Asphalt Re-recycling

Abstract

Use of Reclaimed Asphalt Pavement (RAP) is beneficial to both road owners and builders as it allows for significant raw material cost reduction, while potentially maintaining expected pavement service life. In upcoming decades, the recycling of previously recycled pavements (i.e., re-recycling) will become widespread. There is currently little technical knowledge on how or how many times asphalt pavement can be recycled while sustaining its expected durability.
A novel asphalt binder aging method involving thin layers, heat, water spray, and UV radiation was developed to simulate approximately 20 years of in-service aging. The aged binder was recovered and blended with a softer, virgin binder. The blend was subjected to the next aging cycle. The process was repeated four times to simulate four recycling cycles (80 years) at 25 percent RAP addition. Three virgin binders were tested: standard Performance Grade (PG), one grade softer PG, and standard PG softened with paraffinic oil to one PG softer binder. Very detailed chemical & rheological analyses were performed to understand the impact of multiple recycling on irreversible chemical changes and evolution of rheological properties over the time. Results indicated that at moderate recycling levels, re-recycling is a viable option if an appropriate virgin binder is used.

Full text

Asphalt Re-recycling
Pavel Kriz, Ph.D., P.Eng.
Americas Asphalt Technical Leader
Imperial Oil Limited
Sarnia, Ontario
Bennett J. Tardiff, Ph.D.
Lead Researcher
Imperial Oil Limited
Sarnia, Ontario
Stephanie R. Sta. Maria
Research Technologist
Imperial Oil Limited
Sarnia, Ontario
Ralph D. Shirts
Asphalt Technology and Business Support Section Head
ExxonMobil Fuels, Lubricants & Specialties Marketing Company
Spring, Texas
Acknowledgements
Ms. Selena Lavorato and Mr. Steve Manolis of Coco Asphalt Engineering in Toronto, Ontario are
gratefully acknowledged for providing Reclaimed Asphalt Pavement (RAP) samples for the study. Dr.
Tiffany MacDougall of Imperial Oil Limited is acknowledged for FTIR spectra analysis. Dr. John
Brownie of Imperial Oil Limited is acknowledged for proofreading this manuscript. Imperial Oil Limited
and ExxonMobil Corporation are acknowledged for providing support for this study.
© Canadian Technical Asphalt Association 2017

372 ASPHALT RE-RECYCLING
ABSTRACT
Use of Reclaimed Asphalt Pavement (RAP) is beneficial to both road owners and builders as it allows for
significant raw material cost reduction, while potentially maintaining expected pavement service life. In
upcoming decades, the recycling of previously recycled pavements (i.e., re-recycling) will become
widespread. There is currently little technical knowledge on how or how many times asphalt pavement
can be recycled while sustaining its expected durability.
A novel asphalt binder aging method involving thin layers, heat, water spray, and UV radiation was
developed to simulate approximately 20 years of in-service aging. The aged binder was recovered and
blended with a softer, virgin binder. The blend was subjected to the next aging cycle. The process was
repeated four times to simulate four recycling cycles (80 years) at 25 percent RAP addition. Three virgin
binders were tested: standard Performance Grade (PG), one grade softer PG, and standard PG softened
with paraffinic oil to one PG softer binder. Very detailed chemical & rheological analyses were performed
to understand the impact of multiple recycling on irreversible chemical changes and evolution of
rheological properties over the time. Results indicated that at moderate recycling levels, re-recycling is a
viable option if an appropriate virgin binder is used.
RÉSUMÉ
L'utilisation de granulats bitumineux recyclés (GBR) est bénéfique tant pour les propriétaires de routes
que pour les constructeurs, car elle permet une réduction significative des coûts des matières premières,
tout en conservant potentiellement la durée de vie prévue des chaussées. Dans les prochaines décennies, le
recyclage des revêtements préalablement recyclés (re-recyclage) se généralisera. Il existe actuellement peu
de connaissances techniques sur la façon dont ou le nombre de fois où la chaussée bitumineuse peut être
recyclée tout en maintenant sa durabilité attendue.
Une nouvelle méthode de vieillissement du bitume impliquant des couches minces, de la chaleur, de l'eau
pulvérisée et des rayons UV a été développée pour simuler environ 20 ans de vieillissement en service. Le
liant vieilli a été récupéré et mélangé avec un liant vierge plus souple. Le mélange a été soumis au
prochain cycle de vieillissement. Le processus a été répété quatre fois pour simuler quatre cycles de
recyclage (80 ans) à 25% d'addition de GBR. Trois liants vierges ont été testés: grade de performance
standard (PG), un grade PG plus souple et un PG standard ramolli avec de l'huile paraffinique pour obtenir
un liant PG plus souple. Des analyses chimiques et rhéologiques très détaillées ont été effectuées pour
comprendre l'impact du recyclage multiple sur les changements chimiques irréversibles et l'évolution des
propriétés rhéologiques au fil du temps. Les résultats indiquent qu'à des niveaux de recyclage modérés, le
re-recyclage est une option viable si un liant vierge approprié est utilisé.
© Canadian Technical Asphalt Association 2017

KRIZ, TARDIFF, STA. MARIA & SHIRTS 373
1.0 INTRODUCTION
1.1 Background and Overview of the Issues
Reclaimed Asphalt Pavement (RAP) material represents a considerable economic value. RAP has been
widely used in road construction layers as well as a component of new Hot Mix Asphalt (HMA)
pavements since the oil crisis of 1970s. Significant growth in RAP usage has occurred in the last decade
due to increasing economic and environmental pressures. On average about 20 percent RAP is added to
asphalt mixtures in the U.S.A. today, with Canadian usage being somewhat lower [1]. RAP is generally
obtained by milling of failed asphalt pavements and thus this material needs to be reused with care in the
new asphalt mix. In service, an asphalt binder is exposed to the elements for a number of years, which
significantly alters its chemical and rheological properties. Aging results in a substantial increase in RAP
binder consistency, reduction in stress relaxation ability, and decrease in cohesive strength. These changes
generally reduce RAP binder ability to resist fatigue and low temperature cracking.
When RAP is added to the new asphalt mix, the properties of the aged RAP binder have to be accounted
for in the design. At higher RAP levels, one or more grades softer Performance Graded (PG) binder is
typically used to compensate for high stiffness and poor relaxation properties of RAP binder. Blending
kinetics between virgin and RAP binders is important as it significantly influences the mix performance
[2], [3]. In addition to using a softer PG, softening agents such as paraffinic & aromatic oils, biomass
derived oils & other bio-products, re-refined engine oil bottoms, etc. have been used to provide softer
binders to compensate for RAP properties. The chemical composition of a binder in RAP-containing
pavement is thus quite different from a traditional refinery produced PG. It contains RAP binder, virgin
binder and/or additional softening agents sometimes of a very different chemical nature. Field aging of
such a complex mixture may be quite different from aging of refinery produced asphalt grades and aging
protocols in the current American Association of State Highway and Transportation Officials (AASHTO)
specifications may be insufficient to prevent premature pavement failures due to excessive aging.
Extended aging protocols, such as 40-hour Pressure Aging Vessel (PAV) are suggested [4], [5].
As recycling has become widespread during the last decade, it is reasonable to assume that RAP
containing pavements will be milled and re-used in decades to come. Oxidation of asphalt molecules is a
permanent and irreversible process, addition of a softening agent only adjusts the viscosity and/or re-
disperses molecular aggregates, but does not reverse the permanent changes to chemical composition.
There is currently little technical knowledge on how many times asphalt pavement can be recycled while
sustaining its expected durability. To the Authors’ best knowledge there is currently only one project
specifically looking at multiple asphalt mix recycling. It was commissioned in France with an overall
budget of €4.7M and first outcomes are expected in 2018 [6]. The intention here is to initiate the
discussion on asphalt re-recycling in Canada, as this will become an important issue in the near future.
1.2 Case for Action and Study Objectives
Increased use of RAP and artificial softening agents alters binder chemistry in recycled pavements. Binder
aging may thus differ from aging of traditional, refinery produced, binders. Economic and environmental
pressures are assumed to drive high RAP use in upcoming decades and therefore pavement re-recycling
needs to be studied today. The following objectives are proposed:
1. Develop a realistic laboratory binder aging protocol to better reflect in-service binder aging;
2. Understand the impact of multiple recycling cycles on binder chemistry and rheology; and
3. Understand what type of binder is the most suitable for maintaining durability with re-recycling.
© Canadian Technical Asphalt Association 2017

374 ASPHALT RE-RECYCLING
2.0 MATERIALS AND METHODS
2.1 Asphalt Samples
Ontario sourced RAP (Field RAP) was provided by Coco Asphalt Engineering in Toronto, Ontario to
obtain a typical RAP binder from Southwestern Ontario for comparative purposes. The RAP binder was
extracted by TriChloroEthylene (TCE) according to ASTM International D1856. The chemical analysis
confirmed that the recovered RAP binder was unmodified and originated from Western Canadian heavy
crudes, such as Cold Lake bitumen.
Three virgin asphalts were selected for a laboratory re-recycling experiment:
1. PG 64−22 sourced from Imperial Oil Limited Strathcona refinery, Edmonton, Alberta ;
2. PG 58−28 sourced from ExxonMobil Billings refinery in Billings, Montana (USA); and
3. Laboratory prepared blend of PG 64−22 from ExxonMobil Billings refinery softened to meet PG
58−28 by adding 12.5 weight-percent of very heavy paraffinic oil (viscosity at 60 °C of 5.4 Pa·s).
Specification properties of Field RAP and virgin binders are presented in Table 1.
Table 1. Specification properties of asphalt samples used in the study.
Test
Test Method
Field RAP
PG 64−22
PG 58−28
PG 64−22 Oil
Original Asphalt
Density at 15 °C, g/cm3
ASTM D70
1.0540
1.037
1.0305
1.0281
Softening Point, °C
ASTM D36
65.4
43.8
41.8
42.0
Flash Point, °C
AASHTO T 48
ND
279
312
310
Penetration at 25 °C (100g/5s)
ASTM D5
22
92
126
118
Viscosity at 60 °C, Poise
ASTM D2171
65963
1860
874
1220
135 °C, cSt
ASTM D2170
1761.8
395.6
259
332
Solubility in TCE, %wt.
AASHTO T 44
99.76
99.90
99.95
99.60
n-Heptane insolubles, %wt.
ASTM D3279
26
15
13
14.4
DSC wax, %wt.
Internal method1
2.1
0.7
2.2
2.4
Viscosity at 135°C, Pa·s
AASHTO T 44
2.010
0.392
0.253
0.33
DSR T at |G*|/sin δ = 1.00 kPa, °C
AASHTO T 315
87.7
65.6
59.3
59.9
Rolling Thin Film Oven Residue
AASHTO T 240
DSR T at |G*|/sin δ = 2.20 kPa, °C
AASHTO T 315
97.7
67.2
60.9
60.5
Pressure Aging Vessel Residue
AASHTO R 28
DSR T at |G*|sin δ = 5000 kPa, °C
AASHTO T 315
30.4
19.3
18.9
15.3
BBR T at S = 300 MPa, °C
AASHTO T 313
−15.2
−18.1
−19.8
−24.3
BBR T at m = 0.300 MPa/s, °C
AASHTO T 313
−2.2
−18.8
−18.6
−19.2
HTPG, °C
AASHTO R 29
87.7
65.6
59.3
59.9
LTPG, °C
AASHTO R 29
−12.2
−28.1
−28.6
−29.2
Performance Grade
AASHTO R 29
82−10
64−28
58−28
58−28
1
Wax determined by integrating melting enthalpy from Differential Scanning Calorimetry (DSC) data obtained by
heating at a rate of 10 °C per minute. To obtain a wax percentage, the melting enthalpy of pure crystallisable fraction
was estimated as a function of temperature: 1.371T + 141.5 J/g for −54 ≤ T ≤ 34 °C and 188 J/g for T > 34 °C.
© Canadian Technical Asphalt Association 2017

KRIZ, TARDIFF, STA. MARIA & SHIRTS 375
2.2 Weather-Ometer Aging Method (WOM)
The purpose of developing an alternative long-term aging methodology was to incorporate more realistic
aging conditions in terms of temperature, film thickness, and climatic conditions (i.e., moisture, UV light,
etc.) and to induce somewhat harsher aging versus the current protocols for paving asphalts that may
underestimate in service aging [4]. Although exposure of binder to UV light is limited to a very top
pavement layer and not the bulk, it is important with respect to top-down, low temperature cracking. UV
light also increases the severity of aging and greatly speeds-up the process in WOM. For these reasons it
was adopted in the aging protocol.
2.2.1 Instrument
The R. B. Atlas Ci4000 Xenon Arc Weather-Ometer (WOM) was used. It is a weathering instrument
capable of simulating exposure to the elements. The instrument was equipped with a xenon lamp that
produced irradiance simulating sunlight; in the infrared spectrum above 750 nm, in the visible spectrum
400 – 750 nm, and in the ultraviolet spectrum 200 – 400 nm. The spectral output from the lamp was
altered by placing type “S” Borosilicate filters around the lamp so that only certain portions of each
wavelength band can transmit in order to simulate outdoor sunlight at the Earth’s surface. The lamp
operated at a power output between 2500 – 7500 W, and has irradiance ranges in the broad band range of
29 to 141 W/m2 for a wavelengths between 300 – 400 nm, in the narrow band ranges of 0.25 to 1.26 W/m2
at 340 nm and 0.59 – 2.76 W/m2 at 420 nm, and in the illuminance control/Lux range of 307 – 1356 W/m2
at 400 – 750 nm. The WOM simulated rainfall by spraying deionized filtered water on specimens as they
rotated on a rack around the lamp. The water temperature was held constant at 7°C.
2.2.2 WOM Aging Sequence
Parameters of the instrument were set up per ASTM D4798, Standard Practice for Accelerated Weathering
Test Conditions and Procedures for Bituminous Materials (Xenon-Arc Method), Cycle B-2. In this
method thin films of asphalt are spread onto aluminum plates and then placed onto racks inside the WOM
chamber where the plates are rotated around the xenon lamp and exposed to a sequence of light exposure
and filtered water spray. The chamber has a rotating rack with three levels, the top and bottom levels are
angled to provide uniform exposure over all specimens. For cycle B-2 the instrument is set to control
irradiance at 340 nm at 0.35 W/m2.
Samples placed in the WOM chamber go through four stages: Stage 1 exposes the samples to 60 minutes
of light combined with water spray on the front and back of the specimens; Stage 2 is a 90 minute light-
only exposure with the chamber temperature controlled at 60°C; Stage 3 exposes the sample to light and
water spray on the front and back of the specimens for 120 minutes; Stage 4 is a 990 minute light-only
exposure with the chamber at 60°C. Cycle B-2 also includes a stage of 180 minutes of cold exposure,
which was excluded from this study. The four stages are completed in one day.
2.2.3 WOM Sample Preparation and Aging
The three virgin asphalts were aged in the Rolling Thin Film Oven (RTFO) first to simulate hot mix plant
aging. Then the thin films of asphalt were prepared on aluminum plates and aged in WOM to simulate
approximately 20 years of service. To prepare samples for the WOM, 0.6 mm thick aluminum plates were
taped off to leave a rectangular area (50 by 110 mm) in the centre of the plate. Asphalt samples were
heated at 140°C until fluid, poured onto the exposed area of the aluminum plate, and covered with a sheet
of Teflon. A press equipped with two temperature controlled platens was set to approximately 150°C. The
© Canadian Technical Asphalt Association 2017

376 ASPHALT RE-RECYCLING
aluminum plate, asphalt, and the Teflon sheet were pressed for 5 - 10 seconds between the platens with a
force of approximately 3500 psi to obtain asphalt films of 0.65 mm thick. The specimen was removed
from the press and allowed to cool to room temperature. The Teflon sheet was peeled off and the sample
plate was placed into a lukewarm water bath. This allowed the excess asphalt to be peeled off with the
tape, leaving a clean rectangle of sample on the aluminum plate.
A total of 20 to 25 plates for each sample were placed into WOM holders and positioned on the bottom
level of the sample rack. Each plate underwent three complete WOM sequences (three days) as described
in the Section 2.2.3, with each plate being inverted up-side-down after each cycle. Since the ASTM
method is developed for a stiffer roofing coating grades, inverting plates was necessary in this study, as a
softer paving samples had a tendency to flow down the plate. When the aging process was finished, the
plates were placed into an oven at 100°C to heat until the sample could be scraped off using a Teflon
scraper, and combined into a single sample container.
2.3 Laboratory Asphalt Re-recycling
A portion of WOM-aged sample was tested and the other portion was used to prepare the next so called
softened blend, which would be tested and then used in a subsequent WOM-aging to simulate a next ‘in-
service’ period. The term “softening” versus more commonly used “rejuvenation” is preferred here for a
process when an aged asphalt sample is brought to an original viscosity/modulus by blending with softer
virgin asphalt or a variety of oils and other lower viscosity products. The blending process does not
reverse the effects of chemical oxidation, which are permanent changes to molecular structures. Therefore
the process of aging cannot be reversed by blending to restore binder properties to a former state.
The following sample preparation process was followed: The aged binders prepared in WOM were
softened with fresh virgin asphalt binder (from which they originated) at a mass ratio of 25 percent Aged
to 75 percent Virgin. Both aged and softened samples were taken through a complete set of testing before
each cycle would be repeated (Figure 1). A total of four “re-recycling” cycles were performed and
samples termed Aged 1-4 and Soft 1-4 were generated from each of the three virgin source binders.
Figure 1. Testing protocol to generate laboratory aged and softened samples.
The re-recycling experiment thus resulted in 24 laboratory prepared samples. One WOM-aging protocol
was estimated to represent aging equivalent to approximately 20 years of pavement in-service aging. This
assumption was made by comparison to Field RAP (cf. Section 3.1). Four re-recycling cycles performed
in this study are thus assumed to represent 80 years or pavement service at 25 percent RAP addition
Start with
Virgin Binder
Specification,
Rheological &
Chemical
Testing
i=0
RFTO Aging
i+1
WOM Aging à
Samples Aged i
Specification,
Rheological &
Chemical
Testing
Softening:
25%wt. Aged i
+ 75%wt. Virgin à
Samples Soft i
Cycle repeated
four times
i={1,2,3,4} i<4?
Y
EndN
© Canadian Technical Asphalt Association 2017

KRIZ, TARDIFF, STA. MARIA & SHIRTS 377
(binder replacement) every cycle. It is a very rough assumption for the sake of discussion as an exact age
of the Field RAP was unknown.
2.4 Specification Testing
Asphalt specification testing was performed by the AASHTO Material Reference Laboratory (AMRL)
accredited Imperial Oil Limited laboratory in Sarnia, Ontario according to respective ASTM and
AASHTO standards.
2.5 Rheological Testing
A Thermal Analysis (TA) AR 2000ex rheometer equipped with a liquid nitrogen cooling system was used
for small strain dynamic oscillatory testing. Parallel plate geometry was used in all tests. Gap settings and
plate diameter selection are listed in Table 2. The minimum test temperature ranged from −10 to 0°C,
depending on the sample stiffness and brittleness. The maximum test temperature varied according to
sample stiffness and ranged between 80 and 100°C. The temperature increment for isothermal tests was
10°C. A strain sweep was performed at 100 rad/s for each test temperature and each sample to determine
the linear viscoelastic region and appropriate strain level for subsequent frequency sweeps (Appendix,
Table 4). Dynamic oscillatory frequency sweeps were subsequently performed from 0.1 to 100 rad/s at
each temperature. Mastercurves were merged graphically in the TA Data Analysis software. The reference
temperature was 50°C. Mastercurves were obtained for the material functions  . Williams-
Landel-Ferry model was used for time-temperature domain transformation [7]. The discrete relaxation and
retardation spectra were calculated using a custom Excel® routine based on the approach of Baumgaertel
and Winter [8].
Table 2. Geometry, temperature ranges and gap settings.
Geometry
Temperature
Gap
40 mm Parallel plate
70 to 100 °C
0.5 mm
25 mm Parallel plate
30 to 70 °C
1 mm
8 mm Parallel plate
–10 to 30 °C
2 mm
2.6 FTIR Testing
Fourier-Transform Infrared Spectroscopy (FTIR) data was collected using a Thermo Nicolet 4700 ATR
equipped with a Smart Orbit Diamond accessory. A diamond crystal approximately 2 mm in diameter was
covered with a small amount of sample and the sample was pressed onto the crystal to allow a good
contact. Sample absorbance was determined between wavelengths of 4000 and 500 cm−1.
3.0 RESULTS AND DISCUSSION
3.1 Validation of WOM Aging Method
Specification, chemical and rheological properties of the three virgin binders aged in RTFO and WOM
were determined and compared to properties of a Field RAP binder recovered from SW Ontario RAP,
which was not laboratory-aged in RTFO or PAV prior to testing.
© Canadian Technical Asphalt Association 2017

378 ASPHALT RE-RECYCLING
Superpave and traditional asphalt specification test results are presented in Figure 2. The experimental
results indicate that the WOM-aging method is able to provide aged samples with specification properties
comparable to recovered Field RAP binder. The difference between limiting temperature for flexural
creep stiffness and m-value,    or 
 where  stands for critical, was slightly negative for all
specimens, with the lowest value observed for oil-softened PG 64−22. As m-value is related to stress
relaxation, this indicates the loss of relaxation properties during aging was somewhat faster in this sample.
Figure 2. Comparison between Aged 1 binders and recovered Field RAP binder. Low temperature
properties (left): Fraass breaking point, Bending Beam Rheometer (BBR) test results presented as
limiting temperatures; High temperature properties (right): Dynamic Shear Rheometer (DSR) test
as limiting temperature, Softening point and Penetration (100g, 5s) at 25 °C.
The chemical composition of PG 64−22 Aged 1 and Field RAP binders was compared by FTIR analysis.
It has been well-established that oxidation of asphalt is a major contributor to pavement embrittlement and
cracking. Asphalt is an inherently complex system, featuring a wide range of chemical functionalities,
many of which are subject to oxidation due to atmospheric oxygen, UV light and/or elevated temperatures.
The oxidation is generally considered to take the form dehydrogenation (unsaturation, aromatization) and
introduction of oxygen containing functionalities into asphalt molecules. Molecular changes during
asphalt aging generally result in reduced molecular mobility and increased polarity and level of
intramolecular associations. Asphalt stiffens and embrittles as a consequence of these changes [9, 10].
To verify the effectiveness of WOM aging, FTIR data obtained for Field RAP and PG 64−22 Aged 1
samples were compared (Figure 3). The two spectra are similar. Peaks observed at 1700 and 1000 cm−1
are indicative of C=O and C−O functionalities associated with oxidation. The peak at 1000 cm−1 was more
prominent in the Field RAP spectrum when compared to laboratory aged samples. Moreover, the spectrum
for the Field RAP shows significant baseline drift, which may possibly be due to the higher concentration
of condensed aromatic rings. FTIR data indicate that the Field RAP is slightly more oxidized than the
WOM-aged asphalt. FTIR data for PG 58−28 and PG 64−22 + Oil Aged 1 samples were not collected.
Lastly, the effectiveness of the WOM-aging method to simulate in-service aging was studied through
means of detailed rheological analysis. The complex modulus as a function of temperature at 10 rad/s
isochrone for Field RAP and the three Aged 1 samples is presented in Figure 4. The complex modulus
function is comparable for all four binders, with Field RAP showing slightly higher temperature
susceptibility as demonstrated by steeper slope at higher temperature. Also the Field RAP is slightly
stiffer at intermediate temperatures. This suggests that WOM method slightly underestimates the extent of
field aging in this case.
-24
-18
-12
-6
0Fraass, °C T, S=300
MPa, °C T, m=0.3
MPa/s, °C T (S-m), C
Field RAP PG 64-22 Aged 1
PG 58-28 Aged 1 PG 64-22 + Oil Aged 1
0
20
40
60
80
100
T, |G*|/sin 
kPa, CSoftening point,
°C Pen 25 °C, dmm
Field RAP PG 64-22 Aged 1
PG 58-28 Aged 1 PG 64-22 + Oil Aged 1
© Canadian Technical Asphalt Association 2017

KRIZ, TARDIFF, STA. MARIA & SHIRTS 379
Figure 3. FTIR spectra, comparison between Field RAP and WOM-aged samples.
Figure 4. Comparison of dynamic material functions for Aged 1 and Field RAP samples. Complex
modulus versus temperature at 10 rad/s reference frequency (top); Black diagram, i.e. complex
modulus versus phase angle (bottom).
0
0.2
0.4
0.6
0.8
1
1.2
1.4
5001000150020002500300035004000
Relative Absorbance
Wavenumber, cm
PG 64-22 Aged 1
Field RAP
0
2.5
5
7.5
10
-40 -20 0 20 40 60 80 100 120 140 160 180
log |G*|, Pa
Temperature, C
Field RAP
PG 64-22 Aged 1
PG 58-28 Aged 1
PG 64-22 + Oil Aged 1
0
2.5
5
7.5
10
010 20 30 40 50 60 70 80 90
log |G*|, Pa

Field RAP
PG 64-22 Aged 1
PG 58-28 Aged 1
PG 64-22 + Oil Aged 1
© Canadian Technical Asphalt Association 2017

380 ASPHALT RE-RECYCLING
A Black diagram (Figure 4 bottom) is often used to finger print chemical composition of binders through
means of rheological analysis. It has been shown that Black diagrams are useful to trace the extent of
oxidation in binders [10]. All four binders are quite similar in Black diagram, however, PG 64−22 + Oil
Aged 1 sample is more elastic at higher complex modulus levels than the other three samples. It has been
shown [10] that this indicates a higher degree of oxidation.
The relaxation modulus and equilibrium compliance functions were compared next (Figure 5). These
functions were calculated from relaxation and retardation spectra, respectively. For details on how these
functions are calculated see [10]. The relaxation properties of Field RAP and PG 64−22 + Oil Aged 1
sample differ from other two Aged 1 samples, with relaxation times shifted to longer times. Relaxation is
very sensitive to asphalt oxidation and this indicates that PG 64−22 + Oil Aged 1 aged somewhat more
severely during the experiment in comparison with the two straight-run asphalts.
Figure 5. Relaxation modulus versus time (top) and equilibrium compliance versus time (bottom).
Similar observation was made when the creep compliance corrected for the contribution of viscous flow,
 , versus time function was studied. It has been shown that for neat and mildly oxidized
asphalts or lower molecular weight polymers, this quantity levels off and approaches constant equilibrium
-15
-10
-5
0
5
10
-10 -8 -6 -4 -2 0 2 4 6 8
log G(t), Pa
log t, s
Field RAP
PG 64-22 Aged 1
PG 58-28 Aged 1
PG 64-22 + Oil Aged 1
-10
-7.5
-5
-2.5
0
-10 -8 -6 -4 -2 0 2 4 6 8
J(t)-t/0, 1/Pa
log t, s
Field RAP
PG 64-22 Aged 1
PG 58-28 Aged 1
PG 64-22 + Oil Aged 1
© Canadian Technical Asphalt Association 2017

KRIZ, TARDIFF, STA. MARIA & SHIRTS 381
compliance at longer times [10], [11]. This was observed for PG 64−22 Aged 1 and PG 58−28 Aged 1
samples. Absence of equilibrium compliance is typical for severally oxidized asphalts or high molecular
weight polymers (Mw > 100,000 Da.). Although oxidized asphalts do not necessarily have such high
molecular weight, increased polarity due to aging leads to intra- and inter-molecular interactions and
associations. Larger degree of association in oxidized samples may slow down Brownian motion resulting
in the formation of a weak structure of a very high apparent molecular weight and may prevent induced
longer range molecular rearrangements similar to that observed in high molecular weight polymers. This
behaviour was observed for Field RAP and for the PG 64−22 + Oil Aged 1 sample.
In summary, the WOM method appears to provide quite realistic aging for paving binders. Thin asphalt
film thickness, aging temperatures within realistic pavement temperatures, and exposure to water and UV
light provided an aging process able to mimic some 20 years of in-service aging in Southwestern Ontario.
However, there are some limitations. The method is very complex and time consuming, suitable for
research activities but impractical for routine use. Although the rheological and chemical properties were
similar to Field RAP, certain differences exist, suggesting the WOM-aging is not as severe as in-service
aging. The UV aging is also not exactly representative of in-service aging. The UV light was found to be a
strong factor in WOM aging method as it significantly accelerates certain types of oxidation reactions;
however it is established that UV light does not penetrate deep in the pavement. Therefore, WOM method
is thus likely more relevant to the aging in the very top pavement layer rather than in the bulk.
Nevertheless, the objective here was to develop more realistic laboratory aging method to understand
whether asphalt binder properties are compromised by multiple recycling and also to distinguish among a
variety of virgin asphalt binders used to soften the RAP binder. With that regard, the method aging
capability was found satisfactory.
3.2 Effect of Re-recycling on Specification Properties
3.2.1 Testing Overview
Field RAP, virgin binders and their respective Aged 1-4 and Soft 1-4 samples were subjected to
specification testing. No artificial laboratory aging (i.e., PAV) was done prior to the testing. Samples were
tested as received after WOM-aging (Aged 1-4) or after subsequent softening with virgin binder (Soft 1-
4). Also the original virgin binders and Field RAP samples were not laboratory-aged and all properties
were measured on the original samples in order to allow for comparison. That includes BBR and DSR at
intermediate temperature. The BBR limiting temperatures are reported as interpolated values from actual
test temperatures, i.e. −10 °C factor as prescribed in AASHTO was not applied. Eight charts were
prepared for the three datasets, as follows:
1. Dynamic shear rheometer (DSR) test according to AASHTO T 315 to determine the limiting
temperature in °C at  
 ;
2. DSR test according to AASHTO T 315 at intermediate temperature. Results are reported as 
  in MPa;
3. BBR test according to AASHTO T 313 to determine the limiting temperature in °C at 
, i.e. Stiffness limiting temperature;
4. BBR test according to AASHTO T 313 to determine the limiting temperature in °C at  
 , i.e. m-value limiting temperature;
© Canadian Technical Asphalt Association 2017

382 ASPHALT RE-RECYCLING
5. Difference between flexural creep stiffness and m-value limiting temperatures as obtained from
BBR test, i.e.    in °C;
6. Penetration at 25 °C according to ASTM D5;
7. Softening point according to ASTM D36; and
8. Fraass breaking point according to EN 12593 (European Norm).
3.2.2 Laboratory Re-recycled PG 64−22 – Specification Testing
Evolution of both Superpave and conventional specification properties with re-recycling for PG 64−22
and its Aged and Soft samples is presented in Figure 6. With each recycling cycle a slight hardening and
degradation of relaxation properties was observed. This is most remarkable by an increase in m-value
limiting temperature or by a decrease in   . After the fourth cycle the    drops below
−5°C which has been suggested as a performance limit with regards to excessive fatigue cracking in field
samples [4]. The reason is likely a phase separation due to uneven polarity increase across the asphalt
solubility fractions and inability to effectively disperse the oxidized asphaltene fraction. Data suggest that
the use of original straight run asphalt PG to recycle the RAP binder at a 25 percent binder replacement
rate, i.e., not one PG softer or artificial softened binder, can yield a satisfactory performance up to three
re-recycling cycles which should approximate 60 years of pavement service.
3.2.3 Laboratory Re-recycled PG 58−28 – Specification Testing
Next, a one grade softer PG was used in re-recycling process. Specification properties are presented in
Figure 7. The PG 58−28 Aged samples are approximately by 6°C softer than the PG 64−22 aged samples
at the high temperature PG. They are quite comparable at the low temperature PG, which can be attributed
to excellent low temperature properties of Strathcona PG 64−22 (meets marginally PG 64−28). No trends
were observed with re-recycling as PG 58−28 kept satisfactory properties during all four cycles. That
includes    which still showed satisfactory results after the fourth cycle. The    ranged
between −2 and –3°C indicating good phase stability during the re-recycling. This observation suggests
that PG 58−28 can be re-recycled for four or possibly more times at a 25 percent binder replacement rate.
3.2.4 Laboratory Re-recycled PG 64−22 + Oil – Specification Testing
An alternative to PG 58−28 (one PG softer) is softening the standard PG 64−22 with oil to PG 58−28.
Specification properties for PG 64−22 + Oil samples are presented in Figure 8. Although the evolution of
high temperature properties with re-recycling is comparable to PG 58−28, the low temperature properties
degrade rapidly, with    reaching less than −5°C as soon as after the second cycle. This may
indicate a loss of relaxation properties due to a phase separation. Softening oil can improve asphalt low
temperature properties initially, however after aging the benefits are quickly lost and artificially softened
asphalt is outperformed by straight run refinery asphalts. That includes PG 64−22. There are, however
many softening agents available on the market and their chemical and rheological properties differ. Their
performance in the re-recycling process may differ from that of the oil used in this study.
The purpose here was to highlight that artificial softening may not be the optimal solution, and future re-
recycling may aggravate the problem. Research is needed to understand compatibility of softening agents
in asphalt especially after long term aging. Until this is fully understood and field validated, the use of
softer refinery grades (e.g. PG 58−28) appears as the best solution for recycling followed by the use of
refinery grades applicable to a given climate (e.g. PG 64−22).
© Canadian Technical Asphalt Association 2017

KRIZ, TARDIFF, STA. MARIA & SHIRTS 383
Figure 6. Specification properties of laboratory re-recycled PG 64−22. Field RAP properties
displayed for comparison.
64
70
76
82
88
94
T, |G*|/sin kPa, C
0
1
2
3
4
5
|G*|·sin at 25 C, MPa
-30
-24
-18
-12
T, S=300 MPa, C
-30
-24
-18
-12
T, m=0.3 MPa/s, C
-6
-4
-2
0
2
4
T (S-m), C
0
20
40
60
80
100
Pen 25 C, dmm
0
20
40
60
80
Softening point, C
-20
-15
-10
-5
0
Fraass Break Pt., C
© Canadian Technical Asphalt Association 2017

384 ASPHALT RE-RECYCLING
Figure 7. Specification properties of laboratory re-recycled PG 58−28. Field RAP properties
displayed for comparison.
58
64
70
76
82
88
T, |G*|/sin kPa, C
0
5
10
15
|G*|·sin at 19 C, MPa
-30
-24
-18
-12
T, S=300 MPa, C
-30
-24
-18
-12
T, m=0.3 MPa/s, C
-4
-2
0
2
4
T (S-m), C
0
20
40
60
80
100
120
Pen 25 C, dmm
0
20
40
60
80
Softening point, C
-20
-15
-10
-5
0
Fraass Break Pt., C
© Canadian Technical Asphalt Association 2017

KRIZ, TARDIFF, STA. MARIA & SHIRTS 385
Figure 8. Specification properties of laboratory re-recycled PG 64−22 + Oil. Field RAP properties
displayed for comparison.
58
64
70
76
82
88
T, |G*|/sin kPa, C
0
5
10
15
|G*|·sin at 19 C, MPa
-30
-24
-18
-12
T, S=300 MPa, C
-30
-24
-18
-12
T, m=0.3 MPa/s, C
-8
-6
-4
-2
0
2
T (S-m), C
0
20
40
60
80
100
120
140
Pen 25 C, dmm
0
20
40
60
80
Softening point, C
-25
-20
-15
-10
-5
0
Fraass Break Pt., C
© Canadian Technical Asphalt Association 2017

386 ASPHALT RE-RECYCLING
3.3 Effect of Re-recycling on Chemical Properties
Figure 9 shows IR spectra for the straight run PG 64−22 asphalt brought through the re-recycling process.
Comparing the virgin PG 64−22 asphalt with the Aged samples, there is an increase in the intensity of a
peaks at approximately 1700 cm−1 and 1000 cm−1, indicative of an increase in C=O and C−O bond
formation, respectively. However, it should be noted that there is not an appreciable change in the IR
spectra between Aged 1 and Aged 4. FTIR spectra for PG 58−28 and PG 64−22 + Oil and their aged
products were not obtained for this study.
Figure 9. FTIR spectra for PG 64−22 (Virgin & Aged 1-4).
In summary, the changes in chemical composition observed in this study directionally support the well-
established knowledge on asphalt aging. The resolution of quantifiable chemical composition parameters
are likely not sufficient to study changes in chemical composition during re-recycling. Other techniques
with better qualitative or quantitative insights (FT-ICR-MS
2
or SAR-AD
3
) [12], [13] would be useful in
studying trends in chemical composition during re-recycling. Early work performed by the current authors
with FT-ICR-MS show promising results in this area.
3.4 Effect of Re-recycling on Rheological Properties
3.4.1 Tested Material Functions
Aged and Soft samples were subjected to dynamic testing in DSR. Material functions that were
determined and analyzed are summarized in Table 3. Two comparative datasets for each of the three
virgin binders and their respective Aged and Soft samples were prepared containing following series:
1. Virgin binder, Soft 1 & 4; and
2. Field RAP, Aged 1 & 4.
2
FT-ICR-MS = Fourier Transform Ion Cyclotron Resonance Mass Spectrometry.
3
SAR-AD = Saturates-Aromatics-Resins-Asphaltene Determinator™.
0
0.2
0.4
0.6
0.8
1
50075010001250150017502000
Relative Absorbance
Wavenumber, cm1
Aged 4
Aged 3
Aged 2
Aged 1
Virgin
© Canadian Technical Asphalt Association 2017

KRIZ, TARDIFF, STA. MARIA & SHIRTS 387
Table 3. List of Examined Material Functions.
Symbol
Property
Units
Independent variable
Units
 or 
Storage Modulus or
Compliance
Pa or 1/Pa
Temperature or reduced frequency
°C or rad/s
 or 
Loss Modulus or Compliance
Pa or 1/Pa
Temperature or reduced frequency
°C or rad/s
 or 
Complex Modulus or
Compliance
Pa or 1/Pa
Temperature or reduced frequency
°C or rad/s
 or  
Phase Angle
°
Temperature or reduced frequency
°C or rad/s

Dynamic viscosity
Pa.s
Temperature or reduced frequency
°C or rad/s

Complex viscosity
Pa.s
Temperature or reduced frequency
°C or rad/s

Complex Modulus
Pa
Phase Angle
°

Loss Modulus
Pa
Storage Modulus
Pa

Relaxation Modulus
Pa
Time
s

Creep Compliance
1/Pa
Time
s
 
 corrected for the
contribution of viscous flow
1/Pa
Time
s
Samples Soft and Aged 2 & 3 are not presented in the charts to increase legibility in B&W print as the
difference were small and a lot of overlap was observed. Due to a large number of datasets and material
functions studied, only a select set of charts is presented herein.
3.4.2 Evolution of Rheological Properties in Softened Samples with Re-recycling
Rheological properties of virgin binder were compared to samples Soft 1 – 4 to understand how the
properties of softened samples evolve with each cycle. Storage and loss moduli and phase angle as
functions of temperatures are presented in Figure 10. As expected, softened samples have generally higher
moduli and higher degree of elasticity when compared to the virgin binder as they contained 25 weight-
percent of WOM-aged material. No gradual increase in stiffness or elasticity was observed in a series of
softened samples (Soft 1 – 4), indicating that the effect of re-recycling may not be apparent in the softened
samples. Similar observation was made for all three virgin binders and their softened samples.
The relaxation and retardation properties were compared next. Examples of relaxation modulus and creep
compliance corrected for a contribution of viscous flow are presented in Figure 11. Relaxation modulus of
softened samples is shifted to longer relaxation times especially at lower modulus values when compared
to virgin binder. The plateau observed for softened samples may be an indication a phase separation as the
two amorphous phases relax stresses at different rates. No significant gradual change in relaxation spectra
was observed for the series of softened samples. The situation is somewhat different when creep
compliance corrected for contribution of viscous flow function is examined (Figure 11). The virgin asphalt
shows equilibrium compliance at longer times (constant), while softened samples do not. Also with every
recycling cycle the softened samples gradually increase the equilibrium compliance threshold suggesting
an increase in polarity and degree of molecular association.
© Canadian Technical Asphalt Association 2017

388 ASPHALT RE-RECYCLING
Figure 10. Comparison of isochrones (ω=10 rad/s) for softened samples. Example for  for PG
64−22 series (top),  for PG 58−28 series, (middle) and phase angle, , for PG 64−22 + Oil
series (bottom).
-2.5
0
2.5
5
7.5
10
-50 -25 0 25 50 75 100 125 150
log G', Pa
T, C
PG 64-22
PG 64-22 Soft 1
PG 64-22 Soft 4
0
2.5
5
7.5
10
-50 -25 0 25 50 75 100 125 150
log G'', Pa
T, C
PG 58-28
PG 58-28 Soft 1
PG 58-28 Soft 4
0
10
20
30
40
50
60
70
80
90
-50 -25 0 25 50 75 100 125 150
, 
T, C
PG 64-22 w Oil
PG 64-22 w Oil Soft 1
PG 64-22 w Oil Soft 4
© Canadian Technical Asphalt Association 2017

KRIZ, TARDIFF, STA. MARIA & SHIRTS 389
Figure 11. Comparison of relaxation modulus (PG 64−22 + Oil, top) and creep compliance corrected
for contribution of viscous flow function (PG 64−22, bottom) for virgin and softened binders.
3.4.3 Evolution of Rheological Properties in Aged Samples with Re-recycling
Rheological properties of Field RAP binder were compared to samples Aged 1 – 4 to understand how
much the laboratory aged samples would differ from Field RAP and also whether properties of aged
samples gradually evolve with each cycle. Storage and loss moduli and phase angle as functions of
temperatures are presented in Figure 12. Properties of aged samples and Field RAP were found
comparable especially for Aged 4 samples. Slight gradual increase in stiffness and elasticity was observed
in Aged samples with each cycle. This was further confirmed when zero shear viscosity, steady-state
compliance, crossover frequency, and rheological index were compared (Figure 13). The first two
parameters were calculated from relaxation and retardation spectra, respectively [10]. Crossover frequency
and rheological index were calculated by fitting with the Christensen-Anderson model [14]. Zero shear
viscosity gradually increases with re-recycling indicating material stiffening with every cycle, while
steady-state compliance decreases, indicated reduced ability to flow. The same effect is observed with
crossover frequency (decreases) and rheological index (increases).
-5
-2.5
0
2.5
5
7.5
10
-10 -8 -6 -4 -2 0 2 4 6
log G(t), Pa
log t, s
PG 64-22 w Oil
PG 64-22 w Oil Soft 1
PG 64-22 w Oil Soft 4
-10
-7.5
-5
-2.5
0
-10 -8 -6 -4 -2 0246
log J(t)-t/0, 1/Pa
log t, s
PG 64-22
PG 64-22 Soft 1
PG 64-22 Soft 4
© Canadian Technical Asphalt Association 2017

390 ASPHALT RE-RECYCLING
Figure 12. Comparison of isochrones (ω=10 rad/s) of dynamic material functions for aged samples.
Example for  for PG 64−22 + Oil series (top),  for PG 64−22 series (middle), and ,
for PG 58−28 series (bottom).
-2.5
0
2.5
5
7.5
10
-25 0 25 50 75 100 125 150
log G', Pa
T, C
Field RAP
PG 64-22 w Oil Aged 1
PG 64-22 w Oil Aged 4
0
2.5
5
7.5
10
-25 0 25 50 75 100 125 150
log G'', Pa
T, C
Field RAP
PG 64-22 Aged 1
PG 64-22 Aged 4
0
10
20
30
40
50
60
70
80
90
-25 0 25 50 75 100 125 150
, 
T, C
Field RAP
PG 58-28 Aged 1
PG 58-28 Aged 4
© Canadian Technical Asphalt Association 2017

KRIZ, TARDIFF, STA. MARIA & SHIRTS 391
Figure 13. Gradual changes in zero-shear viscosity, , (top-left), steady-state compliance, 
, (top-
right), crossover frequency, , (bottom-left), and rheological index, R, (bottom-right) for in Aged
samples during re-recycling.
The example of relaxation modulus and creep compliance corrected for a contribution of viscous flow is
presented in Figure 14 - top. The gradual increase of relaxation times was observed in aged samples with
each cycle. The relaxation modulus for Aged 4 samples was comparable to that of Field RAP. Equilibrium
compliance was only observed for Aged 1 samples (Figure 14 - bottom), starting with Aged 2 samples
equilibrium compliance was not observed suggesting increased level of molecular association and higher
apparent molecular weight. The compliance functions were comparable for Aged 2 – 4 and Field RAP
samples.
3.5 Performance of the Three Virgin Binders in Re-recycling
There are number of criteria when performance of a binder in asphalt re-recycling is considered. Those
would include a variety of rheological and chemical parameters. In order to do the comparison here, two
criteria were considered: 1) viscoelastic properties, e.g., moduli, degree of elasticity and relaxation
properties, and 2) rate of aging, i.e., how fast the asphalt properties change during aging. The first criterion
determines whether the properties of the asphalt are acceptable for service, the second criterion determines
whether the long term performance is predictable. Both are important.
0
0.2
0.4
0.6
0.8
1
1.2
RAP Aged 1 Aged 2 Aged 3 Aged 4
log Je0, 1/Pa
PG 58-28 PG 64-22 PG 64-22 + Oil
4
4.25
4.5
4.75
5
RAP Aged 1 Aged 2 Aged 3 Aged 4
log 0, Pa.s
PG 58-28 PG 64-22 PG 64-22 + Oil
0
2000
4000
6000
8000
10000
12000
14000
RAP Aged 1 Aged 2 Aged 3 Aged 4
0, rad/s
PG 58-28 PG 64-22 PG 64-22 + Oil
0
0.5
1
1.5
2
2.5
RAP Aged 1 Aged 2 Aged 3 Aged 4
Rheological Index
PG 58-28 PG 64-22 PG 64-22 + Oil
© Canadian Technical Asphalt Association 2017

392 ASPHALT RE-RECYCLING
Figure 14. Comparison of relaxation modulus (PG 58−28, top) and creep compliance corrected for
contribution of viscous flow function (PG 64−22, bottom) for Field RAP and aged binders.
Figure 15. Comparison of penetration at 25 °C (left) and temperature at |G*|/sin δ = 1 kPa (right)
for the Aged 1-4 and Soft 1-4 samples originated from the three virgin binders.
-5
-2.5
0
2.5
5
7.5
10
-10 -8 -6 -4 -2 0246
log G(t), Pa
log t, s
Field RAP
PG 58-28 Aged 1
PG 58-28 Aged 4
-10
-7.5
-5
-2.5
0
-10 -8 -6 -4 -2 0246
log J(t)-t/, 1/Pa
log t, s
Field RAP
PG 64-22 Aged 1
PG 64-22 Aged 4
0
25
50
75
100
125
150
Pen 25 C, dmm
PG 58-28 PG 64-22
PG 64-22 + Oil
58
64
70
76
82
88
94
T, |G*|/sin kPa, C
PG 58-28 PG 64-22 PG 64-22 + Oil
© Canadian Technical Asphalt Association 2017

KRIZ, TARDIFF, STA. MARIA & SHIRTS 393
Figure 16. Comparison of temperature at m-value = 0.3 MPa/s (left) and    (right) for the
Aged 1-4 and Soft 1-4 samples originated from the three virgin binders.
Figure 17. Comparison of relaxation modulus (top) and creep compliance corrected for contribution
of viscous flow function (bottom) for Field RAP and Aged 4 binders.
-30
-24
-18
-12
-6
0
T, m=0.3 MPa/s, C
PG 58-28 PG 64-22 PG 64-22 + Oil
-10
-8
-6
-4
-2
0
2
4
T(S-m), C
PG 58-28 PG 64-22 PG 64-22 + Oil
-5
-2.5
0
2.5
5
7.5
10
-10 -8 -6 -4 -2 0 2 4 6 8
log G(t), Pa
log t, s
Field RAP
PG 58-28 Aged 4
PG 64-22 Aged 4
PG 64-22 w Oil Aged 4
-10
-7.5
-5
-2.5
0
-10 -8 -6 -4 -2 0 2 4 6 8
log J(t)-t/, 1/Pa
log t, s
Field RAP
PG 58-28 Aged 4
PG 64-22 Aged 4
PG 64-22 w Oil Aged 4
© Canadian Technical Asphalt Association 2017

394 ASPHALT RE-RECYCLING
The specification properties of aged and softened samples were compared. Penetration and DSR limiting
temperature (Figure 15) are comparable for PG 58−28 and PG 64−22 + Oil series, while for PG 64−22
series penetration is lower and DSR limiting temperature is higher. This is expected as the two former
binders are PG 58−28 and thus softer at the high temperature end. The low temperature properties (Figure
16) are more important when re-recycling is considered. The m-value limiting temperature is quite
comparable for the three series. It is the    parameter which shows significant differences among
the three binders. For PG 64−22 + Oil sample, the limit of −5 °C is approached after the first aging cycle
and exceeded after the second. As mentioned above, −5 °C was suggested as limit to prevent excessive
pavement cracking and binders with    less than −5 °C would be considered inferior. PG 64−22
binder reached the    limit at fourth cycle, while PG 58−28 still performed satisfactorily at
fourth cycle and possibly beyond.
The relaxation properties, which are critical when fatigue and low temperature cracking resistance is
considered, were compared next (Figure 17 – top). The relaxation times increased in this order: PG 58−28
Aged 4, PG 64−22 Aged 4, PG 64−22 + Oil Aged 4, and Field RAP. The longer the times the slower the
relaxation and the more likely is the material to crack as imposed stresses are not relaxed fast enough as
the material strength limit is approached. Therefore samples exhibiting shorter relaxation times are
expected to perform better with respect to cracking. Creep compliance corrected for the contribution of
viscous flow function is presented in (Figure 17 – bottom). Only PG 58−28 Aged 4 shows equilibrium
compliance suggesting the aging did not significantly increased apparent molecular weight in this sample.
PG 58−28 therefore should perform the best with respect to cracking, followed by PG 64−22, and PG
64−22 + Oil.
The rate of aging was evaluated next. The differences in properties before and after WOM aging were
compared in Figure 18. DSR parameter, temperature at   , increased by ~15°C after the
first aging cycle, while it increase by ~18°C after the fourth aging cycle. This suggests slight acceleration
of aging with each recycling. All three binders performed comparably. BBR parameter, temperature at m-
value = 0.3 MPa/s, showed significant difference among the three binders. Both refinery grades,
PG 58−28 and PG 64−22, lost about 4 - 5°C of m-value limiting temperature after every aging cycle, with
slight increase to 6 and 8°C after the fourth cycle. PG 64−22 + Oil samples lost between 10 to 12°C after
every cycle, so the rate of aging is significantly higher for this sample. Such a high loss in low temperature
properties may result in inferior properties and the potential low temperature performance may be grossly
overestimated.
Figure 18. Gain in high temperature properties (left) and loss in low temperature properties (right)
after each aging cycle.
0
5
10
15
20
25
After 1 After 2 After 3 After 4
Gain T,|G*|/sin kPa, C
PG 58-28 PG 64-22 PG 64-22 + Oil
0
2
4
6
8
10
12
14
After 1 After 2 After 3 After 4
Loss in T, m=0.3 MPa/s, C
PG 58-28 PG 64-22 PG 64-22 + Oil
© Canadian Technical Asphalt Association 2017

KRIZ, TARDIFF, STA. MARIA & SHIRTS 395
3.6 Rate of RAP Addition – Implication to Re-recycling
Trends in binder behaviour were observed with increasing number of recycling cycles, however the trends
were relatively subtle with exception of the PG 64−22 + Oil series. The binder replacement rate of
25 weight-percent was used in this study, because it represents somewhat typical RAP usage across the
continent. Figure 19 shows how the representation of virgin binder and aged binders which originated
from previous cycles evolves.
Figure 19. Evolution of proportion of virgin and aged material which underwent aging onetime
(RAP 1), two-times (RAP 2), three-times (RAP 3) and four-times (RAP 4) at 25% wt. recycling rate
(left) and 50 %wt. recycling rate (right).
For an example a pavement which was constructed with virgin materials is assumed. After 20 years it is
milled and 25 weight-percent (binder replacement) of the original pavement material was reused in the
new mix. The binder composition in this mix would be 75 percent of virgin binder and 25 percent of
binder 20 years old (by weight). After another 20 years (40 years in total), the same process is repeated.
The composition of the binder would be 75 percent of virgin binder, 18.75 percent of binder 20 years old
and 6.25 percent of binder 40 years old (by weight). Fast-forward, at the end of the fourth cycle the
composition would be 75 percent virgin, 18.75 percent of binder 20 years old, 4.6875 percent of binder 40
years old, 1.1719 percent of binder 60 years old, and 0.3906 percent of binder 80 years old (again by
weight).
It is apparent that during re-recycling the RAP binder content is always equal to the recycling rate;
however, the composition is dependent to the previous cycles (rate of recycling and number of re-
recycling). It was interesting to observe throughout this study that even a relatively small amount of
significantly aged material (e.g. 0.39 weight-percent four-times through WOM in fourth cycle) can impact
the specification, chemical and rheological properties of the binder. It is reasonable to assume that if such
a small quantity can influence properties of 99.6 percent of remaining material, it is likely extremely
oxidized, difficult to disperse highly polar fraction with very high stiffness and poor relaxation properties.
Very high polarity would also mean such fraction would be very prone to further oxidation and thus
accelerate further aging in subsequent cycles.
It is important to note that when the recycling rate doubles to 50 weight-percent (Figure 19 – right), the
amount of the most aged material, RAP 4 after the fourth cycle in this example, increases from 0.39 to
6.25 weight-percent, i.e. by 16-fold in this particular case. The concentration of the most aged material,
0
0.2
0.4
0.6
0.8
1
1 2 3 4
w/w
Cycle No.
RAP 4
RAP 3
RAP 2
RAP 1
Virgin
0
0.2
0.4
0.6
0.8
1
1 2 3 4
w/w
Cycle No.
RAP 4
RAP 3
RAP 2
RAP 1
Virgin
© Canadian Technical Asphalt Association 2017

396 ASPHALT RE-RECYCLING
i.e., material which originated from the first ‘virgin’ pavement, at the end of a given recycling cycle can
be calculated by Equation 1:
   (1)
Where:  is the fraction of the most aged material;
 is the recycling rate fraction; and
 is the cycle No.
It is apparent that the impact of re-recycling on binder properties increases exponentially with increased
recycling rate. High concentration of binder that has seen several life-cycles can make utilization of such
RAP in the new mix increasingly difficult. In order to make recycling sustainable, it is important to
understand that binder aging is an irreversible process and its properties will eventually be compromised.
Some binders, like PG 64−22 + Oil may only perform for a few cycles, while high quality refinery
produced binders may yield good performance for several cycles. Nevertheless, high recycling rates may
reduce the number of viable cycles significantly and thus make recycling unsustainable in future.
4.0 CONCLUSIONS
The following conclusions based on experimental evidence and with respect to study objectives are
drawn:
1. The Weather-Ometer (WOM) aging method developed here can provide realistic asphalt binder
aging simulation when traditional and Superpave specifications are considered. When detailed
chemical and rheological analysis is performed, the WOM-aging was slightly less severe than the
one observed for Field RAP binder.
2. The evolution of properties with multiple re-recycling at 25 weight-percent binder replacement rate
was studied. The addition of 75 weight-percent virgin binder was able to effectively soften 25
weight-percent of previously aged material every cycle. Despite the properties of softened samples
did not evolve significantly with each recycling cycle, a degradation of performance was observed
with increasing number of cycles on aged samples, namely for low temperature (   and m-
value) and relaxation properties ( and  ). The trend was attributed to altered
chemical composition especially for the material which underwent the highest number of aging
cycles. The most aged material is the most polar, most sensitive to further oxidation and accelerates
aging with increasing number of recycling cycles.
3. Two straight run refinery produced asphalts clearly outperformed asphalt softened with oil when re-
recycling is considered. Initial small differences were magnified with each recycling cycle. Asphalt
softened with oil aged faster and    exceeded −5°C limit as soon as after the second
recycling. The recommendation to achieve the best long-term durability when recycling is
considered is to prefer compositionally well balanced straight-run asphalts over artificially softened
asphalts, even when refinery produced asphalt is one PG harder.
4. In order to make recycling sustainable, it is important to understand that binder aging is an
irreversible process and binder properties will eventually be compromised. High recycling rates and
use of poorly performing virgin binders may reduce the number of viable cycles significantly and
thus make recycling unsustainable in the future.
© Canadian Technical Asphalt Association 2017

KRIZ, TARDIFF, STA. MARIA & SHIRTS 397
REFERENCES
[1]
McDaniel RS, Shah A, Huber GA, Copeland A. “Effects of Reclaimed Asphalt Pavement Content
and Virgin Binder Grade on Properties of Plant Produced Mixtures”, Road Materials and Pavement
Design, 13 (1), 161-182 (2012).
[2]
Kriz P, Grant DL, Veloza B, Gale MJ, Blahey AG, Brownie JH, Shirts RD, Maccarrone S.
“Blending and Diffusion of Reclaimed Asphalt Pavement and Virgin Asphalt Binders”, Road
Materials and Pavement Design, 15 (1), 78-112 (2014).
[3]
Kriz P, Grant DL, Gale MJ, Lavorato S, Pahalan A. “Reclaimed Asphalt Pavement - Virgin Binder
Diffusion in Asphalt Mixes”, Proceedings, Canadian Technical Asphalt Association, 59, 375-400
(2014).
[4]
Reinke G, Hanz A, Anderson M, Ryan M, Engber S, Herlitzka D. “Impact of re-refined engine oil
bottoms on binder properties and mix performance on two pavements in Minnesota”, Proceedings,
6th Eurasphalt & Eurobitume Congress, Prague, Czech Republic (2016).
[5]
Asphalt Institute, “The use of REOB/VTAE in Asphalt, State of the knowledge”, IS-235, Asphalt
Institute, Lexington, Kentucky (2016).
[6]
Insitute pour la recherche appliquée et l'expérimentation en génie civil. “MURE - MUlti-Recyclage
des Enrobés tièdes”, http://www.pnmure.fr/en/ (August 18, 2017).
[7]
Williams ML, Landel LF, Ferry JD. “The Temperature Dependence of Relaxation Mechanisms in
Amorphous Polymers and Other Glass-Forming Liquids”, J Am Chem Soc, 77 (14), 3701-3707
(1955).
[8]
Baumgaertel M, Winter HH. “Determination of Discrete Relaxation and Retardation Time Spectra
from Dynamic Mechanical Data”, Rheol Acta, 28, 511-519 (1989).
[9]
Petersen JC. A Review of the Fundamentals of Asphalt Oxidation, Transportation Research Board,
Washington, D.C. (2009).
[10]
Kriz P, Brownie JH, Blahey AG, Sokol KL, Grant DL, Cointe F, Boussad N, Shirts RD.
“Rheological Assessment of Neat Paving, Air Rectified Paving and Oxidized Roofing Bitumen”,
Proceedings, 5th Eurobitume & Eurasphalt Congress, Istanbul, Turkey (2012).
[11]
Ferry JD, Viscoelastic Properties of Polymers, Wiley, New York (1980).
[12]
Qian K, Robbins WK, Hughey CA, Cooper HJ, Rodgers RP, Marshall AG. “Resolution and
identification of elemental compositions for more than 3000 crude acids in heavy petroleum by
negative-ion microelectrospray high-field Fourier transform ion cyclotron resonance mass
spectrometry”, Energy Fuels, 15 (6), 1505-1511, (2001).
[13]
Schabron JF, Boysen RB, Kalberer EW, Rovani JF. “Hydrocarbon Separation and Analysis
© Canadian Technical Asphalt Association 2017

398 ASPHALT RE-RECYCLING
Apparatus and Methods”, United States Patent US9353317 B2 (May 31 2016).
[14]
Christensen DW, Anderson DA. “Interpretation of Dynamic Mechanical Test Data for Paving Grade
Asphalt”, Journal, Association of Asphalt Paving Technologists, 61, 67-116 (1992).
© Canadian Technical Asphalt Association 2017

KRIZ, TARDIFF, STA. MARIA & SHIRTS 399
APPENDIX
Table 4. Test strain used in frequency sweeps as determined from strain sweeps.
Geometry
Temperature, °C -10 0 10 20 30 30 40 50 60 70 70 80 90 100
% Strain
Field RAP 0.068 0.068 0.200 0.25 0.34 0.1 0.4 1.0 2.0 4 8 10 15 25
PG 64−22 0.060 0.200 0.500 1.00 1.30 1.0 2.5 5.0 10.0 20 40 40 100 140
Aged 1 0.070 0.100 0.100 0.25 0.70 0.2 0.4 1.6 2.5 4 10 13 25 40
Soft 1 0.060 0.200 0.540 1.00 1.30 1.0 2.5 5.0 10.0 20 40 40 100 140
Aged 2 0.060 0.200 0.200 0.50 0.80 0.2 0.4 1.6 2.5 4 18 18 25 40
Soft 2 0.060 0.200 0.500 1.00 1.30 1.0 2.5 5.0 10.0 20 40 40 100 140
Aged 3 0.060 0.200 0.200 0.45 0.86 0.2 0.4 1.2 2.5 4 10 13 25 40
Soft 3 0.016 0.025 0.090 0.21 1.20 1.8 2.3 4.5 5.6 9 40 40 100 140
Aged 4 0.010 0.050 0.150 0.20 0.40 0.2 0.4 1.6 2.5 4 10 13 25 40
Soft 4 0.060 0.200 0.540 1.00 1.30 1.0 2.5 5.0 10.0 20 40 40 100 140
Avg Aged 0.05 0.14 0.16 0.35 0.69 0.20 0.40 1.50 2.50 3.88 12 14 25 40
Avg Soft 0.05 0.16 0.42 0.80 1.28 1.20 2.45 4.88 8.90 17.23 40 40 100 140
PG 58−28 0.150 0.300 0.300 0.40 1.00 0.4 0.8 4.0 8.0 20 25 40 100 250
Aged 1 0.068 0.068 0.200 0.25 0.34 0.2 0.5 0.5 3.0 6 18 18 50 100
Soft 1 0.150 0.300 0.300 0.40 1.00 0.4 0.8 4.0 8.0 13 25 40 100 250
Aged 2 0.150 0.200 0.200 0.25 0.34 0.2 0.5 0.5 3.0 6 18 18 50 100
Soft 2 0.150 0.300 0.300 0.40 1.00 0.4 0.8 4.0 8.0 13 25 40 100 250
Aged 3 0.150 0.200 0.200 0.25 0.34 0.2 0.5 1.5 3.0 6 18 18 50 100
Soft 3 0.150 0.300 0.300 0.40 1.00 0.4 0.8 4.0 8.0 13 25 40 100 250
Aged 4 0.068 0.068 0.200 0.25 0.34 0.2 0.5 1.5 3.0 6 18 18 50 100
Soft 4 0.150 0.200 0.200 0.40 1.00 0.6 1.0 1.6 4.0 6 25 40 100 250
Avg Aged 0.11 0.13 0.20 0.25 0.34 0.20 0.50 1.00 3.00 6.00 18 18 50 100
Avg Soft 0.15 0.28 0.28 0.40 1.00 0.45 0.85 3.40 7.00 11.25 25 40 100 250
PG 64−22 + Oil 0.060 0.100 0.180 0.30 0.63 0.8 1.2 3.0 5.0 10 25 40 80 200
Aged 1 0.005 0.009 0.050 0.09 0.15 0.3 0.6 0.6 2.0 4 6 6 20 50
Soft 1 0.060 0.100 0.180 0.30 0.63 0.8 1.2 3.0 5.0 10 25 40 80 200
Aged 2 0.015 0.030 0.060 0.09 0.15 0.3 0.6 0.6 2.0 4 6 6 20 50
Soft 2 0.060 0.100 0.180 0.30 0.63 0.8 1.2 3.0 5.0 10 25 40 80 200
Aged 3 0.015 0.030 0.060 0.09 0.15 0.3 0.6 0.6 2.0 4 6 6 20 50
Soft 3 0.060 0.100 0.180 0.30 0.63 0.8 1.2 3.0 5.0 10 25 40 80 200
Aged 4 0.015 0.030 0.060 0.09 0.15 0.3 0.6 0.6 2.0 4 6 6 20 50
Soft 4 0.060 0.100 0.180 0.30 0.63 0.8 1.2 3.0 5.0 10 25 40 80 200
Avg Aged 0.01 0.02 0.06 0.09 0.15 0.25 0.60 0.60 2.00 4.00 6 6 20 50
Avg Soft 0.06 0.10 0.18 0.30 0.63 0.80 1.20 3.00 5.00 10.00 25 40 80 200
8 mm Parallel Plates (PP)
25 mm PP
40 mm PP
© Canadian Technical Asphalt Association 2017