycling In this method. for instance. In-place recycling is divided into four sub-categories (Karlsson and Isacsson. In addition. 1980). soft bitumen. 1 . and also commercial recycling agents. There is also a variety of materials used as rejuvenator. If rejuvenator is used. Depending on the processing temperatures. Since that moment. foamed bitumen. The deteriorated pavement is heated and milled before being spread over again on the road surface. The reclaimed material is mixed with new materials before being paved and compacted. The first recorded asphalt pavement recycling project was in 1915 (Epps et al.  Repaving process is almost the same as remixing except no fresh asphalt is added to the reclaimed material. reclaimed asphalt pavement (RAP) is mixed with new materials in the mixing plant.. the reclaimed material is mixed with rejuvenator before repaved. warm recycling (70-120oC) and cold recycling (below 70oC). in-plant recycling is divided into hot recycling (above 120oC). 2003d):  Remixing is a process in which approximately 30-50 mm thick of the deteriorated asphalt surface is milled and scarified.  Cold in-place recycling is a process in which bitumen emulsion.1 Background Recycling pavements has been used for many years as a rehabilitation technique in the highway industry. bitumen fractions.

pavement recycling philosophy has certain advantages compared to conventional pavement construction. The design step will determine the type and proportion of materials in the mixture. which is a combination of aged binder and aggregate. In the design process. 2007). shoulders and embankments. mineral aggregates. These benefits of recycling pavements can be summarized as follows (Kandhal and Mallick. materials and production process. type and amount of rejuvenator are selected based on the viscosity mixing equations. the problems encountered in the recycling asphalt industry include not only those found in conventional asphalt but also issues associated with RAP sizes. In the US. and bitumen binder Preservation of road geometry Reduce the construction time delay Due to the benefits brought to society by pavement recycling techniques. the proportion of RAP binder. 2004). However. The popularity of recycling asphalt pavements has made RAP the most recycled material in the US in terms of both percentage and tonnage. 1997): Reduce costs of new construction and rehabilitation projects Being consistent with environmental sustainable development in terms of conservation of energy. Hence.2 Problem statement The quality of an asphalt mixture is primarily affected by the design method. mixing methods and diffusion mechanisms. roadbeds. president of the National Asphalt Pavement Association. almost all the States allow reuse of RAP in the surface course with percentage of RAP from between 10 and 30% (United States Department of Transportation. In some aspects. The production process assures the mixture has the level of homogeneity as required in the design step by using proper mixing methods such as mixing temperature and duration. The fundamental philosophy of these 2 . 1. Mike Acott. recycled mixture is different from conventional asphalt mixture as the input materials include RAP. reported that approximately 73 million tons of RAP were reused every year (NAPA. Full depth reclamation is a process that allows the reconstruction of the whole pavement structure using existing pavement materials. up to this moment 80% of the asphalt pavements removed each year from widening and resurfacing projects are put back on roads.

The other issue is whether the results of these viscosity mixing equations are reliable. On the contrary. the longer the time for the heat to penetrate and break the RAP materials into separated pieces. existence of RAP materials as agglomerates. up to this moment. Size of RAP is also a problem as the bigger the size of RAP. The question here is whether the short mixing duration in the plant mixer can produce the recycled mixture with the level of homogeneity similar to that of the laboratory procedure product. Although there is a wide range of RAP sizes handled in the asphalt pavement recycling industry. 2007). In the laboratory. mixing temperature. RAP binder is extracted and recovered before being deliberately blended with rejuvenator without any interventions. the other problem is whether the complete blending between RAP binder and rejuvenator assumed in these equations actually occurs in recycled asphalt mixture. In addition. The methods of preparing the recycled mixtures in the laboratory also indicate a shortcoming as these methods could not represent the mixing mechanism in the field. the maximum RAP size is 50 mm. for instance. 3 . there is a lack of research to investigate the effect of RAP sizes on the properties of recycled mixtures. If the complete blending cannot occur in the industry but only in the laboratory. the mixing process between RAP binder and rejuvenator in the asphalt mixing plants is affected by many factors such as the presence of aggregate. especially when two bitumen binders with complicated chemical composition are mixed together. In the laboratory. RAP material is conventionally preheated for a long time before being mixed with virgin aggregate and rejuvenator. the output and input of these viscosity mixing equations rely primarily on just the viscosity values and proportions of each bitumen constituents. and even 75 mm is allowed in asphalt hot recycling process (United States Department of Transportation. the laboratory procedure has overestimated the properties of the recycled mixture. RAP material at ambient temperature is mixed with superheated virgin aggregate in the plant mixer for a really short time. On the contrary. and efficiency of the mixer. Even when the results of these viscosity mixing equations are accurate. filler. The long preheating time might coincidently soften RAP and enhance the mixing between RAP and virgin materials.equations is that RAP binder and rejuvenator binders are completely blended.

complete blending (CB). RAP is mixed with superheated virgin aggregate (215oC) for different durations before this combination is blended with virgin bitumen. To investigate the effects of mixing methods and RAP materials on homogeneity of hot recycled asphalt mixtures.4 Research methodology To study the effects of mixing process on the properties including homogeneity and mechanical performance.1. RAP and virgin binder are fully interacted in CB case. RAP is preheated at 110oC for two hours before being mixed with virgin aggregate and binder. The RAP/superheated virgin aggregate mixing duration starts from short mixing time where RAP still exists at approximately original size and gradually increases until the change in RAP lump size is insignificant. The mixing procedures for BR and CB methods conform to those of conventional asphalt mixtures. 1. The primary difference between these methods is the mixing mechanism. and a newly developed field simulation method (FS). the mixing procedure duplicates what occurs in the asphalt mixing plant. On the contrary. The BR case implies the situation in which there is no interaction between RAP and virgin binder. To investigate the effect of mixing methods and RAP materials on mechanical properties of recycled mixture including stiffness modulus. and resistance to fatigue damage. the SHRP procedure.3 Research objectives The objectives of this research are as following: To better understand the hot recycling asphalt technique. In the SHRP method. The RAP/virgin binder proportion is 4/6. To develop a protocol to prepare the hot recycled mixture in the laboratory that duplicates the production mechanism in the industrial asphalt mixing plant. recycled asphalt mixtures are manufactured by different mixing procedures. On the contrary. To correlate the homogeneity and mechanical properties of hot recycled asphalt mixture. However. resistance to permanent deformation. the binder for CB case is the blend between RAP and virgin binder. the binder for BR case is pure virgin bitumen. In FS method on the contrary. The mixing procedures include black rock (BR). 4 .

resistance to fatigue damage. Therefore.5 Scope of work The scope of work in this thesis includes: Chapter 1: Introduction The content of this chapter briefly demonstrates current problems of asphalt recycling technique that leads to the research objectives. The assessment indicators include stiffness modulus. The virgin binder is obtained by mixing clear binder (Shell Mexphalt C 160/220 Pen) with iron oxide pigment. This is to prevent the variability of RAP aggregate gradation. The proportion of pigment is 10% by weight of the binder making this binder red. This is aimed to correlate the homogeneous level and mechanical properties of recycled hot asphalt mixtures. the use of artificial RAP is used to assure that every single RAP particle is an agglomerate of RAP aggregate and binder. As the clear binder is dyed red by 10% by weight of iron ioxide. two RAP sizes are used to study the effects of RAP materials on properties of hot recycled asphalt mixture. The analysis of these surfaces in vertical order allows qualitative identifying of the locations of RAP material in a 3D manner. the effects of mixing methods and RAP sizes on mechanical performance of hot recycled asphalt mixtures are further investigated using normal straight run bitumen 160/220 Pen as virgin binder. This might affect the mixing process and rejuvenation between virgin and aged binder. 1. and resistance to permanent deformation. Surfaces of slices cut from compacted recycled specimens are photographed by digital camera. In addition. Stiffness modulus values of samples for homogeneity assessment are also determined by indirect tensile stiffness test. The use of virgin binder with a different colour from that of RAP binder helps to clearly differentiate the locations of RAP and virgin materials.In addition. The homogeneous level of hot recycled asphalt mixture is examined by positioning the locations of RAP and virgin materials. binder content and origin. Chapter 2: Literature review This chapter contains up-to-date knowledge on hot asphalt recycling 5 . the proportion of the pigment alters the flow characteristic of the binder. RAP material is artificially aged and processed in the laboratory.

techniques and related issues. The homogeneity of recycled mixtures is studied by using virgin binder with different colour from that of RAP binder. Chapter 7: Effects of mixing methods and RAP materials on fatigue life of recycled asphalt mixtures Chapter 8: Effect of mixing methods on permanent deformation resistance of hot recycled asphalt mixtures Chapter 9: Conclusions and Recommendations 6 . and 8. This might affect the mixing process and rejuvenation between virgin and aged binder.7. The content of this chapter also includes the correlation between homogeneity and stiffness distribution of hot recycled asphalt mixtures. the effects of mixing methods and RAP sizes on mechanical performance of hot recycled asphalt mixtures are further investigated using normal straight run bitumen 160/220 Pen as virgin binder. Chapter 4: Zero shear viscosity and the accuracy of viscosity mixing equations This chapter contains the evaluation of different viscosity mixing equations using zero shear viscosity. Chapter 5: Effect of laboratory mixing methods on homogeneity of recycled asphalt mixtures This chapter presents the effects of different RAP/superheated virgin aggregate mixing durations and RAP sizes on homogeneity of hot recycled asphalt mixture. Chapter 6: Effect of mixing process and RAP materials on stiffness of recycled asphalt mixtures The proportion of the pigment certainly alters the flow characteristic of red virgin binder (Chapter 5). The proportion of pigment is 10% by weight of the binder making this binder red. resistance to fatigue damage and resistance to permanent deformation of hot recycled asphalt mixtures are presented in Chapters 6. The data and analysis on the effect of mixing method and RAP materials on stiffness modulus. The red colour of virgin binder is obtained by mixing clear binder with iron oxide pigment. Chapter 3: Laboratory RAP production This chapter presents the purposes and procedure for laboratory RAP production. Therefore.

Different from gap-graded. the strength of aggregate skeleton is based on the interlock among aggregate particles. to bond aggregate particles together and improve the performance of the mixture (Read and Whiteoak.2 Literature review 2. In gap-graded. Normally. Once asphalt mixture is compacted to required air void content. 1996). and porous asphalt (BS-EN:13108-7. Depending on the required aggregate gradation and binder content. However. The high void content due to single size of coarse aggregate will be filled with sand. Hence. there normally exist many different aggregate size fractions in continuous-graded mixtures with the hypothesis that the smaller particles will fill up the voids generated by the bigger particles (Roberts et al. and pedestrian streets. the particle size distribution is not continuous (Figure 1). 2006). mineral aggregate with different size fractions plays a role as skeleton to provide the strength. will act as an adhesive. 2003). 2006).2 Classification of asphalt mixture Depending on the proportion and particle size distribution. as the required amount of binder content in gap-graded mixture is higher due to using fine aggregate as sand and filler.. hot rolled asphalt (BS-EN:13108-4. there is one single size of coarse aggregate. 1991). 2.1. filler and bitumen. parking areas. Bitumen. This makes the continuous-graded mixture better at deformation resistance than gap-graded one. a residual fraction of fractional distillation process of crude oil. gap-graded and continuous-graded mixtures. stone mastic asphalt (BS-EN:13108-5. The structural strength of gap-graded mixtures is built by the mortar of sand.1 Asphalt mixture 2. 2006). gap-graded mixture is classified into mastic asphalt (BS-EN:13108-6. 7 .1. asphalt mixtures are divided into two main categories. 2006). bitumen and filler. This kind of mixture is normally used for construction of highway pavement layers. this type of mixture has better fatigue resistance (Read.1 Definition of asphalt mixture Asphalt or bituminous mixture is a combination of bitumen and mineral aggregate.

The proportions of each component rely primarily on the temperature and the loading time. On the contrary. As asphalt mixture is a visco-elastic material. the asphalt mixture will behave elastically. 1996) 2. the stiffness of asphalt mixture normally includes elastic and viscous components.Figure 1: Comparison on gradation of gap-graded and continuous-graded asphalt mixture (Read. The stiffness of asphalt mixture can be determined by the following equation: E   (1) Where: E : stiffness modulus  : applied stress 8 . 2003). Under low temperature and short loading time.1. the relation between stress and strain will follow viscous manner under high temperature and long loading time (Read and Whiteoak.3 Properties of asphalt mixture Stiffness modulus Stiffness is the resistance to deformation under applied stress conditions.

the elastic component of the strain will recover instantaneously. This will cause rutting phenomenon in the pavement structure. Once the load is released. : strain caused by applied stress Permanent deformation Permanent deformation is the phenomenon that unrecoverable strain is accumulated after the load is released in each loading cycle. cannot recover (Perl et al. The strain starts increasing when the load is applied. 1983). Figure 2: Strain response due to applied stress of Visco-Elasto-Plastic Constitutive model (Perl et al. the permanent deformation. Figure 2 illustrates the strain response to the applied load. However.. the accumulation will become large after millions of loads (Figure 3). due to plastic characteristic of asphalt mixture. There is also a component called visco-elastic strain which will recover with time. Although this viscous and plastic deformation is really small after each loading cycle.. 1983) 9 .

1996) Fatigue characteristic of asphalt mixture Fatigue can be defined as: “The phenomenon of fracture under repeated or fluctuating stress having a maximum value generally less than the tensile strength of the materials” (Pell. However. the fluctuation of surrounding temperature.Figure 3: Visco-elastic response to millions of load application (Read. It consists of two main phases. The relationship defining the fatigue life of bituminous mixture based on crack initiation is as follows (Brown. 2000). The empirical data shows that the tensile range from 30 to 200 microstrain is the condition that fatigue damage might possibly occur (Brown. tensile stress induced in the pavement is not only due to traffic loading but also the effect of surrounding environment. for instance. 1988). 2000):  1 N f  c   f     m (2) Where 10 . and is caused by tensile strains generated in the pavement by not only traffic loading but also temperature variations and construction practices”. crack initiation and crack propagation. Read (1996) also defined the fatigue as: “Fatigue in bituminous pavement is the phenomenon of cracking.

The polar molecules own unevenly distributed electrical charges and tend to interact with the others. Oxidation is the phenomenon when chemical molecules in bitumen are oxidised by oxygen in the atmosphere and form polar groups containing oxygen.2 Durability of asphalt mixture 2. Therefore. the durability of asphalt pavement is defined as follows (as cited in Scholz. which were: 1.2. 2.2. In order to satisfy the performance demand. some of the causes were just listed but not verified by experimental data. However. Changes in chemical composition as chemical molecules of bitumen react with oxygen (oxidation). carbonyl and carboxylic groups (Read and Whiteoak. Loss of oily components of bitumen by volatility or absorption by mineral aggregate. 3. Depending on 11 . the ageing mechanism of asphalt mixture is understood as that of the bitumen. in the context of a given amount of traffic loading. for instance. hydroxyl. aging and temperature variations. Molecular structuring causing thixotropic effects (steric hardening).1 Definition of durability Asphalt pavement in general has to carry the traffic under certain climatic conditions. Traxler (1963) studied the causes of ageing or hardening phenomenon in asphalt binder and concluded there were 15 factors which might cause ageing in bitumen (Table 1). 2003). m : factors depending on the composition and properties of the mixture 2. 2. a pavement must have the ability to withstand any damage during the whole service life. Petersen (1984) suggested the fundamental factors caused hardening in asphalt materials. 1995): “Durability as it applies to bituminous paving mixtures is defined as the ability of the materials comprising the mixture to resist the effects of water.2 Mechanism of ageing in asphalt mixture As asphalt mixture is a combination of bituminous binder and a skeleton of mineral aggregate.N f : number of applications of load to initiate a fatigue crack  t : initial value of tensile strain c. without significant deterioration for an extended period”.

oily component is also absorbed by porosity when bitumen is in contact with aggregate. and the resin will turn to asphaltene. nitrogen base (N). In addition. resin. 2003). Due to ageing. The increasing proportion of asphaltene plus the fact that the maltene phase necessary to disperse asphaltene is insufficient will increase the viscosity of bitumen. the loss of volatiles is also attributed to long term exposure of asphalt to the environment. Ageing is also due to the loss of volatiles in bitumen. The viscosity of bitumen is mainly attributed to the asphaltene component. Oily proportion of bitumen primarily volatizes due to high temperature. the proportion of nitrogen base and first accidaffin change greatly. The accidaffin group will keep the peptized asphaltene solvated. Bitumen composition was divided into asphaltene (A). Rostler and White (1962) studied the compositional change of 85/100 penetration grade bitumen. and paraffins (P) fractions. 12 . first accidaffin (A1). second accidaffin (A2). first accidaffin changes to nitrogen base. the oil will convert to resin. and nitrogen base turns to asphaltene. Each fraction had its particular function. During the ageing process.the strength of the bond. In addition. 1993). However. the hardening due to these phenomenon is not considerable compared to ageing by oxidation (Read and Whiteoak. depending on the mineralogy of aggregate. these polar molecules will from a network and comprise a wide range of molecular types and sizes (John. and oil. This solution will be gelled by paraffins. Nitrogen base has the function of peptizing inert asphaltene. The resulting proportion of each fraction in bitumen would cause incompatibility (or sysneresis) and substantially affect the durability of bitumen. Noureldin (1995) also studied the ageing effect of bitumen and reported that bitumen is a combination of asphaltene.

Table 1: Mechanism of bitumen aging (Traxler. 1963) 13 .

2. Especially in the condition of above 100oC. and paraffins (P). During ageing process. the mixture was conditioned under infrared light of 7 days at 60oC. asphaltene (A). expressed as ratio N/P. Each blend was subjected to syneresis analysis. Temperature extremely affects the properties of bitumen. and Alberta were fractioned into five basic components. This ratio should be greater than 1 for the bitumen to be free of syneresis. first accidaffin (A1). The aim of filter paper test was to identify the separation of oil phase. second accidaffin (A2). White et al. Read and Whiteoak (2003) stated that the oxidation rate increases twofold for each increment of 10oC. Venezuela. The higher the temperature bitumen is exposed to. Data in Figure 4 illustrates that for 30s mixing time. The test results showed that durable bitumen must have the compositional parameter. The mixture was produced by mixing Ottawa sand and bitumen at 160oC for 6 minutes. The test was carried out by shaking 2 gram pellet in a square bottle. The loss in milligrams after 500 revolutions of the pellet was recorded. 14 .2. Arkansas.5oC in mixing temperature will elevate the softening point by 1oC. Temperature Temperature seriously affects the ageing or hardening rate of bitumen. Durability of each blend was also studied by Pellet abrasion test before and after aging. Bitumen blends were produced by blending different components in various proportions.3 Factors affecting ageing mechanism Chemical composition of bitumen binder White et al. the more bitumen ages. (1970) claimed that compatibility of bitumen relied primarily in the proportion of Nitrogen bases and Paraffins. or syneresis parameter. The bitumen content of the mixture was 2% by weight of the sand. The homogeneity of each blend was studied by microscope. (1970) studied the effects of chemical composition on durability of bitumen. expressed as (N+A1)/(P+A2). nitrogen base (N). California. Four bitumen binders from different origins.4. The pellet was made by compressing a mixture of Ottawa sand and bitumen. above 0. a raise of 5. The compatibility of each blend was also evaluated by filter paper test under ultraviolet light.

2003) 15 . 2003) Air void content in total mixture The air void contents after compaction presents highway engineers with a paradox. Most highway agencies prefer the range of air voids after compaction from 3 to 5% (Roberts et al.Figure 4: Relationship between the temperature of the mixture and change in softening point (Read and Whiteoak. the asphalt binder coating aggregate can be oxidized faster. the high air void content will cause faster ageing speed. The stiff oxidized bitumen will easily rupture. Although low air void content reduces the ageing rate.. Vice versa. if the void content is higher than 9%. if the compaction quality is not well controlled. Figure 5 shows the ageing of pavement made of bitumen that has penetration of 100 dmm at 25oC. 1991). it might increase rutting phenomenon. the penetration of bitumen after five years in service is almost the same as the initial value. if the air void content is increased. If the air void content is less than 5%. the pavement is extremely aged as the penetration reduces from 70 to 20 (Read and Whiteoak. On the contrary. The penetration after mixing is 70 dmm. However. Water can get in and destroy the bond between bitumen film and aggregate and reduce the tensile strength of the mixture.

the less viscosity of bitumen increase after short and long term ageing (Figures 6 and 7). 2003) Bitumen film thickness During the mixing process.Figure 5: The effect of void content on the hardening of bitumen on the road (Read and Whiteoak. The loss of bitumen penetration during mixing is approximately 30% (Read and Whiteoak. Another study by Kandhal and Chakraborty (1996) also showed that the thicker the bitumen film. Bitumen will be extremely aged due to oxidation and loss of volatiles. 16 . 2003). the bitumen is exposed to extremely high temperature in condition of very thin layer approximately 5 to 15µm thick.

1996) Figure 7: Asphalt film thickness versus viscosity after long term ageing (Kandhal and Chakraborty. 1996) 2.2.Figure 6: Asphalt film thickness vs viscosity after short term ageing (Kandhal and Chakraborty.4 Consequences of ageing in bituminous mixture The compositional changes in bitumen due to ageing will result in the increase of bitumen viscosity. Kandhal and Koehler (1984) carried out a study on durability of dense-grade 17 .

The data measured from 1961 to 1976 (Figure 8) shows that the longer time the pavement is exposed to environment. the more prone to cracking. Research by Daniel et al. The experimental results (Figure 10) demonstrated that the increase of ageing time would result in the increase of fracture temperature. the decrease of loss modulus plus the increase of viscosity of bitumen due to ageing will result in the pavement being more prone to cracking at low temperature. Figure 8: Effect of ageing time on bitumen viscosity extracted from pavements (Kandhal and Koehler. 18 . (1996) studied the effect of ageing on fracture temperature. the more aged the pavement. (1998) shows that the longer the ageing time. the lower the phase angle of bitumen (Figure 9). 1984) Ageing will also change the visco-elastic properties of bitumen.pavements using different types and sources of bitumen in Pennsylvania. the higher the pavement viscosity. Kliewer et al. US. In fact. In fact.

5 cm before and after ageing procedure. studied the effect of lime on durability of bitumen by comparing the resilient modulus of cylindrical specimens with diameter of 4 cm and thickness of 2. 19 . conditioning in the oven 150oC for 1 hour.. The results showed that mixtures using lime treated bitumen had better resistance to ageing.. 1998) Figure 10:Fracture temperature versus ageing time (Kliewer et al. Both normal and lime treated bitumen were used in this research.Figure 9: Phase angle versus ageing levels (Daniel et al.6 MPa. 1996) 2. The ageing procedure included: pressing the mixture of sand and bitumen in specified dimension mould at 150oC under pressure of 27. (1976).5 Laboratory tests simulating field ageing Plancher et al.2.

Characteristics before and after ageing are determined by resilient modulus and indirect tensile tests. Von Quintas in 1988 simulated the hardening of bitumen during production. is simulated by conditioning loose mixture at 135 or 163oC during 1. Not only short term. temperature. 6. 2003). although scattered. The ageing effect was obtained by comparing the viscosity of recovered bitumen with that of original bitumen. for instance. air void content. The most comprehensive study on ageing of aggregate mixture is the project A-003A by Bell et al. and production. by conditioning the loose mixture at 135oC in a force draft oven for 8. the ageing during production process has been ignored. long term ageing was also studied by conditioning specimens for 2 days at 60oC followed by 3 days at 107oC (Airey. The aged mixture is then compacted to 4 and 8 percent air void after STOA. 20 . - Hugo and Kenedy (1985) implemented the accelerated ageing procedure in both dry and moist conditions. Both types required the slab with thickness of 4 cm to be conditioned under temperature of 100oC for 4 to 7 days. many factors affecting the ageing mechanism were considered. the affect of different void content and temperature on hardening is not mentioned in this research. These mentioned scholars studied the ageing of bitumen in compacted mixtures. cooling at 25oC for 72 hours then testing the resilient modulus for aged specimen. Short term ageing. (1994a) under the Strategic Highway Research Program (SHRP). The laboratory data. Although it is advantageous to study the ageing in moisturised condition. compacted specimens are preconditioned for 2 days at 40 or 60oC to assure the stability of specimens. generally showed approximately the same ageing level as in the field. and 15 hours (STOA). After ageing. For long term ageing. 24. the compacted specimens are conditioned in force draft oven for 2 or 7 days at constant temperature of 107oC. An open reservoir is put in the oven in order to maintain the relative humidity higher than 80% in moisturised condition. the bitumen was recovered and subjected to Shell plate viscosity testing at different temperatures. However. 16. In this research.- cooling at 25oC for 72 hours then testing the resilient modulus for un-aged specimen. or short term ageing. Before LTOA. and 36 hours. The results from the tests showed that in some mixtures. storing at 150oC for 5 hours. presenting the hardening of bitumen during production process.

the overall objective of recycling pavement is. to restore the properties of existing deteriorated pavements materials. 1980) is to choose proper modifiers to: 21 . modifying.3 Design methodology for recycling of bituminous pavements 2.1 Objectives and design procedure of pavement recycling There are different objectives depending on the organizations in charge or related to the pavement recycling job. to such a level that can satisfy the service requirements. or 9 years in dry-freeze climate. (1994b) concluded that: STOA: is the simulation of bitumen hardening in pavement during production and construction plus less than two year performance. cannot be used without any modifications. and aromatic oils (Karlsson and Isacsson. softening agents. due to the chemical changes during service life of asphalt pavement.3. For instance. recycling modifiers. After this field validation. This wide definition includes reclaiming. The government is also interested due to the saved money from construction hence the problem of constrained budgets for highway maintenance can be solved. The bitumen. fluxing oils. Airey (2003) suggested the condition of 4 days at 85oC to simulate LTOA in pavement for 15 years performance in the condition of wet-no-freeze climate. LTOA (8 days at 85oC): is the simulation of bitumen hardening in pavement for 18 years performance in the condition of wet-no-freeze climate. 2003d). The recommended accelerated ageing procedures were then verified by field practice. rejuvenators..the resilient modulus is 4 times higher than the initial value after STOA and 6 times for LTOA. Bell et al. Hence. or 7 years in dry-freeze climate. the contractors want to recycle the pavements as the costs of construction are cheaper due to the reuse of existing materials (RAP). recycling. 2. the accelerated ageing procedure is recommended as follows: STOA: loose mixture is conditioned at 135oC for 4 hours. LTOA: compacted mixture is conditioned at 85oC for 5 days or 100oC for 2 days. In term of highway engineers. extender oils. it is demonstrated that this condition is equal to LTOA during 5 days at 85oC. However. The purpose of asphalt pavement recycling (Epps et al. in AASHTO PP2 (1994). Any materials that can alter the properties of RAP binder are defined as modifiers.

1977). milled.3. recommended the design procedure for asphalt pavement recycling includes: Determine the properties of existing pavement (RAP) including bitumen content. in-plant and in-situ recycling. The reclaim agent must reduce the viscosity of aged binder to the desired level and improve the durability. there has been a wide variety of pavement recycling techniques. In-situ recycling is also divided into hot in-situ recycling which includes remixing and repaving and cold in- 22 . Depending on the mixing temperatures required of the recycled mixture. The first recycling pavement project recorded was in 1915 (Epps et al. 2. ripped. in-plant recycling is further divided into cold. Analyse and use data for design. and hot recycling (Karlsson and Isacsson. Restore the chemical composition of aged bitumen for durability. hot recycled mixture with 10 to 30% of RAP can have the same performance compared to virgin mixture (Kandhal and Mallick. (Davidson et al. the structural layers which will benefit from recycling as surface. bitumen demand for aggregate.. consistency of bitumen. 1997). warm. and crushed into required sizes before being blended with virgin bitumen and aggregate in the mixing plant. Since that moment. recycling techniques are categorized into two main types. the reclaimed pavements are transported to asphalt production plant. 1980). Regarding in-plant recycling. Different from in-plant recycling. These approaches are normally classified based on the materials to be recycled for instance. and the procedure as well as equipment for recycling purpose. Provide sufficient binder to coat the new aggregate added and satisfy the stability requirement of the mixture.2 Methods for recycling bituminous pavements The concept of recycling or reusing the existing pavement material in new construction or rehabilitation project has lasted for many years. deformation and cracking. aggregate gradation. Select the reclaiming agent. base or sub base.. To obtain the overall objective of recycling bituminous pavements. In addition.- Restore the consistency of aged bitumen. Generally. in-situ methods are processed on site. hot recycling proves to be the most advantageous due to the ability to correct most of the pavement surface defects. Apart from the other in-plant recycling methods. 2003d). bituminous or portland cement pavement.

Dunning and Mendenhall (1978) suggested that the flash point of modifier should be enough to produce the blend with flash point of 205oC. (1970) also concluded that the properties of bitumen were governed predominantly by the composition of the maltene phase. (1977) studied the recycling aspect with a wide range of 6 aged binders and 12 reclaiming agents. White et al. and the syneresis parameter (Section 2.2. A new surface is then laid and compacted on the new recycled base (Kandhal and Mallick. It was suggested that the compositional parameter of reclaiming agent should be in the range of 0.situ recycling or full depth reclamation. a new layer is laid on top of the recycled pavement to increase structural strength of the pavements. there were segregations with all the blends with asphaltene coming from Alberta. four bitumens from different sources. Arkansas.4 to 1. Venezuela.4 to 0. all the surface and a part of base course is milled. and mixed with bitumen emulsion. In addition. The procedure in repaving method is almost the same except no new materials are added. The different asphatene fractions with the same proportion were in turn mixed with different maltene phases. the viscosity at 60oC of modifier should be in range of 90 to 300 cP.3 Selection of rejuvenators Davison et al. preferably 0. and Alberta. In the study by White et al.8. 25. In remixing method. The percentages of aged bitumen for each combination were in turn 0.3. All bitumen binders were fractionated into five basic fractions. scarified. However. California. existing pavement is milled. The molecular weight of asphaltenes 23 . In addition. (1970). The other blends with California asphaltene showed no sign of heterogeneity or syneresis. the effect of asphaltene origin was also investigated by using one reclaiming agent with different asphaltenes. The consistencies as well as the chemical fractions of recycled mixture before and after ageing were compared. In full depth reclamation. 2. 1997). The origin where bitumen came from also affects the durability of recycled materials. 50%. This is actually the ability to peptize asphaltenes by mean of the maltenes phase.3) should be higher than 1 to assure the durability improvement of aged bitumen binders. mixed with new materials before being laid and compacted. 5. When the asphaltene and maltene from California and Alberta were cross blended. foamed bitumen or soft bitumen to produce a stabilized base. were used.

The aged bitumen was artificially produced from SHRP bitumen ABF in an air bubbling apparatus (denoted as ABF-AB1). industrial supercritical fractions (ISCF-A. 1995). The properties of recycling agents are illustrated in (Table 2).6 20. AAF.4 3.0 0. ISCF-C).2 2.7 1.7 68. followed by the blends with 24 . Viscosity Saturate Asphaltene Aromatic (dPa.3 58. YBF is the popular AC-20 and there is no clear origin recorded for this bitumen.4 0.4 0. AAF and ABM in turn originated from West Texas Sour and California Valley (Mortazavi and Moulthrop. ABM) in a supercritical pilot plant at Texas A&M University at a temperature of 221oC and pressure of 49.0 30.8 0. The results (Table 3) indicated that the blends with bitumen fractions had lowest ageing index. The aging index was the ratio between viscosities before and after artificial aging by TFOT and PAV method. (1997) by assessing the potential of some rejuvenators for asphalt pavement recycling.3 Bar (Bullin et al.9 86. The amount of aged bitumen was determined by ASTM 4887 (2003) so the final viscosities of the blends were approximately the same as that value of SHRP AAF-1 (approximately 2000 P at 60oC).0 28.5 ISCF A ISCF B ISCF C CRA A CRA B CRA C 1997) Table 2: Industrial supercritical fractions and commercial rejuvenating agent properties (Chaffin et al. His experimental data indicated that the blends with higher molecular weight asphaltenes had higher viscosity than blends with lower molecular weight asphaltene.6 1. CRA-B.s) (wt%) (wt%) (wt%) 17.7 90.4 8.2 12. 1993). The SHRP bitumens come from different crude sources. commercial recycling agents (CRA-A. Meanwhile.5 71. All the blends as well as AAF-1 were subjected to TFOT and PAV ageing to determine the ageing index. The durability of recycled bitumen was also studied by Chaffin et al. CRA-C) as well as bitumen fractions were appraised. In this study..5 434.0 11. The blends were produced by mixing aged ABF-AB1 with different rejuvenating agents.4 85. Bitumen fractions are fractions F3 extracted from SHRP bitumen (YBF.7 0.3 79. ISCF-B.also affected the viscosity of the blend..

21 N/A 3. The design viscosity of the recycled binder is usually obtained by blending different proportions of aged binder and modifier until the desired viscosity is obtained (Epps et al.53 3.85 2. The SHRP AAF-1 had the highest ageing index. The first step is to draw a line that connects the viscosity values of aged and virgin binder (points 3 and 5)..59 PAV AI 12.68 N/A 1.96 1.67 1. Figure 11 is an illustration of how to use the viscosity blending chart.70 1. The point that this vertical line crosses the X-axis is the proportion of virgin binder.4 Estimation of the consistency of the aged bitumen – modifier blend A crucial phase of the design procedure for recycling asphalt pavement is to predict the viscosity of recycled bitumen binder.30 4.industrial supercritical fractions and commercial rejuvenating agents. 25 . Bitumen is a compound of many different chemical substances.46 5.67 1.85 1. The viscosity blending chart can also be used to determine the viscosity of recycled blend or for the selection of rejuvenators.s) 1890 1900 2140 2080 1840 1850 1900 2000 2090 1670 TOFT AI 2.93 SHRP AAF-1 AAF-AB1/ISCF A AAF-AB1/ISCF B AAF-AB1/ISCF C AAF-AB1/CRA A AAF-AB1/CRA B AAF-AB1/CRA C AAF-AB1/YBF F3 AAF-AB1/AAF F3 AAF-AB1/ABM F3 (Chaffin et al. Final step is to draw the vertical line from the point that the first two lines cross each other (point 2). 1997) Table 3: Ageing index (AI) after TFOT and PAV ageing of recycled blend with different rejuvenators 2.. Composition Asphalt/agent N/A 72/28 61/39 43/57 81/19 83/17 83/17 61/39 67/33 44/56 Viscosity (dPa.50 1.42 4.80 1. (1977) built a viscosity blending chart to simplify the viscosity estimation process (Figure 11).00 3.89 4. 1980). Davidson et al. To determine the proportion of virgin binder in the blend. and consistency of each is far different from the others. This process is time consuming due to many blending trials and complicated chemical constituents of bitumen. The construction of this nomograph was based on the trial blending data of 6 aged binders and 12 reclaiming agents.3. the next step is to draw a horizontal line at desired viscosity value of the blend on Y-axis.

could not express the interaction between two liquids in the blend. 1995). due to its simplicity. ln  mix   x1 ln 1   x 2 ln  2  Where: (3)  mix : viscosity of the mixture of two binary liquids 1 .. may alter the characteristic of the blend. In these equations. This is due to the complicated chemical composition of bitumen itself and especially in case of recycled binder which is a mixture of at least two bituminous materials coming from different crude oil sources. 2003d). the interaction between two liquids. The 26 . 1977) Viscosity of bitumen blend between aged and virgin binders can also be predicted using viscosity mixing equations. 2 : viscosity of both liquids x1 . However. This equation was then developed to ASTM D4887 (2003).. x 2 : volume percentages of both liquids The Arrhenius equation. whether in chemical or physical form. Arrhenius (1887) suggested the following equation. one of the most popular tools to predict the viscosity of recycled binder (Karlsson and Isacsson.Figure 11: Nomograph for predicting 60oC viscosity of recycled asphalt (Davidson et al. aged and virgin bitumen binders are considered as liquids and the viscosity of the bitumen blend can be estimated approximately by mixing theory for binary liquids (Chaffin et al.

In this research. AC-5 from Shell in Deerpark. The viscosity of binary liquid mixture is expressed as follows: ln  mix   x1 ln 1   x 2 ln  2   x1 x 2 G12 Where: (4) G12 : is a characteristic of the system of two binary liquids or interaction parameter Epps et al. and AAF-1. ABH from SHRP. One advantage of this viscosity mixing rule over Arrhenius equation is the interaction between two liquids in the blend. the viscosities of recycled blends are still different (White et al. softening agents are illustrated in Table 4.. and AAV. (1980) also introduced an equation to estimate the viscosity of a mixture between two liquids. A wide range of softening agents was used. This equation is expressed as follows: ln ln  mix   x1 ln ln 1   x 2 ln ln  2  (5) Actually. and supercritical fractions. Texas. Grunberg and Nissan (1949) also introduced an equation to estimate the viscosity of binary liquids. AAA-1. Viscosities of aged bitumens. the Arrhenius equation is a special case of Grunberg and Nissan viscosity mixing rule where G12 is equal to zero. (1995) carried out a study to verify the effectiveness of different viscosity mixing equations. 1970). three SHRP bitumens. commercial recycling agent. Commercial recycling agents were Sun Hydrolene 125. Witco Cyclogen. Grunberg and Nissan equation was the best equation to estimate the viscosity of binary liquids. Chaffin et al. The consistencies of aged 27 . Low viscosity bitumen included AC-3. ABM-1. The viscosity of binary liquid is based on the log log relationship with the log log viscosity of each individual liquid and the mass percentage in the whole mixture. low viscosity bitumen binders. AAF-1.difference in origin might lead to the fact that although two rejuvenators have the same viscosity. Supercritical fractions were extracted from AC-20 (YBF). and ABM-1 were used to produce aged bitumen by using pressurised oxygen vessel and air bubbled reaction apparatus. for instance. Texas. A study by Irving (1977) concluded that generally. after mixing with aged binder. Exxon Nuso 95 and Mobil Mobisol 120. the predicted viscosity could be less than 30% difference from actual value. Irving (1977) also claimed that if constant parameter G12 was used universally. AAA-1. AC-5 from Diamond Shamrock (DS) in Duma.

Using Grunberg and Nissan equation. The finding of Chaffin et al. 1995) 28 . interaction parameter G12 for each pair of aged bitumen/softening agent was calculated.bitumens and softening agents were measured by rheological apparatus at 60oC and frequency of 1.. Table 4: Viscosity at 60oC of aged bitumen and softening agents (Chaffin et al. (1995) is also in agreement with Irving (1977). Viscosity values estimated by different viscosity mixing equations were then compared to experiment data.6Hz. Aged bitumens were blended with low viscosity bitumen. using a constant value of G12 would result in considerable errors in viscosity estimation. The results of G12 for the whole experiment (Table 5) showed that the interaction parameters G12 of each pair of aged bitumen/softening agent varied considerably. softening agents including commercial agents and supercritical fractions at increments of 20%. In fact.

. 1995) The variation of G12 was attributed to the viscosity difference between softening agents and aged bitumen.softening agent Grunberg interaction parameter G12 (Chaffin et al. the greater the absolute value of interaction parameter G12 . Grunberd and Nissan equation is mathematically transformed as follows: DLV  Where:    G12  2 ln  m / 1   G12  x2    1    ln  /    x 2 ln  2 / 1   ln  2 / 1   2 1     (6) 29 . Using DLV. (1995) introduced the dimensionless log viscosity (DLV). Chaffin et al.Table 5: Aged bitumen . The larger the difference in viscosity between aged bitumen and rejuvenator. To eliminate the viscosity effects.

the viscosity mixing equation by Chaffin et al. both methods have the same philosophy. This will lead to the fact that the method performs well with some types of blend and not with others. In this case. the viscosity of the blend has a second order polynomial relationship with the proportion of aged binder. On the other hand. Although the ways to find the solutions are mathematically different. Meanwhile. (1995) expressed the second polynomial relationship between DLV and proportion of aged binder. (1995) came up with the following equation to estimate the viscosity of bitumen blend: 2 DLV  0. The result indicated DLV method performed well when supercritical and commercial recycling agents were used. the penetration and softening point of the mixture is estimated by those of individual bitumen comprising the mixture and the mass proportions. pen2 : penetrations of both binders comprising mixture a . 2 x2 Viscosity of mixture. to use one constant interaction parameter universally. m . (1995) compared the predicted viscosity values using different method to the experimental data. The penetration of the mixture is estimated by the following equation: log penmix   a log pen1   b log pen2  Where: penmix : calculated penetration of the mixture (8) pen1 . Chaffin et al. b : mass proportions of each bitumen in the mixture ( a  b  1 ) The softening point of the mixture is estimated by the following equation: 30 . Arrhenius or ASTM 4887 is the best to estimate viscosity of the blend.26 x as  0.1 . rejuvenator and aged binder Proportion of aged binder in the mixture Based on experimental data.73 x as (7) Where: x as : the proportion of aged binder in the blend In Grunberg and Nissan method. Mixing rule is also expressed in BS EN 13108 part 1 (2006). Chaffin et al. the deviation is considerably larger than those obtained by the other mixing rules when soft asphalt was used as recycling agent.01  0. In this document.

the conventional method is modified to prevent the “blue smoke” phenomenon. some modified features should be made to the conventional mixing plant to incorporate RAP materials to produce the hot recycled asphalt mixtures (Roberts et al. there are problems using conventional method to produce hot recycled asphalt mixture. With existence of RAP. The burner is in charge of drying aggregate to remove the moisture content and preheat aggregate to the temperature required for mixing. there will be environmental problems as the RAP bitumen is burned directly by the flame in the burner causing blue smoke (Hughes.. The way RAP is processed is different with different modified mixing plants. 2. Hence. 2. 31 .Tmix  aT1  bT2 Where: Tmix : calculated softening point of the mixture.1 Batch plant The schematic of a batch mixing plant is illustrated in Figure 12. each component has its own function. (9) T1 . Virgin aggregate is superheated and the excessive thermal energy will be used to heat up RAP material from ambient to required mixing temperature and remove the RAP moisture content. The materials for recycled mix are not just virgin aggregate and bitumen but also RAP. After storing in hot bin. However.4.4 Hot recycled mixture production The conventional plants for producing hot asphalt mixture are batch plant and drum plant. 1978). If the RAP material is dried the same way as the virgin aggregate. After drying process. the heated aggregate is conveyed to the screening deck to determine the quantities necessary and stored in the hot bin. For conventional asphalt mixture. 1991). T2 : softening point of each individual binder in the mixture. RAP is stored separately before being introduced to virgin aggregate in the hot bin. the combination of RAP and virgin aggregate will be mixed with virgin bitumen to produce the recycled mixture. Bitumen is also preheated and weighed before sprayed and mixed with heated aggregate in the pugmill.

2005): 32 . 2000) There are different methods of introducing RAP to the superheated virgin aggregate (Brock and Richmond.Figure 12: Batch Plant (MAPH-2.

it will increase the ageing speed of RAP bitumen (Shell bitumen handbook). if the drying temperature is higher than 100oC. cannot pass the screen decks. On the other hand. the heat from superheated virgin aggregate is not well transferred to the RAP materials to remove all the moisture content and heat up RAP.4 mm will be blind due to the RAP binder filling up those tiny holes of the screens.5 mm). The ageing process might also be affected by the steam generated due to RAP moisture content during the mixing process in the pugmill. weighed. In addition. With bitumen. especially sizes smaller than 6. the other might be combined of many 1 mm particles. and stored.- Method 1: RAP is mixed with superheated virgin aggregate before being screened. In the second method. the time for heating will be longer and might seriously affect the production rate as well as the cost of asphalt production. RAP materials do not have the same characteristics as virgin aggregate as RAP is a combination of aged bitumen and aggregate. sometimes considerably different. - - In the first method. bigger RAP size cannot be used in this process as the bigger size. As RAP is mixed with virgin aggregate at extremely high temperatures. Hence. The only problem with the third method is the production rate and the cost. ½ inch (12. one might be made of only one ½ inch (12. as the RAP is introduced to superheated virgin aggregate and stored in the hot bin. for instance. Method 2: RAP is weighed and screened before introduced and stored with superheated virgin aggregate in hot bin Method 3: RAP is dried separately and then mixed with heated virgin aggregate and bitumen. the properties of RAP binder might change. for example. However. the blend is not well mixed. the bottom deck screen. Two RAP particles have the same size. it is really difficult to control the gradation of the recycled mixture as the size of RAP and the size of RAP aggregate are not similar.5 mm) aggregate. 33 . size 2 inches (50 mm). In addition. if low drying temperature is employed.

However. The mixing between aggregate is implemented in the pugmill. the movement of materials is opposed to the direction of the flame in counter drum mixer (Brock and Richmond. The drying and heating time of aggregate is twice as long as the mixing time between aggregate and bitumen.2. 2005) There are two main types of drum facility. aggregate is heated and mixed with bitumen in the same drum. The procedure in the drum facility is quite different from batch plant. materials go in the same direction with the flame. 2003). the conventional drum facility must be modified for hot recycled asphalt production. In batch plant. Both types are different regarding the way that materials are introduced into the mixer. Vice versa. the process of drying and heating aggregate is separate from the mixing aggregate with bitumen. the modifications are different. In parallel mixer. In this facility. Bitumen is introduced and mixed with heated aggregate at the two third point of the overall time. Figure 13: Drum mixer with RAP centre inlet (Brock and Richmond. Depending on what type of drum mixer.2 Drum facility (Drum mixer) The schematic of drum mixer for producing hot asphalt mixture is demonstrated in Figure 13. Drying and heating aggregate are processed in the burner. In order to intake RAP materials. with conventional drum facility. RAP is mixed with superheated virgin aggregate before being mixed with virgin 34 .4. parallel and counter drum mixer. The first version of parallel drum mixer has a centre RAP inlet (Figure 13). Mixing time of aggregate with bitumen is normally 30 seconds (Read and Whiteoak. 2005).

2005) 35 .binder. Figure 14: RAP in parallel drum mixer with isolated area (Brock and Richmond. 2005) In parallel type. RAP is introduced with superheated virgin aggregate before entering isolated mixing area where the blend is mixed with virgin binder. the mixer is designed separately and called added continuous mixer (Figure 15). The design of kicker flight in the middle of the drum mixer is aimed to form a dam of virgin aggregate so that the direct exposure of RAP to the flame is prevented. Figure 15: RAP in parallel drum facility with added continuous mixer (Brock and Richmond. In this facility. Sometimes. there is also a drum mixer with separated mixing area (Figure 14).

RAP material. the energy RAP absorbs from superheated virgin aggregate also increases RAP temperature and weakens the bitumen bonds between RAP aggregate particles. RAP materials are separated and mixed with virgin aggregate. Generally. the exposure of RAP to direct flame of the burner is absolutely prevented. As drying and mixing compartment are separated by the shell of the drum. 2005) Except for the parallel drum facility with isolated area in which. Figure 16: Double Barrel Drum facility (Brock and Richmond. The heat transferred from superheated virgin aggregate will help to remove the RAP moisture content. In addition.The latest type of drum facility is called double barrel mixer. the water steam is removed considerably from mixing area. The combination of RAP and 36 . An advantage of drum mixer is big sizes of RAP can be processed as the RAP is incorporated directly into the mixer. Under the mechanical mixing. In this facility. for instance. Hence. RAP material at ambient temperature is introduced with superheated virgin aggregate. the ageing after mixing with superheated aggregate due to the direct exposure to high temperature and steam condition. in both types of mixing plant. the shell of the drum is used as a shaft of the coater (Figure 16). the characteristics of RAP binder might be changed. virgin aggregate and virgin bitumen are mixed in the condition of high temperature and steam. An example of preheating temperature for RAP in drum mixer is illustrated in Table 6.

Table 7 illustrates the maximum sizes of RAP allowed for batch and drum mixing plants in relation 37 . the production cycle is approximately the same except for double barrel mixer at about 90 seconds (MAPH-2. Superheat Temperature Required (oC) 116 C Mix o RAP Content (%) RAP Moisture Content (%) 127 oC Mix 138 oC Mix 149 oC Mix 10 20 30 40 50 0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5 132 134 137 140 143 146 144 151 157 163 169 175 162 173 183 194 204 215 186 218 234 250 266 282 216 240 264 289 313 338 144 147 149 152 155 158 158 164 171 177 183 189 178 188 199 209 220 231 203 219 235 251 267 283 238 262 287 311 336 360 156 159 162 164 167 170 172 178 184 191 197 203 166 315 214 225 236 246 221 237 253 269 286 302 260 284 309 333 358 382 168 171 174 177 179 182 186 192 198 204 211 217 209 219 230 241 251 262 239 256 272 288 304 320 282 309 331 356 380 404 Table 6: RAP preheating temperature required in Drum Mixer (Brock and Richmond. In drum mixer.virgin aggregate is then mixed with virgin binder. 2000). 2005) 2.3 RAP sizes used for production of recycled mixture There is a variety of RAP sizes used for production of RAP in recycled mixtures regarding the equipment used for recycling and the percentages of RAP in the mixture. The total duration for a production cycle in batch plant is about 60 seconds (Read and Whiteoak.4. 2003).

for instance. 2007) 2. the aim of a mixing procedure is to reduce the scale and intensity of segregation.5 Mixing mechanism 2. 38 . RAP % .Drum Plants Base 50 30 70 50 15 40 50 60 40 40 Open 50 50 Open 50 30 30 40 Open 40 50 50 30 50 50 Open 50 50 25 Binder 50 30 70 50 15 40 50 50 40 None Open 25 50 Open 50 30 30 40 Open 40 50 50 30 50 50 Open 50 50 25 Surface 15 30 70 50 15 40 30 None 40 None Open 15 20 Open 50 30 None None Limit 10 50 30 15 50 10 Open 15 15 10 Top Size for RAP 2 in 1. gases and liquids.5 in Specs 2 in Table 7: State DOT specification requirements for the use of reclaimed asphalt pavement (RAP) in hot asphalt paving mixtures (United States Department of Transportation. there are some States. for instance. Arkansas.to RAP proportion of some States in America.5.1 Mechanical mixing There are many aspects that scientists have to cope with in the mixing industry. liquids.5 2 in 2 in Specs 2 in 1. State Alabama Alaska Arizona Arkansas California Colorado Connecticut Delaware Florida Georgia Hawaii Idaho Illinois Indiana Iowa Kansas Kentucky Louisiana Maine Maryland Massachusetts Michigan Minnesota Mississippi Missouri Montana Nebraska Nevada New Hampshire New Jersey Max. The maximum size of RAP used is normally less than 2 inches (50 mm). mixing of solid particles.75 in Specs 3 in 2 in 1. However. and Minnesota which use up to 3 inches (75 mm) RAP sizes (United States Department of Transportation.5 in 2 in 2 in 1. RAP % .5 in 2 in Specs 2 in 1 in Specs 0.5 in 3 in 2 in 1. Generally. 2007).Batch Plants Base 40 30 70 50 15 40 35 60 25 30 Open 50 50 Open 50 30 30 40 Open 20 50 59 30 50 50 Not Used 50 35 25 Binder 40 30 70 50 15 40 35 50 25 None Open 25 50 Open 50 30 30 40 Open 20 50 50 30 50 50 Not Used 50 35 25 Surface 15 30 70 50 15 40 25 None 25 None Open 15 20 Open 50 30 None None Limit 10 50 30 15 50 10 Not Used 15 15 10 Max. and cohesive powders.5 in 2 in Specs 2 in 1.

rotational. In addition. type of particles... van der Waal’s forces. and mixing time. there will be movements of small free flowing agglomerates of aggregate particles (Figure 18). 2001). the surface roughness will reduce the contact areas hence reduce the bonding forces.for example. size segregation. making the most homogeneity of mixing ingredients (Harnby et al. Figure 17: Ideal mixture with homogeneity (Harnby et al. 39 . vibrating effects must be satisfied to break any bonds that exist between particles and relocate these particles in the mixture. In this research. In term of interparticulate forces among particles. Particle types also attribute to the magnitude of liquid bridge bonding. different sizes of particle also cause segregation as these particles will have different velocities during mixing process (Harnby et al. and liquid bridge bonding. the external forces. For instance.. The ideal mixture with homogeneity is illustrated in Figure 17. If these particulate forces are not deactivated during mixing process. type of mixing machine. for instance. interparticulate forces. There are many factors affecting the mixing process that lead to segregation. Otherwise. segregation will occur and cause adverse effects to the quality of the mixture. for instance. if the surfaces of particles are not smooth. 2001) In order to obtain the most homogeneity of the mixture. there are electrostatic bonding. 2001). just the aspect of mixing cohesive powders is considered as asphalt mixtures (due to bitumen having a liquid state at high mixing temperature causing interparticulate forces among aggregate particles) possess relatively comparable characteristics to those of cohesive powders.

Figure 18: Mixture with “self-loving” particles (Harnby et al. as the size of most aggregate particles is larger than 75µm. breaking up all the bonds between aggregate particles. If the mixing machine and procedure cannot overcome these issues. and blending hot virgin aggregate with bitumen binder. The mixing mechanism of recycled asphalt mixture is quite different from that of conventional asphalt mixture. heating aggregate. due to moisture. In addition.. electrostatic charging. for instance. the combinations 40 . This is to make a homogeneous blend of different particle sizes of aggregate. relocating all the particles so the mixture can reach the homogeneous status. segregation will occur. During the whole process. heating and the rotational effects on aggregate create aggregate movement. 2001). In addition. even filler. hence just the bitumen contributes to the bond among aggregate particles. van der Waal’s forces. before mixing with bitumen so binder can cover each particle surface to improve the bonding between particles in the final mixture. However. 2001) Production of conventional asphalt mixture comprises weighing. as the cohesion among particles due to existence of water is not considerable compared to that of bitumen. Both types of asphalt mixtures are finally. all the materials are rolled centrifugally in a mixing vessel. breaking the bond between aggregate particles. In addition.. the flow of every single aggregate particle is considered free due to the fact that electrostatic charging and van der Waal’s forces among particles bigger than 75µm are inconsiderable and can be ignored (Harnby et al. the drying process also separates all the particles. The heating time of aggregate is aimed to dry the material.

2000). the aggregate is not just virgin aggregate but also RAP. this ultimate difference between two types of mixture makes the production. This proportion of aged bitumen makes RAP not conform to any requirements of normal design methods for virgin mixture.5 mm) size. mixing mechanism quite different. All the RAP materials were processed to less than ½ inch (12. RAP Source FL CT AZ Viscosity at 60oC. low stiffness RAP from Florida (FL). Hveem. inert as black rock at ambient temperature.. in fact. PG 52-43 and PG 64-22 were used as rejuvenators (Table 9). for instance. or Superpave. Two soft virgin binders. The production of recycled asphalt includes the following steps: superheating virgin aggregate blending with RAP mixing the blend of virgin aggregate and RAP with virgin bitumen “Black rock” concept The ultimate characteristic distinguishing RAP (reclaimed asphalt pavement) from virgin materials is the content of aged bitumen. a blend of aged bitumen and aggregate as well. The proportions of RAP in the recycled mixture were 10. conventional asphalt mixture is a combination of virgin aggregate and bitumen binder while in recycled mixture. three types of RAP. the level that a proportion of aged bitumen might be incorporated with virgin binder during the mixing process and through service life of pavement has still been ambiguously identified. Although the RAP aggregate covered by aged binder is considered stiff. and high stiffened RAP from Arizona (AZ) were used. Poise 23760 65191 124975 Table 8: Viscosity at 60oC (Poise) of RAP binder (McDaniel et al.. Marshall. Viscosity of RAP binder is showed in Table 8. 2000) 41 . Research has been carried out to investigate whether RAP material acts as black rocks once accommodated in the recycled mixture (McDaniel et al. and 40%.of the same type of materials. However. bitumen binder and aggregate. medium stiffness RAP from Connecticut (CT). In fact. In this study.

9 54.6 -7. 42 .1 -16. repeated shear at constant height (RSCH). This total blending simulates what occurs during the design process.3 33.8 25. Stiffness BBR S BBR m-value Actual (Critical Temperature) MP1 (Performance Grade) PG 52-34 53. For the black rock case. indirect tensile creep (ITC) and indirect tensile strength (ITS) are employed to identify the differences in properties of three blending situations. all the virgin aggregate. the compacted sample is stored in condition of 85oC for five days continuously. Preheated temperatures of virgin binder rely on the performance grade: 155-160oC for PG 64-22 and 135-140oC for PG 52-34. These tests cover a wide range of asphalt properties. actual blending.Virgin Binders Aging RTFO Property High Temp.4 PG 82-24 PG 82-22 AZ 89 85. RAP and virgin aggregate are mixed directly before being blended with virgin bitumen.6 11. Stiffness PAV PG Table 9: Critical temperature and performance grades of virgin and recovered RAP binders Test samples are designed to simulate three possible cases with different levels of blending. with both 10 and 40% of RAP. In the total blending.7 -18.3 -15. and black rock.8 66. Superpave tests including frequency sweep (SF). the loose mixture is held in the oven at 135oC for 4 hours successfully to simulate the short term ageing. and RAP aggregate in case of black rock condition. are heated overnight at 150oC. To prepare for the samples.9 -16. In case of actual practice. rutting. RAP material is also preheated at 110oC for the duration of 2 hours before mixing.2 75. simple shear (SS). After mixing.7 -25.6 21. the aged bitumen is extracted and recovered from the RAP before being mixed with RAP and virgin aggregate. 143-148oC for PG 64-22. actual blending. total blending. total blending.4 19.1 PG 89-15 PG 88-10 Original High Temp.1 -14.9 PG 53-33 PG 52-28 PG 64-22 67. fatigue. and low temperature cracking. In terms of long term ageing.4 75.5 -23. All the specimens are compacted by Superpave gyratory compactor at different temperatures. and black rock. and 122-130oC for PG 5834.4 PG 82-25 PG 82-22 CT 82. just RAP aggregate is extracted and mixed up with virgin aggregate and binder.8 -5. Stiffness Intermediate Temp.2 PG 66-26 PG 64-22 Recovered RAP Binders (Unaged) FL 82.1 -15.

repeated shear load is applied on the test specimen to produce a horizontal strain of 0. reaches 5000 or the permanent strain exceeds 5%.005%. 10 Hz and temperatures of 20. The test result also indicates that with 10% of RAP. in terms of temperature cracking. Axial stress is also applied to keep the specimen height constant. Indirect tensile test (AASHTO TP9) is used to analyze thermal susceptibility and low temperature cracking of asphalt mixture. plastic shear deformation of black rock case is considerably higher than those of total blending and actual practice with 40% RAP content. On the contrary with the black rock situation.1 sec loading time and 0. the stress controlled shear load is applied to specimen until the number of loading cycles. However. the shear load is increased at the rate of 70kPa/sec until it reaches the specified shear load relevant with test temperatures of 4. The results at frequencies of 0. This is similar with increasing test temperature. In simple shear at constant height test. To evaluate rutting phenomenon of asphalt mixture. Plastic shear strain of mixture is determined under given loading mode and temperature (Procedure C. deformation. AASHTO TP7-94). Maximum deformation results from the test indicates that with the same RAP. the repeated shear at constant height test is employed. each consisting of 0. Horizontal and vertical deformations of the specimen due to the static compression load applied across the diametric plane of specimen are recorded over a period of time (240 seconds) to calculate the creep compliance. and 40oC (Procedure D of AASHTO TP7-94). On the other hand. Test temperature is 58oC for the PG 64-22 and 52oC for the PG 52-34 binder. specimen is tested at -10oC by applying the load with strain rate of 12. 40oC show that stiffness of the samples increases in accordance with the increase of RAP proportion for total and actual blending cases. the recycled mixture with softer virgin binder has higher plastic shear strain. the maximum deformation increases if the softer virgin binder is applied. and -20oC. During the test.5 mm/minute until fracture 43 . 20.01.6 sec for rest period. and loading time was studied at three temperatures 0. In this test.Frequency sweep at constant height is used to determine the complex shear modulus and phase angle of asphalt mixture. stiffness of samples is the same with 10 and 40% of RAP. -10. With the same virgin binder. the plastic shear strain is not influenced substantially with different sources of RAP and blending situation. The load is held constant for 10 seconds before being released at a rate of 25kPa/sec. The relation of load magnitude.

total practice and black rock cases. Laboratory procedures for recycled mixture preparation In general. the procedures for preparing the recycled mixture include heating the RAP to a specified temperature for a certain period. The test result shows that with 10% RAP contents. Data from all the tests showed that with 10% RAP. It has been argued whether the complete blending obtained in the laboratory can actually occur in an industrial mixer. However. In fact. and mixing with preheated rejuvenator before 44 . total practice and black rock were approximately the same. RAP binders interacted with virgin binder and the mixture generally had approximately the same properties as that of complete blending case. there were 36 cases in which the test results of actual blending. Actually. the proportion of RAP was only 15%. and black rock were different. and only 6 cases in which actual blending resembled the black rock. (2000) also carried out a study to compare the laboratory mixing procedure with that of an industrial mixer. 3 cases with 40% of RAP. In addition. where 6 cases actual blending resembled the black rock with 10% RAP. 9 cases that actual practice and total blending were almost similar. There are 12 cases where the results from total blending. there were 3 cases where actual blending resembled the black rock. The test results with 40% RAP showed inconsistent trend. The test results showed that the laboratory mixture processed the same characters as those of a mixture mixed by real mixing plant. the stiffness of the actual practice has tendency of being between black rock and total blending cases. The overall trend indicated that the there is no considerable differences between actual blending. McDaniel et al. Due to the results from the tests. actual practice. McDaniel et al. although there is not substantial variation. However. There were 21 cases where the test results showed no considerable variations between total blending and actual practice. among 66 combinations. the laboratory mixing procedure used in this research is quite different from that of the mixing plant. The long RAP preheating time might enhance the interaction between RAP and virgin binder.occurs. (2000) concluded that RAP did not work as “Black Rock” or inert component in the recycled mixture. especially in the case where high proportion and larger size of RAP is used. 10 out of 12 cases occurred with PG 64-22.

(2000). The properties of original RAP binders were compared to those of the RAP after different preheating time. The loose mixture was then conditioned at 60oC for 15 hours. The requirement of heating RAP is to make the RAP workable so that it can be mixed with rejuvenator (McDaniel and Anderson. McDaniel et al. Stephens et al.compaction. There were also samples with the same aggregate gradation and virgin binder for comparison.5. 2001). (2001) concluded that the increase of indirect tensile and unconfined compression strength was attributed to the long RAP preheating time. the higher the strength of the recycled mixtures (Figure 19). the effect of preheating on the recycled mixture would be insignificant. Carpenter and Wolosick (1980) heated the RAP to 116oC before mixing with rejuvenator. and Mobilsol-30) at 82oC. In fact. For instance. Two types of RAP taken from Arizona and Florida were subjected to different preheating times and temperatures. the purpose of preheating RAP is to soften the RAP binder to break RAP into separate pieces so the rejuvenator can cover the RAP for diffusion process. Noureldin and Wood (1987). the lump of RAP was totally heated through and broken down during mixing for complete blending. Effect of RAP preheating duration Stephens et al. However. 2001). The test results indicated that the longer the RAP preheating time. (2000) investigated the effect of preheating time on the properties of RAP binder. This was not a firm conclusion as the increase of indirect tensile and unconfined compression strength might also be accredited to the RAP being hardened due to exposure at high temperature during the long preheating time. heated the RAP to 110oC for 2 hours mixing with rejuvenator at required mixing temperature.. All the samples were subjected to indirect tensile and unconfined compression tests. AE-150. Preheating time was also used for preventing the effect of RAP moisture content on the properties of the recycled mixture (Stephens et al. (2001) studied the effect of RAP preheating time on the strength of recycled mixture with the hypothesis that if RAP acted as black rock. heated the RAP at 116oC for 30 minutes before mixing with preheated rejuvenator (AC 2. McDaniel el al. Samples were prepared with the same 15% RAP content except with preheating times from 0 to 540 minutes. The test results (Figures 20 and 21) indicated that the complex modulus of RAP after two hours preheated 45 .

(2001) during the first two hours of preheating. Figure 19: RAP preheating time versus indirect tensile and unconfined compression strength (Stephens et al.. This supports the conclusion of Stephen et al. 2001) Figure 20: Arizona RAP binder complex modulus versus preheating time and temperature (McDaniel et al. 2000) 46 . the increase in indirect tensile and unconfined compression strength of the mixture is attributed to complete blending with virgin binder as the RAP lump is heated through and completely separated..at 110 and 150oC did not change considerably.

2000) 2. Diffusion mechanism There have been some studies carried out in the past to investigate the diffusion mechanism of virgin binder into RAP bitumen. However. The mechanical mixing only supports the fact that virgin binder can coat the particles covered by RAP bitumen. The reclaimed asphalt was taken from a city street in Champaign. the mechanism of diffusion is not well understood even though asphalt recycling is not a new aspect in the industry. RAP bitumen was then extracted and tested. The properties of RAP bitumen and the grading of RAP aggregate are presented in Tables 10 and 11. 47 . 1976 and crushed until passing 12. Carpenter and Wolosick (1980) undertook research to evaluate the effects of modifiers on the performance of recycled asphalt mixture..Figure 21: Florida RAP binder versus preheating time and temperature (McDaniel et al.2 Diffusion process or chemical mixing The performance of recycled asphalt mixtures is not only affected by the quality of mechanical mixing but also the interaction between RAP and virgin binder. Meanwhile.5 mm sieve.5. the requirement of the recycling process is for the virgin binder to be well mixed and penetrate or diffuse into RAP bitumen so as to reduce the viscosity and bring back its expected requirements. Illinois in June.

s) Penetration at 25 C (dmm) Penetration at 4oC (dmm) Softening point (oC) Asphalt content (%) Specific gravity (g/cm3) o Value 4490 26 22 63 5. were prepared. viscosity at 60oC of 234 mm2/s (0. Resilient modulus. RAP bitumen was extracted and blended with 48 .5 mm 9.028.3 mm 4. the amount of rejuvenator is 20% of the weight of RAP bitumen. With rejuvenated samples. rejuvenator and % of rejuvenator.Properties Viscosity at 60oC (Pa. 1980) Sieves Size 12.23 Pa. Both had the same material.5 mm 6.00 mm 850 µm 425 µm 150 µm 75 µm Percent Passing 100 81 78 68 58 34 23 13 9 Table 11: Grading of RAP aggregate (Carpenter and Wolosick. 1980) Carpenter and Wolosick used 100% of RAP for this research and Paxole 1009. The result of the blending trial is a graph showing the relation between viscosity of aged bitumen. In order to study the influence of diffusion mechanism on the properties and performance of recycled mixture.3 1.s) and specific gravity of 1. bitumen and air void content but different methods of preparation.75 mm 2. In order to get the recycled binder with target viscosity of 100 Pa.s. creep compliance.198 Table 10: Properties of RAP binder (Carpenter and Wolosick. rejuvenated and recycled. was employed as rejuvenator. The amount of rejuvenator was determined after trial blending with different proportions of rejuvenator over RAP binder. and permanent deformation tests were implemented in this research. two types of samples.

Hence. Figure 22: Effect of diffusion on resilient modulus (Carpenter and Wolosick. all the rejuvenator diffused into RAP bitumen. Differently in recycled sample. viscosity of the aged binder coated aggregate decrease and that of outer layer (rejuvenator) increases. In the rejuvenated samples. the rejuvenator forms a low-viscosity layer covering the RAP particles 2. The difference between the test results of recycled and rejuvenated mixture was studied to evaluate the influence of diffusion on the recycled samples. Carpenter and Wolosick (1980) simply developed the diffusion model of rejuvenator into RAP bitumen. the viscosity of rejuvenator and aged binder remain approximately the same. The diffusion process included the following steps: 1. 4. the test result is almost the same with different testing times. The blend of rejuvenator and aged bitumen reaches equilibrium In step 1. simultaneously softening the aged bitumen and reducing the amount of rejuvenator 3. was mixed directly with rejuvenator to simulate the mixing procedure in the asphalt industry. The samples then were tested at predetermined time intervals. there is almost no interaction as rejuvenator just coats the RAP particles (Figure 23). The proportion of rejuvenator that 49 . 1980) The test results showed that the resilient modulus of the recycled mixture critically decreased during the first two weeks before starting to increase again (Figure 22). Rejuvenator then starts to diffuse into aged binder. The phenomenon did not happen in the case of rejuvenated samples. rejuvenator starts to penetrate into RAP bitumen. after being heated at 116oC. reclaimed materials.rejuvenator before being mixed with RAP aggregate.

The process will progress until equilibrium status where all the rejuvenator diffuses into aged binder and generates a homogeneous blend (Step 4). This situation leads to the fact that the 50 . Figure 24 shows after compaction. the binding force between aggregate particles is due to the high-viscosity RAP bitumen. Actually. the rejuvenator at this moment has been located in the voids among particles. Figure 23: Schematic of modifier coating an aggregate particle during recycling process (Carpenter and Wolosick.covers the RAP particle will gradually decrease. the viscosity of the remaining rejuvenator increases (Steps 2 and 3). Simultaneously. 1980) This simple diffusion mechanism can be used to interpret the critical duration in which the resilient modulus decreases and permanent deformation increases. The mutual interaction between rejuvenator and aged binder results in the reducing viscosity of the outer layer of aged binder. as the penetration has not started.

outer layer and inner layer.resilient modulus of recycled mixture is higher than that of rejuvenated sample. As the diffusion or penetration of rejuvenator into RAP bitumen starts. The resilient modulus keeps decreasing until no rejuvenator remains. the binder is divided into two layers. 51 . the consistency of each layer will be different in relation to different testing time intervals. the resilient modulus decreases as the outer layer of RAP bitumen begins to be softened by rejuvenator. At this point. 1980) Staged-extraction process Carpenter and Wolosick (1980) also carried out an extra experiment to verify the developed diffusion model. If the diffusion of rejuvenator into RAP bitumen exists. the viscosity of outer layer keeps increasing until the blend between rejuvenator and RAP bitumen reaches equilibrium. This phenomenon is also due to the remaining low-viscosity rejuvenator starting to bind the aggregate particles together. In this experiment. Figure 24: Diffusion model (Carpenter and Wolosick.

The mixture is prepared with the same procedure as that for the performance tests. However, the mixture is left uncompacted. In order to divide the bitumen coat into two layers, a sample of loose mixture is immersed in trichloroethelyne for 3 minutes. The bitumen recovered by Abson method will represent the outer layer. The inner layer is achieved by washing and recovering all the bitumen. The consistency of these recovered bituminous materials will be tested. The whole procedure will be repeated at different times. The results (Figure 25) show that simultaneously, the penetration of inner layers increases and that of outer layers decreases until both penetrations are the same.

Figure 25: Penetrations of outer and inner layers as function of time (Carpenter and Wolosick, 1980)

The research carried out by Carpenter and Wolosick (1980) did not take into account the influence of virgin aggregate during mixing process. This research dealt only with the diffusion mechanism of the mixture of 100% RAP and rejuvenator. This might not represent the phenomenon that occurs in recycled mixture containing virgin aggregate. Depending on the efficiency of mixing process, there might be not only RAP particles coating by rejuvenator but also virgin aggregate particles coating by rejuvenator or blend of RAP binder and rejuvenator in recycle mixture containing virgin aggregate. Noureldin and Wood (1987) also studied the diffusion of rejuvenator into RAP bitumen. In this research, not only the mixture of RAP and rejuvenator but also the mixture of RAP, rejuvenator and the addition of virgin aggregate were considered. The existence of virgin aggregate in the mixture aimed to simulate the real situation in the recycling asphalt industry. RAP material was milled from road US-52 in Indianapolis (Indiana). Properties of RAP binder are showed in Table 12. 52

Three rejuvenators, AC 2.5, Mobisol 30, and AE 150, were used. The amount of rejuvenators and percentage of RAP bitumen were estimated based on the Asphalt Institute design recycled asphalt mixture method (Arrhenius viscosity mixing equation). The target was that after being recycled by rejuvenator, the recycled bitumen must have the viscosity at 60oC approximately similar to that of AC 20 (from 190 to 240 Pa.s). There were three combinations of RAP bitumen and rejuvenator, 40% RAP bitumen/60% AC 2.5, 45% RAP bitumen/55% AE 150, and 85 %RAP bitumen/15% Mobilsol 30. Properties Penetration at 25 oC, (dmm) Viscosity at 60 oC (Pa.s) Kinematic Viscosity at 135 oC (cSt) Softening Point (oC) Bitumen content (%) Value 28 2089 726 60 6

Table 12: Properties of RAP bitumen (Noureldin and Wood, 1987)

Noureldin and Wood (1987) also used the staged extraction method. However, different from Carpenter and Wolosick (1980), the bitumen coat was divided into four microlayers. One advantage of this method is to show the non-uniform ageing pattern of RAP material. After four microlayers extraction of RAP materials, the results (Table 13) indicated that RAP bitumen from the outer two layers were seriously hardened. On the contrary, the two inner layers close to the aggregate surface were slightly aged, the consistency were almost the same as those of the original bitumen AC 20. Solvent increment (mL) 200 200 300 700 Binder (% by weight) 55.5 26.5 11.2 6.8 Penetration (dmm) 24 33 65 57 Viscosity at 60 oC (Pa.s) 2400 1500 250 330

Table 13: Test results on reclaimed staged-extraction of RAP (Noureldin and Wood, 1987)

To prepare for the sample, RAP was heated at 115oC for 30 minutes and rejuvenator at 82oC before mixing together with virgin aggregate for 2 minutes. Virgin aggregate was also

53

heated at 115oC for 30 minutes. The loose mixture was then preserved at 60oC for 15 hours. To obtain the bitumen of each microlayer, the sample of 1200g was in turn immersed in 200, 200, 300, and 700 mL of trichloroethylene for 5 minutes. Bitumen of each microlayer was recovered by Abson method and its consistency was determined. In the case of mixing only RAP with rejuvenators, results from staged-extraction tests showed that all the rejuvenators could restore the consistency of the two outer microlayers. However, the other two inner microlayers showed almost unchanged tendency (Table 14). Due to the test being carried out just at one point in time, the result could not show the changing tendency of each layers consistency. Solvent Binder Increment (mL) 200 60% AC 2.5 40% RAP binder 200 300 700 200 55% AE 150 45% RAP binder 200 300 700 200 15% Mobilsol 30 85% RAP binder 200 300 700 Binder (% by Weight) 67.5 21.5 7 4 69 16.5 8.5 6 71 18 6 4 Penetration at 25oC (dmm) 67 68 59 50 75 70 62 49 75 69 63 48 Viscosity at 60oC (Pa.s) 167.4 188.0 239.4 300.0 168.3 201.0 229.0 302.0 186.4 198.0 204.0 315.2

Table 14: Test results on reclaimed, staged-extraction, no virgin aggregate (Noureldin and Wood, 1987)

In the case where the recycled mixture was a combination of RAP material, rejuvenator, and virgin aggregate, the amount of aggregate was estimated so the recycled binder accounted for 6% by weight of the mixture. The grading of virgin aggregate was also selected hence the gradation of the whole mixture satisfied the requirement of Indiana specification. In accordance to 6% bitumen over total weight of the mixture, the amount of aggregate added were in turn 60, 55, and 15% in relation with mixture using rejuvenator 54

AC 2.5, AE 150, and Mobilsol 30. The staged-extraction test results showed that only the mixture using rejuvenator AE 150 had the same tendency as that of mixture using only RAP and rejuvenators. The other two using rejuvenators AC 2.5 and Mobilsol 30 had different trends (Table 15). The viscosities of inner layers were higher than those of outer layers. The shortcoming of this research was that the effect of time on the diffusion process was not considered. In addition, the use of a single diffusion pattern could not fully describe the diffusion mechanism that occurs in recycled asphalt mixture. The inconsistent viscosity pattern of micro-layers compared to that of the mixture using only RAP and rejuvenator indicated the segregation of the bitumen phase in recycled mixture. This substantiates the fact that in recycled mixture, there exist RAP particles covered by rejuvenator, virgin aggregate particle covered by rejuvenator, and RAP or virgin aggregate particles covered by blend of aged binder and rejuvenator. Each situation has its own diffusion mechanism. Solvent Binder Increment (mL) 200 60% AC 2.5 40% RAP binder 200 300 700 200 55% AE 150 45% RAP binder 200 300 700 200 15% Mobilsol 30 85% RAP binder 200 300 700
1987)

Binder (% by Weight) 72 19 5.5 3.5 71 19 6 4 74 17.5 5.5 3.5

Penetration at 25oC (dmm) 60 51 52 130 70 67 60 50 73 80 90 100

Viscosity at 60 oC (Pa.s) 210.0 289.2 247.0 80.9 197.2 173.4 242.4 361.6 204.9 166.4 126.0 124.0

Table 15:Tests results on reclaimed, staged-extraction, virgin aggregate used (Noureldin and Wood,

55

. The recovered rejuvenated RAP bitumen is then subjected to rheological testing to identify the difference between layers. However. Sieves Size No.7 mm 9. containing No 4 (4.2 19.50 No. Huang et al.5 mm 25. 200 % Pass 100 81 46 30 23.75 mm % Pass 100 97. However. just the fine RAP.6 77. 2005) Sieves Size 37. The process is repeated three times for the first three microlayers. The schematic of extracted layers is illustrated in Figure 26.8 No.7 35. The layer in contact with aggregate surface was obtained by washing the remaining rejuvenated RAP with solvent.A staged-extraction process was also used by Huang et al. rejuvenator. The gradation of RAP and virgin aggregate are showed in Tables 16 and 17. (2005) used the combination of RAP. and virgin aggregate. Limestone was used as virgin aggregate. virgin bitumen. all the particles passing size No 4 were removed before mixing with RAP.75mm) passing RAP particles was used.3 Table 16: Properties of RAP aggregate (Huang et al. 2005) The rejuvenated RAP was then recovered under staged-extraction process.4 mm 19 mm 12.5 mm 4.3 1. 56 . the rejuvenated RAP was easily separated from the whole mixture due to size difference. Rejuvenated RAP was soaked in solvent trichcloethylene for 3 minute. 100 No.30 No. and aggregate were mixed together at 190oC. The binder was then recovered by Abson method. (2005) to investigate the effects of rejuvenator on RAP material.. 20% RAP.9 Table 17: Properties of Virgin Aggregate (Huang et al.3 14.4 No. In this research. The mixing process simulates the procedure in the practical industry. After mixing.

2005) The test results for viscosities of bitumen from extracted layers at different temperatures showed uniform tendency (Figure 27).. Huang et al. Figure 27: Viscosity at 135oC of different micro-layers coated RAP particles (Huang et al. 2005) 57 . approximately 6-6.. The outer layer due to rejuvenation had lower viscosities at 135oC than those of the inners layers. The layer in contact with surface of aggregate was the stiffest.8% of RAP binder was transferred from RAP materials to virgin aggregate. It was concluded that after mixing.Figure 26: Layers extraction process (Huang et al. By washing all the bitumen on virgin aggregate. (2005) also demonstrated that during mixing process. about 40% of RAP binder was blended with rejuvenator.

The absorption value at certain time will be automatically calculated by WinFirst software using wavelength and relevant energy absorbed.Investigating the diffusion mechanism using marker Karlsson and Isacsson (2003) also carried out a study to examine the diffusion of bitumen rejuvenator. the diffusion was not characterized by accessing the consistency or rheology of rejuvenated binder microlayers. A Mattson Infinity 60 AR spectrophotometer was used in this research (Karlsson and Isacsson. The number of scans performed during each recorded interval is in turn 64 or 256 (Karlsson and Isacsson. Each analyte has particular absorption wavelength or wave number (Table 18). the movement of the analyte also means the penetration of that bitumen layer into the bottom one. During 72 hours. the changes in energy absorption are scanned with a resolution of 4 cm-1 and recorded every 1 or 5 minutes. with different slot width are stuck together on the surface of the prism. The thicknesses of each frame are in turn 200 and 500 µm in accordance with the thickness of top and bottom layers. 58 . each bitumen layer is scraped into the mould by two scrapers that fit into the frames. As the top bitumen layer contains analyte substance. Figure 29 is an example of absorbance difference versus time in relation with various wavelengths. In this study. This equipment includes a non-absorbing trapezoidal prism made of ZnSe (Figure 28). 2003c). The diffusion mechanism was investigated by measuring the variations in energy absorption capability of bitumen layers over a period of time by FTIR-ATR (Fourier Transform Infrared Spectroscopy by Attenuated Total Reflectance). To prepare for the specimen. Two frames. 2003a) The philosophy of this method is to measure the movement of chemical analytes through bitumen layers. Figure 28: Schematic of FTIR – ART (Karlsson and Isacsson. 2003a).

50. The diffusion process is assumed to occur in constant pressure and temperature. 2003a) Based on the calculated absorption results. c  2c  D.α)L. the diffusion coefficient is estimated by mathematical expression of Fick’s Law (Karlsson and Isacsson. 100. and 925 minutes during diffusion of rejuvenator into A-B180 at 60oC (Karlsson and Isacsson. Both layers have total thickness of L in which rejuvenator thickness accounts for (1. 1986) Figure 29: Spectra obtained after 10.Table 18: Approximate absorption wave length of functional groups (Petersen. Fick’s model is simply described in Figure 30. Initially. 2 t x c: concentration in terms of time (10) 59 . the concentration of bitumen is 0 and that of rejuvenator is co. 2003b). 500.

e n  n1   L       n  Dt  L (11) There is a wide variety of materials used by Karlsson and Isacsson (2003). . the result is the mathematical relation between the concentration of rejuvenator at position x and time t as follows: 2 c x. 2003b) 2c0 sin(n )  nx  . 2003b) To solve the equation of Fick’s Law with boundaries expressed in Figure 30. The results (Figure 31) indicate that the diffusion coefficients of 60 . cos  .D: diffusion coefficient t: time x: position Figure 30: Schematic of Fick’s Law diffusion model (Karlsson and Isacsson. and the chemical composition. penetration 85 (C-B85) B60 from Venezuela. The bitumen in this research includes: Rejuvenator V115 from Nynas (R115) which is a heavy naphthetic petroleum distillate B180 from Mexico having penetration of 180 (A-B180) B180 from Saudi Arabia. B-B180 and C-B60. Results from the test indicated that the diffusion of rejuvenator was influenced by many factors. Influence of type of bitumen on diffusion coefficient Rejuvenator R115 is used together with three bitumens A-B180. temperature. t   (1   ). rejuvenator. penetration of 180 (B-B180) B85 from Venezuela. penetration 60 (C-B60) The study was carried out under conditions of different testing temperatures and thickness of bitumen and rejuvenator layers. The thickness of bitumen and rejuvenators are equally 500 µm.c0  (Karlsson and Isacsson. type of bitumen. for instance. The experiment is carried out at different temperatures.

However. However. the test is repeated with different temperatures and four combinations of bitumen-rejuvenator layer thickness. Figure 31: Influence of bitumen type on diffusion coefficient (Karlsson and Isacsson. 2003b) 61 . and 500/500 µm. 2003b) Influence of temperature on diffusion coefficient In this aspect. 200/500 µm. The test results (Figure 32) indicate a minor effect of layer thickness on the diffusion coefficient. the diffusion coefficient in relation with different layer thickness deviates slightly. In general. below 90oC. 500/200 µm. rejuvenator R115 is used together with A-B180. This phenomenon could be attributed to the fact that bitumen B-C60 has lower penetration than those of bitumens A-B180 and B-B180. the experimental results show the increasing tendency of diffusion coefficient if the test temperature increases.rejuvenator into bitumen A-B180 and B-B180 are almost the same and higher than B-C60. 200/200 µm. Figure 32: Influence of temperature on diffusion coefficient (Karlsson and Isacsson.

Bitumen A-B180 mixed with different markers are in turn applied on top of pure A-B180. will act like 62 . 2003b) 2. Actually.Influences of chemical composition of rejuvenators on diffusion coefficient To investigate the effect of composition on diffusion mechanism.6 Segregation and consequences 2. The test results (Figure 33) demonstrate that generally. the diffusion process between these two binders cannot take place.6. the diffusion coefficient decreases if the molecular weight of marker increases. a wide variety of markers are used together with bitumen A-B180. segregation can be perceived as the concentration of coarse or fine materials in some areas of the paved mat in conventional asphalt mixture (Brock et al. In asphalt recycling aspect. 2003). This happens in the case where the virgin bitumen or rejuvenator is not well distributed in the whole mixture or even when these two bitumens are well distributed. RAP material generally has different characteristics from that of pure virgin aggregate or bitumen. aged bitumen normally with high viscosity. there is not only the concentration of certain sizes of aggregate in one area but also the concentration of different binders in some areas. Figure 33: Influence of chemical composition of markers on diffusion coefficient (Karlsson and Isacsson. At ambient temperatures.. The proportion of marker is 3% by weight of the bitumen.1 Segregation Segregation is defined traditionally as non-homogeneity of the asphalt mixture. 2003). The same result has been reported by (Qiu and Bousmina.

for instance. 30mm. 50mm. This mechanism will repeat during the mixing process in order to get a homogeneous blend between RAP and virgin aggregate. However. This is aimed to destroy the bond among aggregate to relocate the position 63 . Therefore. two particles will be separated and relocated. and 20mm. The first issue is that there are many sizes of RAP available. Once the viscosity of RAP binder reaches the critical point that the bond cannot hold two particles together. According to heat transfer theory. the longer the time for heat transfer (Cutnell and Johnson. Energy transferred from superheated aggregate will heat up the RAP and soften the bond between aggregate particles as the higher the temperature. the time for the heat transferred from virgin aggregate to soften and break RAP into separate pieces for the relocation is quite different. each size of RAP may not be made of one particle but probably comprise many smaller-sized particles. the ideal homogeneous mixture is illustrated in Figure 17. after RAP at ambient temperature is blended with superheated virgin aggregate. Hence. 2004). RAP material will change from a mixture of solid agglomerates into a mixture of solid particles with different sizes and aged liquid bitumen. If the assumption is made that each RAP particle is black. there exist not only inter particulate forces as in virgin aggregate case but also the bridge force due to aged bitumen among aggregate particles. while blending virgin aggregate with RAP. The aim of heating RAP is also to use thermal energy transferred from superheated aggregate to soften the bond between aggregate particles as the higher the temperature. the lower the viscosity of bitumen binder. The mixing process transforms from mixing among solid particles to mixing between solid particles and liquid bitumen. the energy transferred from virgin aggregate will heat up RAP. In addition. with white colour for virgin aggregate. Hence. The increase of RAP temperature will gradually change the aged bitumen from solid to liquid state.a solid and RAP has the same characteristics as those of black rock. RAP material will no longer act like black rock but be the same as the blend of aggregate and liquid bitumen. this mechanism is affected by many factors. During the recycled asphalt production. when the temperature is high enough to turn aged bitumen into liquid state. However. the larger the size the particle is. the lower the viscosity of aged bitumen.

of each particle. If the bitumen bond between aggregate is strong enough. the data showed there were 36 cases that test results from total blending. there was apparently not enough support for this conclusion with 40% of RAP in the recycled mixture (Huang et al. However. there might still exist agglomerates moving in the mixture.. In the whole mixture. the virgin bitumen. 2001). and black rock were almost the same. If segregation occurs. For instance. In 2 out of 3 cases.. certainly there will be some agglomerates of different particle sizes of RAP. some particles are coated with aged bitumen which is stiff. If this situation exists. there were 12 cases in which test results of total blending. The segregation was also identified in research by Noureldin and Wood (1987). The existence of small amount of RAP. the others with soft rejuvenator. actual practice and. If the blending time among RAP and virgin aggregate is not enough to break RAP into separate pieces. (2000) maintained that the actual practice and total blending were almost the same. actual practice. particular 10% in this research. AE 150. Even with 10% of RAP. or rejuvenator cannot completely interact and recover the properties of aged bitumen. especially with higher percentages of RAP. after stageextractions and consistency tests. Mobilso 30. There would be segregation during the mixing process. the system will be dominated by the free flowing of agglomerates of aggregate particles (Harnby et al. This conclusion could be applied for the case of 10% RAP as there was a consistent trend of the results. Depending on the sizes and proportion of aggregate in the mixture.5.. did not affect substantially the properties of the recycled mixture. This will lead to the fact that in the mixture. AC 2. The same phenomenon also happens even when the duration and temperature of mixing process are enough to make the whole RAP bitumen become liquid. with 40% RAP. 2005). some aggregates are 64 . agglomerates of fillers and liquid bitumen. black rock cases were different (McDaniel et al. for instance. This mixture is demonstrated in Figure 18. The black agglomerates comprising different aggregate particles will be moving with virgin aggregate. the fact that total blending and actual practice were the same might not be firmly proved as in this case. where RAP was mixed with 3 rejuvenators. The study by McDaniel el al. the inner layers had lower viscosities and higher penetrations than those values of the outer ones. there will be no considerable adversity to the performance of the recycled mixture. 2000).

covered by soft rejuvenator. dynamic modulus. resilient modulus. using sections between 80 and 160 meters long. Gardiner and Brown (2000) implemented a study to assess the adverse effects of segregation on the quality and performance of asphalt mixture. tensile strength at dry and wet. there would certainly be some aggregate covered by soft rejuvenator first.6. Increased rutting and raveling. The segregation consequently occurs due to the existence of bitumen with different consistency in the mixture. After this combination is mixed with rejuvenator. Increased moisture damage due to high air voids caused by segregation. 65 . This phenomenon is also reported in Huang et al (2005). the rejuvenator must cover the RAP aggregates. This situation will also reduce the capability of diffusion process as in order to diffuse efficiently. Gardiner et al. and low temperature condition. (2000) stated that the existence of segregation in the mixture can cause substantially: Decreased fatigue life in areas that have high concentration of coarse aggregate. All the tests show the adverse effects of segregation to the performance of mixture (Table 20). the RAP bitumen transferred from RAP aggregate to virgin aggregate was 6 to 6. The data of segregation was gathered from the field. Parameters used to evaluate the effects of segregation to the performance of mixture were permeability.2 Consequences of segregation on the performance of asphalt mixture Segregation can cause adverse effects on the quality of the mixture as well as the performance of the pavement during service life. and the others covered by stiff aged binder in the mixture. (2005) reported that. during mixing RAP and virgin binder. Huang et al. The influence of segregation to fatigue life was also evaluated. some still covered by stiff RAP bitumen. RAP binder and rejuvenator are not well mixed.8%. the level of segregation compared to Job Mix Formula was classified as followed: Non-segregation: the percent passing any sieve differed less than 5% Low segregation: at least two sieves with a change more than 5% Medium segregation: at least two sieves with a change more than 10% High segregation: three sieves with a change more than 15% The test samples then were prepared in the laboratory in accordance with the difference to Job Mix Formula (Table 19). 2. especially with high volumes of traffic. Based on the data collected.

2000) Table 20: Summary of the influence of segregation on mixture properties (Gardiner and Brown. It might reduce the strength of the mixture or pavement after 66 . and target air voids of laboratory-simulated segregation mixtures (Gardiner and Brown. beside those kinds of segregation normally occuring in conventional mixtures.Table 19: Gradations. 2000) In the recycled mixture. there also chemical segregation as the rejuvenator cannot diffuse into RAP binder immediately. bitumen contents.

being paved and compacted. 67 . There is also binder segregation due to the rejuvenator not being well distributed in the mixture due to improper mixing. Carpenter and Wolosick (1980) reported that the resilient modulus of recycled mixture reduced during the first two weeks and then started to increase to the equilibrium value due to the occurrence of diffusion process.

The aggregate is supplied by Dene limestone quarry. 6 mm and dust.2 Materials 3. The viscosity of 40/60 Pen bitumen is zero shear viscosity extrapolated from DSR data using Cross model at 60oC (Section 4. Each time RAP is acquired from the industry.s) BS EN 1426 (2000) BS EN 1427 (2000) BS EN ISO 3838 (1996) DSR-ZSV 50. the homogeneity of RAP is still not assured.1). In addition.3 Laboratory RAP production 3. Gradation of each nominal size is in Table 22. The variability includes not only the gradation of RAP but also the RAP binder content and origin. The shaded areas present the aggregate sizes used in BS: 68 . Even if the amount of RAP necessary for the whole research is obtained. The fact that RAP contains materials with different or unknown origin might seriously affect the result of the research. The properties of 40/60 Pen are shown in Table 21. Although the task is time consuming.1 Bitumen The bitumen used for laboratory RAP production is 40/60 Pen bitumen supplied by Shell. the properties of RAP might be different. 3. Penetration at 25oC (dmm) Softening Point (oC) Density (g/cm3) Viscosity at 60oC (Pa.2. an agglomerate of RAP aggregate with different sizes bound together by RAP binder.6 56 1.2 Aggregate The aggregate for laboratory RAP production primarily comprises three nominal sizes. one of the objectives of this research is to study the effect of RAP sizes on mechanical properties of recycled asphalt mixture. laboratory RAP production helps to control RAP aggregate gradation as well as RAP binder content and origin.03 440 Table 21: Properties of bitumen 40/60 Pen 3.2. The purpose of laboratory RAP production is to eliminate the problematic variability of RAP materials.1 Introduction One of the issues related to using RAP from the industry is the RAP variability. The laboratory RAP production also helps to assure that every RAP piece is a lump.3. 10 mm.

98 3.48 11.4987-1 (2005).88 0.76 100 100 100 100 98.23 0.31 17.53 95. Sieve Size 31.78 0.00 0.05 0.05 0.47 2.59 23. the gradation of RAP aggregate conforms to the gradation requirement for 10 mm DBM for close graded surface course (BS:4987-1.14 4.91 5.47 81.125 0.5 20 14 10 8 6.03 0.063 Pan Sum mm mm mm mm mm mm mm mm mm mm mm mm mm mm mm Nominal size 10 Pass Retained (%) (%) 0. The physical properties of each nominal size of aggregate are shown in Table 23.69 100.61 100 100 99.00 12.11 0.00 100 87.1 1 Sieve size (mm) 10 100 Figure 34: Design gradation of RAP aggregate 69 .78 0.97 31.26 4.57 1.14 0.76 1.60 1.14 18.22 62.07 0.40 16.14 5.00 0.53 18.02 0.81 85.01 0.07 64.00 0.3 4 2.25 0.10 22.87 43.72 10.40 0.65 1.5 0.00 0.00 0.55 1.01 0.30 0.61 100.19 4.25 2.12 3.76 100.68 0. 2005) (Figure 34).88 40.61 17.01 0.19 4.80 4.07 0.87 8.12 46.21 3.08 43.8 2.32 1.03 1.79 1.11 12. After batching.00 Pass (%) Dust Retained (%) 0.00 Nominal size 6 Pass Retained (%) (%) 0.69 Table 22: Gradation of Dene limestone aggregate 100 90 80 70 Upper Limit Lower Limit RAP aggregate gradation % Passing 60 50 40 30 20 10 0 0.36 2 1 0.

87 2.533 1. For large size of RAP.2% bitumen.673 1. Maximum theoretical density of the mixture containing the designed aggregate gradation and 5.42 Table 23: Properties of Dene limestone aggregate 3. is 2453 kg/m3. the size is visually adjusted therefore the maximum dimension is less than 40 mm. each slab has the dimensions of 305 mm x 305 mm x 40 mm. The target air void content is 8%. after conditioned at 100oC for 1 hour. Based on the target air void content and the maximum density of the mixture.4 Processing RAP The artificial aged slabs are processed into two sizes. large RAP (denoted as LR) and small RAP (SR).575 2.622 1.2% by weight of the total mixture.612 2. artificially aged slabs will be broken manually.3 Procedure for RAP manufacture The bitumen content of RAP material is 5. the amount of bitumen and each aggregate fractions are calculated thus after compaction.500 2. Procedure for manufacture of artificial aged slabs is as follows: Aggregate is heated overnight at 150oC Bitumen is heated at to 150oC for 3 hours Mixing for 2 minutes at 150oC Compact the loose mixture in the mould of 305 mm x 305 mm until the thickness of 40mm is reached. The aim of the conditioning duration is to soften the RAP and eliminate the degradation of RAP aggregate. determined by BS EN 12697-5:2002.443 2. Remove the compacted specimens Condition in force draft oven at 85oC for 120 hours 3. Small RAP 70 .46 2.547 Dust 2.478 10mm 2. During the breaking process.Size Particle density on an oven dried basis g/cm3 Particle density on a saturated surface-dried basis g/cm3 Apparent particle density g/cm3 Water absorption % 6mm 2.

The gap between two jaws of the crusher at static condition is set at 15 mm hence after being crushed.5 mm 31.2 26.74 1.36 mm 1.86 3.6 9.5 mm 20 mm 14 mm 10 mm 6.38 100 Large RAP Pass Retain (%) (%) 0 100 73.38 1 0.74 44.23 0.13 18.85 1.82 15.69 32. the maximum dimension of small RAP material is about 20 mm.35 mm 2. Sieve size 50 mm 37.92 47. Figure 35 illustrates the appearance of both small and large RAP.material is obtained by crushing large RAP material using a jaw crusher.3 mm 4 mm 3.38 0. The gradation of RAP materials are presented in Table 24.51 2.18 mm Pan Sum Small RAP Pass Retain (%) (%) 0 100 0 100 0 100 7.53 13.2 0 100 Table 24: Gradation of processed RAP materials Figure 35: Appearance of RAP materials 71 .26 92.8 26.

01 RAP aggregate gradation after being crushed 0.8% to assure visco-elastic properties of RAP binder. 2005). after being extracted and recovered from RAP material (BS-EN:12697-4. is subjected to penetration test at 25oC.2 RAP binder RAP binder. 100 90 80 70 RAP aggregate gradation Upper Limit Lower Limit % Passing 60 50 40 30 20 10 0 0.1 1 Sieve size 10 100 Figure 36: Gradation of RAP aggregate before and after processing 3.1 RAP aggregate Although the degradation of RAP is deliberately eliminated by 1 hour conditioning at 100oC. the RAP aggregate gradation might still alter during the crushing process. The strain is 0. The properties of RAP binder are presented in Table 25. 72 . Viscosity of RAP binder is zero shear viscosity extrapolated from DSR data using Cross model at 60oC. The result shows the gradation of RAP aggregate after processing is almost the same as that of original (Figure 36). The test temperatures range from 4 to 45oC when 8 mm plate is used with 2000µm thick specimen and 20 to 80oC for 25mm plate and 1000µm thick specimen.3. After being extracted and recovered from small RAP material.5. RAP aggregate is subjected to particle size distribution test (BS-EN:933-1. softening point and density test.5 Determine RAP properties 3.5. Rhelogical properties are also determined by dynamic shear rheometer (DSR) under strain-controlled mode. 1997).

Penetration at 25oC (dmm) Softening Point (oC) Density (g/cm3) Viscosity at 60oC (Pa.03 1859 Table 25: Properties of RAP binder 73 .s) BS EN 1426 (2000) BS EN 1427 (2000) BS EN ISO 3838 (1996) DSR-ZSV 31 58 1.

Arrhenius. Chaffin et al. However. the more unstable 74 . The dynamic viscosity  * of all materials was measured at 60oC.05 or 0. Absolute viscosity can be conventionally determined by capillary method. at high temperatures.. it is sometimes impossible to obtain the  o of stiff bitumen due to the limit capability of the equipment (Chaffin et al. angular frequency of 0. for instance at 60oC or higher. In addition. and could * be used instead of absolute viscosity. Especially with stiff bitumen. (1995) studied the efficiency of three mixing rules. 1995). Grunberg and Nissan. for instance.1 rad/second under the geometry condition 25 mm plate and 500 µm gap. Inaccurate viscosity input might result in substantially erroneous prediction. However. in practice. Figures 37 and 38 illustrate the complex viscosities of bitumen 160/220 Pen and 100/150 Pen at 60oC versus different frequencies. 0. The viscosity of bituminous binder has been normally determined at an arbitrary temperature of 60oC. The softer the bitumen.1 rad/second. the viscosity values of soft bitumen are not stable under low shear rates. the difference between  o and absolute viscosity depends on the type of bitumen. independent of frequency or shear rate. and Epps.4 Zero shear viscosity and the accuracy of viscosity mixing equations 4. * However. on 47 bituminous materials including straight run bitumen. 2006). the difference might be considerable. Viscosity can be also determined by dynamic shear rheometer (DSR) at low frequency. bitumen fractions and commercial recycling agents. Chaffin et al (1995) claimed that the  * at this specified frequency and temperature could be considered as a low * frequency limiting complex viscosity  o . The estimation accuracy relies not only on the efficiency of viscosity mixing rules but also on the viscosity of aged and virgin binders. the limits of this approach are that it is time consuming and requires calibrations (Malkin and Isayev.1 Introduction To estimate the viscosity of the blend between aged and virgin binder is a vital part in the recycled asphalt mixture design process.

50E+03 3. it is possible to increase the angular frequency.20E+03 1.50E+03 1. this value might not represent the viscosity of bitumen.0E-01 1.0E+00 1.00E+03 Complex Viscosity (Pas) 8. 1.50E+03 2.00E+00 1.00E+02 0. To eliminate the above limitations of DSR in determining bitumen viscosity.0E-04 1.0E-05 1.0E-01 1.00E+03 4. rather than using the arbitrary temperature 60oC.0E-03 1.0E-02 Frequency (Hz) 1.0E+01 Figure 37: Complex viscosity of bitumen 160/220 Pen versus different frequency at 60oC 5.50E+03 4.00E+03 5.00E+03 Complex Viscosity (Pas) 3.0E+01 Figure 38: Complex viscosity bitumen 100/150 Pen versus different frequency at 60oC The inaccurate viscosity determination might affect the efficiency of viscosity mixing equations.0E-02 Frequency (Hz) 1. In addition.00E+02 4. the temperature can also be reduced. and using the Cross model to obtain the 75 .00E+03 2.the complex viscosity values at low frequencies.00E+03 1.0E-03 1.00E+00 1.0E-04 1. If one value of low frequency is used to * determine  o .00E+02 0.0E+00 1.0E-05 1.00E+02 6.00E+02 2.

(1995) is correct.02 g/cm3 1.low frequency limiting complex viscosity (ZSV) by mathematical extrapolation.03 g/cm3 1. The first purpose of this experiment is to use zero shear viscosity to evaluate the efficiency of different viscosity mixing equations. This means the bigger the difference between viscosity of aged and virgin binder. Therefore. Chaffin et al. Two rejuvenators are soft bitumen 160/220 and 100/150 Pen. 76 .2.02 g/cm3 Table 26: Properties of bitumens There are two mixes in this study. For both Mix A and Mix B. Based on experiment data. If the hypothesis of Chaffin et al. Bitumen Aged binder 100/150 Pen 160/220 Pen Penetration at 25oC (dmm) 31 119 192 Specific Gravity 1. In addition. 2005) and binder 40/60 Pen with target air void content of 8%. The mixture is then conditioned at 85oC for 120 hours for LTOA (Airey. the higher the absolute value of G12 . 2003). Mix A represents blends of aged binder and 160/220 Pen and Mix B represents those of aged binder and 100/150 Pen. The properties of bitumens are in the Table 26. the more the viscosity predicted by Arrhenius mixing rule deviates from the Grunberg and Nissan value.2 Experiments 4. Arrhenius is a special case of Grunberg and Nissan equation when G12 is equal to zero.1 Materials One aged bitumen and two rejuvenators are used in this study. The bigger the viscosity difference. (1995) claimed that the interaction parameter G12 in Grunberg and Nissan equation has relation with the difference between viscosities of rejuvenator and aged binder. In addition. 4. it is realized that the difference between viscosities of rejuvenator and aged binder is dependent on the temperature. 2005). RAP was produced from 10 mm DBM (BS:4987-1. the purpose of this experiment is also to evaluate the effect of temperatures on efficiency of Arrhenius and the other equations. the adjustment in testing temperatures might improve the efficiency of Arrhenius equations. The aged bitumen is extracted and recovered from artificial RAP by fractionating column method (BSEN:12697-4.

it is impossible to obtain those values by current laboratory methods such as rotational viscometry and rotational dynamic rheometry due to equipment limits. m : dimensionless material constant 77 . The complex viscosities of bitumen are measured over temperature range from 20 to 80oC under strain-controlled mode. 1965).the increment of 20% aged binder produced a total of 8 pairs of aged binder/rejuvenator in this project. The strain is chosen at 0. the Cross model for pseudo-plastic materials is used for extrapolation of the zero shear rate viscosity (Cross.2. 4. This model describes the relationship between shear rate and apparent viscosity by the following equation:     Where: o    1 k     m (12)  : viscosity  : shear rate  o : zero shear viscosity    : viscosity at infinite shear rate k : material constant.2 Rheological testing Rheological properties of each blend are determined using dynamic shear rheometer with 25 mm parallel plates and internal gap of 1000 m. All the blends are mixed manually by a small paddle for 30 seconds at 160oC to assure the homogeneity.1 Zero shear viscosity (ZSV) Zero shear viscosity or zero frequency complex viscosity is a theoretical concept. In practice.3.8% to assure bitumen working in the visco-elastic regime under frequency range from 0.1 to 10 Hz. 4.001gram. The weight of each blend is approximately 10 grams. Hence.3 Results and discussion 4. Each blend is produced by pouring the determined proportions of liquid bitumen into a glass tin by using digital balance with the accuracy of 0.

except for high temperatures.As the rotational rheological device cannot cover sufficient range of shear rate. 2002). or shear rate (Cross. for instance. At these high temperatures. The results show that the Cross model is fitted with the experimental data as almost all the R square values of the regression analysis are substantially high. rejuvenators and those blends with low percentages of aged binder are soft and the complex viscosity values are almost independent of angular frequencies.. Based on the rheological data. higher than 60oC. b : material constant (14)  : frequency 78 .   was complex viscosity at infinite frequency and  was oscillation angular frequency (Anderson et al.  0  a  b Where: a. the assumption was made that     (Sybilski. The results of extrapolation work are presented in Table A-1 for blends of Mix A and Table A-2 for blends of Mix B with different proportion of aged binder at different test temperatures (Appendix A). For these situations. Figure 39 is an example of the extrapolation work. Hence. Figure 40 illustrates the dynamic viscosities versus frequency of a blend of Mix A (20% RAP and 80% 160/220 Pen) at 70oC. a linear polynomial relation between complex viscosity and frequency (Equation 14) is used for zero frequency viscosity extrapolation instead of Cross model which requires the viscosity of material to be dependent on frequency. 1996). the Cross model was simplified as followed:  o m 1  k  (13) Cross model (Equation 13) was also used for extrapolation of zero shear viscosity from the dynamic oscillation data where  o was zero shear viscosity. the theoretical zero shear viscosity is extrapolated by using Cross model with the help of Matlab software. 1965).

Figure 39: Extrapolated ZSV of Mix A blends at temperature of 25oC 79 .

0 Complex Viscosity (Pas) 10.0 Frequency (Hz) Figure 40: Complex viscosity versus frequencies of blend (20 % RAP and 80% 160/220 Pen) at 70oC 4. For the Grunberg and Nissan equation. The difference between experimental and predicted data is evaluated by following indicators.4).3.3. For both Mixes A and B. Residuals (R).0 0.0 1.2 Efficiency of viscosity mixing equations Four viscosity mixing equations are evaluated in this study.0 4. Figure 41 illustrates the differences between experimental and predicted values using four different viscosity mixing equations. the viscosity of each blend is calculated by each of four mixing equations with the same proportion of RAP binder.0 1. The Residual represents the difference from the predicted to the experimental values and is calculated by the following equation: Ri  p i  ei (15) 80 . Arrhenius (ASTM 4887).0 2.0 5. The experiment data and predicted viscosity using four mixing equations for Mix A and Mix B are illustrated in Tables A-3 and A-4 (Appendix A).0 9. parameter G12 is determined by fitting this equation into the experiment data at different testing temperatures using Matlab.0 7. Each pair of predicted and experimental data is compared by using regression analysis with the help of Matlab software. R square value.100. ZSVs of RAP binder and rejuvenator.0 6.0 8.0 3. Grunberg and Nissan. and DLV (Section 2. and Root Mean Square Error (RMSE) (Tables 27 and 28).0 10. Epps.

the better fit. the closer the R value to zero. The closer the RMSE to 0. the better the fit. Zero Shear Viscosity (P) 10 7 Experiment ASTM 4887 G&N Epps DLV 10 6 0 10 20 30 40 50 60 70 Proportion of virgin binder (%) 80 90 100 Figure 41: Experiment and predicted viscosity using different viscosity mixing equations of Mix A (Blends of different proportion of aged binder and 160/220 pen) at 20oC The Root Mean Square Error (RMSE) is the standard deviation between predicted and experimental data sets.Where: Ri : Residual of data point i pi : predicted value ei : experimental value R is also presented by the percentage difference from the predicted and experimental value. RMSE is determined by the following equation: RMSE  p i 1 n i  ei  2 n (16) Where: 81 .

higher than 55oC. If the G12 is negative. The higher the absolute value of G12 . All viscosity mixing equations show the tendency of improving accuracy when increasing the temperature (Tables 27 and 28). 1995). This also substantiates the fact that DLV method does not perform well when soft bitumen is used as rejuvenator (Chaffin et al. the viscosity predicted using DLV method are the most deviant from actual values. This supports the finding of Irvin (1977). DLV method could estimate the viscosity of the bitumen mixture within 50% of the actual value. the residual generated by Grunberg and Nissan and Epps equation are almost similar. Maximum Residual values generated by using Grunberg and Nissan and Epps equations are generally less than 20%. the RMSE generated by Grunberg and Nissan equation are smallest at almost every test temperatures. At high temperatures. 1995).. Arrhenius equation (ASTM 4887) is a special case of Grunberg and Nissan when G12 is equal to zero. Table 29 presents the values of G12 of Mix A and Mix B blends at different temperatures. Generally. The R square values of regression analysis between four equations and experiment are almost the same. the predicted viscosities by Arrhenius equation (ASTM D4887) are within 30% of the experimental data. approximately 10%. the greater the difference the viscosity predicted by Arrhenius equation from the G&N value. However in general. Arrhenius equation overestimates the viscosity of the blend and vice versa. generates the second highest residuals. Chaffin et al (1995) 82 . The results demonstrate that on the average. Grunberg and Nissan equation again generates the lowest values of residual. these values from Epps and Grunberg and Nissan equation are almost identical. for instance. The RMSE values generated by DLV and Arrhenius equations (ASTM D4887) are in turn the first and second largest. the most popular equation for estimating the viscosity of bitumen mixture (Chaffin et al. In term of Residual values. Arrhenius equation.n : the number of data points evaluated pi : predicted value ei : experiment or actual value The results from regression analysis for blends of Mix A and Mix B are presented in Tables 27 and 28.. At temperatures higher than 45oC.

Figure 42 presents the difference between viscosity of aged binder and rejuvenator versus G12 at different temperatures. Therefore. the interaction occuring inside each blend of different bitumen binders is different from the others.. The interaction is also different in mixtures comprised of the same two bitumen but different proportions (White et al. However. The results indicate there is no relation between viscosity difference and the G12 parameter. 83 . The viscosity difference is expressed as the ratio of aged binder viscosity over that of rejuvenator. the fact that one constant value of interaction parameter G12 is used universally would result in substantial errors in viscosity estimation. Therefore. The inaccurate estimation using Arrhenius viscosity mixing equations is probably caused by interaction between aged and virgin binder. the deviation of viscosity predicted by Arrhenius equation from actual values is not attributed to the viscosity difference.claimed the variation is due to the difference between viscosity of aged binder and rejuvenator. 1970). As the bitumen has complicated chemical composition. the results in this experiment demonstrate that the difference between viscosity of aged binder and rejuvenator is not the only reason for the imprecise viscosity estimation.

0E+02 5.82 0.8 0.996 7.03E+02 75 0.3 20.8E+02 11.999 2.999 1.3 0.3E+02 29.4E+01 7.8 35.991 9.5 0.9 42.9768 2.53E+03 55 0.971 2.999 2.68E+04 50 0.963 3.66 0.999 2.856 1.9E+03 15.20 0.948 2.8E+03 10.000 2.09E+03 65 0.3 32.9E+01 3.9E+05 25.9E+04 12.9 47.5E+03 60 1.996 9.5E+05 17.997 1.1 23.993 8.2 0.8 0.9126 3.1E+03 19.999 1.8E+05 16.5E+04 16.82 0.3E+05 22.0E+01 11.1 0.5 0.48E+06 25 0.8 0.0E+04 13.997 1.3 0.9484 6.39 0.000 7.998 1.886 3.9E+04 11.991 5.999 1.1E+04 10.4 0.7 0.8 0.4 0.000 3.997 3.0E+02 6.7 0.7 0.3 0.0 0.8E+04 10.2E+03 7.1E+02 70 0.994 2.7 0.5 0.4E+01 8.997 1.993 1.6 0.999 1.9 0.6E+05 35 1.9736 5.6E+06 17.7E+05 36.0E+02 7.8936 5.5 40.2 0.9 25.5 0.3 53.5E+04 9.994 5.4 0.1 0.999 2.19E+01 Table 28: Mix B .5E+04 8.0E+04 50 0.9E+01 10.995 2.4E+03 10.1E+02 4.1E+05 22.93E+03 60 1.5 31.24E+02 70 0.981 9.3E+05 21.9719 2.1 0.4E+02 75 1.6 24.971 1.985 4.989 2.000 1.000 1.4E+01 9.8 30.996 1.6 0.998 2.5 0.5 0.3E+03 8.2E+04 8.8 0.999 1.07 0.994 2.4E+04 16.5E+02 6.943 3.8 0.2E+02 20.1 39.9 44.0E+05 27.2 0.4E+03 23.7E+06 20.9E+05 14.Regression analysis between experiment and predicted values using different viscosity mixing equations at different temperatures Grunbrerg and Nissan ASTM 4887 Epps DLV RMax Res RMax Res RMax Res Ro C square RMSE (%) square RMSE (%) square RMSE (%) square RMSE 20 0.998 1.991 2.991 5.993 1.4E+01 11.1 0.9E+03 30.997 6.951 3.8 0.2 27.1 48.9847 9.0E+06 30 0.4E+05 20.923 7.2 0.999 1.992 2.2E+01 26.973 1.4E+06 20.2 0.5E+02 26.9702 8.957 6.6E+05 25.997 7.5 0.3E+03 8.8 0.5E+05 40 0.000 1.990 1.990 6.996 6.972 1.993 3.999 1.3 0.977 1.2 0.999 9.94 3.4E+05 14.6 27.19 0.0E+03 65 0.7E+01 12.933 1.07 1.993 3.5E+03 16.23E+02 80 0.819 6.967 1.2E+01 Table 27: Mix A .8 46.7 0.983 1.93E+05 40 0.6E+06 21.998 7.997 2.3E+03 11.996 7.7 0.Grunbrerg and Nissan ASTM 4887 Epps DLV RMax Res RMax Res RMax Res Ro C square RMSE (%) square RMSE (%) square RMSE (%) square RMSE 20 0.0E+01 9.0E+03 55 1.7E+04 45 1.4 0.2 36.30E+06 30 0.0 0.1 0.9 33.3E+05 10.9E+01 7.2 0.2 0.996 9.49 0.7 0.999 1.943 8.4 0.997 5.9365 1.1E+06 25 1.4 0.Regression analysis between experiment and predicted values using different viscosity mixing equations at different temperatures Temperature Temperature Max Res (%) 48.5 0.0E+01 4.9782 1.5E+02 18.7 0.964 6.999 1.91E+04 45 0.5 33.5E+01 5.93 1.4E+01 15.980 4.996 4.4E+01 8.8 0.8E+02 11.951 3.979 2.991 3.5 0.5 Max Res (%) 37.9 0.77 0.5E+04 7.8E+01 12.5E+04 13.8E+02 23.000 1.9 0.9 0.0E+02 12.971 7.000 8.8 0.990 1.7E+01 10.3E+03 26.0 0.7E+03 22.1 84 .9621 1.8 0.996 7.000 3.8E+01 18.4 28.999 4.5 39.0 0.2 36.4E+01 80 0.5E+01 31.1E+04 10.0E+02 6.76 0.6 0.96E+05 35 0.4 0.996 8.998 1.79 0.995 2.7E+02 8.8 0.1 0.2E+03 8.987 1.

0552 -0.3950 -0.1 20 10 0.Temperature 20 25 30 35 40 45 50 55 60 65 70 75 80 Mix 1 0.4 50 Viscosity Difference 0.2260 Mix 2 0.0574 Table 29: G12 parameter of Mix A and B at different temperatures 70 0.0434 0.2 0.2253 -0.2242 -0.4 0.0916 -0.1172 -0.4509 0.2582 0.5 Mix A-V 60 Mix B-V Mix A-G Mix B-G 0.2624 -0.2160 0.3776 -0.0 Temperature (Degree Celcius) Figure 42: Viscosity difference and G12 versus temperatures G12 85 .1 0 20 25 30 35 40 45 50 55 60 65 70 75 80 0.3 0.5 0.0507 -0.3 40 30 0.2 0.0141 -0.1357 -0.3276 -0.1719 0.1206 0.2477 -0.2724 -0.3121 -0.0995 -0.2512 -0.2128 -0.

During this process. the combination of RAP and virgin aggregate is mixed with virgin binder in Step 3.5 Effects of laboratory mixing methods on the homogeneity of hot recycled asphalt mixture 5. 5. Then. The heat source for separating RAP lumps also comes from the mixer maintained at mixing temperature. The homogeneity of recycled mixtures is studied by using virgin binder with different colour from that of RAP binder. this step is also to ensure that the rejuvenated binder is well distributed all over the mixture and coats 86 .Step 1: Virgin aggregate is superheated to predetermined temperature.2 Development of laboratory mixing protocol It is essential that the laboratory mixing protocol should duplicate the mixing mechanism occurring in the actual asphalt mixing plants. The purpose of Step 2 is to use the heat from superheated virgin aggregate to heat up and soften RAP materials.1 Introduction This chapter contains the development of a new mixing protocol that simulates the mixing mechanism between RAP and virgin materials in asphalt mixing plants. The proportion of pigment is 10% by weight of the binder making this binder red. . The effects of different RAP/superheated virgin aggregate mixing durations and RAP sizes on homogeneity of hot recycled asphalt mixture are considered. The laboratory mixing process should include the following steps: .Step 2: Superheated virgin aggregate is mixed with RAP material at ambient temperature in the mixer maintained at mixing temperature.Step 3: the combination of superheated virgin aggregate and RAP material is blended with virgin binder in the mixer maintained at mixing temperature. The red colour of virgin binder is obtained by mixing clear binder with iron oxide pigment. In addition. The content of this chapter also includes the correlation between homogeneity and stiffness distribution of hot recycled asphalt mixtures. Step 1 is to enable virgin aggregate enough thermal energy to heat up the RAP from ambient to mixing temperature. separating RAP lumps into single particles covered by RAP binder. The aim of this step is to assure that virgin binder can incorporate and rejuvenate the RAP binder. RAP bitumen is also transferred onto the surfaces of virgin aggregate particles. .

The amount of heat (Cutnell and Johnson. The efficiency of the laboratory mixer is quite different from that of the industrial mixer for asphalt production. the bulk of materials is moving circularly under mechanical effect of mixing paddles. on the contrary. The laboratory production of recycled asphalt mixture requires the lid of the mixer to be opened several times during manufacture process for material intakes. the superheated temperature can be estimated.2. besides mechanical mixing effects. In actual asphalt mixing plants. 5. this duration is enough for virgin binder to coat all the aggregate particles. there is primarily horizontal and inconsiderable vertical movement. In fact. the movement of each RAP particle due to gravity also enhances the separating progress of RAP lumps. RAP.1 Estimation of superheated temperature of virgin aggregate The amount of heat required to raise the RAP at ambient temperature to the mixing temperature is equal to the amount of heat dispersed from virgin aggregate so the temperature drops from superheated to mixing value. Once the heat transferred from superheated virgin aggregate weakens the bitumen bond. Once the heat transferred from superheated virgin aggregate softens the RAP binder. The mixing time for Step 3 is 2 minutes due to laboratory experience with the mixer.every single aggregate particle. RAP particles are separated due to the mechanical effect of mixing paddles and external friction among surface of particles. The breaking progress is not enhanced by the vertical movement due to gravity of RAP materials. single virgin aggregate and RAP particles experience centrifugal force until this force is smaller than gravity and materials start to fall downward. In the actual asphalt mixing plant. the continuous production process also assures the heat conductivity between superheated virgin aggregate and RAP material being more efficient than that in the laboratory. Based on the quantities of virgin aggregate. specific heat of RAP and virgin aggregate. For the manufacture of conventional asphalt mixture. 2004) required to raise the temperature of a mass is: Q  mcT2  T1  (17) 87 . In the laboratory mixer.

might not be applicable in the laboratory mixer.2. 6. 6 and 8 minutes for large RAP mixture and 1. Figure 43 illustrates that at 2 minutes mixing time.2 Determine the superheated virgin aggregate/RAP mixing duration Due to the efficiency difference. The adjustment method is to start from short mixing time and gradually increase the mixing time until the change in RAP lumps size is insignificant. magg : Quantities of RAP and virgin aggregate in the mixture (kg) cRAP. Ta : superheated temperature (oC) for virgin aggregate. 2. 4. for instance. thermal energy transferred from superheated virgin aggregate and from the mixer will soften the bitumen bond and separate RAP lumps into single particles. mixing temperature. the superheated virgin aggregate/RAP mixing duration is adjusted based on the mixer used for the research. 4.Where: m : the mass in quantity (kg) c : specific heat (kJ/kg oC) T1 . During the mixing duration. Tm .T2 : current and desired temperatures (oC) The amount of heat required to raise temperature of RAP from ambient to mixing temperature is: Q1  m RAP c RAP Tm  Ta  (18) The amount of heat dispersed from superheated virgin aggregate so the temperature reduces from superheated to mixing temperature is: Q2  magg c agg Ts  Tm  (19) Where mRAP. the RAP still remains approximately the original size. 88 . 5. To investigate the effect of mixing on the properties of hot recycled mixture. The mixing temperature will be determined in the next section. cagg : specific heat of RAP and virgin aggregate (kJ/kg oC) Ts . the short mixing time in the real industrial mixer. Consequently. the RAP/virgin aggregate mixing duration is decided to be 2. Figure 44 is for 8 minutes mixing time when the change in RAP size under mechanical mixing is negligible. 8 minutes for small RAP mixture. 60 second. ambient temperature of RAP.

Figure 43: RAP size after 2 minutes RAP/superheated virgin aggregate mixing time Figure 44: RAP size after 8 minutes RAP/superheated virgin aggregate mixing time 89 .

The method using infrared scanner can only classify between the bitumen-rich and aggregate-rich areas due to the difference in thermal properties (Gardiner and Brown. Although the density of iron oxide is quite different from those of clear and RAP binder. and 180oC. 2000). The temperature at 0. new material and additive if applicable. The optimum mixing temperature is the temperature at which the viscosity of virgin binder has the viscosity value of 0. the segregation that occurs in recycled asphalt mixture includes not only the different concentration of aggregate sizes and bitumen but also the distribution of RAP material. aggregate and bitumen due to the significant difference in density of these components. Unfortunately.. the segregation detection method based on density. X-ray scanner. Similarly. Hence. The proportion of the pigment is 10% by weight of the clear binder.5. 150. 1973).s (Read and Whiteoak. for instance. Meanwhile. Therefore. The viscosity of virgin binder is determined by Brookfield viscosity test at 120. Figure 45 illustrates the images taken by X-Ray photo machine and normal digital camera of RAP binder and Shell clear binder (Mexphalte C 160/220 pen) dyed by red iron oxide pigment.2. the X-ray scanner image cannot show any difference between RAP binder and clear binder dyed by iron oxide pigment.25 g/cm3 to 1. these kinds of equipment cannot identify the position of RAP and new binder as the density of these binders are approximately the same.03 g/cm3.2 Pa. 5. 2003).3 Method for segregation evaluation The primary characteristic attributed to the difference between conventional asphalt and recycled asphalt mixture is reclaimed material (RAP). RAP binder 90 .3 Determine the mixing temperature The mixing temperature in Step 3 is determined based on the virgin binder used as rejuvenator. most measuring and segregation detection techniques for conventional asphalt mixture are not applicable or not efficient for recycled asphalt mixture. the segregation level might considerably affect the quality of recycled asphalt mixture and is determined mainly on how well new and RAP materials are mixed (Tia et al. 1980). In this experiment.s viscosity is interpolated by linear relationship between double log viscosity and temperature (Heukelom.2 Pa. a combination of aged binder and aggregate. mixing temperature is 135oC 5. can only identify the location of air void. However.

The result of this method is only correct in cases where the detected areas contain RAP binder and recycled agent. On the contrary. RAP material is artificially made in the laboratory. virgin binder. In this research. However. the blend of synthetic clear binder (Shell Mexphalt C 160/200 Pen) and 10% of iron oxide by weight of the binder. The hypothesis of using virgin binder with different colour is if RAP material is not well mixed with virgin binder. The preparation of testing specimens conforms to the procedures in Section 5.is visually recognized as pure black and Mexphalte C dyed by 10% iron oxide has colour of red (Figure 43).. The colour of RAP binder will be purely black. 1983) can identify the location of recycling agent due to the detection of chemical tracer. In addition. Figure 45: Images taken by X-Ray scanner and normal digital camera of Shell Mexphalte C dyed by 10 % iron oxide and RAP binder The dye chemistry method (Lee et al. the undetected areas contain not only aggregate but also RAP material as RAP lumps. the areas of black (RAP materials) and red (virgin material) are clearly visualized. and the undetected areas contain only aggregate.4 . will have the colour of red. the phenomenon that RAP materials are not fully separated 91 . the RAP should be clearly identified from virgin materials. in undetected areas this method cannot classify virgin aggregate and RAP lumps. Unfortunately. To assess the effect of mixing process on homogeneity of recycled mixture.

2 Mixing procedure The mixing process is implemented in the Mixer A (Section 6. 4. The stiffness data is then statistically analyzed in conjunction with RAP (or virgin material) distribution pattern to characterize the correlation between mixture homogeneity and mechanical properties. denoted as LR and SR. 5. 4. After manufacture.4. surfaces of slices cut from specimens are recorded by digital camera.2). 2. both large and small sizes. 2004). The bitumen content of RAP and recycled mixture are the same. 92 . Samples for visual segregation assessment are also subject to mechanical properties evaluation. The mixing durations for large RAP (LR) are 2.4.4 Specimens preparation 5. 8 and 1. Virgin binder is preheated at 135oC for 2 hours  Step 2: RAP. 5. The mixing procedure is determined as follows:  Step 1: Virgin aggregate is superheated to 215oC for 8 hours.1 Materials The proportion of RAP is 40% in the recycled mixture. The amount of pigment is 10% by weight of virgin binder to make the color of virgin binder red. The purpose is to link the homogeneity level and the mechanical properties of recycled asphalt mixture. 6. 6. are blended with superheated virgin aggregate at ambient temperature in the mixer maintained at 135oC for different durations. The analysis of these surfaces based on vertical order help to understand the distribution of RAP and virgin materials in the mixture in a 3D manner. Virgin binder is prepared by preheating synthesis clear binder (Shell Mexphalt C 160/220 Pen) to 135oC and blending with iron oxide powder. 5.into single aggregate particles covered by RAP binder but exist as RAP lumps in the recycled mixture is also easily identified. 2005).2. Aggregate gradation of recycled mixture is designed the same as that of RAP (BS:4987-1. Stiffness of each sample is measured by Indirect Tensile Stiffness Modulus (ITSM) test at 20oC (BS-EN:12697-26. 8 minutes for small RAP (SR).2% by weight of total mixture.

This is aimed to investigate the relationship between the heterogeneity and stiffness distribution of recycled mixtures with different mixing durations. 5. Two samples with approximately 40 mm thick. a cylinder of 100 mm diameter and 100 mm height is cored from each compacted specimen. RAP binder (black). 5. 5. Figures 50 to 54 show the surfaces of slices from specimens prepared from small RAP with different mixing times. The amount of materials is estimated so the final cylindrical specimen has the diameter of 150 mm and 100 mm height. The surfaces of these samples are photographed by digital camera for visual segregation assessment.4. 2004).5 Results and analysis 5. Surface image of each slice is also recorded by digital camera for further visual analysis. Before being poured into the mixer.1 Visual assessment for segregation Figures 46 to 49 illustrate the surface characteristic of specimens manufactured with large RAP and different mixing times. are cut from this core for stiffness evaluation (BS-EN:12697-26.4.3 Compacting procedure The loose mixture is compacted by gyratory compactor at 130oC until the targeted density is obtained (BS-EN:12697-31. Both surfaces of specimens for stiffness evaluation and surface of slices cut from these specimens after stiffness evaluation are included. Due to the colour difference between that of virgin (red). the locations of these components are easy to visualize. samples 1 and 2. Step 3: the blend of RAP and virgin aggregate is mixed with virgin binder for 2 minutes.5. 2004).4 Machining specimens for segregation assessment After being de-moulded the next day. After stiffness measurement. The advantage of using virgin binder with different colour from that of RAP binder is to help identify the location 93 . The surfaces analysis of slices in vertical order will give a clear picture of how RAP lumps are distributed in the cylindrical specimen. and aggregate. The slices are in order from the bottom to top. virgin binder is well stirred manually to prevent the settlement of iron oxide at the bottom of the container. The targeted air void content of each specimen is 4%. these samples are sliced up into slices approximately 10 to 15mm thick.

RAP lumps are gradually separated into single pieces. This helps to better understand whether virgin binder can incorporate and rejuvenate the properties of RAP binder or RAP might work as inert lumps in the recycled mixture. In addition. the thermal energy transferred from hot virgin aggregate will increase the temperature of RAP materials. for instance. In addition. Figures 46 to 49 show that when the mixing duration increases. the bonding strength among RAP aggregate particles will be weakened. RAP lumps are almost inert and scattered in specimen at approximately their original size. 2004). in the other slices. mixing temperature and mechanical effect.of RAP lumps which is impossible with the other methods. especially recycled mixture with large RAP. However. including filler stuck together by bitumen. The homogeneity of recycled mixture is clearly improved once the RAP/superheated virgin aggregate mixing duration is extended. Therefore. The effect of mechanical mixing process plus the weakening of bitumen bond between RAP aggregate particles slowly disintegrates the RAP lumps into smaller sizes. In fact. for instance. the factors that manipulate the efficiency of RAP separating process are. Once RAP material is mixed with superheated virgin aggregate. Actually. depending on size of material and contact surface (Cutnell and Johnson. the more thermal energy is transferred from superheated virgin aggregate. For recycled mixture with large RAP and short mixing time. these areas contain primarily virgin materials. Under the mechanical effect of the mixing paddles. the presence of RAP is negligible (Figures 46 a and h). which is the agglomerate of single aggregate particles. the efficiency of mixing process is also attributed to mixing duration. This phenomenon is as expected as the longer the mixing time. the size of RAP areas on the surface of slices are large (Figure 46). In some slices. for instance 2 minutes. This means the density of RAP are extremely high. This is because the mixing duration (energy) is not enough to degrade RAP lumps into separate particles. due to the fact that the energy transfer process occurs gradually. the size of RAP areas in the surface of each slice gradually becomes smaller. The area containing just black binder and aggregate will be definitely RAP materials. With large RAP recycled mixtures. (Figures 46 b and c). the occurrence of RAP is intensive. RAP materials exist as lumps. the mixing 94 . The surface images show that RAP/superheated virgin aggregate mixing duration seriously affects the homogeneity level of recycled mixtures. Due to the inherent properties of bitumen. RAP lumps are not well distributed in the specimen.

Figure 46 shows the surface images of recycled mixture made of large RAP. the thermal energy transferred from superheated virgin aggregate and the mixer increases the temperature of RAP binder and progresses the rejuvenation between RAP and virgin binder. the mixing time or the thermal energy from superheated virgin aggregate is not enough to 95 . the more single pieces of RAP are separated from RAP lumps. Hence. the colour of virgin binder will somehow be changed. 2 minutes mixing duration and Figure 50 for mixture composed of small RAP. RAP binder cannot be diffused efficiently by virgin binder unless RAP binder is heated up enough or activated using thermal energy from virgin aggregate and mixer during the mixing duration. it is assumed that if the RAP binder is rejuvenated by virgin binder or these two binders are integrated. 2003b). the areas of pure black and red binder are also reduced. The surface images show the same tendency for both large and small RAP recycled mixtures. 60 seconds mixing duration. the higher the temperatures. The red colour of virgin binder will be darker due to the black effect of RAP binder. During the mixing process. In addition. This indicates more integration between RAP and virgin binder. The colour of virgin binder almost remains unchanged as original red. This is because the diffusion of virgin into RAP binder is significantly influenced by the temperature. This is as expected as the longer the mixing time. The surface images also demonstrate that RAP/superheated virgin aggregate mixing duration significantly affects the integration between RAP and virgin binder. the longer the RAP/superheated virgin aggregate mixing duration. The rejuvenating process is enhanced not only by the total exposed surface of RAP particles but also the temperature of RAP binder itself.efficiency is enhanced as more active RAP particles are involved in the mixing process and well distributed around the mixture. the darker the red color of the virgin binder. During the RAP/superheated virgin aggregate mixing duration. the total exposure surface of RAP particles for rejuvenating process increases and more RAP binder is rejuvenated by virgin binder. In fact. Visually. the separation of RAP lumps also offers the chance for RAP binder to be rejuvenated by virgin binder. the more efficient the diffusion process (Karlsson and Isacsson.

This circumstance might be worse under really short mixing time. there is still a considerable portion of RAP works as lumps (Figures 55 and 56). Hence. all the RAP lumps can be disintegrated as single pieces covered by RAP binder and well distributed over recycled mixture. for instance 2 minutes. the more thermal energy and mixing effort required for separating RAP lumps. It is reasonable as the larger the RAP size. However. 96 . The surface images show that the size of RAP seriously affects the homogeneity level of recycled mixture. the areas of inert RAP lumps of samples composed of large RAP are considerably larger and more intensive than those of small RAP (Figures 46 and 51). It is expect that when the mixing time is extended.activate the RAP binder for diffusion. RAP binder is inert and there is not considerable integration between RAP and virgin binder. With the same RAP/superheated virgin aggregate mixing duration. in the industrial mixer. This validates the fact that “complete blending” between RAP and virgin binder assumed in the design process never occurs during mixing. about maximum 90 seconds total. even when small RAP is used with 8 minutes RAP/superheated virgin aggregate mixing duration. the mechanical properties of recycled mixture might be not as consistent as predicted.

RAP lump a b c d e f g Figure 46: LR mixture – 2 minutes mixing time h 97 .

a b c d e f g Figure 47: LR mixture – 4 minutes mixing time h 98 .

a b c d e f g Figure 48: LR mixture – 6 minutes mixing time h 99 .

a b c d e f g Figure 49: LR mixture – 8 minutes mixing time h 100 .

a b c d e f g Figure 50: SR mixture – 1 minute mixing time h 101 .

a b c d e f g Figure 51: SR mixture – 2 minutes mixing time h 102 .

a b c d e f g Figure 52: SR mixture – 4 minutes mixing time h 103 .

a b c d e f g Figure 53: SR mixture – 6 minutes mixing time h 104 .

a b c d e f g Figure 54: SR mixture – 8 minutes mixing time h 105 .

Figure 55: Large version of Figure 54 g 106 .

Figure 56: Large version of Figure 54 h 107 .

the surfaces of specimens for stiffness evaluation demonstrate there is considerable segregation in recycled mixture. The stiffness of each specimen is measured by indirect tensile stiffness modulus test (ITSM) (BS-EN:12697-26. specimens composed of large RAP with different mixing durations were also subjected to mechanical properties evaluation. the stiffness values of each specimen are measured by four directions along the circumference (Figure 57) with the hypothesis that the specimens with higher level of heterogeneity might have greater variations among stiffness at different directions. 2004). The aim is to link together the homogeneity level and mechanical performance of recycled asphalt mixture. it is also realized that stiffness variation with different measured directions of most specimens are not satisfied by the standard. This issue might be attributed to the heterogeneity of recycled mixture.5. 124 milisecond risetime and 5 m horizontal diametral displacement to ensure that the specimen responds as an elastic material. The test is carried out at temperatures of 20oC. However. 1 2 Specimen 3 4 Figure 57: Stiffness measurement scheme 108 . stiffness of each specimen is the mean value of stiffness measured at two perpendicular directions unless the difference between these two values is greater than 10% of the mean. In addition. Therefore.5. Conventionally.2 Mechanical assessment Together with visual assessment for segregation.

is 1540 MPa. Likewise.6 8 5. 1452 MPa. For sample 2 (2 minutes mixing time). For short mixing time. for 2 minutes mixing time.Mixing Time (mins) 2 4 6 8 1 2 1 2 1 2 1 2 Samples Air void Content % 4. for sample 1 (4 minutes mixing duration).9 5. the variations in stiffness values do occur not only in the same specimen but also in different specimens with similar mixing time. The coefficient of stiffness variation is 20%.5 Stiffness in Different Directions (Mpa) 1 1813 1306 1452 2324 1690 2147 1977 1924 2 2232 1197 2816 1959 1591 2106 2098 1950 3 1778 1374 2004 1774 1408 2002 1999 1881 4 2078 1844 2183 1491 1472 1850 1873 2175 Mean Stiffness (MPa) 1975 1430 2114 1887 1540 2026 1987 1983 COV (%) 11 20 27 18 8 7 5 7 Table 30: Stiffness versus different RAP/superheated virgin aggregate mixing duration Table 30 presents the results of stiffness values versus different mixing times. the stiffness value difference is not only attributed to the variation of air void content. Stiffness decreases once the air void content increases (Tayebali et al. 1994). For instance. significantly different from that of sample 2. mean stiffness of sample 1 is 1975 MPa compared to 1430 MPa of sample 2. and the coefficient of variation is 27%. The data demonstrates that mixing time significantly influences the stiffness variation. However. 2026 MPa. Figure 58 also shows the tendency that stiffness generally decreases once the air void content increases. The reason for the stiffness variation might be due to the difference of air void content of specimens.5 6. For example. the differences among stiffness values obtained from the same specimen are considerable. sample 2 (2 minutes mixing time) has stiffness value of 1844 MPa (measuring direction 4) which is considerably higher than regressed 109 . Similarly. the mean stiffness of sample 1 (6 minutes mixing time).8 6 5..5 6. In addition. maximum and minimum stiffness values are in turn 2816 MPa. the maximum stiffness value is 1844 MPa while the minimum is 1197 MPa.

RAP and virgin materials are not mixed together vertically.7316 2500 Stiffness Modulus (MPa) 2000 1500 1000 Mixing duration 2 minutes 4 minutes 6 minutes 8 minutes 8 9 500 4 5 6 7 Air Void Content (%) Figure 58: Stiffness modulus versus air void content The variation in stiffness values is attributable to the heterogeneity of recycled mixture. Figures 46 and 47 show that although came from the same compacted specimen. RAP lumps are not fully disintegrated into single particles coating by RAP binder. specimen 1 (at the bottom) tends to 110 . although the stiffness of sample 1 (2 minutes mixing time) is expected to be high with 4. 3000 y = -203. This situation is exaggerated by the manufacturing process. This results in the situation that when the loose mixture is transferred to the mould for compaction. the virgin binder that should be in contact and rejuvenate the RAP binder could not be in the right place and fulfill its function. it is quite low with 1813 MPa (measuring direction 1) and 1778 MPa (measuring direction 3).stiffness of 8% air void content. For short mixing duration.1 R2 = 0.6% air void content. During mixing process. RAP lumps tend to move first and settle down at the bottom of the mould. On the contrary. This results in some areas with high concentration of RAP while the others contain primarily virgin materials. There is a lack of vertical movement of material in the mixing bowl due to the properties of the mixer. RAP lumps tend to move up to the surface due to bigger sizes (Figure 59). sample 1 (4 minutes mixing time) has stiffness of 2816 MPa (measuring direction 2) which is considerably higher than expected with 5.66x + 3110. Therefore. Similarly.5% air void content.

Figure 59: Small particles tend to move downward to the bottom during mixing process It has been maintained that the uneven distribution of RAP materials results in the variation of stiffness values.have considerably higher proportion of RAP than specimen 2 (on top). In addition. This is because the material distribution patterns are not consistent along the vertical direction of the specimen. the characteristics of these surfaces are all different. For different mixing durations and RAP 111 . the stiffness of specimen 1 is also considerably higher than that of specimen 2 (Table 30). An effort has been made to investigate and quantify this relationship by correlating:   the stiffness: determining the stiffness of specimen at four measuring directions and RAP (or virgin materials) distribution pattern: by visual surface assessment of the stiffness-measured specimen and slices machined from this specimen. there is no clear correlation between the distribution of RAP (or virgin materials) and stiffness values. However. Unfortunately. Figures 46 a and d illustrate the surfaces at both ends of specimen 1 (made from large RAP and 2 minutes mixing time) and b and c are for the surfaces of the slices machined from this specimen.

The stiffness data demonstrates the tendency that the homogeneity of recycled mixture is substantially enhanced once the RAP/superheated virgin aggregate duration is extended. large and small. The mixing mechanism includes the following steps:   Virgin aggregate are superheated RAP material at ambient temperature is blended with superheated virgin aggregate. the more likely and effective the rejuvenating process. The heat transferred from superheated virgin aggregate helps to soften RAP agglomerate. Table 31 shows that the general coefficient of variation significantly decreases from 22% to 5% once the mixing duration increases from 2 to 8 minutes.. Under mechanical mixing effort. If the relation between stiffness values and materials distribution patterns are evaluated based only on the surfaces at both ends of specimen (Lee et al. weakening the bitumen binding among RAP aggregate particles. the result will be certainly different. these RAP lumps will be disintegrated and blended with virgin aggregate. Mixing time (minutes) 2 4 6 8 Mean Stiffness (MPa) 1702 2000 1783 1984 Standard Deviation 374 449 286 105 Coefficient of Variation (%) 22 22 16 5 Table 31: Mean stiffness versus RAP/superheated virgin aggregate mixing duration 5. 112 . the phenomenon is the same.materials. Not only the coefficient of stiffness variation in each specimen decreases (Table 30) but also does the general coefficient of variation. the more the RAP material is disintegrated.  The RAP/superheated virgin aggregate blend will be mixed with virgin bitumen or rejuvenator. 1983).6 Summary The newly developed mixing method has duplicated the mixing mechanism that really occurs in the industrial asphalt mixing plant.

especially RAP binder. RAP material. the relation between mixing effort and homogeneity level (1) could not be quantified and neither could that between homogeneity level and mechanical properties (2). Hence. 113 . However. homogeneity and mechanical properties of recycled asphalt The relation among mixing effort. is not fully blended or rejuvenated by virgin binder. stiffness measurement proves to be potential tool to investigate the effect of mixing effort on mechanical properties of hot recycled mixtures. This phenomenon is different from the assumption in the recycled mixture design process where RAP and virgin binder are fully blended. In this preliminary experiment. the more RAP and virgin materials are incorporated and hence the homogeneity level of recycled mixture is enhanced. Each factor will mutually influence the others. The visual assessment also demonstrates that even when small RAP size is used and at considerable mixing duration. and mechanical properties of hot recycled mixture is illustrated in Figure 60. The more mixing effort. there is a relation between mixing duration and stiffness values. Mixing Effort 1 3 Homogeneity Level 2 Mechanical Properties Figure 60: Relation between mixing effort. the more homogeneity and less variation in mechanical properties. homogeneity level.Different mixing efforts (different RAP/superheated virgin aggregate mixing durations) have depicted a clear picture of mixing mechanism between RAP and virgin materials. Due to the fact that the RAP and virgin materials distribution do not follow any consistent pattern. the relation between mixing effort and mechanical properties (3) can be quantified. There is no clear numerical parameter or values to characterize these relations. The longer the mixing time.

The preliminary investigation of the effect of mixing process on mechanical properties of hot recycled asphalt mixture shows the likely tendency that the more mixing. especially RAP and virgin binder. In addition. the less variation in stiffness values. Not enough mixing effort results in some places containing primarily RAP and the others are dominated by virgin materials. However. This might be due to the heterogeneity of recycled mixture. this preliminary statement needs to be further investigated and verified as the number of testing samples is limited to 2 for each RAP/superheated virgin aggregate mixing duration. 114 . due to visual assessment. The difference among RAP and virgin materials properties. contributes to the considerable variation in stiffness values. the virgin binder is dyed by iron oxide (10% by weight of virgin binder) which might affect the mixing characteristic of virgin binder and the stiffness distribution.

115 . the proportion of the pigment certainly alters the flow characteristic of binder. Specimens are manufactured using the Mixer A (Section 6. In FS method.2. 6. The stiffness distributions of recycled asphalt mixtures manufactured by different methods are statistically compared to each other and those of BR and CB mixtures to investigate the effect of mixing procedures and RAP materials.1 Effects of mixing protocols and RAP sizes on stiffness of hot recycled asphalt mixture To investigate the effect of mixing protocols on stiffness distribution. Recycled asphalt mixtures are manufactured by different methods. In field simulation method (FS). and the conventional method (from SHRP). The stiffness of recycled specimens are measured by indirect tensile stiffness test (ITSM) (BS-EN:1269726. recycled specimens are manufactured by different methods. The summary of experiments to investigate the effect of mixing protocols on stiffness is shown in Table 32. large and small are used (denoted as LR and SR). the SHRP procedure. Both sizes of RAP.1 Introduction As the clear binder is dyed red by 10% by weight of iron oxide.2. black rock (BR). complete blending (CB). this chapter also investigates the effect of different mixing methods and RAP sizes on stiffness of recycled asphalt mixtures. 2004). RAP/superheated virgin aggregate mixing duration varies from 2 to 8 minutes. The mixing methods include the newly developed method that duplicates the mixing mechanism in the industrial mixer (denoted as FS: field simulation) with different RAP/superheated virgin aggregate mixing durations.6 Effects of mixing procedures and RAP materials on stiffness distribution of hot recycled asphalt mixtures 6. normal straight run bitumen 160/220 Pen is used as virgin binder.1.2.2 Experiment design 6. and the field simulation method (FS) developed in Section 5. This might affect the mixing process and rejuvenation between virgin and aged binder. The procedures are explained in Section 6.3.2). the RAP/superheated virgin aggregate mixing durations range from 2 to 8 minutes. However. Both large and small RAP materials are used in this experiment. Therefore.

The aim is to compare the mechanical properties of control mixtures with recycled mixtures manufactured by different protocols. due to the rotating axis of mixer B being not vertical (90o) but 60o compared to the ground plane. the movement of material 116 .2. However. There is a thermocouple attached to the oil to control the heat supply hence the mixer can be maintained at the required temperature. and “Complete Blending” (CB) assumed in the design process in which RAP binder is fully interacted with virgin binder or rejuvenator. Mixer B The schematic of mixer B is presented in Figure 62. The difference in mechanical properties compared to those of control mixtures will demonstrate the effects of mixing procedures and properties of RAP materials on the quality of hot recycled asphalt mixtures. are employed. The characteristics of each mixer are as follows: Mixer A The schematic of mixer A is illustrated in Figure 61. mixers A and B.2 Effect of mixing equipment on stiffness distribution of hot recycled asphalt mixtures Two mixers with different mechanical mixing effects. There are two mixing paddles moving with different orbits that help to drive and blend materials in the mixing bowl. The heat supply of this mixer is maintained by heating the oil that moves between the external and internal walls of mixing bowl. RAP/Virgin aggregate mixing duration (minutes) LR FS SHRP SR FS SHRP 2 × 4 × × 2 × 4 × × RAP/Virgin aggregate mixing duration (minutes) 6 × 8 × 6 × 8 × Table 32: Test plan to study the effects of mixing method on stiffness 6. There are four mixing paddles used to guide and blend materials in the mixing bowl.The experiment also includes the stiffness measurement of control mixtures that present the “Black Rock” (BR) case where RAP binder is inert.

inside the mixing bowl consists of not only horizontal but also vertical direction. the heat supplied for the mixer B is controlled by the thermocouple that measures the air inside the mixer. Primary Axis Lid Mixing Bowl Rotating Orbit Secondary Axis Thermocouple Mixing Paddle Materials Figure 61: Schematic of Mixer A Lid Rotating Orbit Thermocouple Mixing Paddle Materials Figure 62: Schematic of Mixer B 117 . Mixer B also allows reverse rotation. Also different from mixer A.

RAP binders are extracted and recovered from RAP materials. This might alter the properties of RAP binder.3 Effect of mixing protocols on RAP binder properties Different mixing protocols will generate conditions that RAP materials are exposed at high temperature for a certain period of time. Mixer Type A B RAP Sizes LR SR LR SR Mixing time (minutes) 2 × × × × 6 × × × × Table 33: Test plan to study the effects of mixer on stiffness of recycled asphalt 6. Depending on the exposure condition (Section 6. The rheology properties of these conditioned RAP materials are studied by Dynamic Shear Rheometer (DSR) and compared to that of original RAP binder to quantify the effect of mixing procedure on properties of RAP.3.2.The summary of experiments to study the effect of mixers on stiffness distribution of hot recycled mixtures is presented in Table 33.1). RAP/Virgin aggregate mixing duration (minutes) LR FS 2 × SHRP × RAP/Virgin aggregate mixing duration (minutes) SR FS 2 × SHRP × 8 × 8 × Table 34: Test plan to study the effects of mixing methods on binder properties 118 . The summary of experiments to investigate the effect of mixing protocol on properties of RAP binder is shown in Table 34.

6.021 64 70/100 Pen 83 47. This is to eliminate the effect of different crude oil sources on mechanical properties of recycled asphalt mixture. both large (LR) and small size (SR).3 Materials and specimens manufacture 6.1 Material preparation and mixing procedure Hot recycled mixtures In this experiment. all the bitumen binders used for the whole research come from the same crude oil source. the properties of recycled binder are almost similar to those of the 70/100 Pen bitumen. 160/220 Pen Penetration at 25oC (dmm) Softening Point (oC) Density (g/cm3) Viscosity at 60oC (Pas) 192 37. Shell bitumen 160/220 Pen is used as rejuvenator.2% and the aggregate gradation of RAP is the same as that of virgin aggregate (Figure 34). The amount of RAP in recycled mixture is 40% thus the viscosity of recycled blend is approximately similar to that of 70/100 Pen bitumen (Table 35). Field Simulation method (FS)  RAP materials. The result shows that if aged and virgin binder is completely mixed. The proportion of RAP in the recycled asphalt mixture is estimated using Grunberg and Nissan viscosity mixing rule at 60oC. are conditioned at room temperature.4 1. As bitumen content of RAP material is 5. large and small.028 192 Table 35: Properties of bitumen 160/220 Pen and 70/100 Pen The materials for recycled mixtures conform to the requirements for surface course 10 mm DBM (BS:4987-1.2% of the total weight of virgin materials. The properties of 160/220 Pen bitumen are showed in Table 35. amount of rejuvenator required is also 5. Both sizes of RAP.3. Although the origin of bitumen is not allowed to be revealed due to Shell Global policy. The room temperature is maintained by thermal control system at 25oC. 119 . Figure 63 illustrates the master curves of 70/100 Pen and recycled blend with aged/virgin binder ratio of 4/6.2 1. 2005). are used and denoted as LR and SR.

The combination of RAP and virgin aggregate is then blended with virgin binder for 2 minutes. The mixer temperature is maintained at 135oC.0E+05 Log Reduced Frequency Figure 63: Master-curves of 70/100 Pen and recycled blend with 40% RAP and 60% 160/200 Pen bitumen 120 . RAP material is mixed with superheated virgin aggregate for 2. 8 minutes. Rejuvenator is preheated at 135oC for 2 hours.0E+00 1. The combination of RAP and virgin aggregate is then blended with virgin binder for 2 minutes. are conditioned at 110oC for 2 hours. 6.0E+05 1.0E-07 40% RAP Recycled binder .0E-03 1.Phase 20 Angles 10 70/100 Pen . Preheated RAP material is mixed with preheated virgin aggregate for 30 seconds in the mixer maintained at 135oC.0E+06 1.0E+04 1. SHRP method      RAP materials.0E+03 1. both large (LR) and small size (SR). Virgin aggregate is conditioned at 150oC for 8 hours.0E+01 1. 1. 4.Complex Modulus 40 30 Phase Angle (Degree) 1.0E+07 Complex Modulus (Pa) 90 80 70 60 50 40% RAP recycled binder Complex Modulus 70/100 Pen .0E-05 1.0E+09 1.0E+08 1.0E+03 1.0E+01 1.     Virgin aggregate is superheated at 215oC for 8 hours.0E+02 1.0E-01 1. The mixer temperature is maintained at 135oC.Phase Angles 0 1.

the same as the proportion of RAP material in hot recycled mixture. instead of batching recovered and virgin aggregate. 2005). The preparation and mixing procedure for Black Rock mixture are as follows:     Batched aggregate is preheated at 135oC for 8 hours.2% by weight of total mixture. The preparation and mixing procedure for Complete Blending (CB) mixture are as follows: 121 . If recovered RAP aggregate is used. Hence. The proportion of RAP/virgin binder is 4/6. RAP binder is not fully extracted and recovered from RAP materials. Therefore. The bitumen content is 5. However. Virgin binder (160/220 Pen) is preheated at 135oC for 2 hours. After batching.. The bitumen content is 5. after binder recovery process (BS-EN:12697-4.2% by weight of total mixture and the preparation of rejuvenated binder for complete blending case is as follows:   RAP binder is extracted and recovered from RAP materials RAP binder is blended with virgin binder (160/220 Pen) at 160oC by mechanical mixer to produce homogeneous blend. the result is probably altered as rejuvenator might interact with remaining RAP binder. even if RAP material is solvated in methylene chloride solution and soaked overnight. after being extracted and recovered. the Black Rock mixture is manufactured from total virgin aggregate instead and virgin binder 160/220 Pen. Mixer is maintained at 135oC. RAP binder is assumed to be fully blended with rejuvenator. “Complete Blending” Mixture (CB) In Complete Blending case. is fully rejuvenated by virgin binder before blending with batched aggregate. there is an assumption that there is no interaction between RAP and virgin binder. 2000). If RAP aggregate is used. The same phenomenon occurs as for recovery process.Control mixtures “Black Rock” mixture (BR) In Black rock case. RAP binder still exists in recovered RAP aggregate. the remaining RAP binder on recovered aggregate will deviate the assumption of Complete Blending case. Due to the fact that original RAP and virgin aggregate are the same. Complete Blending mixture in this case is made of pure virgin aggregate. RAP binder. RAP binder is normally extracted and recovered from RAP material. the combination of recovered RAP and virgin aggregate are then mixed with pure virgin binder or rejuvenator (McDaniel et al. Preheated aggregate and binder is mixed in the mixer for 2 minutes.

6. All the specimens are then conditioned at 20oC for 15 days before ITSM test. The material preparation and mixing procedure are the same as those for CB mixture.2% by weight of total mixture. denoted as CB-V. The bulk specific density of each specimen is determined (BS-EN:12697-6. the hypothesis is that if RAP lumps are fully separated and well blended with virgin material.1). The target height of compacted slab is 60 mm. 2003).3. Due to RAP and virgin binder properties being different. the bulk specific gravity at target air void content 4%. To eliminate the surface defects due to compaction. 6. On the contrary. there is also a mixture made of virgin aggregate and bitumen 70/100 Pen. the amount of materials is determined based on maximum density. each specimen is cut at both ends to achieve the thickness of 40 mm.3 Machining and storage of specimens Compacted slabs are demoulded the next day and 5 specimens are cored from each slab. 2003) to estimate the air void content (BS-EN:12697-8.2 Compaction The loose mixtures are compacted straight away after mixing by roller compacter (BSEN:12697-33. 6. and the volume of the slab after compaction. For both recycled and control mixtures.3.    Batched aggregate is conditioned at 150oC for 8 hours Rejuvenated binder is conditioned at 150oC for 2 hours Mixer is maintained at 150oC Aggregate and rejuvenated binder are mixed in the mixer for 2 minutes The target of rejuvenated binder with 40% RAP binder is bitumen 70/100 Pen (Section 6. 2003). Hence. The bitumen content of this mixture is 5. the specimens are stored at 5oC for further testing. The scheme of coring process is illustrated in Figure 64. the heterogeneity of recycled mixture will result 122 . stiffness values of different measuring directions will be approximately the same. This is to appraise the quality of recycled mixture compared to that made from virgin materials.4 Assessment method The assessment method is primarily based on the comparison among stiffness data of recycled mixtures manufactured by different mixing efforts. After stiffness determination.3. The internal dimensions of the mould are 305 mm x 305 mm.

Anderson Darling statistic test. To eliminate the effects of diffusion on mechanical properties. The stiffness distribution is then analyzed by statistical tools. 124 millisecond rise-time to ensure specimens are in the linear elastic regime (BS-EN:12697-26. 40 stiffness values are generated for each case. for instance 2 minutes RAP/superheated virgin aggregate mixing duration. as testing specimens are cored from roller-compacted slabs. Stiffness distributions relevant to different mixing efforts are compared to each other and also to those of “Black Rock” and “Complete Blending” case to study the effect of mixing on mechanical properties of hot recycled mixtures. 305 mm S1 S2 305 mm 100 mm S3 40 mm S4 S5 Figure 64: Coring and cutting scheme for compacted slabs The total number of specimens tested for each case. 5m strain. the quality of mixing is also revealed by the stiffness values between different samples.in the considerable variance in stiffness values where mixing duration is not long enough. descriptive statistics. In addition. hypothesis test. is ten. 123 . 2004). As the stiffness values of each sample are measured in four directions at 45oC angular increments. for instance. stiffness measurement is implemented 15 days after the day of compaction. Stiffness is determined at 20oC. Air void characteristic of testing specimens is also considered due to its effect on stiffness values.

213 0.903 0. The summary of the statistical t-test is presented in Table 37.108 0.250 0.645 1. in order to compare stiffness values of different asphalt mixtures.218 5.172 0.103 LR SR FS-2 FS-4 FS-6 FS-8 SHRP FS-2 FS-4 FS-6 FS-8 SHRP CB CB-V BR Table 36: Air void content of recycled mixture manufactured by different mixing method The effect of air void content on stiffness of asphalt mixture has been investigated by several researchers (Tayebali et al..481 4.106 0.214 0. 1996). 2.696 4.313 5.6.086 5.353 3.917 0.101.114 0. It is difficult to control the air void content of each specimen. Pairs of air void content sets that do not conform to the hypothesis of t-test are highlighted.5 Results and analysis 6. The standard deviation (SD) or coefficient of variation (COV) indicates the scatter of air void content around the mean of each set of specimens. Therefore.123 0. the t-test is used.5.472 0. Mean (%) 5.143 0. the air void contents of these materials must be approximately the same. The hypothesis of the test is if there is no significant difference between the means of two data sets.270 0.1 Air void contents Five cylindrical specimens are cored from roller-compacted slab.640 1. Although the target air void content is set at 4% volume.835 5.101].088 0.533 COV 0. The air void data is presented in Table 36.263 4.626 4.512 0.692 0. there is some scatter in air void content data.270 0.493 5. 1994) (Read.171 0.634 6.460 0.088 5.101 0. To compare whether the mean air void content of each data set are identical.769 0.185 SD 1. The results show that almost all mean air void contents are approximately similar to the others except control mixture 124 . Specimens at the centre of slab tend to have lower air void content.377 0. the standard error of the difference in means t must be smaller than the critical value tcritical for the relevant confidence level. The confidence level is 95% hence tcritical must be in the range [-2.

In general.made of bitumen 70/100 Pen (CB-V) and recycled mixture made from small RAP by SHRP mixing method (SR-SHRP). the air void contents of these two sets of asphalt mixtures are considerably lower than those of the others. 125 .

LR

SR

Control Mixes

FS-8 FS-6 FS-4 FS-2 SHRP FS-8 FS-6 FS-4 FS-2 SHRP BR CB CB-V

2.052 -1.009 0.356 1.304 -0.195 -1.381 -1.059 -0.14 5.756 0.234 1.758 2.333

-3.462 -1.266 -1.169 -1.972 -2.701 -3.603 -2.577 3.735 -2.339 -0.711 0.476

FS-8 FS-6 Table 37: Air void comparison by t-test

1.176 2.906 0.646 -0.812 -0.027 1.551 8.606 1.594 3.508 3.613 LR FS-4

0.584 -0.49 -1.514 -1.214 -0.305 3.833 -0.224 0.913 1.563 FS-2

-1.307 -2.285 -3.082 -1.76 6.729 -1.499 0.629 1.571 SHRP

-1.15 -0.68 0.345 4.916 0.419 1.677 2.242 FS-8

0.806 1.598 4.853 1.638 2.554 2.963 FS-6

1.666 9.154 1.701 3.723 3.732 SR FS-4

9.22 0.15 2.5 2.79 FS-2

-8.25 -6.35 -2.65 SHRP

2.17 2.559

1.174 Control Mixes BR CB CB-V

126

6.5.2 Effects of mixing time on stiffness
The summary of stiffness data is presented in Table 38. The mean stiffness comparison among different sets of asphalt materials by t-test is shown in Table 39. The hypothesis is if two means of stiffness are similar with confidence level of 95%, the calculated t value must be in the range [-1.991; 1.991]. Each pair of stiffness sets that conforms to the hypothesis is highlighted. The results show clearly that the mean stiffnesses of different sets of specimen are quite different each other. For BR case, the difference is understandable as the viscosity of bitumen 160/220 Pen is extremely low compared to that of the other bitumen. However, except the BR case, the materials for the manufacture of the other recycled asphalt mixtures are exactly the same. In fact, in these cases, if the RAP binders are fully extracted and mixed with virgin bitumen (bitumen 160/220 Pen), these result blends will have similar characteristics. It has been argued that the stiffness difference among sets of recycled asphalt mixtures is attributed to difference in air void characteristics. For instance, as a result of air void analysis (Tables 36 and 37), the mean air void content of CB-V and SR-SHRP asphalt are considerably lower than those of the other asphalt mixtures. However, the air void comparison also demonstrates that apart from CB-V and SR-SHRP, the air void contents of the other asphalt mixtures are statistically similar. Hence, if the stiffness values of these asphalt mixtures are different, the reason must be due to different mixing characteristics.

LR

SR

FS-2 FS-4 FS-6 FS-8 SHRP FS-2 FS-4 FS-6 FS-8 SHRP

CB CB-V BR

Mean 1262 1412 1610 1720 1614 1732 1733 1808 1866 1774 2294 2409 752

SD 483 314 197 66 102 240 77 96 139 47 144 210 117

COV 0.383 0.222 0.122 0.038 0.063 0.139 0.044 0.053 0.075 0.026 0.063 0.087 0.156

Table 38: Summary of stiffness values (MPa) of recycled mixtures manufactured by different mixing methods

127

LR

SR

Control Mixes

FS-8 FS-6 FS-4 FS-2 SHRP FS-8 FS-6 FS-4 FS-2 SHRP BR CB CB-V

3.367 6.685 6.212 5.532 -5.964 -4.743 -0.739 -0.316 -4.162 45.569 -22.973 19.823

3.379 4.457 -0.12 -6.704 -5.709 -3.661 -2.705 -5.109 23.683 -17.747 17.569

FS-8 FS-6 Table 39: Stiffness comparison by t-test

1.852 -3.874 -8.355 -7.62 -6.27 -5.394 -7.205 12.465 -16.158 -16.705 LR FS-4

-4.765 -7.756 -7.274 -6.345 -5.902 -6.934 6.29 -13.232 -14.047 FS-2

-9.207 -8.723 -5.847 -3.244 -8.966 35.049 -24.389 -21.553 SHRP

2.158 5.309 3.441 3.971 38.709 -13.548 -13.66 FS-8

3.888 2.149 2.035 44.02 -17.785 -16.483 FS-6

0.03 -2.909 44.276 -21.836 -19.181 SR FS-4

-1.277 26.329 -14.264 -14.654 FS-2

51.194 -21.812 -18.72 SHRP

-52.615 -43.641 BR

-2.866 Control Mixes CB CB-V

128

The most significant feature that separates recycled asphalt from virgin asphalt is the recycled asphalt mixtures contain RAP materials, agglomerate lumps of RAP binder and aggregate. In order to manufacture homogeneous recycled asphalt, RAP lumps should be fully separated into single pieces and uniformly distributed in the whole mixture. Therefore, virgin binder can rejuvenate and alter the properties of RAP binder. If this cannot occur, then there will be some areas with high concentration of RAP and vice versa, other areas are dominated by virgin material. Due to the fact that properties of RAP binder are entirely different from those of virgin binder, there will be a substantial variation in stiffness values. Figures 65 to 68 illustrate the relation among stiffness values, the location of specimens cored from roller-compacted slabs, and the stiffness measured at different directions for recycled asphalt mixtures manufactured from large RAP (LR) with different mixing durations. The results show that mixing time between RAP and superheated virgin aggregate significantly affects the homogeneity of recycled asphalt mixtures. With short mixing duration, for instance 2 minutes, the degree of stiffness fluctuation is really high (Figure 65). The stiffness values are not only different from specimen to specimen but also different between measured directions in the same specimen. Tables 40 to 43 show the stiffness values measured at different directions for specimens made of large RAP with different mixing durations. The data demonstrates that the homogeneity level of hot recycled asphalt mixture is considerably improved once the RAP/superheating virgin aggregate mixing duration is extended. Stiffness difference between specimens are considerably reduced and this phenomenon is substantiated by the fact that the general coefficient of variation significantly decreases from 38.3 to 3.8% (Table 38) once the mixing duration increases from 2 to 8 minutes. In addition, the stiffness coefficient of variation in each specimen is also reduced to less than 10% once the mixing duration is increased to 8 minutes (Table 43). The longer the mixing time, the closer the homogeneity of recycled mixture approaches that of complete blending (CB) mixture (Table 44). The mixing duration between RAP and superheated virgin aggregate is an important factor that determines the quality of hot recycled mixture. During the mixing process, the heat transferred from superheated aggregate will soften the RAP lumps. Under mechanical effects of mixing paddles, RAP lumps are separated into single pieces covered by RAP 129

not well distributed in the mixture. If the mixing time is not long enough. this will result in the considerable variation in stiffness values. Consequently.binder from circumference to the centre and distributed all over the mixture. 3000 2500 Stiffness (MPa) 2000 1500 1000 S3 500 S5 0 1 Meas ured Dir S4 Sa mp les S2 2 ect ions S1 3 4 3000 2500 Stiffness (MPa) 2000 1500 1000 S8 500 S9 0 1 2 S7 3 4 S10 S6 Sa mp les Meas ure dD ir ection s Figure 65: Stiffness versus core location and measuring direction of LR FS-2 mixture 130 . RAP lumps cannot be separated and hence.

3000 2500 Stiffness (MPa) 2000 1500 1000 S5 500 S3 0 1 Meas ured Dir S4 Sa mp S8 S7 S9 S2 2 ect ions S1 3 4 3000 2500 Stiffness (MPa) 2000 1500 1000 500 0 1 2 3 4 S10 S6 Meas ur ed Dir ection s Figure 66: Stiffness versus core location and measuring direction of LR FS-4 mixture Sa mp les les 131 .

3000 2500 Stiffness (MPa) 2000 1500 1000 S2 500 S5 0 1 2 S1 3 4 S4 S3 s le p m a S Meas ur e dD ir ection s 3000 2500 Stiffness (MPa) 2000 1500 1000 S9 500 S7 0 1 2 S8 3 4 S10 S6 Sa mp les Meas ured Dir ect ions Figure 67: Stiffness versus core location and measuring direction of LR FS-6 mixture 132 .

3000 2500 Stiffness (MPa) 2000 1500 1000 S5 500 S3 0 1 Meas ur e dD S4 Sa mp les S2 2 ir ections S1 3 4 3000 2500 Stiffness (MPa) 2000 1500 1000 S10 500 S8 0 1 Meas ured Dir S9 S7 2 ection s s le p m a S S6 3 4 Figure 68: Stiffness versus core location and measuring direction of LR FS-8 mixture 133 .

9 S5 6.8 3.5 S4 4.3 7.4 S10 3.4 S7 5.7 S4 5.7 Air Void Samples Content (%) S1 5.1 S8 5.2 23.2 S7 7.7 13.3 S3 5.7 18.8 S2 4.1 6.8 20.8 Table 40: Stiffness values of LR FS-2 specimens Stiffness in Different Directions (MPa) 1 2 3 4 1063 881 872 1035 1623 1530 1183 1068 1854 1703 1239 1285 1925 1517 2012 1405 1613 1620 1218 1387 1140 1062 952 1168 1692 1648 1298 1256 1914 1933 1503 1540 1543 1663 1250 1186 1150 1745 1114 1699 Mean Standard Stiffness Deviation (MPa) 963 100 1351 267 1520 305 1715 299 1460 194 1081 97 1474 228 1723 233 1411 229 1427 341 COV (%) 10.1 17.1 S6 5.3 9.4 13.0 S6 4.4 S9 4.7 S5 4.0 11.3 S9 6.9 S8 5.4 Stiffness in Different Directions (MPa) 1 2 3 4 1499 1007 1047 731 722 920 753 768 2049 2023 1792 1840 1205 1144 1168 1660 978 597 827 666 1676 1908 1598 1768 823 603 691 756 1928 1529 1723 1829 1665 1621 1583 1555 944 973 862 740 Mean Standard Stiffness Deviation (MPa) 1071 318 791 88 1926 129 1294 245 767 171 1738 133 718 94 1752 171 1606 48 880 104 COV (%) 29.9 Table 41: Stiffness values of LR FS-4 specimens 134 .9 22.0 S10 5.Air Void Samples Content (%) S1 5.4 19.3 S2 7.0 15.5 16.1 9.7 11.5 13.1 S3 5.

7 8.1 S10 4.4 S6 6.1 S5 4.9 1.9 S10 5.7 Table 42: Stiffness values of LR FS-6 specimens Stiffness in Different Directions (MPa) 1 2 3 4 1727 1687 1697 1660 1666 1623 1629 1669 1768 1759 1744 1794 1795 1786 1853 1825 1751 1712 1657 1662 1806 1678 1744 1724 1747 1718 1690 1728 1780 1699 1685 1696 1637 1620 1651 1612 1823 1865 1721 1745 Mean Standard Stiffness Deviation (MPa) 1693 28 1647 24 1766 21 1815 30 1696 45 1738 53 1721 24 1715 44 1630 17 1789 67 COV (%) 1.6 S3 4.6 3.1 S4 6.6 S7 4.3 4.3 S9 6.6 1.6 S2 3.7 Table 43: Stiffness values of LR FS-8 specimens 135 .5 1.8 S5 6.0 1.6 S4 4.6 S8 5.2 1.3 S3 5.0 3.7 1.9 3.9 6.7 3.0 2.4 2.Air Void Samples Content (%) S1 4.2 S9 3.3 S8 4.1 0.0 S6 4.7 2.4 S7 4.2 3.5 Air Void Samples Content (%) S1 3.5 Stiffness in Different Directions (MPa) 1 2 3 4 1524 1328 1270 1232 1957 1974 1647 1911 1868 1863 1835 1839 1415 1447 1462 1480 1850 1701 1604 1700 1559 1584 1509 1486 1431 1505 1390 1406 1610 1512 1623 1568 1760 1888 1847 1779 1528 1606 1452 1461 Mean Standard Stiffness Deviation (MPa) 1339 130 1872 152 1851 17 1451 28 1714 102 1535 45 1433 51 1578 50 1819 60 1512 71 COV (%) 9.6 S2 5.

3 S2 5.9 S9 4.7 S6 5.5 S4 4.2 S10 4 Stiffness in Different Directions (MPa) 1 2 3 4 2406 2415 2373 2386 2221 2255 2103 2186 2095 2103 2077 2084 2324 2233 2283 2317 2286 2202 2204 2282 2265 2358 2259 2269 2483 2470 2268 2488 2130 2119 2099 2076 2406 2356 2346 2259 2531 2562 2536 2496 Mean Standard Stiffness Deviation (MPa) 2395 19 2191 65 2090 12 2289 42 2244 47 2288 47 2427 106 2106 24 2342 61 2531 27 COV (%) 0.1 4.8 2.1 2.6 1.Air Void Samples Content (%) S1 4.0 S7 4.1 S3 5.1 Table 44: Stiffness values of CB specimens 136 .6 1.1 2.4 1.0 0.6 S8 4.8 3.7 S5 4.

137 . If the rejuvenation can only occur once new virgin binder is in contact with RAP binder. the more exposed area of RAP materials to virgin binder. from 1200 MPa to 1700 MPa once the mixing durations are extended from 2 to 8 minutes.03 1232 1000 1612 Mean 872 500 597 Max Min 0 0 2 4 6 8 10 Mixing Time (minutes) Figure 69: Stiffness versus mixing time of LR FS mixtures Figure 69 shows the relation between RAP/superheated virgin aggregate mixing durations and stiffness of recycled mixtures made with large RAP (LR).28 Stiffness (MPa) 1500 1412. When the mixing time is increased. the means stiffness is far different from that of black rock (BR) case. The data show that the mean stiffness increases considerably.Longer mixing duration not only advances the distribution of RAP materials all over the recycled mixture but also enhances the rejuvenation process between virgin and RAP binder. The data also shows that even at very short mixing duration. The top and the bottom dashed lines are for the maximum and minimum stiffness values versus mixing durations. the better this process. As more RAP binder is incorporated with virgin binder.26 1244.83 1610. the stiffness values of recycled mixture generally increase. the fact that more pieces are separated from RAP lumps will enlarge the total RAP exposed area for rejuvenation. for instance 2 minutes. 2500 2049 2000 2012 1974 1865 1720. 751 MPa. The three continuous centre lines present the mean stiffness and the boundaries of the mean with 95% confident level.

this is not the only reason as the increase of stiffness values might be also due to the distribution of RAP materials. In addition. This statistical test is used to validate whether the stiffness values of recycled asphalt mixture match the standard distribution with chosen confidence interval. Most importantly if RAP material is merely distributed. However. distribution or rejuvenation. stiffness values are considerably different from specimen to specimen. which could be dominantly the reason for the increase of stiffness. On the contrary. In addition. stiffness data is further investigated using the Anderson-Darling test (Anderson and Darling. the incorporation between RAP and virgin binder also reduces the variation among stiffness values. credited to the increase of stiffness values as the distribution of RAP over the mixture and incorporation between RAP and virgin binder occurs simultaneously. This situation is exaggerated in the case where the mixing time is not sufficient to separate RAP lumps and these lumps are randomly distributed over the mixture. the stiffness values measured at different directions of each specimen are also different. Virgin binder stiffened by RAP bitumen will increase the minimum value of stiffness. It is sometimes impossible to identify the causes. Anderson-Darling parameter (AD) presents the difference between empirical data and theoretical normal distribution and is expressed as following: AD 2   N  S (20) 138 .The increase in stiffness value might be attributed to the incorporation or rejuvenation between virgin and RAP binder. as RAP lumps start to separate and more RAP binder is rejuvenated by virgin binder. The confidence level for the test is 95%. If the increase in stiffness is attributed to the distribution of RAP lumps. When the mixing time is increased. 1952). However. the specimens with higher concentration of RAP will have higher stiffness. The theory of the test is to plot the empirical continuous distribution function (CDF) of the data and compare it to that of the standard distribution. it is possible to recognize between distribution and rejuvenation. As the stiffness of RAP is considerably higher than that of virgin binder. Anderson-Darling test for standard distribution To better understand the reason for the increase of stiffness values. the properties of the mixture will be dominated by the properties of RAP binder where RAP material predominates. the properties of the mixture will be dominated by the characteristic of virgin material. the properties of both RAP and virgin binder have changed.

4 40 2297 143.6 40 751. BR. the better the data fits into normal distribution (Stephens.6 117. CB Normal 35 30 25 Frequency 20 15 10 5 0 Variable C B-V BR CB Mean StDev N 2409 209.8 40 500 1000 1500 2000 Stiffness (MPa) 2500 3000 Figure 70: Stiffness distribution of control mixtures 139 . and CB-V. the AD parameter ranges from 0. CB. standardized sample value The hypothesis of the Anderson-Darling test is if the data set conforms to the standard distribution.787 and P value must be higher than (1-Confidence level). Stiffness distribution of Control Asphalt Mixtures The stiffness distributions of three control asphalt mixtures. The results from the Anderson-Darling test demonstrate that stiffness values of three control asphalt mixtures conform to the normal distribution (Figure 71). 1974). The mean stiffness values of CB and CB-V are approximately the same. Anderson-Darling parameter has to be smaller than 0.3 to 0. With 95% confidence interval.5. BR. The result is as expected as the rheological properties of recycled binder (CB) with 40% RAP binder is almost identical to that of bitumen 70/100 Pen. are illustrated in Figure 70. The smaller the AD value is. Histogram of CB-V. Coefficient of variation (COV) values are less than 15%.S  i N (2i  1) log F (Yi )  log(1  F (YN 1i )) N (21) Where: N: number of samples F: assumed normal distribution function with estimated µ and δ Yi: the sorted. less than 10% difference.

BR.8 314.68 N 40 40 40 40 AD 1. CB Normal . Results from the Anderson-Darling test show that only at 8 minutes mixing time. At lower mixing 140 .698 0.95% CI 99 Variable CB-V BR CB Mean StDev N AD P 2409 209.310 95 Cumulative Probability (%) 90 80 70 60 50 40 30 20 10 5 1 500 1000 1500 2000 Stiffness (MPa) 2500 3000 Figure 71: Probability plot of stiffness values – control mixtures Probability Plot of LR FS-2.236 P <0.8 40 0.546 0.363 0. LR FS-6. LR FS-4.Probability Plot of CB-V. the AD value is the same level as that of the control asphalt.506 751.0 65.775 95 Cumulative Probability (%) 90 80 70 60 50 40 30 20 10 5 1 0 500 1000 1500 2000 Stiffness (MPa) 2500 3000 Figure 72: Probability plot of stiffness values – LR FS mixtures Figure 72 presents the stiffness cumulative probability of recycled asphalt mixtures composed of large RAP manufactured with different RAP/superheated virgin aggregate mixing durations.6 117.063 0.0 197.95% CI 99 Variable LR FS-2 LR FS-4 LR FS-6 LR FS-8 Mean 1262 1412 1610 1721 StDev 482.4 40 0.421 0.150 0.005 0. LR FS-8 Normal .6 40 0.329 0.547 0.425 2297 143.

The stiffness distribution 141 . In addition. This group possesses the same characteristics as that of the BR mixture. If the stiffness values of 2 minute mixing time are analyzed by these two groups. this expresses a considerable interaction between RAP and virgin binder hence. G1 and G2. the mean stiffness of this group is 860 MPa. It is quite clear that rather than conforming to the normal distribution. the stiffness distribution does not conform to the standard distribution. If the comparison is based on only the mean stiffness. 20 15 Frequency (%) 10 5 0 600 900 1200 1500 Stiffness (MPa) 1800 2100 Figure 73: Stiffness histogram of LR FS-2 mixture The stiffness histogram of recycled mixture composed of large RAP and 2 minute mixing duration (LR FS-2) is shown in Figure 73. However.times. the mean stiffness of LR FS-2 is about 1262 MPa compared to 751 MPa of the BR case. G1 whose stiffness values range from 500 to 1200. and G2 from 1400 to 2200 MPa. the stiffness distribution shows that group 1 of LR FS-2 has 22 values varying from 500 MPa to 1200 MPa. In fact. especially with 2 minutes mixing time. the distribution of stiffness values in each separated group conforms to the standard distribution (Figure 74). The AD parameter for each group is almost the same level as that of control mixtures (Figure 75). stiffness modulus values gather primarily into two main groups. the properties of LR FS-2 are far different from that of the BR case. insignificantly different from that of the BR case. AD values are high. Figure 76 shows the inter-quartile stiffness ranges of LR FS-2 group 1 and the BR case.

546 860.914 95 Cumulative Probability (%) 90 80 70 60 50 40 30 20 10 5 1 0 500 1000 1500 2000 Stiffness (MPa) 2500 3000 Figure 75: Stiffness probability plot of LR FS-2 – Stiffness grouping analysis 142 .indicates that there is insubstantial interaction between RAP and virgin binder and the properties of LR FS-2 is almost dominated by virgin binder.2 179. 40 30 Frequency (%) 20 10 0 600 900 1200 1500 Stiffness (MPa) 1800 2100 Figure 74: Stiffness histogram of LR FS-2 mixture – Stiffness grouping analysis Probability Plot of LR FS-2. LR FS-2 G2 Normal . LR FS-2 G1.0 22 0. the bondage between RAP aggregate pieces are neither weakened nor deactivated. Therefore.005 0. or RAP acts as “Black Rock”.343 0.95% CI 99 Variable LR FS-2 LR FS-2 G1 LR FS-2 G2 Mean StDev N AD 1262 482.173 P <0. RAP lumps are not heated up.395 1753 175.5 18 0.8 40 1. Mechanical mixing therefore only distributes RAP at approximately original size all over the mixture. If the mixing duration is insufficient.

1400 1200 1000 800 600 400 200 0 Stiffness (MPa) LR FS-2 G1 BR Figure 76: Inter-quartile stiffness ranges of LR FS-2 group 1 and BR mixture 2250 2000 1750 Stiffness (MPa) 1500 1250 1000 750 500 LR FS-2 LR FS-4 LR FS-6 LR FS-8 Figure 77: Inter-quartile stiffness ranges of LR FS mixtures When the mixing time is increased. separated pieces will be moved around and mixed up with virgin aggregate and binder. the more interaction between 143 . RAP material no longer acts as “Black Rock” but starts to integrate with virgin binder. The more RAP pieces are disintegrated. Under mechanical effects. RAP lumps will start to disintegrate first from outside to centre once the bituminous bond between RAP aggregate particles is prevailed by mechanical mixing effort. the RAP binder is heated up and becomes softer.

The bottom line presents the end of the first quartile data and the centre line is for the median. The data shows the minimum stiffness value increases from 597 to 1612 MPa when the mixing time increases from 2 to 8 minutes. Figure 77 illustrates the inter-quartiles stiffness range for recycled mixtures composed of large RAP (LR) by field simulation method (FS) with different mixing durations. it only slightly influences the stiffness values of small RAP recycled mixtures. The mechanical mixing effort could not disintegrate those RAP lumps composed of fine aggregates and fillers or the coarse aggregate covered by filler mastic.RAP and virgin binder. the original size of small RAP only changes slightly due to small RAP being already crushed into smaller size before use. However. the mean stiffness values of SR recycled mixtures change unnoticeably. In addition. the data also indicates the first quartile of stiffness data and the stiffness median increases considerably once the RAP/superheated virgin aggregate mixing duration is extended. the stiffness of LR recycled mixture increases due to more virgin binder interacting with RAP binder. the stiffness values generally increase. the original size of large RAP is significantly reduced. Therefore. the increase in mixing time considerably increases the homogeneity of recycled mixture 144 . Due to the incorporation between RAP binder and rejuvenator. However. Figure 78 shows that when the mixing duration increases from 2 to 8 minutes. This is because under the mechanical mixing efforts. 2250 2000 1750 1500 1250 1000 750 500 Stiffness (MPa) SR FS-2 SR FS-4 SR FS-6 SR FS-8 Figure 78: Inter-quartile stiffness ranges of SR FS mixtures While the mixing duration considerably affects the stiffness values of recycled mixture composed of large RAP.

the minimum stiffness of SR mixture is 1497 MPa compared to 597 MPa of LR mixture. the recycled mixtures composed of small RAP material have higher levels of homogeneity than the mixtures composed of large RAP. the stiffness in Table 45 also shows that stiffness variation in each specimen of SR FS-2 mixture is also lower. The larger the total exposed area.5. Figure 78 shows that the inter-quartile stiffness range of SR FS mixtures is considerably narrowed once mixing time is extended.2% of LR mixture and the stiffness variation in each specimen is a maximum of 6% compared to 24% of LR mixture (Tables 38 and 46).9 compared to 38. at 2 minutes mixing duration. For 4 minutes mixing duration. the location of specimens cored from roller-compacted slabs. and the stiffness measured at different directions of recycled asphalt mixtures composed of small RAP (SR) with 2 and 4 minutes mixing duration. Stiffness coefficient of variation for SR mixture is 4. and the mean stiffness is 1806 MPa compared to 1262 MPa. Size of RAP affects not only the homogeneity level but also the interaction between RAP and virgin binder. maximum 10% compared to 30% of LR FS-2 mixture. For instance. In addition. This total surface area is also the exposed area of RAP material to virgin binder for rejuvenation.3 Effects of RAP sizes on stiffness The size of RAP lumps substantially affects the homogeneity of the recycled mixtures. For the case of 2 minutes mixing time. For the same RAP/superheated virgin aggregate mixing duration. the properties of recycled mixture composed of large RAP are primarily dominated by virgin binder. The phenomenon is as expected as the interaction between RAP and virgin binder is controlled by the total surface area of RAP materials. In addition. The interaction is indicated by the mean and the minimum stiffness values of SR mixtures being far different from those of LR mixtures (Figure 81) for the same mixing duration. the phenomenon is the same. As small RAP materials have larger total surface area. the stiffness coefficient of variation for SR mixture is 13. At 2 minutes mixing duration. Figures 79 and 80 illustrate the relation among stiffness values. On the contrary. the recycled mixtures composed of 145 . the total surface area is inversely related to the size of RAP materials. the higher the probability that the virgin binder can coat and penetrate into RAP binder. 6.3% of LR mixture (Table 38).4% compared to 22. RAP acts as “Black Rock”.composed of small RAP materials. In fact. there is a considerable interaction between RAP and virgin binder in recycled mixture composed of small RAP materials.

3000 2500 Stiffness (MPa) 2000 1500 1000 S3 500 S2 0 1 2 S4 3 4 S1 S5 Sa mp S10 500 S8 0 1 2 S6 3 4 S7 S9 Sa m p le s Meas ur ed Dir ect ions 3000 2500 Stiffness (MPa) 2000 1500 1000 Meas ur ed Dir ection s Figure 79: Stiffness versus core location and measuring direction of SR FS-2 mixture les 146 .small RAP tend to have higher stiffness values due to more interaction between RAP and virgin binder.

3000 2500 Stiffness (MPa) 2000 1500 1000 S5 500 S3 0 1 2 S1 3 4 S2 S4 Sa mp S10 S9 S8 0 1 2 S6 3 4 S7 Meas ured Dir ection s 3000 2500 Stiffness (MPa) 2000 1500 1000 500 s le p m a S Meas ur ed Dir ection s Figure 80: Stiffness versus core location and measuring direction of SR FS-4 mixture les 147 .

2000 1800 1600 1437 1400 1412 Stiffness (MPa) 1732 1733 1550 1808 1582 1610 1866 1653 1721 1612 1200 1232 1000 800 600 597 400 200 0 0 2 4 6 8 10 Mixing Tim e (m inutes) 1244 872 LR FS Mean LR FS Min SR FS Mean SR FS Min Figure 81: Stiffness range of LR and SR recycled mixtures manufactured by FS methods 148 .

5 S10 5.Air Void Samples Content (%) S1 5.0 S7 6.1 0.8 3.2 S5 6.1 S6 6.4 S4 5.9 3.6 S2 5.7 4.6 S8 4.9 S10 5.4 3.1 S8 4.6 S4 5.7 S3 5.6 Table 45: Stiffness values of SR FS-2 specimens Stiffness in Different Directions (MPa) 1 2 3 4 1860 1702 1664 1698 1703 1681 1677 1667 1790 1674 1550 1661 1782 1766 1659 1673 1782 1619 1745 1632 1734 1712 1697 1601 1766 1706 1721 1698 1835 1779 1834 1889 1764 1714 1769 1803 1816 1831 1818 1831 Mean Standard Stiffness Deviation (MPa) 1731 88 1682 15 1669 98 1720 63 1695 81 1686 59 1723 30 1834 45 1763 37 1824 8 COV (%) 5.1 S7 5.6 2.2 S5 5.1 0.3 S9 5.4 4.0 7.9 5.4 Table 46: Stiffness values of LR FS-4 specimens 149 .4 0.0 Air Void Samples Content (%) S1 6.4 S3 4.5 1.0 3.5 S6 6.9 4.7 2.3 Stiffness in Different Directions (MPa) 1 2 3 4 1709 1459 1437 1451 1871 1714 1660 1697 2161 2048 2005 2008 1953 1825 1854 1867 1843 1754 1714 1682 2052 1772 1793 1821 1550 1447 1498 1446 1646 1528 1518 1472 1525 1658 1504 1583 1946 1922 1942 1927 Mean Standard Stiffness Deviation (MPa) 1514 130 1736 93 2056 73 1875 55 1748 70 1860 130 1485 50 1541 74 1568 69 1934 12 COV (%) 8.5 2.8 4.6 5.2 S2 5.7 S9 4.

the sizes of RAP contribute almost no effects on stiffness of recycled mixture. In addition. Figure 83 shows that the stiffness values of recycled mixture composed of large and small RAP by SHRP methods are approximately the same.4 Effects of mixing methods on stiffness Mixing methods significantly affect stiffness distribution of hot recycled mixture. RAP sizes significantly affect stiffness values of recycled mixtures manufactured by field simulation method (FS). While stiffness of LR FS-2 mixture is relatively close to that of the BR mixture. 2500 2000 Stiffness (MPa) 1500 1000 500 BR LR FS-2 LR SHRP LR FS-8 CB Figure 82: Inter-quartile stiffness ranges of control and LR mixtures manufactured by different methods In the SHRP method. 58oC.6. in the SHRP method.5. The conditioning temperature is considerably higher than the softening point. especially for the recycled mixtures composed of large RAP. On the contrary. Figure 82 shows that stiffness of recycled mixtures manufactured by SHRP method is quite different from that of mixtures produced by field simulation (FS) method. of RAP 150 . This is due to the mixing mechanisms as these approaches determine how RAP and virgin materials interact with each other. RAP lumps are conditioned in a force draft oven at 110oC. the stiffness of LR SHRP and LR FS-8 are comparatively close to that of the CB mixture.

In this approach. 2000 Stiffness (MPa) 1500 1000 500 0 LR SHRP SR SHRP LR FS-2 SR FS-2 Figure 83: Inter-quartile stiffness ranges of SHRP mixtures and FS-2 mixtures 151 . both small and large sizes of RAP material are entirely heated.binder. However. The superheated temperature of virgin aggregate is 215oC. This is why when using the SHRP method. The mixing mechanism of the field simulation (FS) method. Although this temperature is extremely higher than the softening point of RAP binder. under the conditioning duration of 2 hours. a certain duration or critical duration is required so that the heat can be transferred from superheated virgin aggregate to completely heat up RAP materials. these RAP lumps are disintegrated into smaller pieces and will be blended and rejuvenated by virgin binder. Under the mechanical mixing effort. mixtures composed of large and small RAP have approximately the same stiffness value (Figure 83). on the contrary. is quite different from that of SHRP. RAP lumps exist as unbreakable agglomerates under normal mechanical mixing at ambient temperature. 2004). RAP materials at ambient temperature are mixed with superheated virgin aggregate. The heat that RAP materials absorb during the conditioning duration will soften the RAP binder and weaken the bituminous bond between RAP aggregate particles. under the heat transferred from superheated virgin aggregate. RAP binder will be softened and the bonds among RAP aggregate particles will be deactivated. Although the heat transfer process is significantly influenced by the size of material (Cutnell and Johnson.

recycled mixtures composed of small RAP will have better incorporations between RAP and virgin binder than large RAP mixtures. At this critical point. Due to the heat transferred from virgin aggregate. the smaller the size of RAP. the stiffness of LR FS-8 approaches that of CB mixture. such a long preheating duration is by no means practical in the real industry due to the length constraint of the mixer and economical issues (Lee et al. the bitumen bond between RAP aggregates will be gradually weakened until being overcome by mechanical mixing power. the more RAP aggregate pieces are separated. the longer the critical mixing duration is required. there is a critical period of time which is required to transfer the heat from superheated virgin aggregate to soften RAP materials. 2004). The longer the mixing time. when the mixing time is extended to 8 minutes. The experimental data also indicates that SHRP method tends to overestimate the mechanical properties of hot recycled mixtures. In the case that the mixing duration is not sufficient. is a critical factor that determines the quality of hot recycled mixture. The RAP extensive preheating duration in the SHRP method coincidently advances the incorporation between RAP and virgin materials. the bigger the size of RAP lumps. The data in Table 38 show that stiffness of SHRP mixture is approximately similar to that of FS-4 or FS-6 mixtures. the interaction between RAP and virgin binder of FS mixture will be less 152 . Hence. especially when large sizes of RAP materials are used. or the amount of heat transferred from virgin aggregate. and the more RAP binder will interact with virgin binder. an inconsiderable proportion of RAP will interact with virgin binder and RAP materials will act as black rock. The incorporation between RAP and virgin binder in this case will depend primarily on RAP size. Due to the total surface area. mixing duration in FS method.. Due to the fact that heat transfer process is significantly influenced by the sizes of material (Cutnell and Johnson. field simulation (FS) method would better exemplify what occurs in the real industrial mixers. Hence. RAP lumps start to disintegrate. the more interaction between RAP and virgin binder occurs. If the RAP/superheated virgin aggregate mixing duration is shorter than the required duration. 1983).As the mechanical mixing efforts remains constant during the mixing process. In addition. RAP materials will be distributed all over the mixture at approximately original size. This is because in FS method. If the mixing duration is the same. If the large size of RAP is used.

In this case. Figure 84 and Table 38 show that the mean stiffness of SHRP mixture. If RAP lumps are entirely disintegrated into single pieces of aggregates. coated by RAP binder. and under mechanical mixing. 2500 2000 Stiffness (MPa) 1500 1000 500 BR SR SHRP SR FS-8 CB Figure 84: Inter -quartile stiffness ranges of control and SR mixtures manufactured by different methods Theoretically. SHRP and FS. is far different from that of complete blending mixture. even at 8 minutes mixing time which is never practical in the industry. the stiffness 153 . including also fine aggregate particles and filler. the mean stiffness is 1865 MPa. It is interesting that both methods. stiffness of FS mixture is certainly lower than that of SHRP mixture. By FS method. Therefore. even when small RAP is used. (2000) where the properties of actual blending mixture (similar to SHRP method in this research) is the same as that of complete blending. 1773 MPa compared to 2294 MPa. the complete blending might occur. could not produce the recycled mixtures that possess the stiffness values approximately the same as that of the complete blending case. the complete blending situation will exist once all the RAP binder is activated and rejuvenated by virgin binder. This finding is quite different to the result of McDaniel et al. virgin binder could cover every single piece of RAP binder-coated aggregate.than that of the SHRP mixture.

filler from filler mastic. On the contrary.08 3. this situation is hardly happens in reality. the temperature in mixer B is controlled by the temperature of the air inside the mixing apartment.09 5. LR FS-2 Mean Mixer A Mixer B 5.of recycled mixture will be approximately the same as that of complete blending mixture. This is because mechanical mixing can disintegrate coarse aggregate from the RAP lumps. the loose recycled mixtures manufactured by Mixer B always have considerably higher temperature than those produced by Mixer A.1 14.6 11.75 SR FS-2 Mean COV (%) 8. although the temperatures are set the same. In this experiment. a small experiment was carried out with the aim to eliminate the effects of differences in mixing temperature on the mechanical properties of recycled mixture. The result shows that the air void contents are approximately similar except LR FS-6 and SR FS-6 mixtures produced by Mixer B which have slightly lower volumes of air void content. Therefore. the mean stiffness values of large RAP mixtures manufactured by both Mixer 154 .2 Table 47: Air void summary of recycled specimens manufactured by different method and equipments Table 47 presents the summary of air void content and coefficient of variation of recycled mixtures manufactured by both Mixers A and B. 6.47 LR FS-6 Mean COV (%) 17. although the original size of RAP is reduced. the temperatures of the loose recycled mixtures after mixing are also different.5.8 4. there will be a proportion of RAP binder that is not activated and rejuvenated by virgin binder.0 20.9 5.49 3. Due to the difference in mixing operation.5 Effect of mixing equipment on stiffness The heat supply in mixer A is controlled by the thermocouple attached on the mixing bowl. or especially.21 4. The results show that for 2 minutes mixing time. Consequently. However. However.15 COV (%) 25.46 SR FS-6 Mean COV (%) 27.4 6. Figures 55 and 56 show that after 8 minutes mixing time by the FS method. The summary of stiffness values are shown in Table 48.1 19. this process does not separate fine aggregates from RAP lumps. During the manufacture process. The results indicate that the setting temperature of Mixer B should be about 20oC lower than that of Mixer A in order to produce the loose mixture with the same temperature. temperature is recorded by external thermocouple attached directly to the loose mixture during the mixing process. RAP materials still exist as agglomerates.

3 12. the minimum stiffness value of Mixer B mixture is 1533 MPa compared to 1232 MPa of mixture manufactured by Mixer A. In addition. Not only the general stiffness coefficient of variation but also the stiffness variations in each specimen (Table 50 and Figure 86) are lower than those of Mixer A mixture. the location of specimens cored from roller-compacted slabs.2 11.2 3. the homogeneity of Mixer B mixture is also much better. the stiffness values distribution has almost the same pattern as that of mixture manufactured by Mixer A. the stiffness values measured in different directions for the same specimens also vary substantially. The mean stiffness of this group.5 Max 2049 1974 2161 2002 2048 1839 1896 1947 Min 597 1232 1437 1582 752 1533 1631 1702 Median 1156 1576 1714 1794 1219 1679 1779 1827 Table 48: Stiffness (MPa) of recycled specimens manufactured by different methods and equipment When the mixing time is extended from 2 to 6 minutes. is quite close to 751 MPa of Black Rock (BR) mixture.A and B are slightly different. Besides the general variation coefficient of stiffness being 29. The tilt axis of Mixer B allows the material in the mixing apartment to move not only horizontally but also 155 . 1576 MPa (Table 48). 1610 MPa by mixer A and 1690 MPa by mixer B. this mixture has 20 stiffness values lower than 1200 MPa. Table 49 and Figure 85 illustrate the relation among stiffness values.2 4. This is because Mixer B is more efficient in than Mixer A. the interaction between RAP and virgin binder of large RAP mixture manufactured by Mixer B is far better than that of Mixer A. and the stiffness values measured in different directions of LR FS-2 mixtures produced by Mixer B. Method LR FS-2 Mixer A (MA) LR FS-6 SR FS-2 SR FS-6 LR FS-2 Mixer B (MB) LR FS-6 SR FS-2 SR FS-6 Mean 1262 1610 1732 1808 1342 1690 1771 1825 SD 483 197 204 96 392 77 74 63 COV (%) 38.6 4. 1026 MPa. In addition.3 29.8 5. In addition. indicates that there is inconsiderable interaction between RAP and virgin binder. Although there is insignificant difference in the mean stiffness.2%. The minimum stiffness value of Mixer B mixture is almost the same as the median stiffness of Mixer A mixture.

However. the effect of mechanical mixing can only distribute the RAP material all over the mixture. the longer the critical duration. The bigger the size of RAP. when the mixing duration is adequate (reaches critical point). mechanical mixing will separate RAP lumps and enhance the interaction between RAP and virgin binder.vertically. Mixer B has a function that can reverse the mixing direction also enhances the homogeneity of the recycled mixture. The more efficient the mechanical mixing. The length of critical duration depends primarily on RAP size. In addition. the higher the homogeneity level of recycled asphalt mixtures which are manufactured. The data indicates that when the mixing time is not sufficient to deactivate all the bitumen bonds in RAP lumps. 156 .

9 21.8 Air Void Samples Content (%) S1 3.3 4.Air Void Samples Content (%) S1 6 S2 6.7 S4 3.3 S3 4.9 S3 3.7 5.4 6.1 2.4 S6 3.5 S6 5.1 4.8 S9 2.2 1.2 S8 5.8 Table 49: Stiffness values of LR FS-2 specimens – Mixer B Stiffness in Different Directions (MPa) 1 2 3 4 1832 1738 1739 1576 1566 1563 1533 1570 1727 1669 1655 1654 1744 1639 1662 1603 1733 1679 1648 1660 1714 1636 1634 1787 1772 1748 1634 1651 1647 1760 1678 1687 1839 1787 1754 1815 1762 1742 1730 1630 Mean Standard Stiffness Deviation (MPa) 1721 107 1558 17 1676 35 1662 60 1680 36 1693 73 1701 69 1693 48 1799 37 1716 59 COV (%) 6.1 24.6 S7 3.1 15.1 3.7 S4 6.9 19.4 Table 50: Stiffness values of LR FS-6 specimens – Mixer B 157 .1 3.1 2.7 S5 3.3 S10 3.4 S8 2.2 S9 3.5 S7 5.3 S5 5.8 2.1 Stiffness in Different Directions (MPa) 1 2 3 4 996 934 1422 1415 813 752 1044 969 1703 1082 1448 1994 1018 1622 1349 1261 1294 1034 1042 1245 1179 1070 1115 1043 1093 1078 1194 1022 1002 1044 1967 1996 2028 1934 1473 1271 2048 1931 2005 1752 Mean Standard Stiffness Deviation (MPa) 1192 263 895 135 1557 387 1313 249 1154 135 1102 59 1097 72 1502 554 1677 363 1934 131 COV (%) 22.6 36.6 6.0 S10 3.8 S2 4.6 2.0 11.

3000 2500 Stiffness (MPa) 2000 1500 1000 S5 500 S3 0 1 2 S1 3 4 S2 S4 Meas ured Dir ect ion 3000 2500 Stiffness (MPa) 2000 1500 1000 S10 500 S8 0 1 2 S6 3 4 S7 S9 Sa mp les Meas ured Dir ect ion Figure 85: Stiffness versus core location and measuring direction of LR FS-2 mixture – Mixer B Sa mp les 158 .

3000 2500 Stiffness (MPa) 2000 1500 1000 S5 500 S3 0 1 2 S1 3 4 S2 S4 Sa mp les Meas ur e dD ir ection 3000 2500 Stiffness (MPa) 2000 1500 1000 S10 500 S8 0 1 2 S6 3 4 S7 S9 Sa mp les Meas ur ed Dir ection Figure 86: Stiffness versus core location and measuring direction of LR FS-6 mixture – Mixer B 159 .

Master-curves of these processed RAP binders are constructed from rheological data and compared to that of original RAP binder. RAP lumps including both small (SR) and large sizes (LR). 160 .5. Rheological testing is carried out under 0. The test frequencies range from 0. After being extracted and recovered.8% strain to ensure bitumen responds in linear visco-elastic region.6 Effects of mixing methods on RAP binder properties The increase of stiffness may also be attributed to the aging of RAP binder. RAP lumps are conditioned in the force draft oven at 110oC for 2 hours. are mixed with superheated virgin aggregate for different mixing durations. The procedures are the same as those in the manufacture process of field simulation method (FS). For the SHRP method. In this experiment. the RAP binder extracted from small RAP lumps shows noticeably more ageing compared to original RAP binder.1 to 10 Hz. This is because during the mixing process. Testing temperatures range from 5 to 45oC for 8 mm plate and from 20 to 80oC for 25 mm plate.6. Figures 87 and 88 show the complex modulus and phase angle versus log reduced frequency of original RAP and processed RAP binders. The data indicates there is no significant alteration to RAP binder after RAP lumps are mixed with superheated virgin aggregate by FS method. The thickness of testing specimen is 1000 m for 25 mm plate and 2000 m for 8 mm plate. For the SHRP method. RAP binder might be aged more due to the exposure to high temperature from superheated aggregate as well as the mixer for a long period of time. An experiment has been carried out to investigate if there is any alteration to RAP binder properties during the mixing process. rheological properties of processed RAP binders are studied by Dynamic Shear Rheometer (DSR).

1.0E+00 -6 -5 -4 -3 -2 -1 0 1 2 3 4 Log Reduced Frequency (Hz) Figure 87: Complex modulus versus log reduced frequency of RAP binder before and after processed by different mixing methods 90 80 70 Phase Angle 60 50 LR FS-2 LR FS-8 SR FS-2 40 SR FS-8 RAP 30 SR SHRP LR SHRP 20 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 Log Reduced Frequency (Hz) Figure 88: Phase angle versus log reduced frequency of RAP binder before and after processed by different mixing methods 161 .0E+07 1.0E+03 SR FS-2 SR FS-8 1.0E+04 1.0E+06 Complex Modulus (Pa) 1.0E+01 LR SHRP 1.0E+05 LR FS-2 LR FS-8 1.0E+08 1.0E+02 RAP SR SHRP 1.

6. The increase in mixing duration significantly improves the homogeneity level of recycled mixture. The smaller the original size of RAP materials. the RAP lumps are separated thereby increasing the contact areas between RAP and virgin binders. In this case.6 Summary Stiffness values of cylindrical specimens measured in different orientations indirectly express the heterogeneity of recycled mixture. the mixing condition favourably enhances the interaction between RAP and virgin binders. RAP is mixed with superheated virgin aggregate. The thermal energy transferred from virgin aggregate will help to increase the RAP temperature and weaken the bitumen bond between RAP aggregates. The variation in stiffness values at different measured directions will be substantial for a heterogeneous mixture and minor in the case where the recycled mixture is homogeneous. The homogeneity level is also substantially affected by the sizes of RAP material. For the same mixing effort. the point that the bitumen bond is overcome by mechanical mixing effort is defined as a critical point. the better the interaction between RAP and virgin binder. As the thermal transfer process is time dependent. the mechanical mixing effort only distributes the RAP all over the mixture at its approximately original size. RAP lumps under mechanical mixing will start to disintegrate into separate pieces of aggregate coated by RAP binder. the recycled mixtures composed of small RAP are generally more homogeneous than those made from large RAP. The long preheating condition at a temperature considerably higher than the softening point of the RAP binder coincidently deactivates the bitumen bond between RAP aggregate particles. Therefore. At this critical point. In the SHRP method. the stiffness of 162 . On the contrary in the FS method. the incorporation between RAP and virgin binder depends primarily on the original size of RAP. Under mechanical mixing. If the RAP/superheated virgin aggregate duration does not exceed the critical duration. The mixing methods considerably affect the reaction between RAP and virgin binders as the mixing mechanisms determine how RAP and virgin binder are blended together. The more homogeneous the recycled mixture. Not only the stiffness variation among different specimens but also the variation of stiffness values measured in different directions for the same specimen significantly reduces. the more interaction between RAP and virgin binder.

Quantitatively. the stiffness values of recycled asphalt mixtures are lower than those of Complete Blending mixture. the complete blending between RAP and virgin binder assumed in the design process would never exist in the production of recycled asphalt mixtures. even at favourable conditions of considerably long mixing times compared to those in the real asphalt mixing plant. Qualitatively. Although the increase of mixing duration has positive effects on the homogeneity. as RAP binder is not completely blended with virgin binder. 163 . there still exist a considerable proportion of RAP as lumps.recycled mixture generally increases once the RAP/superheated virgin aggregate mixing duration is extended.

stiffness under certain stress level is measured at the direction that specimen has the lowest stiffness modulus.1 Introduction This chapter investigates the effect of different mixing methods and RAP sizes on fatigue life of recycled mixtures. Fatigue life of recycled mixtures manufactured by different mixing methods are determined by indirect tensile fatigue test (BS-EN:12697-24.3. The target stress levels are selected to obtain a wide range of fatigue lives. 2004). 7. The rise time is 124 milisecond with the cyclic load pulse of 0. After stiffness measurement. 2004). However for Black Rock case with lower stiffness modulus. Normally. However. The maximum is at least ten times greater than the minimum fatigue life (DD-ABF. the range of stress level is from 50 to 200 kPa. LR FS-2. 1993). the stiffness is the average of two stiffness values at two perpendicular directions of specimen. Before fatigue testing. The failure criteria is the point where total vertical deformation of specimen reaches 9 mm (BS-EN:12697-24. This is also the 164 . The indirect tensile fatigue test is carried out under stress-controlled mode by Nottingham Asphalt Tester (NAT) at 20oC.1). stiffness under the same stress level that specimen will experience in the fatigue test is determined by indirect tensile stiffness test (DD-213. 1996). Conventionally. The schematic of the indirect tensile fatigue test is illustrated in Figure 89. for this experiment. is 10. The fatigue lives of recycled mixtures are compared based on the parameters of the fatigue equations and the number of loading cycles to fatigue failure at 100 microsrain.67 Hz.2 Materials preparation and testing plan The materials for this experiment are those specimens used for ITSM (Section 6. these specimens are subjected to indirect tensile fatigue test (ITFT) to determine the number of loading cycles to failure.7 Effects of mixing methods on fatigue life of hot recycled asphalt mixtures 7. The materials used in this chapter are the same as those in Chapter 6. for instance. the target stress levels vary from 100 to 400 kPa. due to the stiffness modulus values of recycled specimen varying considerably with different measuring directions. The number of specimens tested for each set of material.

max is the maximum tensile stress at the centre of specimen (kPa) is Poisson’s ratio (assumed to be 0.direction for the fatigue test due to the assumption that failure will occur initially at the weakest direction. max  1  3   1000 Sm (22)  x . linear regression analysis using Least Squares method is applied to obtain the best-fitted equation for fatigue life.35) is the indirect tensile stiffness modulus at  x .max (MPa)  Sm Based on the testing data of 10 specimens for each set of material. The 165 . Figure 89: Schematic of ITFT test The initial maximum tensile strains at the centre of specimen are plotted against the relevant numbers of loading cycles to failure on logarithmic scales. 1996):  x . The initial maximum tensile strain at the centre of specimen is calculated as follows (DD-ABF.max  Where:  x.

7. 1996). due to the fact that 160/220 Pen is considerably softer than fully blended recycled binder and 70/100 Pen bitumen. K 2 The fatigue lives of recycled asphalt mixtures are then compared to each other and to those of control mixtures to study the effects of different mixing protocols and RAP materials on fatigue life of hot recycled asphalt mixtures. The K2 values of CB and CB-V are approximately similar and extremely lower than that of the BR mixture. This is because the rheological properties of recycled bitumen for CB mixture is almost the same as that of 70/100 Pen binder (Section 6. The results show that regressed fatigue lines of CB and CB-V mixtures are almost statistically the same. the mixture with the stiffer bitumen will provide longer fatigue life. As fatigue failure normally occurs between 30-200 microstrain range (Read.3. CB and CBV. BR. 166 . The R square values indicate that the empirical relationship (Equation 23) between number of loading cycles to failure and initial maximum strain at the centre of specimen fits the experimental data. 7.empirical relationship that is used for regression analysis is expressed as follows (Pell.1). CB and CB-V have longer fatigue life than that of BR mixture (Table 52). the fatigue life of BR mixture is more stress dependent than that of CB and CB-V mixtures.3 Results and analysis The parameters of fatigue equations of control mixtures and recycled mixtures manufactured by different methods are summarized in Table 51.3. 1973): N f  K 1  i  Where: Nf K2 (23) Number of load applications to failure at particular level of initial strain Initial tensile strain Material Coefficients i K1 . the numbers of loading cycles to failure at 100 microstrain of different mixtures are also extrapolated for comparison purposes (Table 52).1 Control mixtures Figure 90 illustrates the regressed fatigue lines of three control mixtures. In addition. The results confirm the finding of Cooper and Pell (1974) that under stress-controlled mode. However.

3. The data indicates that a small increase in stiffness value can be associated with a tremendous extension of the fatigue life of recycled asphalt mixture. The results show that for both LR and SR mixtures. Actually. the data also demonstrates that the fatigue line of recycled mixture is gradually less dependent on the initial maximum tensile strain or stress level. the variation in stiffness values is considerably reduced when the mixing duration is increased (Section 6.081 to -3. This results in the substantial increase of fatigue life.2). The results indicate that to increase mixing effort also means to improve the reliability of fatigue line of recycled asphalt mixture. In fact. while the stiffness increases slightly. The data clearly shows that when the mixing time increases. the boundaries with 95% confidence of these regressed lines are extremely different. especially with LR mixture.2 Recycled mixtures Figures 91 and 92 present the regressed fatigue lines of LR and SR recycled mixtures manufactured by different mixing protocols. the dependence of fatigue life on stress level gradually transforms from that of black rock to complete blending when the mixing time increases. while the stiffness increases linearly.5. This is due to the homogeneous level of recycled asphalt mixture. Even for recycled mixture composed of small RAP material.899 when the mixing time increases from 2 to 8 minutes (Table 51). Figure 94 illustrates the correlation between stiffness and fatigue life at 100 microstrain versus mixing times of recycled mixtures composed of large RAP material. approximately five times once the mixing time is extended from 2 to 8 minutes (Figure 95). the fatigue life increases exponentially. 167 . Figure 93 shows that the boundary with 95% confidence of fatigue line of LR recycled mixture is significantly narrower once the mixing time increases from 2 to 8 minutes. the fatigue lines are slightly improved once the mixing time is extended.7. the confidence with such a large extrapolation will be certainly low. there is a noticeable increase in fatigue life. For recycled mixture composed of large RAP material. the mean K2 values reduced from -2. In addition. the stiffness of recycled mixture generally increases once the mixing time is extended. The phenomena will be more pronounced if the fatigue lives of the recycled asphalt mixtures are extrapolated at 30 microstrain. However. Although there are not considerable differences between these fatigue lines. As more RAP and virgin binder can be interacted.

23E+11 2.446 SR FS-6 0.96 3.311 SR SHRP 0.15E+12 -3.896 -3.47E+12 -3.78E+11 6.934 SR FS-8 0.631 SR FS-4 0.21E+11 5.18E+12 -3.71E+10 1.69E+09 1. R K1 K2 Square Min Mean Max Min Mean Max LR FS-2 0. The results from fatigue testing also demonstrate that with the same mixing effort.95 4. The fatigue lives of SHRP mixtures are approximately the same as those of mixtures manufactured by FS method with 4 to 6 minutes mixing time (Table 52).35E+09 2.634 -3.07E+14 -4.10E+12 -3.278 -3.45E+13 2.117 -2.94E+07 1.31E+13 -4.907 -3.28E+05 2. the homogeneity and stiffness of recycled mixture are mutually correlated.536 LR FS-8 0.The size of RAP also significantly affects the number of loading cycles to fatigue failure of hot recycled mixtures.18E+12 5.619 -2.088 -3.75E+14 -4.20E+12 1.49E+12 1. In fact.067 LR FS-6 0.26E+12 -3.73 8.00E+09 1.04E+13 4. This is because to produce a hot recycled asphalt mixture with a certain level of homogeneity. with the same mixing efforts.899 -3.65E+13 -4.753 -2.01E+08 -1.468 -2.98 6.892 -1.5. This supports the statement in Section 5.94 2.066 LR FS-4 0.29E+11 3.81E+10 2.698 -3.757 -3.03E+11 -3.275 -2.462 -3.07E+13 6.73E+06 1.99 1.97 8.705 -3.58E+11 2.081 -1.268 -3.039 SR FS-2 0. the bigger the size of RAP material.34E+15 -4. recycled mixtures composed of large RAP have much lower fatigue life than that of small RAP mixtures (Figure 96).314 -2.441 -1. The data from fatigue test is in an agreement with stiffness data.85 1. the more mixing effort is required.98 6.097 -2. the fatigue lives of recycled asphalt mixtures are substantially lower than those of the complete blending (CB) case.971 CB-V 0.699 -3.91 2. In addition. recycled mixtures composed of large RAP generally have lower stiffness than that of small RAP mixtures (Section 6. the fatigue life of SR FS-8 mixture is just about half of the CB mixture.71E+11 2. Mixing methods also have considerable effects on fatigue life of recycled mixture as the mixing mechanisms determine how RAP and virgin materials are blended with each other.96 7.331 BR 0.519 LR SHRP 0.5.694 -3.94 2.79E+11 -3.65E+10 5.65E+08 1.1 that the complete blending between RAP and virgin binder assumed in the design would never occur in the industry.610 -1.328 Table 51: Parameters of fatigue equation at 95% confidence of control and recycled asphalt mixtures manufactured by different mixing methods 168 .95 5.3).132 -2.85E+11 5.101 -2.108 CB 0. Even when small RAP material is used with 8 minute mixing time.94E+07 2.79E+09 1.92E+08 1. In addition.

0E+03 1.0E+05 Number of Loading Cycles to Failure Figure 90: Fatigue lines of control mixtures 169 .0E+02 BR CB 1.LR FS-2 LR FS-4 Fatigue life at 100 microstrain 20038 30807 LR FS-6 74896 LR FS-8 170696 LR SHRP 53345 SR FS-2 67056 SR FS-4 75659 SR FS-6 137838 SR FS-8 247131 SR SHRP 143872 CB-V 377413 CB 327959 BR 11713 Table 52: Extrapolated fatigue life at 100 microstrain of recycled asphalt mixtures manufactured by different mixing methods 10000 1000 Microstrains 100 CB-V 10 1.0E+04 1.

0E+05 1.0E+02 1.0E+04 SR SHRP 1.0E+02 LR FS-4 LR FS-6 LR FS-8 LR SHRP 1.0E+04 Number of Loading Cycles to Failure Figure 91: Fatigue lines of LR mixtures manufactured by different methods 10000 1000 Microstrains 100 SR FS-2 10 SR FS-4 SR FS-6 SR FS-8 1.0E+05 1.0E+03 1.0E+03 Number of Loading Cycles to Failure Figure 92: Fatigue lines of SR mixtures manufactured by different methods 170 .10000 Microstrains 1000 100 LR FS-2 10 1.

0E+02 1.8E+05 1.4E+05 1800 1700 1600 1500 1400 1300 1200 Fatigue Stiffness 1100 1000 LR FS-2 LR FS-4 LR FS-6 LR FS-8 1.0E+04 0.0E+04 6.0E+04 2.0E+05 Number of Loading Cycles to Failure Figure 93: Fatigue lines with boundaries of 95% confidence interval of LR FS-2 and LR FS-8 mixtures 2000 2.0E+03 LR FS-8 1.0E+00 Figure 94: Fatigue life at 100 microstrain and stiffness versus different mixing time of LR mixtures 171 Stiffness (MPa) Fatigue Life .0E+04 1.0E+05 1900 1.0E+05 8.2E+05 1.0E+04 4.10000 Microstrains 1000 100 LR FS-2 10 1.6E+05 1.

2E+05 1.6E+05 1.2E+05 2.4E+05 1.0E+04 2.0E+00 SR FS-2 SR FS-4 SR FS-6 SR FS-8 Fatigue Stiffness 1400 1300 1200 1100 1000 Figure 95: Fatigue life at 100 microstrain and stiffness versus different mixing time of SR mixtures 2.4E+05 2.0E+05 8.2E+05 2.6E+05 1.0E+04 4.8E+05 2000 1900 1800 1700 1600 1500 Fatigue Life 1.0E+04 2.2E+05 1.4E+05 2.0E+04 6.0E+04 0.0E+05 8.0E+04 0.4E+05 1.0E+05 1.6E+05 2.6E+05 2.0E+04 4.0E+00 LR SR FS-2 FS-4 FS-6 FS-8 SHRP Figure 96: Relation between RAP sizes and fatigue life at 100 microstrain of recycled mixtures manufactured by different mixing methods 172 Stiffness (MPa) .2.0E+05 1.8E+05 Fatigue Life 1.0E+04 6.

2 Materials preparation and testing plan The materials for determining the resistance to permanent deformation of hot recycled mixtures are summarized in Table 53.1 Introduction This chapter studies the effect of different mixing methods on resistance to permanent deformation of recycled mixtures.8 Effects of mixing methods on permanent deformation of hot recycled asphalt mixtures 8. permanent deformation determination is implemented 15 days after the day of compaction. The material preparation is the same as that in Chapter 6. Only large RAP material is used in this experiment. RAP/superheated virgin aggregate varies from 2 to 8 minutes. specimens are stored in a cabinet at a temperature of 20oC. The magnitude of stress level is 100 kPa. RAP/Virgin aggregate mixing duration (minutes) LR FS SHRP 2 × 4 × × 6 × 8 × Table 53: Test plan to study the effects of different mixing methods on resistance to permanent deformation The resistance to permanent deformation is determined following the procedure in DD 226:1996. The proportion of RAP is also 40%. specimen is 173 . In FS method. The test is conducted at 40oC and under repeated axial dynamic load conditions. The resistance to permanent deformation of recycled mixtures is determined by repeated loading axial test . field simulation (FS) and the SHRP method. Similar to stiffness test. After manufacture. Each load pulse comprises of 1 second for load application and 1 second for the rest period.1. To eliminate the effect of air void content. The number of load applications that the specimen will experience is 3600. Recycled mixtures are manufactured by two different methods.3. 8. Only large size of RAP is used in this experiment. There are also two control mixtures presenting the “Black Rock” and “Complete Blending” cases. The manufactures of theses mixtures are similar to the procedures in Section 6. Before testing. the resistances to permanent deformation of recycled mixtures are compared based on rutting characteristic after densification stage.

In addition. During the test.conditioned at testing temperature for at least 8 hours. to minimize the friction between surfaces of specimen and testing plates. Figure 97: Schematic of RLAT test to determine resistance to permanent deformation 174 . accumulated vertical deformations are recorded after each load application. The deformations of recycled mixtures are also compared to those of control asphalt mixtures to study how different the properties of recycled mixtures are to those assumed in the design process. Permanent deformation patterns of recycled mixtures manufactured by different mixing methods are compared to each other to study the effects of mixing methods on deformation resistance. surfaces of specimen are coated evenly and thinly with silicone grease and graphite flakes. Figure 97 illustrates the schematic of RLAT test to determine the resistance of mixture to permanent deformation. Permanent deformation is then plotted against number of loading applications that the specimen experiences.

 Tertiary stage: the vertical strain per loading cycle increases rapidly again to failure. the results just represent the densification and a part of the secondary stage. although the accumulation of vertical strain increases rapidly. the volume decreases due to the load will be equal to the volume increases in the adjacent areas. the test data shows that deformation caused by densification accounts for a considerable part. the result might be tremendously dependent on the air void content. During this stage. or densification. accumulated permanent deformation increases rapidly. In addition. occurs during the first 2000 cycles. However. generally under 3600 loading applications.. 1991). Figure 98 is an example of the permanent deformation versus the number of load application of LR FS-2 mixture. during the secondary stage (after 2000th load application). the accumulated vertical deformation has a linear relationship with the number of loading cycles. 2006). The primary stage is also called densification in the other literature (Bahuguna et al. During this stage. under the condition of repeated load axial test with a total of 3600 loading cycles. On the contrary. 1991). In this experiment. The data shows that the primary stage.. If the permanent deformations at the 3600th load application are used for the purpose of comparison. the volume of specimen under the load decreases. In this stage. The effect of air void content on permanent deformation has been found in previous literature (Sousa et al.3 Results and discussion The permanent deformation behavior of asphalt mixture under creep test conditions is normally divided into three stages (Bernasconi and Piatti.8. Air void content plays a very important role in the primary stage or densification. 1978):  Primary stage: in this stage.  Secondary stage: the secondary stage starts when the permanent deformation per loading cycle reaches approximately constant value. the air void contents of testing specimens are always hard to control even in 175 . This will consequently cause an increase in the density. the permanent deformation per each loading cycle decreases.. This stage represents the shear deformation and is considered to be the primary deformation behavior of asphalt mixture (Sousa et al.

19 4.51 3.26 4.20 4.28 3.42 4.38 5.83 4.38 5.42 5. Therefore.07 5.83 6.18 3. The higher the value of the rutting 176 .05 2.89 4.54 4. The permanent deformations and air void contents data is summarized in Table 54.83 2.27 4.38 6.77 6.00 2.61 3.03 3.50 4.93 5.12 4.59 5.86 3.20 5.81 3.77 4.46 5.67 2. Sample Number 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 Air Void (%) 4.40 5.18 6.50 6.74 2. air void content will have insignificant effects on secondary stage as the deformation in this stage occurs without volume change.12 3.73 5.18 3.30 3.74 2.77 5. the coefficient of the linear relationship between permanent deformation and number of load application in secondary stage can be used as a rutting indicator.12 3.58 4.34 4.50 5.32 4.29 3.54 5.64 5.91 4.20 3.93 Deformation (%) 3.91 3.24 6.85 LR FS-2 LR FS-4 LR FS-6 LR FS-8 LR SHRP CB BR Table 54: Permanent deformation data of control and recycled specimens manufactured by different mixing methods However.40 3.58 3.58 5.14 3.29 3.laboratory manufacture.53 4.79 6.82 3.00 3.24 2.36 7.19 3.20 3.

the less the susceptibility of recycled mixtures to permanent deformation under repeated loading conditions. for instance. Rutting indicator is the coefficient of the equation expressing the linear relationship between accumulated permanent deformation and number of load cycles after 2000th load applications. The coefficient of variation substantially decreases from 49% to 10% once the mixing time is increased from 2 to 8 minutes. For hot recycled asphalt mixtures. the hypothesis is under mixing effort. there is a proportion of the rutting indicators that are close to that of the BR case. The results indicate that RAP/superheated virgin aggregate mixing duration significantly affects the resistance to permanent deformation of recycled mixtures. comparisons between rutting indicators might reveal the effects of mixing efforts on the resistance to permanent deformation. On the contrary. The longer the mixing time. the lower the resistance to permanent deformation. as recycled mixture become more homogeneous. rutting indicators of recycled mixture mixed for 8 minutes are close to those of the CB case (Figure 99). As all the mixture variables are deliberately controlled the same. For 2 minutes mixing duration. the more integration between RAP and virgin binder results in the increase of mixture stiffness. Actually. the resistance to permanent deformation of recycled mixture never reaches that of the complete blending (CB) mixture.indicator. rutting indicator variation also decreases. The data also indicates that mixing method significantly affects the resistance to permanent deformation of recycled mixtures. even in favourable conditions that never exist in the asphalt production industry. This is because the longer the mixing time. rutting indicator will increase unless under mixing effort. rutting indicator might be altered in relation with the incorporation between RAP and virgin binder. 177 . 2 hour preheating RAP at 110oC of SHRP or 8 minutes mixing time in field simulation (FS) method. In addition. The rutting indicators of control and recycled mixtures manufactured by different mixing methods are shown in Figure 99 and Table 55. integration between RAP and virgin binder is enhanced and vice versa. Recycled mixture manufactured by the SHRP method generally has better resistance to permanent deformation than those of FS mixtures (Figure 99). However.

6 5 Permanent Deformation (%) 4 3 2 1 0 0 500 1000 1500 2000 2500 3000 3500 4000 Number of Loading Circles Figure 98: Permanent deformation versus number of loading application of LR FS-2 specimens 0.0001 BR LR FS-2 LR FS-4 LR FS-6 LR FS-8 LR SHRP CB Figure 99: Inter-quartile rutting indicator ranges of control and recycled specimens manufactured by different mixing methods 178 .0004 Rutting Indicator 0.0003 0.0002 0.0005 0.

32 22.58 21.20E-04 2.71E-04 2.93E-04 2.86E-04 2.03E-04 1.40E-04 2.06E-04 2.39E-04 2.12E-04 2.87E-04 3.66E-04 1.29E-05 COV (%) 49.64E-04 3.46E-04 1.18E-04 3.74E-04 2.83E-04 2.40E-04 3.79E-04 4.78E-04 2.40E-04 4.33E-04 2.84E-04 2.13E-04 1.22E-04 SD 1.78E-04 1.54 Table 55: Rutting indicator data of control and recycled specimens manufactured by different mixing methods 179 .77E-04 4.40E-05 1.10E-04 2.20E-04 2.61E-04 1.62E-04 2.16E-04 Rutting Indicator (s-1) 2.53E-04 3.86E-04 1.83 20.08 10.70E-05 5.15E-04 2.83 12.65E-04 Mean 2.21E-05 5.61E-04 2.05E-05 4.48E-06 3.LR FS-2 LR FS-4 LR FS-6 LR FS-8 LR SHRP CB BR 5.62E-04 1.13E-04 2.60E-04 1.91E-05 2.88E-04 2.21E-04 4.06 10.67E-04 2.

Otherwise. temperature.  RAP material used in laboratory is normally processed to less than ½ or 1 inch. The diffusion model assumes that single RAP aggregate particles coated by RAP binder are covered by virgin binder. The sizes of RAP used in practice are sometimes considerably bigger than that used in the laboratory. RAP at ambient temperature is mixed with superheated virgin aggregate for short time.  The result estimated by available viscosity mixing equations is normally not really close to the actual value.  The current laboratory mixing methods indicate shortcomings as these methods could not depict the mixing mechanism that occurs in the asphalt mixing plant. the mechanical properties of recycled asphalt mixture would be deviated from expected. Laboratory procedure allows RAP to be preheated for a long duration at high temperature before mixing with virgin material.9 Conclusions and recommendations for future research 9. for instance.  The diffusion pattern is supposed to be the same for the whole recycled mixture. The extensive RAP preheating time at high temperature might coincidentally enhance the interaction between RAP and virgin binder. maximum 90 seconds in the asphalt mixing plant.1 Conclusions The principal conclusions which can be drawn from the literature review include:  The philosophy of available viscosity mixing equations assumes that aged binder and virgin binder are completely blended. The question is whether this assumption actually occurs in the recycled asphalt production process. These viscosity mixing equations can provide an approximate value. The diffusion is long term and affected by many factors. Trial experiments with increment proportions of RAP or virgin binder in the blend should be used to obtain the accurate value. Virgin binder starts to diffuse into RAP binder. The question is whether the asphalt mixing plant could generate the recycled mixture with the same homogeneity as that manufactured in the laboratory. However. On the contrary. there is a wide range of RAP sizes that have been used in the highway industry. 180 .

The long RAP preheating time also slightly alters the properties of RAP binder.  There are reciprocal relationships between mixing effort. Using Grunberg and Nissan equation. The mixing mechanism is expressed as follows:  Virgin aggregate is superheated to required temperature 181 . and mechanical properties of recycled mixtures. the relationship between mixing effort and mechanical properties could be quantified. Using this method. the interaction parameter G12 should be determined for each specific bitumen blend.bitumen thickness.  The laboratory mixing method conventionally used to prepare recycled asphalt specimens tends to overestimate the mechanical properties of recycled asphalt mixtures. homogeneity. The fact that one constant value of G12 is used universally would result in substantial errors in viscosity estimation. Therefore. On the contrary. and chemical composition of bitumen. the viscosity could be predicted within 10% of the actual values. The diffusion model does not take into account the effect of RAP lump and existence of virgin aggregate. The long RAP preheating time that never exists in the industry coincidentally enhances the incorporation between RAP and virgin binder. if interaction parameter G12 is properly determined for each specific bitumen blend. could not be evaluated in a quantitative manner. The mutual relationships between mixing effort and homogeneity.  The newly developed laboratory mixing method provides a better means of describing the mixing mechanism between RAP and virgin material in the industrial asphalt mixer. the effect of RAP size is negligible. The other viscosity mixing rules using one universal interaction parameter G12 generally generate considerably high residual errors. or homogeneity and mechanical properties. The principal conclusions which can be drawn from the experimental work presented in this thesis include:  Grunberg and Nissan equation proves to be the most efficient rule for predicting the viscosity of bitumen blends. For instance. the predicted viscosities by Arrhenius equation (ASTM D4887) are within approximately 30% of the actual values and 50% for DLV method.

Hence.  The RAP superheated virgin aggregate mixing duration should be longer than a critical duration in which all the bitumen bonds in RAP lumps are deactivated. the recycled asphalt mixture will possess the properties of the Black Rock mixture. This method also allows the effects of RAP sizes on quality of recycled mixture to be studied.  Measuring stiffness values of a cylindrical specimen in different orientations indirectly expresses the heterogeneity of recycled mixtures. The variation in stiffness values at different measured directions will be substantial for heterogeneous mixtures and relatively minor in the case of recycled mixtures that are homogeneous. In this situation. In this step.  The newly developed method allows the effect of mixing time on homogeneity level to be investigated. the mixtures composed of small RAP 182 . the more efficient the mechanical mixing. the higher level of homogeneous recycled asphalt mixtures are manufactured. besides the fact that RAP disintegration and distribution still progress. For the same mixing effort. RAP material is heated up and softened by the thermal energy transferred from superheated virgin aggregate. there is incorporation or rejuvenation between RAP and virgin binder. The surface analysis of slices from top to bottom clearly depicts 3D distribution of RAP materials in recycled mixture. Otherwise. The homogeneity level is also substantially affected by the sizes of RAP material. This situation is exaggerated if large RAP sizes are used. In addition.  The increase in mixing duration significantly improves the homogeneity level of recycled mixture. mechanical mixing will separate RAP lumps and enhance the interaction or rejuvenation between RAP and virgin binder. virgin bitumen and aggregate. the effect of mechanical mixing can only distribute the RAP material all over the mixture. As an inconsiderable proportion of RAP binder can be rejuvenated by virgin binder. In this step. RAP materials start to disintegrate and are distributed all over the mixture under mechanical mixing  RAP/virgin aggregate blend is mixed with virgin binder. the use of colour binder helps to position accurately the location of RAP binder.

considering the whole service life. However. even at favourable conditions of extremely long mixing time compared to that in the real asphalt mixing plant. However. The mixing condition in laboratory work is also different from the field. A slightly linear increase in stiffness can result in an exponential increase in fatigue life of the recycled mixture. different mixing conditions should be considered in future studies. On the contrary.  Although increase of mixing duration has positive effects on the homogeneity. Qualitatively. 9. Therefore. Therefore. resistance to permanent deformation. The mixing efficiency of a real industry mixer might be totally different from that of a laboratory mixer. Quantitatively. as RAP binder is not completely blended with virgin binder. The more homogeneous the recycled mixture. the more interaction between RAP and virgin binder. for instance under the effects of frost and moisture. The mixing process in this research occurs under conditions of no moisture content. the complete blending between RAP and virgin binder assumed in the design process would never exist in the production of recycled asphalt mixtures. recycled mixtures become stiffer and thus have better resistance to permanent deformation and fatigue failure. the mechanical mixing characteristic of a real asphalt mixer is quite different from that found in the laboratory. Therefore.are generally more homogeneous than those made from large RAP.2 Recommendations for future research The mechanical mixing characteristic substantially affects the homogeneity and mechanical properties of hot recycled asphalt mixture. The movement of material in the laboratory mixer is primarily horizontal. in asphalt mixing plants. a considerable proportion of RAP still exists as lumps. RAP materials with different moisture contents are mixed with superheated virgin aggregate. this includes not only horizontal but also vertical movement. Therefore. Therefore. in the industry. and fatigue life of recycled asphalt mixture are considerably poorer than those of the Complete Blending mixture. recycled asphalt pavement has to work in different climatic conditions. The mixing process between RAP and virgin aggregate might occur under extremely hot and steamy conditions. 183 . This might alter the properties of RAP and virgin binder. In addition. the design methodology for recycled asphalt mixture tends to overestimate the performance of hot recycled asphalt mixture. the stiffness values. research should be carried to validate the finding of the laboratory work.

As a result. the situation might be different with material using gap-graded aggregate. Therefore. How virgin binder diffuses into and recovers the properties of aged binder are not covered by this thesis. The material used in this research is primarily Dense Bitumen Macadam (DBM). a new design method that involves the partial blending between RAP and virgin binder should be considered in future research. 184 . Experiments in this thesis were carried out primarily at a macro scale level. the effect of diffusion process on mechanical properties of recycled asphalt mixture should be considered in future research. However. The gradation of DBM conforms to continuous-graded theory. Once virgin binder is in contact with RAP binder. The rejuvenation might be influenced by many factors. the micro scale level has not been taken into account. These RAP lumps will restrain the interaction between RAP and virgin binder. the presence of filler or proportion of filler in the filler mastic. the rejuvenation or diffusion progresses with time. However. for instance. Future research should take into account the effect of aggregate grading on the incorporation between RAP and virgin binder. temperature. or chemical composition of bitumen binder. there will be a substantial proportion of RAP materials existing as agglomerate after crushing.low temperature cracking resistance and long term durability of recycled asphalt mixture should be also considered in future research. As the complete blending between RAP and virgin material assumed in the design process would never exist in the recycled asphalt production process.

Transportation Research Record. P. A. Y. USA. 185 . M. ANDERSON. Transportation Research Record 1217. A. (2003) State of the Art Report on Aging Test Methods for Bituminous Pavement Materials. Chattanooga.C. BERNASCONI. 23.. Vol 1810. M. T. Vol 4. 193-212. L. BELL. & RICHMOND. D. A. (2005) Milling and Recycling.. G. J. Covallis OR 97311. USA. W. E. ASTM-D4887-99 (2003) Standard Practice for Preparation of Viscosity Blends for Hot Recycled Bituminous Materials. P.. & PAPOULIA. R. D. C. & MARTIN. PANOSKALTSIS. COLLINS. J. BELL.. The International Journal of Pavement Engineering. Transportation Research Board. A. D. (2006) Identification and Modelling of Permanent Deformations of Asphalt Concrete. p: 54-62.. MAY.. J. 4010 Jerome Avenue. D. Y. S. Covallis OR 97311. (2003) Segregation: Causes and Cures. & BROWNFIELD. M. G. V. J. A. Washington D. 123. American Society of Testing Materials. BROCK. (1994b) Laboratory Aging of Asphalt Mixtures: Field Validation. A.C. (2002) Zero Shear Viscosity of Asphalt Binders. & DARLING. J. J. Journal of Engineering Mechanics. The Annals of Mathematical Statistics.References: AIREY. HIR. & FELLIN. USA. J. Oregon State University. 231-239. p: 165-176. R. & PIATTI. Oregon State University. & RENEGAR. D.. National Research Council. Transportation Research Board. BROCK. (1978) Creep of Engineering Materials and Structures. SHRP A-338. ANDERSON. ABWAHAB. G. BAHUGUNA. BROWN. WIEDER. K. R. J. C. USA. CRISTI. National Research Council. G. TN 37407. September 2003. E. p: 1-8. Applied Science Publishers Ltd. D. London. ASTEC Inc. ASTEC Inc. (1989) Investigation of Segregation of Asphalt Mixtures in The State of Georgia. (1952) Asymptonic Theory of Certain "Goodness of Fit" Criteria Based on Stochastic Processes.. L. pp. & SOSNOVSKE. D.. Washington D. (1994a) Selections for Laboratory Aging Procedure for Asphalt-Aggregate Mixture.

Stoppered Pyknometers. 186 .Ring and Ball Method. BS-EN:13108-5 (2006) Bituminous mixtures.BROWN. British Institution. BS-EN:12697-24 (2004) BS EN 12697-24: Bituminous Mixture-Test Methods for Hot Mix Asphalt.Material Specification Part 1: Asphalt Concrete. BS-EN:13108-1 (2006) BS EN 13108-1: Bituminous Mixture . British Institution. British Standard Institution. British Standard Institution. BS-EN:12697-31 (2004) BS-EN 12697-31: Bituminous Mixture . British Standard Institution.Test Method for Hot Mix Asphalt.Test Method for Hot Mix Asphalt. BS-EN:12697-5 (2002) BS-EN 12697-5: Bituminous Mixture . BS-EN:13108-4 (2006) Bituminous mixtures.Part 1: Determination of particle sizes distribution . BS-EN:933-1 (1997) Test for geometrical properties of aggregates .Hot Rolled Asphalt. British Institution. BS-EN:12697-26 (2004) BS-EN:12697-26: Bituminous Mixture Test Methods for Hot Mix Asphalt.Test Method for Hot Mix Asphalt. Part 24: Resistance to fatigue. Part 33:Specimens Prepared by Roller Compactor. BS-EN:12697-33 (2003) BS-EN 12697-33: Bituminous Mixture . Residential Course at The University of Nottingham. British Institution. F. BS-EN:12697-8 (2003) BS-EN 12697-8: Bituminous Mixture . Material Specifications . British Standard Institution. Design and Evaluation. Material Specifications . (2000) Bituminous Pavement: Materials. Section J. BS-EN1427 (2000) Softening Point of Bitumen . Part 5: Determination of the maximum density. BS-EN:1426 (2000) Penetration of Bituminous Materials. Bituminous Materials. Part 6: Determination of bulk density of bituminous specimens. S. Stiffness and Fatigue Cracking. British Standard Institution.Test Method for Hot Mix Asphalt.Sieving method. British Institution. Part 26: Stiffness. Part 4: Bitumen Recovery: Fractionating Column. British Institution. Part 31:Specimens prepared by gyratory compactor. Part 8: Determination of void characteristic of bituminous specimens.Test Method for Hot Mix Asphalt.Stone Mastic Asphalt. British Institution. BS-EN:12697-4 (2005) BS-EN 12697-4: Bituminous Mixture .Test Method for Hot Mix Asphalt. British Institution. British Standard Institution. BS-EN:12697-6 (2003) BS-EN 12697-6: Bituminous Mixture . BS-EN:3838 (1996) Determination of Density Using Capillary . British Standard Institution.

417-437. J.C. H. R. Vol 1630. Journal of Colloid Science.Porous Asphalt. M. C. M. Transportation Research Record 1501. H. Incorporated. (1995) Viscosity Mixing Rules for Asphalt Recycling. USA. LIU. A. DAVIDSON.BS-EN:13108-6 (2006) Bituminous mixtures. Transportation Research Board. p:75-87. & BULLIN. (1965) Rheology of Non-Newtonian Fluids: A New Flow Equation for Pseudoplastic Systems. GLOVER. Vol 36. Report 633. TRRL Laboratory. DANIEL. M. GLOVER. S.. J. (2004) Physics. (1995) Development of Superior Asphalt Recycling Agents. CHAFFIN. J. Material Specifications . J. Crowthorne. E.. KIM. C. Y. R. C. M. & WOLOSICK. W. (1997) Supercritical Fractions as Asphalt Recycling Agents and Preliminary Aging Studies on Recycled Asphalts. R. K. Washington D.. GLOVER. M. British Standard Institution. Transportation Research Board. Berkshire. University of Texas A&M and Institute. (1998) Effects of Aging on Viscoelastic Properties of Asphalt-Aggregate Mixtures. & A.C. CHAFFIN. Specification for constituent materials and for mixtures.. CUTNELL. p: 15-22.C. S. Department of Chemical Engineering. & JOHNSON. Wiley. (1980) Modifier Influence in the Characterization of Hot-Mix Recycled Material. BULLIN.. Transportation Research Record.Mastic Asphalt. & PELL. Journal of Industrial & Engineering Chemistry Research.-J. P. Texas Transportation 187 . J. K. R. COOPER. National Research Council.. J. R. Material Specifications .. M. S. B. J.-S. D. USA. CARPENTER. J. J. R. DAVIDSON. R. Washington D. p: 21-27. BS-EN:13108-7 (2006) Bituminous mixtures. Transportation Research Record 777.. J. p: 656-666. Transportation Research Board. A. Washington D. National Research Council. DAVIDSON. & LEE. R. J.. CROSS. BS:4987-1 (2005) BS 4987-1 (2005):Coated macadam (asphalt concrete) for roads and other paved areas. J. & LIN. John & Sons. British Standard Institution. CHAFFIN. (1974) The Effects of Mix Variables on the Fatigue Strength of Bituminous Materials. M. National Research Council. 20..

GARDINER. EDWARDS. National Highway Cooperative Research Program. HEUKELOM. (1985) Surface Cracking of Asphalt Mixture in Southern Africa. DD-ABF (1996) Method for the determination of the fatigue characteristics of bituminous mixtures using indirect tensile fatigue. (1980) Guidelines for Recycling Pavement Materials. T. British Standard Institution. W. (1978) Problems and Solutions for A Hot Mix Recycling Plant. National Cooperative Highway Research Program. (1977) Practical Aspects of Reconstituting Deteriorated Bituminous Pavements. J. HUGO.C. C. D. (2001) Mixing In The Process Industries. (1977) Viscosities of Binary Liquid Mixtures: The Effectiveness of Mixture Equations. A. p: 454501. Washington D. D. & TERREL. (2000) Segregation in Hot-Mix Asphalt Pavements. Vol 54. 164. & MENDENHALL. (1978) Design of Recycled Asphalt Pavements and Selection of Modifiers. Proceedings. Nature (London). A. HARNBY. & BROWN. M. B. GRUNBERG. J. (1973) An Improved Method of Characterizing Asphaltic Bitumens with the Aid of Their Mechanical Properties.. NHCRP 224. (2005) Laboratory Investigation of Mixing Hot-Mix Asphalt with Reclaimed Asphalt Pavement.. W. No 625. AAPT. M. L. National Research Council. S. pp 62-98. UK Department of Industry. Great Britain. LI. p: 35-46. R. Special Techincal Publication. R. J. & KENNEDY. USA.C. L. DD-213 (1993) Method for determination of the indirect tensile stiffness modulus of bituminous mixtures. No 662. L. Vol 1929. Reed Education and Professional Publishing Ltd. R. VUKOSAVLJEVIC. Transportation Research Board. S. & NIENOW. & NISSAN. National Academy Press. B. W. Vol 42. D. (1949) Mixing Law for Viscosity. CANESSA.. 188 . National Engineering Laboratory. Proc Association of Asphalt Paving Technologist.. S. J.. Vol 30. A. F. p: 17-43 American Society of Testing Materials. HUANG. R. Special Technical Publication. SHU. D. W. HOLMGREEN. HUGHES. USA. E. F. p: 799-800. British Standard Institution. IRVING. Washington D. DUNNING.. Proceeding of Asphalt Association of Paving Technologist. N. American Society of Testing Materials. R. USA. X. P:644-664. L. Association of Asphalt Paving Technologist. N. US. & EGAN. EPPS. & ESCOBAR.DAVIDSON. USA. K. LITTLE. G. B. Transportation Research Record.. H.

U. Alabama. & ISAYEV. Vol 10. KARLSSON. Chemtec Publishing. Transportation Research Record. p: 134-164. Y. (2003a) Application of FTIR-ATR to Characterization of Bitumen Rejuvenator Diffusion. J. S. Washington D. Vol 38.. S. D. & KOEHLER. E. I. U.C. USA. National Research Council. ZENG. & ISACSSON. ASCE. LEE.-C. P. KARLSSON. & MAHONEY. TERREL.. J. S. Strategic Highway Research Program. USA. No 1. Washington D. & MALLICK. 38. (1996) Aging and Low-Temperature Cracking of Asphalt Concrete Mixture. L. No 13. S. Washington DC.S. No3. National Centre for Asphalt Technology.State-of-the-Art. pp: 51-59. Journal of Materials in Civil Engineering. USA. No2. (1984) Significant Studies in Asphalt Durability: Pennsylvania Experience. P. Auburn University. pp: 2835-2844. P. (2003c) Laboratory Studies of Diffusion in Bitumen Using Marker. Vol 15. Federal Aviation Administration. (1993) Understanding How the Origin and Composition of Paving-Grade Asphalt Cement Affect Their Performance. Journal of Materials Sciences. September 1996. R.JOHN. R. Journal of Materials in Civil Engineering. U. and Application. R. Proceeding. Transportation Research Board. R. KANDHAL. Transportation Research Record 999.S. p:41-49. P. W. A. (1983) Test for Efficiency of Mixing of Recycled Asphalt Paving Mixtures. U. & ISACSSON. MAPH-2 (2000) Hot mix asphalt paving handbook. (1997) Pavement Recycling Guidelines for State and Local Government.. United States of America. National Research Council. T. AL 36849. A. ASCE. Kentucky. p: 2835-2844. & CHAKRABORTY. Journal of Cold Regions Engineering. 189 . February 25. S. National Centre for Asphalt Technology. ASCE. Vol 18. (2006) Rheology: Concepts. U. T. & ISACSSON. & VINSON. C. KLIEWER. Methods. (1996) Evaluation of Voids in the Mineral Aggregate for HMA Paving Mixture. Department of Transportation. Lenxington. H. U. Toronto. B.C. & ISACSSON. KARLSSON. KANDHAL. R. MALKIN. R. Texas. R. Vol 911. Auburn University. The University of Texas Austin. (2003d) Material-Related Aspects of Asphalt Recycling . KANDHAL. KARLSSON. Asphalt Association Pavement Technologist. (2003b) Investigations on Bitumen Rejuvenator Diffusion and Structural Stability.

Bituminous Materials. & PETERSON. & ANDERSON. S. M. ANDERSON. UK. MCDANIEL. S. NOURELDIN. H. http://www. Special Technical Publication.. UAS. National Research Council. SHRP-A-646. National Research Council. (1984) Chemical Composition of Asphalt As Related to Asphalt Durability: State of The Art. & MOULTHROP. p:13-30. R.. (1987) Rejuvenator Diffusion in Binder Film for Hot-Mix Recycled Asphalt Pavement. p: 5161. PELL. J. R. PELL.Theory and 190 . A. (1986) Quantitative Functional Group Analysis of Asphalt Using Differential Infrared Spectrometry and Selective Chemical Reaction . S. S. Transportation Research Record 1115. Residential Course at The University of Nottingham. Philadelphia. SHRP. Transportation Research Board. J. C. Symposium on Structural Design of Asphalt Concrete Pavements to Prevent Fatigue Cracking. Washington D.MCDANIEL. Washington D. Design and Evaluation. R. P. S. USA. (2000) Recommended Use of Reclaim Asphalt Pavement in the Superpave Mix Design Method. E.hotmix.C.C. MORTAZAVI. (1993) The SHRP Materials Reference Library. National Research Council. (1983) Visco-Elasto-Plastic Constitutive Law for a Bituminous Mixture Under Repeated Loading. L. PERL. NOURELDIN. J. (2001) Recommended Use of Reclaim Asphalt Pavement in the Superpave Mix Design Method: Technician's Manual. p: J1-J13. National Research Council. S. M. C.. American Society of Testing Materials. R. TURNER. Washington D. Transportation Research Board. M. National Cooperative Highway Research Program.org/recycling. Transportation Research Board. National Research Council. & WOOD.. PETERSEN. Transportation Research Record. Nottingham. Transportation Research Record 999. (1995) Oxidation Of Asphalt Binders And Its Effect On Molercular Sizes Distribution. Fatigue Cracking. A. A. S. P. USA. Transportation Research Board. M. PETERSEN. (1973) Characterisation of Fatigue Behaviour.C.php. Washington DC. Vol 911. NAPA (2004) Asphalt Pavement is the Surprise Leader in the Recycling of Various Materials. & SIDES. R. UZAN. SOLEYMANI. Transportation Research Board. P. No 1241. J. National Research Council. Section J. (1988) Bituminous Pavement: Materials. National Cooperative Highway Research Program.

Thomas Telford. 85-100 Penetration Grade. & PETERSEN. Association of Asphalt Paving Technologist. Nottingham Pavement Engineering Center. T.PhD Thesis. L. D..-Y. 191 . Washington D. J. USA. M. School of Civil Engineering. Lanham. p: 1-24. QC G1K 7P4. W. Nottingham University. M. CRAUS..Application. C. QIU.C. National Research Council. Journal of American Statistical Association. T. & WHITEOAK. National Research Council. Nottingham Centre for Pavement Engineering. ROBERTS. Proceedings. READ. Nottingham. Transportation Research Record 1535. E. E. D. H. AAPT. Canada. J. P. STEPHENS. P:359-377. Vol 45. Maryland. Strategic Highway Research Program. B. (2003) Shell Bitumen Handbook. 69. (2003) Molecular Weight Polydispersity Effects on Diffusion at Polymer/Polymer Interfaces. PLANCHER. F. School of Civil Engineering. p:1-11. C.C. A. J. & DIPPOLD. Montreal). R. V. ROSTLER. Transportation Research Record 1096. Vol 30. (2001) Determination of the PG Binder Grade to Use in a RAP mix. STEPHENS. SYBILSKI.. CRASP (Polytechnique. E. Proceeding of Asphalt Association of Paving Technologist. Sainte-Foy. pp. MAHONEY. & WHITE. D. (1995) Durability of Bituminous Paving Mixtures . & BOUSMINA. AAPT... READ. D. 1991. & MONISMITH. University of Connecticut and the Connecticut Department of Transportation. S. J. Transportation Research Board. J. SCHOLZ. Laval University. USA. Canada Research on Polymer Physics and Nanomaterials. (1962) Composition and Changes In Composition of Highway Asphalts. NAPA Education Foundation. SOUSA. LEE. 730-737. Washington. M. GREEN. J. M. (1974) EDF Statistics for Goodness of Fit and Some Comparisons. Department of Chemical Engineering. F. Mixture Design and Construction. KANDHAL.PhD Thesis. US. R. (1996) Zero-Shear Viscosity of Bituminous Binder and Its Relation to Bituminous Mixture's Rutting Resistance. J. The University of Nottingham. L.. C. S. (1996) Fatigue Cracking of Bituminous Paving Mixtures . H. (1976) Reduction of Oxydative Hardening of Asphalt by Treatments with Hydrated Lime: A Mechanistic Study. BROWN. & KENNEDY. (1991) Summary Report on Permanent Deformation in Asphalt Concrete. (1991) Hot Mix Asphalt Materials. L.

(1970) Fraction Components of Asphalt Compatibility and Interchangeability of Fractions Produced From Different Aphalts.C. National Research Council.. Vol 30. R. WOOD. Association of Asphalt Paving Technologist. TIA. B. (1994) Stiffness of AsphaltAggregate Mixes. C. UNITED STATES DEPARTMENT OF TRANSPORTATION. USA.gov/hnr20/recycle/waste/rap131. Proceeding of Asphalt Association of Paving Technologist. Washington D. M. Vol 49. AAPT. J. (1961) Relation Between Asphalt Composition and Hardening by Volatilization and Oxidation. Transportation Research Board. (2007) Reclaimed Asphalt Pavement User Guidelines. pp: 199-218. D. 192 . H. MITTEN.. Proc Association of Asphalt Paving Technologist. L. R. Vol 32. (1980) The Effect of Added Virgin Binders on the Properties of Cold-Mix Recycled Asphalt Paving Mixtures. R. AAPT. B. R. W. & MONISMITH. US. TAYEBALI. WHITE. Proceedings. Proceedings.tfhrc. Vol 39. (1963) Durability of Asphalt Cement. A. & SKOG.htm. http://www. USA. E. A.. A.p:15-21. p: 492-531. SHRP A-338. M. N. USA. TRAXLER. p: 44-58. F. N. Association of Asphalt Paving Technologist. L. AAPT. TSAI. National Research Council. P:359-377. Washington DC. & HANCHER.-W. TRAXLER. E.

Appendix A Zero shear viscosity extrapolation data 193 .

133 0.320 0.1E-07 1.322 0.339 0.346 0.302 0.9 59.0E+03 5.5E-08 3.000 1.4E+04 9.8 80 100 0 19.997 0.453 0.236 0.1E+02 k 0.435 0.8 80 100 0 19.4E-08 2.402 1.1E+02 2.911 0.392 0.328 0.2 41.715 0.3E-06 4.9E+04 1.3E+05 6.8E-06 8.2E+02 4.0E-04 1.2 41.4E+01 1.2 41.995 0.003 0.653 0.0E-04 5.3E+02 2.1E-06 3.4E+03 1.639 2.1E+03 2.1E+04 1.052 0.598 0.7E-08 1.4E+03 6.0E-04 1.4E-07 1.367 0.9 59.522 0.310 0.7E+02 1.997 0.8E+04 3.5E+04 5.9E+04 4.7E-07 4.992 1.2 41.6E+05 5.000 0.3E-06 7.3E-08 5.992 0.2E+05 2.017 1.8E-09 3.2E-08 2.207 0.2E-10 3.564 0.326 0.440 0.000 0.950 5.2E-07  o (Pa.145 0.2 41.556 0.Table A-1: Cross model parameters of Mix A (blends of different proportion of RAP binder and 160/220 Pen) at different temperatures Temperature (oC) RAP % 0 19.301 0.717 0.405 2.4E+02 1.3E+04 3.0E+03 2.999 1.000 1.9 59.6E+04 7.8 80 100 0 19.158 3.8E-06 3.5E+06 2.8E-05 2.997 0.542 0.4E-04 1.986 0.400 0.9E-05 1.072 0.336 0.762 1.8E+06 2.2E-09 1.233 0.960 0.4E+03 5.3E+03 6.618 1.332 0.6E+02 4.989 0.994 1.492 0.9 59.4E+04 2.724 0.573 0.9E-09 8.997 0.000 1.444 0.526 0.2E+04 1.0E+04 4.9E+03 1.999 1.919 1.404 0.135 0.458 2.5E+05 3.403 0.2E-06 2.7E-05 1.000 1.3E+03 1.0E-03 9.0E-04 2.8 80 100 0 19.601 0.9 59.306 0.224 0.7E-07 9.6E-11 1.339 0.0E-04 1.4E-07 1.446 0.000 1.334 0.950 0.6E+05 4.318 0.000 0.000 1.0E-04 9.s) 20 25 30 35 40 45 50 55 60 6.000 0.999 0.440 0.230 0.358 0.367 0.000 0.395 0.8E-06 1.999 1.000 0.9 59.993 0.1E+03 1.838 1.922 1.102 m 0.535 0.490 0.0E+03 4.080 1.904 1.5E+02 8.8E-07 1.7E+04 5.2E+05 1.8 80 100 0 19.145 1.8E-05 3.390 0.173 1.5E+03 6.392 0.909 2.403 R square 0.7E-06 9.051 0.6E-05 1.992 0.999 0.4E-07 1.7E-12 2.000 1.398 0.9   (Pa.439 1.8 80 100 0 19.4E-08 8.997 0.0E+05 1.0E-04 9.999 0.2E+05 1.8 80 100 0 19.3E+03 2.943 0.993 0.073 0.997 0.2 41.316 0.198 0.5E+04 8.390 0.000 1.997 0.992 194 .430 0.0E-05 1.691 0.999 0.992 0.077 0.0E-04 9.668 1.1E+06 9.009 0.3E-07 1.4E+05 3.2 41.s) 8.051 0.534 0.9 59.000 1.639 0.3E-06 1.2 41.368 0.998 0.998 0.416 0.8E-05 2.9E+04 2.9E+03 1.325 0.988 0.976 1.9 59.405 0.343 1.000 0.470 0.0E-08 1.4E+04 3.0E-04 1.8 80 100 0 19.891 0.380 0.993 1.516 2.000 1.036 3.1E+04 2.821 0.0E-07 4.2E+02 2.2 41.

357 NA NA NA 1.5E+02 2.461 0.910 0.2 41.995 0.524 0.228 0.778 0.567 0.502 NA NA 0.8E+01 3.294 0.049 0.0E+01 3.0E-05 1.9 59.1E+02 2.125 0.3E+01 2.7E-07 NA NA 9.997 0.8E+01 1.722 0.5E+02 1.983 0.0E+02 0.133 0.5E+01 2.1E+01 5.9 59.1E+02 7.4E+01 5.8 80 100 0 19.402 0.999 0.124 0.2 41.0E-04 6.5E+01 9.385 NA NA 0.9E+03 3.622 0.008 0.298 0.0E+02 1.798 0.078 0.6E-06 2.166 0.295 NA NA NA 0.1E+01 3.4E-05 3.2 41.390 NA NA 0.7E-06 3.2E+01 1.451 0.991 0.0E-04 1.0E-09 NA NA NA 1.007 0.3E-07 1.6E-07 NA NA 8.143 0.998 0.1E+02 4.997 195 .507 0.446 0.3E+01 1.0E+02 1.65 70 75 80 59.337 NA NA 0.993 1.9E+00 1.133 0.175 0.366 0.0E-05 4.9 59.999 0.421 0.668 0.284 NA NA NA 1.9 59.5E+01 6.928 0.434 0.7E-09 9.0E-04 1.605 0.8 80 100 0 19.039 0.862 0.000 0.268 NA NA NA 0.024 0.5E-08 5.8 80 100 1.9E+02 4.8E-08 9.070 0.999 0.2E+02 8.608 0.019 0.999 0.8 80 100 0 19.8 80 100 0 19.175 0.1E+01 1.2E+01 5.891 0.6E-05 NA NA NA 4.0E-08 1.9E+02 8.2 41.196 0.

914 1.0E-04 2.2E-08 1.7E+04 1.4E-08 1.000 1.333 0.720 2.000 0.388 0.021 0.990 0.339 0.3E+03 2.338 0.0E-04 1.340 0.3E-07 1.000 1.421 0.1 80.392 0.6E+03 1.438 0.999 0.8E-05 2.997 1.797 2.9 100 0 19.622 0.9E+03 2.419 0.352 0.1E+04 3.000 1.993 0.235 0.999 1.365 0.060 0.1E-07 3.535 0.329 0.338 0.0E-08 4.996 0.998 0.4E+05 2.6E+05 5.9E-05 9.065 m 0.8E-05 9.170 1.099 1.1E+03 1.4E+05 1.1 80.415 1.1E+06 3.4E-07 2.993 1.430 0.348 0.724 0.s) 3.994 0.893 1.883 1.000 0.9 100 0 19.7 41.731 0.1E-08 5.s) 20 25 30 35 40 45 50 55 anh 60 2.653 0.994 0.3E-09 2.9 100 0 19.1E+02 3.Table A-2: Cross model parameters of Mix B (blends of different % RAP binder and 100/150 Pen) at different temperatures Temperature (oC) RAP % 0 19.620 0.071 0.000 1.1 80.128 0.9 60.664 2.994 0.9 100 0 19.5E-06 8.816 2.9   (Pa.1 80.4E-08 8.373 0.344 0.343 0.0E-04 1.9E+05 9.999 1.542 0.1E-07 9.0E+05 7.001 2.8E-05 3.0E+04 1.7 41.000 0.000 0.0E-04 2.000 1.1E-08 9.943 0.791 3.000 1.994 0.336 0.0E-04 3.999 0.7E+02 1.9 60.997 0.1E-09 1.997 0.2E+04 2.6E+05 4.383 0.8E-07 9.9E-08 1.546 2.9 60.418 0.136 1.7E+04 7.9 100 0 19.0E-06 9.999 1.3E+05 3.8E-05 1.8E+04 1.368 0.4E-07 8.2E+04 4.000 0.000 0.1E+06 2.3E+04 2.556 0.9 60.477 0.4E+03 1.6E+03 7.7E+03 7.7 41.233 0.505 0.9E+03 6.1E-10 1.440 0.350 0.9 60.992 0.995 1.4E+05 4.087 1.320 0.361 0.7 41.649 1.994 0.724 0.5E-07 1.555 0.8E+04 3.967 1.0E-04 1.435 1.000 1.0E-04 1.0E+03 1.7E+04 6.368 0.380 0.1E-07 9.335 0.715 0.125 1.999 0.206 0.821 0.991 0.423 0.564 0.3E+03 1.2E+05 2.585 0.649 1.7E-07 9.2E-09 7.538 0.387 0.997 0.372 0.492 0.994 1.9 60.994 0.038 2.571 0.981 196 .7 41.382 1.1E+04 5.891 1.6E+02 4.999 1.5E-07 4.7E+04 1.454 0.8E+05 2.7 41.5E+05 1.8E-05 2.405 0.0E-09 3.7E-08 7.7E+02 1.6E-08 1.4E+03 2.7 41.653 R square 0.9 60.403 0.0E-09 7.999 0.000 0.7E-09 3.5E+02 7.354 0.8E+06 8.2E+02 2.7E+03 4.123 0.000 1.1E+04 5.4E-07 4.000 0.429 0.9 100 0 19.161 5.0E-09 7.000 0.7 41.997 0.204 2.7 41.998 1.1 80.1 80.0E-04  o (Pa.8E-05 1.435 0.0E-04 9.525 0.5E-06 1.424 0.2E-06 1.9E-05 2.0E-04 1.225 0.000 1.995 0.7E+03 4.1 80.992 0.1 80.341 0.995 0.0E+05 3.9 100 0 19.1E+04 2.782 5.1E-06 1.9 100 0 19.9 60.778 0.7E+05 3.0E+05 1.193 0.4E+02 k 1.678 0.537 0.

9E+01 2.9 100 0 19.999 0.067 0.992 0.101 0.2E-09 2.701 0.0E+03 1.096 0.991 0.0E-04 1.862 0.045 0.2E-09 2.337 0.9E-09 NA NA NA 7.7 41.357 NA NA 1.9E+01 7.1E+01 1.000 0.7 41.532 0.363 0.974 0.205 0.0E-06 1.987 0.036 0.025 0.472 0.5E-07 7.999 0.489 0.103 0.9 100 0 19.2E-07 1.5E-05 4.6E-08 1.964 0.6E-05 NA NA 9.9 60.532 0.0E+01 4.2E-06 6.020 0.1E+02 1.1E-05 1.6E+01 1.133 0.268 0.014 0.798 0.008 0.1 80.7E-05 5.1 80.0E-04 3.9 60.2E+02 8.7E-07 9.9 60.636 0.178 0.454 0.7 41.9E+03 6.1E+01 3.3E+01 1.3E+02 1.3E+02 2.65 70 75 80 60.9 100 0 19.5E+02 3.1 80.5E+02 5.268 NA NA 0.9 100 0 19.415 0.0E-04 9.513 0.927 0.983 0.997 0.0E+01 1.9 100 2.267 0.804 0.1 80.966 0.349 0.0E+02 1.997 197 .390 0.295 NA NA NA 0.0E-04 2.149 0.9 60.998 0.589 0.052 0.0E+02 2.980 1.085 0.502 0.7 41.316 0.8E+02 2.010 0.7E-06 1.3E-14 1.998 0.968 0.3E+02 2.284 NA NA NA 0.9E+01 4.581 0.970 0.0E+01 7.670 0.9E-05 1.3E-07 1.049 0.935 0.6E+01 1.0E+02 1.556 0.0E-04 9.404 0.870 0.1 80.999 0.0E+02 0.6E+01 5.5E+02 4.067 0.182 0.421 0.

3 1.91E+04 7.8 3.2 1.7 1.2 4.05E+06 0.69E+05 2.1 9.32E+04 4.40E+05 4.40E+06 8.50E+04 0.87E+05 6.4 3.0 25 41.0 19.9 3.7 6.31E+06 14.31E+05 46.60E+05 36.63E+06 2.3 8.9 3.0 2.0 2.9 59.8 2.44E+05 4.0 0 2.2 5.72E+06 48.0 2.90E+06 12.0 6.31E+05 4.05E+06 0.07E+06 8.60E+05 3.0 9.64E+06 44.6 1.76E+07 2.24E+05 9.52E+04 2.94E+04 4.98E+05 46.0 1.8 7.9 1.8 1.62E+04 4.20E+06 2.21E+04 47.81E+04 28.0 2.51E+06 4.05E+06 21.8 DLV Predicted Viscosity Residue (P) (%) 7.0 9.0 8.06E+07 0.06E+07 0.05E+06 0.8 1.78E+05 0.3 1.7 40 41.33E+06 48.4 5.91E+04 0.1 2.0 1.37E+05 1.0 1.0 1.9 2.50E+05 0.0 2.74E+06 12.0 2.33E+04 0.37E+06 4.05E+06 0.0 3.33E+04 0.2 1.9 4.8 3.8 1.27E+04 1.8 80 5.9 2.87E+05 8.6 80 3.38E+05 29.0 3.50E+05 0.89E+05 0.26E+06 37.53E+06 0.1 35 41.6 5.35E+06 9.25E+05 2.8 9.76E+07 0.8 100 4.37E+06 0.94E+05 2.0 4.06E+07 0.91E+04 0.40E+05 6.99E+05 1.0 3.0 6.03E+05 3.2 1.18E+06 1.91E+04 0.0 3.1 3.0 1.61E+05 3.3 5.8 2.06E+07 0.32E+07 14.3 3.9 1.23E+07 20.Table A-3: Mix A – experiment and predicted viscosity (P) using different viscosity mixing equations at different temperatures Temperature Rap Experiment ASTM Predicted G&N Predicted Epps Predicted Viscosity Residue Viscosity Residue Viscosity Residue % (P) (%) (P) (%) (P) (%) 0 6.89E+05 2.2 1.6 2.8 6.35E+06 10.3 8.0 19.22E+06 17.89E+05 0.7 80 1.7 1.2 2.2 7.4 59.8 9.0 0 3.54E+05 1.6 2.8 6.05E+06 2.9 1.60E+06 2.2 6.53E+05 42.50E+04 43.07E+05 5.6 30 41.48E+05 8.4 5.09E+06 20.62E+06 10.84E+05 10.0 2.27E+04 0.8 2.0 19.13E+05 1.38E+05 9.33E+04 3.76E+07 0.59E+05 10.3 4.62E+06 8.15E+05 3.8 6.9 1.1 1.88E+04 10.54E+06 1.1 1.50E+05 0.37E+06 0.89E+05 0.50E+05 38.74E+04 8.9 100 1.33E+04 0.18E+06 25.1 2.47E+04 4.35E+05 20.01E+06 3.10E+04 13.15E+06 6.26E+06 1.2 3.76E+05 6.1 59.29E+06 9.76E+07 0.2 1.00E+05 16.89E+05 6.91E+05 12.0 0 1.27E+04 0.55E+07 1.5 100 2.27E+04 0.2 1.02E+05 9.8 100 2.37E+05 10.0 4.9 59.4 5.27E+04 7.8 20 41.8 1.88E+06 27.32E+04 46.77E+05 1.10E+04 2.06E+07 1.3 5.9 8.36E+05 5.45E+07 6.33E+06 3.7 4.93E+05 44.76E+07 0.12E+05 41.0 4.98E+05 36.9 1.37E+06 0.7 2.0 19.24E+06 53.76E+04 31.0 0 9.0 1.9 6.7 4.5 2.33E+05 7.33E+06 45.9 4.8 80 2.9 7.5 80 8.50E+05 2.8 1.8 3.0 1.23E+06 2.8 59.79E+04 10.0 19.2 2.05E+05 5.0 3.37E+06 0.72E+05 7.24E+06 1.19E+06 5.9 1.46E+05 31.47E+05 3.3 198 .34E+04 6.91E+04 9.

4 24.09E+03 2.72E+05 5.0 3.50E+03 1.9 22.9 8.8 26.8 2.0 5.09E+04 1.14E+04 2.1 0.0 2.6 7.1 8.9 23.7 0.0 199 .06E+04 8.14E+02 2.24E+03 2.08E+02 3.0 6.3 32.79E+03 1.15E+05 2.0 4.3 3.4 23.56E+03 4.05E+04 2.0 3.5 0.81E+02 1.07E+04 1.22E+05 3.21E+03 4.43E+03 1.72E+05 5.07E+04 2.5 0.47E+02 6.15E+05 2.5 0.6 27.8 24.96E+04 4.47E+02 6.46E+03 2.46E+03 2.8 7.6 30.47E+03 2.32E+05 3.6 23.69E+03 3.18E+04 2.2 41.2 7.5 31.9 0.9 23.58E+02 4.0 4.0 0.0 0.02E+03 4.2 3.24E+03 2.7 22.67E+03 1.44E+04 5.86E+04 3.40E+05 3.4 7.19E+03 4.4 0.2 41.0 0.0 21.46E+03 2.04E+03 8.34E+03 4.5 0.98E+04 5.15E+04 4.69E+03 6.1 3.2 41.65E+03 8.33E+02 0.0 7.2 30.0 3.4 7.7 31.21E+03 4.4 0.0 8.22E+05 1.17E+03 8.0 0.0 3.0 11.22E+05 1.33E+04 3.98E+03 3.61E+04 1.2 33.22E+05 1.86E+04 3.4 33.18E+04 1.8 0.56E+03 4.7 4.1 2.08E+02 3.58E+02 1.45E+04 1.14E+04 1.78E+02 1.68E+03 1.0 0.4 3.9 1.14E+04 4.22E+03 2.6 5.0 0.45E+04 5.66E+02 0.0 2.2 5.39E+03 1.05E+04 2.86E+04 3.15E+05 2.32E+03 2.91E+04 4.2 0.0 0.44E+03 1.2 2.0 0.72E+05 5.39E+04 6.56E+03 5.6 0.40E+02 1.39E+03 1.0 0.39E+04 6.2 41.22E+03 4.2 4.00E+04 1.3 3.0 6.43E+04 4.0 0.29E+03 1.8 80 100 0 19.9 59.22E+05 1.12E+03 1.39E+03 1.69E+02 0.75E+04 3.0 2.0 5.82E+03 1.6 26.46E+03 5.40E+03 1.00E+03 5.1 6.8 80 100 0 19.20E+03 4.1 8.04E+03 8.7 7.65E+03 1.68E+03 1.3 0.65E+03 8.15E+05 2.1 9.7 0.0 0.2 26.9 11.47E+04 3.9 59.0 2.40E+02 1.14E+03 2.39E+04 6.2 0.4 1.2 14.8 80 100 0 19.6 8.08E+02 3.64E+02 1.0 0.10E+03 2.3 15.95E+04 1.9 59.3 39.37E+03 1.9 0.72E+05 5.9 25.09E+04 2.3 39.39E+04 6.80E+03 9.3 5.46E+03 2.13E+02 1.22E+05 1.0 0.58E+03 6.8 22.0 0.00E+03 8.40E+02 1.0 0.72E+05 5.8 33.0 0.31E+03 1.9 0.46E+03 2.39E+03 1.47E+02 5.5 0.0 3.96E+03 1.28E+04 5.47E+02 5.2 7.10E+03 8.29E+05 3.2 25.6 0.0 0.0 0.10E+04 2.63E+03 4.24E+03 2.6 7.0 3.63E+04 5.0 15.9 59.1 30.45 50 55 60 65 100 0 19.58E+02 7.8 80 100 0 19.66E+03 3.0 0.39E+04 6.43E+04 1.6 7.8 80 100 0 19.0 0.08E+02 3.0 0.15E+05 2.48E+03 1.0 0.86E+04 3.15E+03 3.36E+03 3.86E+04 3.47E+03 8.29E+04 2.05E+03 1.88E+03 4.53E+03 1.0 0.0 4.24E+03 2.34E+02 0.0 29.62E+02 8.9 59.9 33.58E+04 5.3 4.56E+03 4.12E+03 2.40E+02 7.8 27.2 41.79E+03 1.40E+02 1.52E+03 8.0 3.96E+04 6.63E+04 1.98E+03 8.2 20.41E+02 1.0 0.87E+03 8.62E+03 7.2 2.0 7.92E+04 4.

9 59.86E+02 1.53E+02 2.13E+03 7.0 0.8 80 100 0 19.77E+02 2.28E+02 1.0 2.83E+02 3.33E+02 1.7 5.9 0.96E+03 1.16E+02 5.8 0.2 3.0 4.99E+02 3.31E+02 1.0 1.0 0.96E+03 1.8 0.6 0.10E+02 1.02E+03 11.63E+02 2.5 0.02E+03 10.35E+02 4.0 0.28E+02 1.0 4.6 17.0 9.05E+03 2.28E+02 2.2 17.0 200 .6 13.96E+03 1.15E+02 6.6 9.9 8.13E+03 7.90E+01 1.02E+03 29.3 18.90E+01 1.54E+02 9.21E+03 2.03E+02 3.97E+03 3.19E+02 2.15E+02 1.2 11.8 20.64E+02 6.49E+02 2.11E+02 5.9 0.8 0.5 3.0 6.04E+03 1.61E+02 1.3 0.51E+02 1.5 21.17E+02 1.4 5.21E+02 5.7 24.29E+02 2.87E+02 1.7 28.09E+03 2.0 0.99E+03 3.10E+03 3.30E+02 3.3 18.8 80 100 0 19.31E+02 1.16E+02 5.0 4.3 25.96E+03 1.07E+03 2.0 0.55E+02 1.8 26.28E+02 2.00E+02 3.9 0.13E+03 7.13E+03 7.32E+02 5.5 27.02E+03 7.0 6.53E+02 1.0 0.8 80 100 5.0 2.2 3.8 10.0 2.00E+02 3.9 59.9 0.21E+03 2.0 23.70 75 80 41.9 18.11E+01 9.0 4.02E+03 22.31E+02 7.4 1.0 10.90E+01 1.56E+02 4.0 1.8 0.2 4.9 10.05E+03 1.86E+01 1.12E+02 1.9 59.18E+02 2.17E+02 1.8 31.19E+02 4.0 0.6 15.2 41.7 5.2 41.24E+02 8.2 12.82E+02 1.55E+02 5.90E+02 3.6 1.96E+03 1.20E+03 3.90E+01 1.13E+03 8.57E+03 3.0 0.

Table A-4: Mix B – experiment and predicted viscosity (P) using different viscosity mixing equations at different temperatures Temperature Rap Experiment ASTM Predicted G&N Predicted Epps Predicted Viscosity Residue Viscosity Residue Viscosity Residue % (P) (%) (P) (%) (P) (%) 0 2.29E+06 2.29E+06 0.0 2.29E+06 0.0 2.29E+06 0.0 19.7 3.64E+06 3.74E+06 2.9 3.99E+06 9.5 3.63E+06 0.2 20 41.9 5.89E+06 6.50E+06 10.5 7.16E+06 21.7 6.20E+06 5.4 60.1 9.35E+06 1.02E+07 9.4 1.12E+07 20.2 9.76E+06 4.3 80.9 2.09E+07 1.72E+07 17.7 1.82E+07 12.5 1.66E+07 20.2 100 2.76E+07 2.76E+07 0.0 2.76E+07 0.0 2.76E+07 0.0 0 8.67E+05 8.67E+05 0.0 8.67E+05 0.0 8.67E+05 0.0 19.7 1.84E+06 1.42E+06 22.7 1.54E+06 16.4 1.37E+06 25.2 25 41.9 2.65E+06 2.47E+06 6.6 2.79E+06 5.5 2.35E+06 11.2 60.1 4.03E+06 3.90E+06 3.1 4.40E+06 9.2 3.71E+06 7.9 80.9 7.46E+06 6.57E+06 11.9 7.10E+06 4.9 6.36E+06 14.8 100 1.06E+07 1.06E+07 0.0 1.06E+07 0.0 1.06E+07 0.0 0 3.07E+05 3.07E+05 0.0 3.07E+05 0.0 3.07E+05 0.0 19.7 5.77E+05 5.18E+05 10.3 6.10E+05 5.7 4.98E+05 13.7 30 41.9 1.05E+06 9.34E+05 11.0 1.20E+06 14.6 8.78E+05 16.3 60.1 1.72E+06 1.51E+06 12.0 1.94E+06 12.9 1.42E+06 17.2 80.9 3.40E+06 2.63E+06 22.7 3.09E+06 9.3 2.53E+06 25.7 100 4.37E+06 4.37E+06 0.0 4.37E+06 0.0 4.37E+06 0.0 0 1.05E+05 1.05E+05 0.0 1.05E+05 0.0 1.05E+05 0.0 19.7 2.14E+05 1.89E+05 11.7 1.92E+05 10.3 1.79E+05 16.2 35 41.9 3.65E+05 3.65E+05 0.0 3.74E+05 2.5 3.36E+05 7.8 60.1 5.96E+05 6.27E+05 5.2 6.42E+05 7.7 5.78E+05 3.1 80.9 1.22E+06 1.16E+06 4.6 1.18E+06 3.1 1.10E+06 9.6 100 2.05E+06 2.05E+06 0.0 2.05E+06 0.0 2.05E+06 0.0 0 3.64E+04 3.64E+04 0.0 3.64E+04 0.0 3.64E+04 0.0 19.7 7.14E+04 6.64E+04 6.9 6.93E+04 2.9 6.25E+04 12.4 40 41.9 1.34E+05 1.31E+05 2.2 1.40E+05 4.4 1.19E+05 11.0 60.1 2.24E+05 2.28E+05 1.7 2.43E+05 8.5 2.08E+05 7.4 80.9 4.68E+05 4.31E+05 7.9 4.49E+05 4.0 4.05E+05 13.4

DLV Predicted Viscosity Residue (P) (%) 2.35E+06 2.5 2.87E+06 21.3 4.24E+06 27.9 6.68E+06 28.6 1.30E+07 37.5 2.76E+07 0.0 8.89E+05 2.5 1.08E+06 40.9 1.61E+06 39.2 2.54E+06 36.8 4.98E+06 33.3 1.06E+07 0.0 3.15E+05 2.7 3.89E+05 32.5 5.92E+05 43.6 9.62E+05 44.1 1.96E+06 42.4 4.37E+06 0.0 1.08E+05 3.0 1.37E+05 35.8 2.19E+05 39.9 3.77E+05 36.7 8.36E+05 31.3 2.05E+06 0.0 3.75E+04 3.1 4.78E+04 33.0 7.74E+04 42.2 1.35E+05 39.7 3.07E+05 34.4

201

45

50

55

60

65

100 0 19.7 41.9 60.1 80.9 100 0 19.7 41.9 60.1 80.9 100 0 19.7 41.9 60.1 80.9 100 0 19.7 41.9 60.1 80.9 100 0 19.7 41.9 60.1 80.9 100 0 19.7

7.72E+05 1.39E+04 2.69E+04 4.74E+04 7.56E+04 1.81E+05 3.15E+05 5.70E+03 1.04E+04 1.88E+04 2.90E+04 6.34E+04 1.22E+05 2.59E+03 4.45E+03 7.72E+03 1.27E+04 2.70E+04 4.39E+04 1.24E+03 2.06E+03 3.40E+03 5.33E+03 1.02E+04 1.86E+04 6.59E+02 1.01E+03 1.78E+03 2.54E+03 5.21E+03 8.46E+03 3.63E+02 5.61E+02

7.72E+05 1.39E+04 2.56E+04 5.13E+04 9.05E+04 1.73E+05 3.15E+05 5.70E+03 1.04E+04 2.06E+04 3.59E+04 6.79E+04 1.22E+05 2.59E+03 4.53E+03 8.48E+03 1.42E+04 2.56E+04 4.39E+04 1.24E+03 2.11E+03 3.86E+03 6.31E+03 1.11E+04 1.86E+04 6.59E+02 1.09E+03 1.92E+03 3.06E+03 5.20E+03 8.46E+03 3.63E+02 5.81E+02

0.0 0.0 4.5 8.1 19.8 4.0 0.0 0.0 0.1 9.5 24.0 7.2 0.0 0.0 1.7 9.9 12.1 5.2 0.0 0.0 2.8 13.4 18.4 8.7 0.0 0.0 7.9 8.2 20.2 0.3 0.0 0.0 3.7

7.72E+05 1.39E+04 2.52E+04 4.98E+04 8.80E+04 1.70E+05 3.15E+05 5.70E+03 9.52E+03 1.79E+04 3.13E+04 6.22E+04 1.22E+05 2.59E+03 4.50E+03 8.42E+03 1.41E+04 2.55E+04 4.39E+04 1.24E+03 1.93E+03 3.35E+03 5.49E+03 1.01E+04 1.86E+04 6.59E+02 1.04E+03 1.78E+03 2.83E+03 4.95E+03 8.46E+03 3.63E+02 5.61E+02

0.0 0.0 6.3 5.1 16.5 5.7 0.0 0.0 8.7 4.7 8.2 1.9 0.0 0.0 1.2 9.1 11.2 5.6 0.0 0.0 6.2 1.5 3.1 0.6 0.0 0.0 2.7 0.3 11.6 5.0 0.0 0.0 0.0

7.72E+05 1.39E+04 2.40E+04 4.61E+04 8.13E+04 1.62E+05 3.15E+05 5.70E+03 9.71E+03 1.84E+04 3.21E+04 6.31E+04 1.22E+05 2.59E+03 4.23E+03 7.64E+03 1.28E+04 2.39E+04 4.39E+04 1.24E+03 1.98E+03 3.47E+03 5.68E+03 1.03E+04 1.86E+04 6.59E+02 1.02E+03 1.73E+03 2.76E+03 4.86E+03 8.46E+03 3.63E+02 5.46E+02

0.0 0.0 10.8 2.8 7.7 10.5 0.0 0.0 6.9 2.1 10.9 0.5 0.0 0.0 4.9 1.0 0.9 11.5 0.0 0.0 3.9 2.1 6.6 1.4 0.0 0.0 1.1 2.3 8.5 6.8 0.0 0.0 2.5

7.72E+05 1.43E+04 1.83E+04 3.00E+04 5.31E+04 1.23E+05 3.15E+05 5.88E+03 7.50E+03 1.22E+04 2.13E+04 4.84E+04 1.22E+05 2.67E+03 3.34E+03 5.21E+03 8.75E+03 1.87E+04 4.39E+04 1.27E+03 1.58E+03 2.42E+03 3.97E+03 8.21E+03 1.86E+04 6.76E+02 8.28E+02 1.24E+03 1.97E+03 3.91E+03 8.46E+03 3.72E+02 4.50E+02

0.0 3.2 31.7 36.8 29.8 32.1 0.0 3.1 28.1 35.3 26.6 23.7 0.0 2.9 25.0 32.4 30.9 30.7 0.0 2.7 23.2 28.8 25.5 19.5 0.0 2.6 18.0 30.2 22.3 24.9 0.0 2.4 19.8

202

70

75

80

41.9 60.1 80.9 100 0 19.7 41.9 60.1 80.9 100 0 19.7 41.9 60.1 80.9 100

1.03E+03 1.33E+03 2.47E+03 3.96E+03 2.11E+02 3.02E+02 4.88E+02 7.13E+02 1.29E+03 2.13E+03 1.28E+02 1.88E+02 2.93E+02 4.02E+02 7.00E+02 1.02E+03

9.88E+02 1.53E+03 2.51E+03 3.96E+03 2.11E+02 3.32E+02 5.55E+02 8.45E+02 1.37E+03 2.13E+03 1.28E+02 1.92E+02 3.05E+02 4.44E+02 6.84E+02 1.02E+03

3.6 15.2 1.6 0.0 0.0 10.0 13.7 18.5 5.9 0.0 0.0 2.3 4.1 10.7 2.3 0.0

9.35E+02 1.44E+03 2.42E+03 3.96E+03 2.11E+02 3.06E+02 4.89E+02 7.46E+02 1.26E+03 2.13E+03 1.28E+02 1.88E+02 2.95E+02 4.31E+02 6.70E+02 1.02E+03

8.8 9.0 2.0 0.0 0.0 1.3 0.2 4.6 2.2 0.0 0.0 0.2 0.8 7.2 4.3 0.0

8.96E+02 1.38E+03 2.35E+03 3.96E+03 2.11E+02 3.12E+02 5.02E+02 7.64E+02 1.28E+03 2.13E+03 1.28E+02 1.82E+02 2.79E+02 4.06E+02 6.45E+02 1.02E+03

12.6 4.4 4.8 0.0 0.0 3.3 2.9 7.1 0.9 0.0 0.0 3.3 4.7 1.2 7.9 0.0

6.55E+02 1.01E+03 1.93E+03 3.96E+03 2.15E+02 2.59E+02 3.73E+02 5.69E+02 1.06E+03 2.13E+03 1.31E+02 1.54E+02 2.13E+02 3.12E+02 5.43E+02 1.02E+03

36.1 23.5 22.1 0.0 2.3 14.2 23.6 20.2 18.0 0.0 2.1 18.1 27.1 22.4 22.4 0.0

203

Appendix B ITFT data

204

Table B-1: ITFT data of LR FS-2 mixture at 20oC Sample Width (mm) E (MPa) mm) S1 40 98 1096 S2 40 98 854 S3 40 98 1849 S4 37 98 1309 S5 39 98 937 S6 42 98 1709 S7 40 98 899 S8 41 98 1515 S9 42 98 1636 S10 40 98 811 Table B-2: ITFT data of LR FS-4 mixture at 20oC Sample Width (mm) E (MPa) mm) S1 41 98 1052 S2 39 98 1165 S3 40 98 1376 S4 40 98 1523 S5 41 98 1346 S6 42 98 1301 S7 41 98 1390 S8 42 98 1768 S9 43 98 1250 S10 38 98 1299 Table B-3: ITFT data of LR FS-6 mixture at 20oC Sample Width (mm) mm) E (MPa) S1 41 98 1178 S2 42 98 1823 S3 42 98 1698 S4 41 98 1273 S5 41 98 1638 S6 41 98 1360 S7 40 98 1195 S8 40 98 1208 S9 41 98 1832 S10 40 98 1050 Table B-4: ITFT data of LR FS-8 mixture at 20oC Sample Width (mm) mm) E (MPa) S1 39 98 1649 S2 41 98 1526 S3 41 98 1516 S4 40 98 1691 S5 42 98 1435 S6 40 98 1403 S7 41 98 1637 S8 40 98 1619 S9 40 98 1422 S10 40 98 1623

kPa) 100 300 400 250 200 100 150 100 250 200

m) 187 720 443 392 438 120 342 135 313 506

N 33537 402 622 670 636 7054 1960 7436 1356 1018

kPa) 100 200 300 350 150 100 250 300 250 200

m) 195 352 447 471 228 158 369 348 410 316

N 9351 1678 971 610 4090 7516 1579 602 1075 2357

kPa) 200 300 350 250 200 150 150 250 200 100

m) 348 337 423 403 250 226 257 424 224 195

N 1544 1166 688 1080 3922 5303 5128 1098 7192 8748

kPa) 150 150 350 300 250 250 200 300 200 350

m) 186 202 473 364 357 365 250 380 288 442

N 20224 9354 399 946 1249 1131 3978 956 2714 592

205

Table B-5: ITFT data of SR FS-2 mixture at 20oC Sample Width (mm) E (MPa) mm) S1 41 98 1413 S2 41 98 1605 S3 40 98 1729 S4 40 98 1669 S5 40 98 1532 S6 40 98 1667 S7 40 98 1587 S8 40 98 1788 S9 39 98 1665 S10 40 98 1515 Table B-6: ITFT data of SR FS-4 mixture at 20oC Sample Width (mm) E (MPa) mm) S1 40 98 1616 S2 41 98 1506 S3 40 98 1565 S4 41 98 1493 S5 41 98 1637 S6 40 98 1448 S7 41 98 1507 S8 42 98 1581 S9 41 98 1593 S10 40 98 1606 Table B-7: ITFT data of SR FS-6 mixture at 20oC Sample Width (mm) mm) E (MPa) S1 40 98 1544 S2 39 98 1751 S3 39 98 1566 S4 40 98 1656 S5 40 98 1681 S6 39 98 1619 S7 40 98 1683 S8 40 98 1605 S9 40 98 1379 S10 40 98 1705 Table B-8: ITFT data of SR FS-8 mixture at 20oC Sample Width (mm) mm) E (MPa) S1 38 98 1560 S2 39 98 1609 S3 40 98 1743 S4 42 98 1514 S5 40 98 1497 S6 39 98 1704 S7 41 98 1740 S8 41 98 1734 S9 39 98 1727 S10 40 98 1808

kPa) 200 350 350 300 250 200 150 250 150 300

m) 290 447 415 368 335 246 194 287 185 406

N 2436 523 691 1046 1649 2950 8990 2437 10847 1261

kPa) 250 200 300 350 150 150 200 300 250 350

m) 317 272 393 481 188 212 272 389 322 447

N 2205 3377 1282 719 12325 8857 1884 1255 1984 565

kPa) 300 350 250 200 150 250 300 350 150 200

m) 398 410 327 248 183 317 365 447 223 240

N 1327 1072 2629 5617 18661 3471 2476 953 9802 7952

kPa) 250 200 350 150 150 250 300 300 200 350

m) 329 255 412 203 205 301 353 355 237 397

N 2445 6589 1250 18133 19370 4404 2437 2815 9840 1549

206

Table B-9: ITFT data of LR SHRP mixture at 20oC Sample Width (mm) E (MPa) mm) S1 40 98 1535 S2 40 98 1528 S3 40 98 1338 S4 39 98 1379 S5 40 98 1366 S6 40 98 1356 S7 41 98 1590 S8 41 98 1579 S9 41 98 1470 S10 40 98 1376 Table B-10: ITFT data of SR SHRP mixture at 20oC Sample Width (mm) E (MPa) mm) S1 41 98 1450 S2 41 98 1584 S3 40 98 1675 S4 42 98 1634 S5 39 98 1640 S6 40 98 1539 S7 40 98 1453 S8 40 98 1481 S9 40 98 1385 S10 39 98 1437 Table B-11: ITFT data of BR mixture at 20oC Sample Width (mm) mm) E (MPa) S1 39 98 744 S2 39 98 655 S3 40 98 791 S4 41 98 754 S5 39 98 764 S6 37 98 853 S7 40 98 878 S8 39 98 818 S9 40 98 815 S10 41 98 919 Table B-12: ITFT data of CB mixture at 20oC Sample Width (mm) mm) E (MPa) S1 39 98 2185 S2 39 98 1805 S3 39 98 1918 S4 39 98 2001 S5 40 98 1708 S6 39 98 1941 S7 38 98 1738 S8 39 98 1822 S9 39 98 1850 S10 38 98 1942

kPa) 350 300 150 200 250 250 300 150 350 200

m) 467 402 230 297 375 378 387 195 488 298

N 709 1116 7945 2275 2001 1788 971 5966 578 2782

kPa) 300 350 250 200 150 150 200 250 300 350

m) 424 453 306 251 188 200 282 346 444 499

N 1621 800 5587 8259 20869 22343 5940 2218 1218 913

kPa) 100 250 50 50 150 100 200 200 250 150

m) 276 782 130 136 402 240 467 501 629 335

N 1963 414 6885 7062 1612 2672 978 930 431 2510

kPa) 350 300 250 200 150 300 150 200 250 350

m) 328 341 267 205 180 317 177 225 277 369

N 2622 2842 9778 14064 37012 3460 33720 16648 6324 2024

207

Table B-13: ITFT data of CB-V mixture at 20oC Sample Width (mm) E (MPa) mm) S1 40 98 1840 S2 39 98 1873 S3 40 98 1757 S4 42 98 2025 S5 40 98 2047 S6 41 98 2251 S7 40 98 2306 S8 40 98 1754 S9 40 98 2210 S10 41 98 1984 kPa) 350 400 350 300 300 250 400 200 250 200 m) 390 438 408 304 300 228 356 234 232 207 N 2267 1201 1774 4891 2936 12190 2497 20100 12239 28312 208 .

209 .