Evaluation of the use of reclaimed asphalt pavement on stone matrix asphalt mixtures [Aug. 2007]

EVALUATION OF THE USE OF RECLAIMED ASPHALT PAVEMENT IN STONE MATRIX ASPHALT MIXTURES
By Donald E. Watson, Research Engineer Adriana Vargas-Nordcbeck, Graduate Student Jason R. Moore, Laboratory Engineer National Center for Asphalt Technology
277 Technology Parkway Auburn University, Auburn, Alabama 36830
Phone: (334) 844-6228 Fax: (334) 844-6248
And Peter Wu, Ph.D., P.E., Assistant State Materials and Research Engineer
David Jared, P.E., Special Research Engineer Georgia Department of Transportation 15 Kennedy Drive Forest Park, Georgia 30297-2534 Phone: (404) 363-7500 Fax: (404) 363-7684
August, 2007

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, D. Jared, and P. Wu

TECHNICAL REPORT STANDARD TITLE PAGE

1.Report No.: FHWA-GA-07-2037

2. Government Accession No.:

3. Recipient's Catalog No.:

4. Title and Subtitle: Evaluation of the Use of Reclaimed Asphalt
Pavement in Stone Matrix Asphalt Mixtures

5. Report Date: August 2007
6. Performing Organization Code:

7. Author(s): Donald E. Watson, P.E.
9. Performing Organization Name and Address: National Center for Asphalt Technology Auburn University 277 Technology Parkway Auburn, AL 36830
12. Sponsoring Agency Name and Address: Georgia Department of Transportation Office of Materials & Research 15 Kennedy Drive Forest Park, GA 30297-2534

8. Performing Organ. Report No.: 2037
10. Work Unit No.:
11. Contract or Grant No.: SPR00-0006-00(357)
13. Type of Report and Period Covered: Final; August 2004-August 2007
14. Sponsoring Agency Code:

15. Supplementary Notes: Prepared in cooperation with the U.S. Department of Transportation, Federal Highway Administration.
16. Abstract: The objectives of this study were to evaluate the effect of various RAP types and proportions on combined
material and performance properties of SMA mixtures. Some of the pertinent conclusions from this study are: 1. The addition of RAP may be beneficial for resistance to moisture damage, and adversely affects only the fatigue performance of the mixtures, especially at high strain levels. 2. Adding RAP up to 30% had little effect on the low temperature PG properties. 3. It is recommended that GDOT specifications be modified to allow up to 20% RAP in SMA mixtures with no change in virgin binder grade. Mixtures will still need to meet the same gradation, volumetric, and performance criteria as virgin mixtures. 4. RAP proportions higher than 20% may be allowed, but the virgin binder grade may need to be reduced to improve fatigue performance properties.

17. Key Words:

18. Distribution Statement:

Stone Matrix Asphalt, reclaimed asphalt pavement,

tensile strength, fatigue, rutting, creep compliance

19. Security Classification (of this report): Unclassified

20. Security Classification
(of this page): Unclassified

21. Number of Pages: 65

22. Price:

Form DOT 1700.7 (8-69) i

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, D. Jared, and P. Wu DISCLAIMER
The contents of this report reflect the views of the authors who are responsible for the facts and accuracy of the data presented herein. The contents do not necessarily reflect the official views or policies of the Georgia Department of Transportation, the Federal Highway Administration or the National Center for Asphalt Technology. This report does not constitute a standard, specification, or regulation.
ii

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared
TABLE OF CONTENTS
EXECUTIVE SUMMARY ............................................................................................ vi
INTRODUCTION............................................................................................................. 1
BACKGROUND AND PROBLEM STATEMENT..................................................... 1 OBJECTIVES ............................................................................................................... 1 SCOPE .......................................................................................................................... 1
LITERATURE REVIEW ................................................................................................ 2
RESEARCH TEST PLAN ............................................................................................... 7
EVALUATION OF MATERIALS ............................................................................... 8 Aggregate Properties ............................................................................................. 8 Asphalt Binder Properties.................................................................................... 10
MIX DESIGNS........................................................................................................... 13 PERFORMANCE TESTS .......................................................................................... 14
Moisture Susceptibility.......................................................................................... 14 Rutting Susceptibility ............................................................................................ 15 Creep Compliance ................................................................................................ 15 Flexural Beam Fatigue ......................................................................................... 18
TEST RESULTS AND ANALYSIS .............................................................................. 19
MATERIAL PROPERTIES ....................................................................................... 19 AGGREGATES .......................................................................................................... 19 ASPHALT BINDER .................................................................................................... 20 DSR RESULTS.................................................................................................. 23 BBR RESULTS.................................................................................................. 23
MIX DESIGNS ......................................................................................................... 264 PERFORMANCE TESTS .......................................................................................... 26
Moisture Susceptibility.......................................................................................... 27 Rutting Susceptibility ............................................................................................ 29 Indirect Tensile Creep Compliance ...................................................................... 29 Flexural Beam Fatigue ......................................................................................... 31
High Strain Results........................................................................................ 31 Low Strain Results......................................................................................... 35 High and Low Strain Comparison................................................................. 38 Summary ............................................................................................................... 38
CONCLUSIONS AND RECOMMENDATIONS........................................................ 42
ACKNOWLEDGEMENTS ........................................................................................... 43
REFERENCES................................................................................................................ 43
APPENDIX A LABORATORY MIX DESIGNS...................................................... 45
APPENDIX B INDIVIDUAL TEST RESULTS ....................................................... 60
iii

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared
LIST OF TABLES
TABLE 1 Rutting of Fine SMA vs. Standard Hot Mix: I-85 Test Section . ...................... 2 TABLE 2 Procedures for the Design of Mixtures Containing RAP................................... 6 TABLE 3 Test Matrix For Mix Variables........................................................................... 8 TABLE 4 Properties of Virgin Aggregates. ....................................................................... 8 TABLE 5 RAP Gradations and Asphalt Content ............................................................... 9 TABLE 6 Gradations of Recycled SMA Mix Using DG1 RAP. ..................................... 14 TABLE 7 Aggregate Properties for Combined Blends. ................................................... 20 TABLE 8 Analysis of Variance for Aggregate Properties............................................... 20 TABLE 9 Aggregate Properties for RAP Material. ......................................................... 20 TABLE 10 Critical Temperatures and Performance Grades of Virgin and Recovered
RAP Binders. .................................................................................................. 21 TABLE 11A Measured Binder Properties of Fractionated RAP Blends. ......................... 22 TABLE 11B Measured Binder Properties of DG1 and DG2 RAP Blends. ...................... 22 TABLE 12 Performance Grades of RAP Blends. ............................................................ 23 TABLE 13 Volumetric Properties of RAP Mixtures. ...................................................... 24 TABLE 14 Ratios of Virgin and RAP Binder Contents for SMA Mixes. ....................... 25 TABLE 15 Savings in Virgin Binder (%) for RAP Mixtures. ......................................... 26 TABLE 16 Tensile Strengths for SMA Mixtures............................................................. 27 TABLE 17 Analysis of Variance for Tensile Strengths. .................................................. 28 TABLE 18 Analysis of Variance for Rut Depths............................................................. 29 TABLE 19 Rutting Susceptibility Results for RAP Mixtures.......................................... 29 TABLE 20 m-values for SMA Mixes. ............................................................................. 31 TABLE 21 Analysis of Variance for m-value.................................................................. 31 TABLE 22 Test Results for High Strain Beams (800 )................................................ 32 TABLE 23 Nf Comparisons for RAP Content Aggregate Source Interaction (800). 34 TABLE 24 Initial Dissipated Energy Comparisons for RAP Contents (800).............. 34 TABLE 25 Average Test Results for Low Strain Beams (400).................................... 35 TABLE 26 Nf and Percent Drop Comparisons for RAP Contents (400 ). ................... 36 TABLE 27 Initial Stiffness Comparisons for RAP Content Aggregate Source
Interaction (400 ). ...................................................................................... 37 TABLE 28 Fatigue Life Comparisons for Strain Levels.................................................. 38
iv

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared
LIST OF FIGURES FIGURE 1 Distribution of Rut Measurements on SMA Pavements. ............................... 3 FIGURE 2 Reduced Reflective Crackingin SMA vs Conventional Mix When Used
as a Concrete Overlay. ................................................................................... 4 FIGURE 3 F/E Particle Content Versus Change In Percent Passing 4.75 Mm Sieve . .... 5 FIGURE 4 Rolling Thin Film Oven . ............................................................................. 11 FIGURE 5 Basics of Dynamic Shear Rheometer ........................................................... 12 FIGURE 6 Indirect Tension Test Creep Compliance Curves......................................... 17 FIGURE 7 Prony Series Fit to Master Creep Compliance Curve. ................................. 17 FIGURE 8 Determination of m, the Slope of the Log Creep Compliance Curve. ......... 18 FIGURE 9 Creep Stiffness Trends for RAP Blends....................................................... 24 FIGURE 10 Ratio of RAP AC to New AC vs RAP Content............................................ 26 FIGURE 11 Effect of RAP Percentage on Tensile Strength. ........................................... 28 FIGURE 12 Creep Compliance Master Curves for RAP Mixtures.................................. 30 FIGURE 13 Number of Cycles to Failure for RAP Mixtures (800). ............................ 32 FIGURE 14 Effect of Aggregate Source on Nf (800). .................................................. 34 FIGURE 15 Relationship between Initial Stiffness and Number of Cycles to Failure
(800 )........................................................................................................ 33 FIGURE 16 Relationship between Drop in Initial Stiffness at 1,000,000 Cycles and
Estimated Nf (400 ). ................................................................................. 36 FIGURE 17 Relationship between Initial Stiffness and Nf (400 ). ............................... 37 FIGURE 18 Number of Cycles to Failure for High and Low Strain Levels..................... 38 FIGURE 19 Effect of RAP Binder on TSR. ................................................................... 369 FIGURE 20 Effect of RAP Binder on Rutting Performance. ........................................... 40 FIGURE 21 Effect of RAP on Low Temperature Properties. .......................................... 40 FIGURE 22 Effect of RAP Binder on Fatigue Life.......................................................... 41
v

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared
EXECUTIVE SUMMARY
The use of reclaimed asphalt pavement (RAP) in asphalt mixes has been encouraged for many years due to economic and environmental considerations. Studies have shown that mixtures with RAP can be expected to perform as well, or better, than conventional virgin aggregate mixes. Benefits of using recycled HMA include lower costs, reduced waste and conservation of natural resources.
When SMA technology was first introduced in the United States in 1990, there was no experience with the use of RAP in this specialty mixture. Due to special requirements for SMA mixes (such as more cubical aggregate, use of polymer-modified asphalt, and fiber stabilizers) the effect of RAP was uncertain, and therefore, its use in SMA mixtures has generally not been allowed. Based on the success obtained with the incorporation of RAP in conventional mixtures, the use of RAP in SMA mixtures needed to be evaluated. This research evaluated the effect of RAP on aggregate, asphalt binder and combined mixture properties.
To accomplish the objectives of this study, a research plan was developed involving extensive laboratory testing. For each blend, performance tests were conducted to evaluate the effect of different RAP proportions and different RAP types with various aggregate sources. A standard mixture of 100% virgin aggregate was used as a baseline for the study and for comparisons of mix performance.
The following general conclusions and recommendations are made based on test results from this research:
Tests on the aggregate properties of the combined blends indicated that addition of RAP generally improved Los Angeles abrasion and flat and elongated (F/E) particle content.
RAP content influenced only the tensile strength and fatigue life (Nf) of the mixtures. The addition of RAP may be beneficial for resistance to moisture damage, and adversely affects only the fatigue performance of the mixtures, especially at high strain levels.
Adding RAP up to 30% had little effect on the low temperature PG properties. It is recommended that GDOT specifications be modified to allow up to 20%
RAP in SMA mixtures with no change in virgin binder grade. Mixtures will still need to meet the same gradation, volumetric, and performance criteria as virgin mixtures. RAP proportions higher than 20% may be allowed, but the virgin binder grade may need to be reduced to improve fatigue performance properties. To maintain a similar dosage rate of polymer modification, it would be necessary to reduce both the high and low binder grades. (For example, reduce the grade from PG 76-22 to PG 70-28.)
Keywords: Stone Matrix Asphalt, reclaimed asphalt pavement, tensile strength, fatigue, rutting, creep compliance.
vi

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, D. Jared, and P. Wu
INTRODUCTION
Background and Problem Statement
The use of reclaimed asphalt pavement (RAP) in asphalt mixes has been encouraged for many years due to economic and environmental considerations. Studies have shown that mixtures with RAP can be expected to perform as well, or better, than conventional virgin aggregate mixes. As a result, asphalt pavement is the most recycled product in the United States, both in terms of tonnage (73 million tons, more than any other material) and in terms of percentage (80 percent of reclaimed asphalt pavement is recycled, a higher percentage than any other substance) (1). Georgia Department of Transportation (GDOT) began using RAP in experimental test projects in the mid 1980s and implemented the use of RAP on a statewide basis in January, 1990.
RAP may be defined as used HMA pavement that has been milled up or crushed. It can be used as a constituent in new mixtures with characteristics similar to those of virgin HMA mixtures. Benefits of using recycled HMA include lower costs, reduced waste and conservation of natural resources.
Although RAP has been successfully incorporated in HMA applications, its use by many agencies has not been extended to the production of open-graded friction courses and stone matrix asphalt (SMA) mixtures. When SMA technology was first introduced in the United States in 1990, there was no experience with the use of RAP in this specialty mixture. Due to special requirements for SMA mixes (such as more cubical aggregate, use of polymer-modified asphalt, and fiber stabilizers) the effect of RAP was uncertain, and therefore, its use in SMA mixtures has generally not been allowed.
Based on the success obtained with the incorporation of RAP in conventional mixtures, the use of RAP in SMA mixtures needed to be evaluated. This research evaluated the effect of RAP on aggregate, asphalt binder and combined mixture properties.
OBJECTIVES
The objectives of this study were to: 1) Evaluate the effect of various RAP types and proportions on combined aggregate
properties such as toughness/abrasion, and flat and elongated particles 2) Evaluate the effect of RAP on asphalt binder properties 3) Determine the feasibility of using SMA mixtures as future RAP sources 4) Evaluate the performance of SMA mixtures containing fractionated RAP and the
potential economical benefits of using this type of material 5) Evaluate the effect of various RAP sources of different gradation, asphalt content and
aggregate properties on the overall performance of SMA mixtures.
SCOPE
To accomplish the objectives of this study, a literature search and review of the information pertaining to the design of SMA mixtures and mixtures containing RAP and their performance was made. Secondly, a research plan was developed involving
1

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

extensive laboratory testing, which included performing mix designs for different aggregate sources, RAP types and RAP proportions. For each blend, combined aggregate properties and resulting asphalt binder properties were determined. A mix design for each blend was conducted to determine optimum asphalt content and volumetric properties. Performance tests were conducted to evaluate the mixtures at different RAP levels. A standard mixture of 100% virgin aggregate was used as a baseline for the study and for comparisons of mix performance.

LITERATURE REVIEW

GDOT placed its first SMA test sections in 1991 on a high traffic volume section of Interstate 85 northeast of Atlanta. At that time, this was the heaviest trucking corridor of the state and had approximately two million Equivalent Single Axle Loads (ESALs) per year. Following the construction of the I-85 test section, rutting measurements were conducted over time to monitor rutting in the SMA and conventional mixes. The results indicated that SMA mixes exhibited significantly less rutting than conventional mixes (Table 1).

Table 1 Rutting of Fine SMA vs. Standard Hot Mix: I-85 Test Section (2)

Year 1993 1994 1995

SMA (mm) 0 2.3 2.5

Standard (mm) 3.0 5.3 6.8

GDOT also conducted a special joint research study with Georgia Tech to learn more about methods of enhancing SMA performance. Findings from this study showed that GDOT SMA mixes undergo up to 40% less rutting than a typical GDOT densegraded surface mix. In addition, it was determined based on laboratory testing that SMA mixes typically have a fatigue life 3 to 5 times that of a conventional surface mix.
The study also indicated that by relaxing the aggregate quality requirements for SMA mixes, important production cost savings could be realized without significantly reducing the performance of the mixes. Based on this research, GDOT implemented a specification requirement that allowed aggregates which have no more than 45% abrasion loss and which have no more than 20% flat and elongated particles when measured at the 3:1 ratio of length to average thickness.
Based on the combination of GDOT and European experience, SMA has proven to have the following intrinsic benefits:
30-40% less rutting than standard mixes 3 to 5 times greater fatigue life in laboratory experiments 30-40% longer service life Lower annualized cost

In 1996, NCAT conducted a review of mix design and performance data from 86 SMA projects involving a total of 140 test sections in 19 different states (3). The study
2

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared
indicated that over 90% of the projects had rutting measurements less than 4 mm, and approximately 25% of the projects had no measurable rutting (Figure 1).
Figure 1 Distribution of rut measurements on SMA pavements (3).
Cracking (thermal and reflective) did not represent a significant problem. SMA mixtures appeared to be more resistant to cracking than dense mixtures, most likely due to the relatively high asphalt content and its resulting high film thickness. There was no evidence of raveling, and the biggest performance problem was the occurrence of fat spots (liquid asphalt flushed to the surface), which is caused by segregation, draindown, high asphalt content or improper type or amount of stabilizer.
The study concluded that SMA mixtures provided good performance in high traffic volume areas and that the increased benefits should compensate for the extra cost of construction.
In 2001, a second study (4) was conducted to evaluate long-term performance on some of the same SMA projects studied in 1996. The survey found that SMA mixtures had given exceptional rut-resistant performance, even when placed on high-traffic volume routes. Only one out of the 11 projects visited exhibited rutting in excess of 6 mm.
Only one project had significant block-type cracking, believed to be caused by the stiff plastic-modified binder. Comparisons between SMA mixes and conventional sections indicated that SMA mixtures may significantly reduce the rate of crack propagation when used as an overlay for concrete pavements.
The fat spots, noted as the major performance problem in the original study (2) had been worn off by traffic over time and were not noticeable during the 2001 review. In general, several projects were still in excellent condition after 9 years of service and based on an overall project condition ratings, SMA mixes could be expected to last up to 25% longer than conventional mixes.
An I-43 project in Wisconsin compared SMA with conventional dense-graded mix as a concrete overlay treatment. The report (4) shows there was moderate transverse reflective cracking in the conventional mix placed on the inside lane while reflective cracking in the SMA mix placed in the outside lane is much less frequent (Figure 2).
3

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared
FIGURE 2 Reduced reflective cracking in SMA vs. conventional mix when used as a concrete overlay.
In 1997, the National Center for Asphalt Technology (NCAT) developed a mixture design procedure for SMA and evaluated material and mixture criteria as part of a National Cooperative Highway Research Project, NCHRP 9-8 (5). Data were collected from a laboratory study conducted with various types of aggregates, fillers, asphalt binders and stabilizing additives. Parameters evaluated included aggregate toughness, flat and elongated particles, aggregate gradation, volumetric mix properties, asphalt binder content, compactive effort and asphalt binder draindown. Results indicated that there was a good correlation between aggregate breakdown and aggregate toughness as measured by the Los Angeles abrasion test for both Marshall (R2 = 0.62) and Superpave Gyratory Compactor (SGC) samples (R2 = 0.84).
To evaluate the effect of flat and elongated particles, mixtures were prepared with 0%, 25%, 50%, 75% and 100% flat and elongated aggregate. Samples were compacted with 50 blows of the Marshall hammer and aggregate breakdown was measured. Figure 3 shows that increased F/E particle content increases aggregate breakdown (R2 =0.89). Increased aggregate breakdown resulted in lower Voids in Mineral Aggregate (VMA). High Los Angeles abrasion values (40% or higher) make meeting the VMA requirements and ensuring a reasonable high asphalt content more difficult. To ensure the formation of stone-on-stone contact, the percent passing the No. 4 (4.75 mm) sieve should be kept below 30 percent.
4

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared
Figure 3 F/E particle content versus change in percent passing 4.75 mm sieve (5). The presence of an adequate aggregate skeleton can be verified by measuring the
voids in the coarse aggregate (VCA) of the mix. As the percent passing the 4.75 mm sieve decreases, the VCA of the mix also decreases.
The design air void range should be kept between 3 and 4 percent. To minimize fat spots and rutting, the air voids in warmer climates should be designed closer to 4 percent. Also, use of polymer modified asphalt produced better rut resistant mixes, while fiber stabilizers were superior in preventing draindown. A combination of stabilizers may provide the best properties in SMA mixes.
During NCHRP Project 9-12, Incorporation of Reclaimed Asphalt Pavement in the Superpave System (6), it was found that even though there was no significant difference at low RAP contents, RAP does not act like a black rock, and blending of the old and new binders occurs to a significant extent. At intermediate RAP contents, these effects can be compensated for by using a virgin binder that is one grade softer on both the high- and low- temperature grades. The tests indicated that high RAP contents increase the mixture stiffness, and therefore, a softer virgin binder must be used to improve the fatigue and low-temperature cracking resistance of the mixture.
In contrast, a study by Huang et al., indicated that only a small proportion of the aged binder would be available to blend with the virgin binder. Results showed that the influence of RAP on the virgin binder was very limited. Only a small portion of RAP asphalt participated in the remixing process while other portions formed a stiff coating around RAP aggregates, and RAP acted as a "black rock" (7).
Kandhal and Foo (8) developed a procedure for selecting the performance grade (PG) of virgin asphalt binder in a recycled HMA mixture based on the Superpave PG
5

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

grading system. Blending charts were constructed and evaluated based on test parameters obtained from the dynamic shear rheometer (DSR); therefore, only high and intermediate test temperatures were considered. As a result of the study, a three-tier system of selecting the PG grade of the virgin asphalt binder was recommended for recycled mixes:
Tier 1: If the amount of RAP in the HMA mix is equal to or less than 15%, the selected PG grade of the virgin asphalt binder should be the same as the Superpave specified PG grade.
Tier 2: If the amount of RAP in the HMA mix is more than 15% but equal to or less than 25%, the selected PG grade of the virgin asphalt binder should be one grade below (both high and low temperature grade) the Superpave specified PG grade. The use of a specific grade blending chart to select the high temperature grade of the virgin asphalt binder is optional.
Tier 3: If the amount of RAP in the HMA mix is more than 25%, use the specific grade blending chart to select the high temperature grade of the virgin asphalt binder. The low temperature grade should be at least one grade lower than the binder grade specified by Superpave.

Guidelines developed by the FHWA Superpave Mixtures Expert Task Group (9) provided a similar three tiered approach. Tier 1 does not require any modification of the mix design process, and the selection of the grade of virgin asphalt binder is based on typical requirements for climatic conditions and predicted traffic. Determination of asphalt binder content in RAP is left to the discretion of the agency. Tier 2 requires the PG grade to be reduced by one grade, or use of a blending chart. For Tier 3, the grade of virgin asphalt binder is either set to one grade lower than that usually selected for given climatic conditions, or selected from a blending chart. Table 2 summarizes the tests required on the RAP and selection of asphalt binder grade.

Table 2 Procedures for the Design of Mixtures Containing RAP (9)

Determine Tier RAP AC

Content

1

(a)

Measure RAP
Gradation
x

Measure RAP AC Stiffness
no

Measure Agg Blend
Properties
x

PG Grade Change
none

2

x

x

3

x

x

(a) At the discretion of the agency (b) Unless blending chart is used (c) Or use blending chart

no (b) yes

x

one grade lower (c)

x

use blending chart

A Minnesota DOT study in 2004 (10) investigated the effect of various types and percentages of RAP on asphalt binder and asphalt mixture properties. Ten mixtures were prepared using two asphalt binders (PG58-28 and PG58-34) and two RAP sources. In addition to the control mixtures, asphalt mixtures were prepared with 20% and 40% of
6

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared
each of the RAP sources. The dynamic modulus, stiffness, and moisture susceptibility results were used to determine the effect of RAP on the asphalt mixture properties.
It was observed that asphalt binder grade and RAP source had a significant effect on mixture stiffness. Moisture susceptibility test data indicated that as the percentage of RAP increased the strength also increased, while the tensile strength ratio decreased. Binder tests showed that the addition of RAP improved the binder grade in terms of high temperature performance, while the low temperature performance did not change significantly except for the case when 40% RAP was added. This means that the resulting binder blends would be more resistant to rutting and equally resistant to thermal cracking compared to virgin binders except at high RAP contents. The tests on the binders indicated that using 20% RAP in asphalt mixtures did not significantly affect the performance, but RAP amounts of 40% did have a significant effect on the performance of the mixtures.
In summary, research has generally found that at low RAP contents (up to 15%) the binder effects are negligible and no modification is required in the design process. At intermediate RAP contents (16% to 25%), these effects can be compensated for by using a virgin binder that is one grade softer on both the high- and low- temperature grades, and at higher RAP contents (over 25%) blending charts should be used to determine the appropriate virgin binder grade. It has also been found that addition of RAP increases the binder stiffness (6), and hence, the mixture stiffness. This may affect low temperature performance and fatigue life (11). On the other hand, increase in mix stiffness resulted in higher indirect tensile strength which improved rutting and moisture resistance.
The information collected suggests that use of RAP in SMA mixtures could produce important benefits in terms of performance. The effect of increased stiffness must be carefully studied, since SMA mixes could be especially vulnerable to distresses associated with this property, such as thermal and fatigue cracking.
RESEARCH TEST PLAN
The research approach was divided into three parts as they relate to the objectives of the study: evaluation of materials, mix designs and performance tests. This study involved evaluating material properties of aggregates, asphalt binder and the combined blend of virgin materials and RAP. The experiment was planned as a 4x4x4 factorial design, with three factors (aggregate source, RAP content and RAP type) at four levels each. This allowed studying the contributions that each of the experimental factors make in regard to performance, as well as the effect of the interaction of treatment factors.
The full factorial design would have required 64 treatment combinations, but due to time constraints and a need to keep research costs in a reasonable range, a one-fourth fraction experiment was selected so that the number of mix designs to be evaluated could be limited without sacrificing the integrity of the experiment. Table 3 shows the test matrix for the fractional factorial design.
7

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

Table 3 Test Matrix for Mix Variables

Aggregate

RAP

Source

Source

0

Regular

X

Mountain

SMA

View

Fine-graded

Coarse-graded

Regular

Lithia

SMA

Springs Fine-graded

X

Coarse-graded

Regular

Camak

SMA Fine-graded

Coarse-graded

X

Regular

Ruby

SMA

X

Fine-graded

Coarse-graded

RAP Content, %

10

20

30

X X
X X
X

X X
X X

X

X X

Evaluation Of Materials Aggregate Properties

Four Georgia aggregate sources were used in this study: Florida Rock at Mt. View (Forest Park), Martin-Marietta at Ruby (Macon), Martin-Marietta at Camak, and Vulcan Materials at Lithia Springs. These sources were chosen because they have been widely used in SMA production in Georgia with positive results. Their properties are shown in Table 4.

Table 4 Properties of Virgin Aggregates

Aggregate Source
Mt. View
Lithia Springs Camak
Ruby

Character of Material
Granite Gneiss/ Amphibolite Granite Gneiss Granite Gneiss Gneiss/ Amphibolite

Specific Gravities Absorption,

Bulk SSD App.

%

2.640 2.659 2.691

0.72

2.591 2.608 2.635

0.62

2.638 2.655 2.682

0.62

2.734 2.746 2.767

0.43

L.A. Abrasion,
% Loss
48
40
37
21

8

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

Four RAP types were also used in this study: conventional RAP, RAP from reclaimed SMA, fine-graded RAP [RAP passing the No. 4 (4.75 mm) sieve], and coarsegraded RAP [RAP retained on the No. 4 (4.75 mm) sieve]. The fine-graded RAP is hereafter referred to as -4 RAP, and the coarse-graded RAP is referred to as +4 RAP.
SMA RAP was included in this study to evaluate the possibility of recycling SMA material back into an SMA mixture. As the first SMA projects reach the end of their service life, it is important to determine if the stiff mastic in an SMA mix might prevent it from being recycled or if the proportion of RAP may have to be reduced. However, when the SMA RAP was tested, it was found that its gradation and asphalt content did not match those of an SMA mix. SMA mixes generally have about 25 percent of the material passing the No. 4 sieve, and in this case, that amount was 77 percent. However, this RAP material was crushed to achieve the 1/2 inch (12.5 mm) NMAS and this undoubtedly affected its gradation.
The asphalt content of the SMA RAP was unusually low with an average of only 4.4 percent based on weight of total mix (the typical asphalt content is about 6 percent). Based on this result, it is likely that the RAP from the SMA project also included a portion of the underlying 3/4 inch (19 mm) Superpave mixture as a result of the milling process. Overall, there was no significant difference between the gradations of conventional and SMA RAP, and from this point on in the report, they will be referred to as dense-graded RAP 1 (DG1 RAP) and dense-graded RAP 2 (DG2 RAP), respectively. Table 5 shows the gradations and asphalt cement content of each RAP source used in this study.

Table 5 RAP Gradations and Asphalt Content

Sieve Size
1" (25 mm) 3/4" (19 mm) 1/2" (12.5 mm) 3/8" (9.5 mm) #4 (4.75 mm) #8 (2.36 mm #16 (1.18 mm) #30 (0.600 mm) #50 (0.300 mm) #100 (0.150 mm) #200 (0.075 mm)
% AC

DG1 RAP 100.0 100.0 99.0 93.0 73.0 58.0 47.0 38.0 29.0 19.0 11.2 5.6

Percent Passing

DG2 RAP

-4 RAP

100.0

100.0

100.0

100.0

100.0

100.0

95.0

100.0

77.0

100.0

61.0

81.0

50.0

65.0

42.0

53.0

32.0

40.0

20.0

25.0

12.0

15.0

4.4

6.2

+4 RAP 100.0 99.0 96.0 84.0 37.0 25.0 21.0 18.0 15.0 10.0 6.2 4.5

The use of fractionated RAP material into coarse and fine-graded stockpiles was also considered in this study. One objective was to see if +4 RAP could be used as a substitute of a portion of the No. 7 stone typically used in high quantities in SMA production. This option would be beneficial in the event that quarries were faced with a supply shortage of No. 7 stone due to its high demand in other HMA and concrete mix applications. Also, -4 RAP was used as a substitute for a portion of the asphalt content,
9

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared
since its higher surface area makes it richer in asphalt cement. This would represent an advantage because asphalt cement is typically the most expensive component of a mixture.
Aggregate toughness was determined by the Los Angeles abrasion test (ASTM C131), which measures the resistance of coarse aggregates to degradation by abrasion and impact. The aggregate is placed in a metal drum along with a charge of steel balls, and the drum is rotated 500 times at a speed of 30 - 33 revolutions per minute (RPM). The inside of the drum is equipped with an angle iron which runs longitudinally. This causes the charge of aggregate and balls to fall with a heavy impact once during each revolution, breaking the aggregate particles into smaller particles. At the completion of the test, the aggregate is sieved over a No. 12 sieve and the amount which passes through the sieve, expressed as a percentage of the total charge, is the Los Angeles abrasion value designated as "percent loss".
Aggregates must be tough in order to prevent crushing and abrasive wear during manufacturing, placing and compacting of HMA. This aggregate property is especially critical in gap-graded mixtures such as SMA because excessive aggregate breakdown will fill void spaces within the mixture and thereby reduce the amount of asphalt cement that would otherwise be needed. As the asphalt content is reduced, the durability of the mixture suffers and results in premature aging and deterioration. Aggregate breakdown will also result in uncoated particle faces which may accelerate any susceptibility to moisture damage.
The flat and elongated property was determined by GDT-129. This characteristic is defined as the percentage by weight of coarse aggregates that have a length in excess of three times its average thickness. The procedure varies from ASTM D 4791 in that the average thickness rather than the maximum thickness is used as one of the measurement criteria. Use of the average thickness means that aggregate would have to be even more cubical in shape than for the ASTM test procedure; however, the average thickness is a subjective visual determination. The flat and elongated test requirements are used to ensure that the aggregate contains cubical particles capable of distributing traffic loads through the stone-on-stone coarse aggregate skeleton of an SMA mix. This also contributes to the improved rutting resistance of SMA mixes as compared to conventional mixtures.
Asphalt Binder Properties
The binder from RAP materials was recovered through Abson recovery tests (ASTM D1856) and its properties were evaluated by means of the Dynamic Shear Rheometer (DSR) and Bending Beam Rheometer (BBR). Asphalt cement from samples of the proposed blends was also extracted and analyzed for rheological properties and performance grade.
Short-term aging of the blended binders was achieved using the Rolling Thin Film Oven (RTFO) procedure according to AASHTO T240. This test procedure is used to simulate the degree of aging that occurs during construction. In this test, a sample of asphalt binder is heated in an oven for 85 minutes at 163 C. The moving film is created by placing the asphalt binder sample in a small jar then placing the jar in a circular metal
10

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared
carriage that rotates within the oven (Figure 4). This rotation is used to continually expose the binder film to hot air.
FIGURE 4 Rolling thin film oven (12). Long-term aging, which simulates several years of exposure to the environment, was achieved using the Pressure Aging Vessel (PAV) in accordance to AASHTO PP1. The PAV is an oven-pressure vessel combination that takes RTFO-aged samples and exposes them to high air pressure (2070 kPa) and temperature (90C, 100C or 110C, depending upon expected climatic conditions) for 20 hours. Engineering properties of the blended binders were obtained through Dynamic Shear Rheometer (DSR) and Bending Beam Rheometer (BBR) testing. The DSR test was used in accordance with AASHTO TP 5 to measure the complex shear modulus (G*) and phase angle () of the blended binders at high and intermediate temperatures. The test uses a thin asphalt binder sample placed between two plates (Figure 5). The lower plate is fixed while the upper plate oscillates back and forth across the sample at 1.59 Hz to create a shearing action. The oscillations at 1.59 Hz (10 radians/sec) are meant to simulate the shearing action corresponding to a traffic speed of about 90 km/hr (55 mph). The physical properties measured with the DSR allow obtaining the rutting and fatigue parameters, which are used to quantify the asphalt binder's contribution in resisting those types of distresses. Rutting is considered a stress controlled, cyclic loading phenomenon. Each traffic loading cycle does work that contributes toward deforming the HMA pavement surface. A part of this work is recovered by elastic rebound of the surface while some is dissipated in the form of permanent deformation and heat (13).
11

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

The work dissipated per loading cycle at a constant stress can be expressed as:

Wc

=





2 o



G



1 / sin



Equation 1

Where:

Wc = work dissipated per load cycle o = stress applied during the load cycle

G* = complex modulus

= phase angle

FIGURE 5 Basics of Dynamic Shear Rheometer (12).

The amount of work dissipated per loading cycle is inversely proportional to

G*/sin, called the rutting parameter. In order to minimize permanent deformation, Wc must be minimized as well. This indicates that higher values of G*/sin correspond to

binders with better rutting resistance

In the case of fatigue cracking, this distress is considered a strain controlled

phenomenon. The work dissipated per loading cycle at a constant strain can be expressed

as:

[ ] Wc = o2 G sin

Equation 2

where is the strain and the other variables are as previously described (13). Fatigue cracking is minimized by decreasing the term G*sin (fatigue parameter).
The BBR test was performed according to AASHTO TP 1 to determine the binder's propensity to thermal cracking. The BBR basically subjects a simple asphalt beam to a small (1,000 mN) load over 240 seconds. Then, using basic beam theory, the BBR software program calculates the flexural creep stiffness (S) and logarithmic creep rate (m) of the asphalt binder (13).

12

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

The creep stiffness of the asphalt binder beam at 60 seconds loading time is given

by:

S (t )

=

PL3
4bh3 (t)

Equation 3

Where: S(t) = creep stiffness at time, t = 60 seconds P = applied constant load, 100 g L = distance between beam supports, 102 mm b = beam width, 12.5 mm h = beam thickness, 6.25 mm (t) = deflection at time, t = 60 seconds

The m-value is the rate of change of the stiffness, S(t), with loading time and is used to describe how the asphalt binder relaxes under load.
Creep stiffness is related to thermal stresses in an HMA pavement due to shrinking while the m-value is related to the ability of an HMA pavement to relieve these stresses. Therefore, asphalt binders with minimum creep stiffness and maximum creep rate are desired in order to resist thermal cracking.
The Superpave asphalt binder specification (AASHTO MP1) is intended to control permanent deformation, low temperature cracking and fatigue cracking in asphalt pavements. The specification accomplishes this by controlling the various physical properties described previously (G*/sin, G*sin, S(t) and m-value). The physical properties remain constant for all performance grades (PG), but the temperatures at which these properties must be achieved vary depending on the climate in which the binder is expected to serve.

Mix Designs

RAP material, virgin asphalt and virgin aggregate were proportioned to produce 12.5 mm SMA mix designs. The 50-blow Marshall procedure, which is most often used by GDOT, was used for asphalt mixture compaction, and PG 76-22 was used as the standard performance grade asphalt. RAP was blended at four proportions (0%, 10%, 20%, and 30%) to determine the effect of RAP over the ranges of anticipated use.
A blend with no RAP was used as a baseline for the study for comparisons of mix performance. A RAP content of 10% represented the least amount that can feasibly be utilized, and a maximum RAP content of 30% was used because it is improbable that blends with greater contents of RAP would be able to meet gradation and volumetric requirements of the mix design. An example of typical gradations for the control and recycled mixes is shown in Table 6 using the DG1 RAP.
In the mix design, the gradations of the SMA mixtures were kept as close as possible to each other to provide a proper comparison among the different aggregate sources and RAP types. The gradations in Table 6 show that there is little variation in the percent passing each sieve size among the different mixtures. Fiber stabilizing additive (cellulose fiber) and an anti-stripping agent (hydrated lime) were included in the mixture as specified by GDOT.
13

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

TABLE 6 Gradations of Recycled SMA Mix Using DG1 RAP

Sieve Size 3/4" 1/2" 3/8" #4 #8 #16 #30 #50 #100 #200

0% RAP 100.0 90.0 65.4 24.4 20.5 16.9 14.6 12.4 10.6 8.6

Percent Passing

10% RAP 20% RAP

100.0

100.0

89.1

90.2

63.4

67.1

27.7

25.9

20.6

21.4

17.2

18.1

15.2

15.7

12.6

13.3

10.7

10.7

8.1

8.4

30% RAP 100.0 89.8 62.5 27.6 22.4 19.1 16.4 13.7 10.6 8.0

Once the blends for each aggregate source and RAP type were determined, replicate samples were prepared for each blend at three different asphalt contents and conditioned in accordance with AASHTO R30. Specimens were compacted using a Marshall hammer, following procedures in AASHTO T-245. The bulk specific gravity of each specimen was determined by AASHTO T 166. The theoretical maximum specific gravity of the loose HMA mix samples was measured in accordance with AASHTO T 209. Percent of air voids in the mix, voids in mineral aggregate (VMA) and voids filled with asphalt (VFA) were calculated for each mixture. As required by GDOT procedures, VMA was determined based on the calculated effective specific gravity of the aggregate blend.

PERFORMANCE TESTS

The performance of the bituminous mixtures was evaluated by subjecting specimens to diametral tensile strength, moisture susceptibility, flexural beam fatigue, rutting susceptibility using the Asphalt Pavement Analyzer (APA), and indirect tensile creep compliance tests.

Moisture Susceptibility

The effect of RAP addition on the moisture susceptibility of the mixtures was evaluated by determining the diametral tensile strength on dry and wet specimens according to GDT-66, Evaluating the Moisture Susceptibility of Bituminous Mixtures by Diametral Tensile Splitting. In this test, internal water pressures in the mixtures are produced by vacuum saturation followed by a freeze and a warm-water soaking cycle. The GDOT procedure differs from AASHTO T 283 in that a saturation range is not specified and the loading rate is 0.065 inches/minute. Six Marshall specimens were prepared using optimum asphalt content and compacted to 7.0 1.0 percent air voids. A subset of three specimens remained unconditioned and was used as the control group. The other subset was partially vacuum saturated with water for 30 minutes and then subjected to a freeze/thaw cycle where the samples are frozen at -
14

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

0.4F (-18C) for 15 hours and thawed at 140F (60C) for 24 hours. Both subsets were then cooled to a constant temperature of 55F (12.8C) and tested for indirect tensile strength at a load rate of 0.065 in/minute. The diametral tensile strength of each specimen was determined by Equation 4. The percentage of retained strength (TSR) was calculated by comparing the properties of dry specimens with water-conditioned specimens as in Equation 5.

S = 2P tD

Equation 4

Where:

S = tensile strength, psi (kPa) P = maximum load, pounds (N) t = specimen height immediately before tensile test, inches (millimeters) D = specimen diameter, inches (millimeters)

Where:

TSR

=



S conditioned S control



100

TSR = percent retained tensile strength Sconditioned = average tensile strength of conditioned subset, psi (kPa) Scontrol = average tensile strength of control subset, psi (kPa)

Equation 5

Rutting Susceptibility

Rutting susceptibility of the mixtures was tested with the Asphalt Pavement Analyzer (APA), according to GDT-115, Determining Rutting Susceptibility Using the Loaded Wheel Tester. The APA is a modification of the Georgia Loaded Wheel Tester (GLWT), and it follows a similar rut-testing procedure. A wheel is loaded onto a pressurized linear hose and tracked back and forth over a set of testing samples to induce rutting. Six samples for each mix type were compacted with a gyratory compactor to 5.0 1 percent air voids and tested at 64C using a vertical load of 100 lbs. and hose pressure of 100 psi for 8,000 cycles.

Creep Compliance

The creep compliance of the mixtures was evaluated according to AASHTO T 322, Determining the Creep Compliance and Strength of Hot-Mix Asphalt Using the Indirect Tensile Test Device, in order to determine if the addition of RAP affected the resistance of the mixtures to thermal cracking. Three replicate specimens for each mixture were compacted with a gyratory compactor to approximately 7% air voids and cut to dimensions of 150 mm diameter by 50 mm height. The tensile creep compliance was determined by applying a static compressive load of fixed magnitude along the diametral axis of each specimen for 100 s. Each specimen was tested at temperatures of -20, -10 and 0C. The horizontal and vertical deformations measured near the center of the
15

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

specimen were used to calculate the tensile creep compliance as a function of time, given

by the following relationship:

D(t) = (t)
0

Equation 6

Where,

D(t) is the creep compliance

(t) is the strain

0 is the stress.

Compliance is a way of characterizing the stiffness of a material. Another term

frequently used is creep stiffness, S(t), which is the inverse of creep compliance as

determined from a creep test:

S (t )

=

1
D(t )

=

0
(t)

Equation 7

An example of creep compliance curves measured at multiple temperatures using the indirect tensile test at low temperature is presented in Figure 6. A nonlinear regression routine is used to determine the master creep compliance curve from the creep compliance curves measured at multiple temperatures. The regression is performed in two steps. First, a regression is performed to simultaneously determine the temperature shift factors (at) and the parameters for the following Prony series (Maxwell model) representation of the master creep compliance curve (14):

( ) N
D( ) = D(0) + Di 1 - e /i
i =1

+ v

Where:

D () = creep compliance at reduced time

= reduced time (= t/at)

at = temperature shift factor

D(0), Di, I, v = Prony series parameters

Equation 8

16

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared
FIGURE 6 Indirect tension test creep compliance curves (14). In essence, the regression finds the best shift factors and Prony series parameters to fit the measured data based upon a least-squares criterion. One of the temperatures is selected as the reference temperature for the master curve (typically -20C), and thus the creep compliance curve at this temperature is fixed in time (at = 1). The regression determines the amount of time (horizontal) shift required for the curves at the remaining temperatures to result in a smooth master curve. Each of these remaining creep compliance curves will thus have a shift factor (at) associated with it. Figure 7 shows the shifted creep compliance data.
FIGURE 7 Prony Series fit to master creep compliance curve (14). The second step in the regression routine is to fit a second functional form to the master creep compliance information. This second functional form is the following power law:
17

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

D( ) = D(0) + D1 m

Equation 9

Where, D() and are as defined previously D(0), D1, and m are the coefficients of the functional form.

The primary purpose for fitting this functional form is to determine the parameter m. This parameter is essentially the slope of the linear portion of the master creep compliance curve on a log-log plot (Figure 8). It has been found to be an important parameter in distinguishing between the thermal cracking performance of different materials.

FIGURE 8 Determination of m, the slope of the log creep compliance curve (14). Flexural Beam Fatigue Fatigue tests were conducted according to AASHTO TP 8, Determining the Fatigue Life of Compacted Hot-Mix Asphalt Subjected to Repeated Flexural Bending, to evaluate the stiffening effect of RAP on the mixture and its impact on the long-term fatigue life of the pavement. Three replicate beams were compacted with a kneading compactor to 6.0 1.0 percent air voids and cut to dimensions of 380 mm long by 50 mm thick by 63 mm wide. The beams were placed in four-point loading and subjected to repeated haversine loads. The deflection caused by the load was measured at the center of the beam. The tests were performed under a constant-strain condition, at strain levels of 400 and 800 micro-strain () and at a temperature of 20C.
18

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

The variables measured include the number of cycles to failure, initial and final stiffness, and dissipated energy. The flexural stiffness is defined as:

S = 1000 t t
Where: S = flexural stiffness, MPa t = maximum tensile stress, kPa t = maximum tensile micro-strain

Equation 10

The initial stiffness is defined as the measured flexural stiffness after 50 cycles. The number of cycles to failure (Nf) is the load cycle at which the specimen exhibits a 50 percent reduction in stiffness relative to the initial stiffness. The dissipated energy is calculated by determining the area within the stress-strain hysteresis loop for each captured data pulse. The cumulative dissipated energy is the summation of the dissipated energy per cycle.

TEST RESULTS AND ANALYSIS
Material Properties
Aggregates
Properties of the virgin and recycled SMA mixes used in the study are shown in Table 7. These data were used to perform an analysis of variance (shown in Table 8), which indicated that combined blend properties such as LA abrasion and flat and elongated particle content are mainly influenced by the aggregate source (p-values < 0.001). At the 95% confidence level, RAP content and RAP type did not have a significant effect on LA percent loss or F/E particle content. As Table 9 shows, there is little variation in the aggregate properties of the RAP materials (3.0% difference for LA loss and 0.8% difference for F/E particle content), which is the reason why RAP type is not significant in this data.
The effect of RAP addition on aggregate properties depended on the quality of the virgin and recycled materials contained in the blend. The differences among aggregate sources were significant (Table 4), and they essentially determined the results for the combined blend. The differences produced by the increase in RAP content were very small as Table 7 shows and were not significant for these data.

19

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

TABLE 7 Aggregate Properties for Combined Blends

Aggregate Mt. View Lithia Springs
Camak Ruby

% RAP
0 10 20 30 0 10 20 30 0 10 20 30 0 10 20 30

RAP Type
DG1 -4
DG2 +4 -4 DG1 +4 DG2 +4 DG2 -4 DG1 DG2 +4 DG1 -4

LA Abrasion, % loss 48.0 48.0* 46.7 47.7 39.6 40.4 41.1 41.0 37.3 38.7 37.3* 40.3 21.1 23.8 26.4 21.1*

F/E particles, % (3:1 ratio)
8.6 4.0 5.4 4.0 11.3 11.8 13.1 17.5 9.3 7.9 9.3 12.9 6.4 4.7 5.7 6.4

*Same as virgin blend because coarse recycled material was not added.

TABLE 8 Analysis of Variance for Aggregate Properties

Factor
Agg. Source RAP Content
RAP Type

LA Abrasion

F-statistic

p-value

763.25

0.000

2.55

0.082

7.24

0.001

F/E Particles

F-statistic

p-value

25.05

0.000

2.89

0.058

1.85

0.167

TABLE 9 Aggregate Properties for RAP Material

RAP Type
DG1 DG2
-4 +4

LA Abrasion, % loss 47.2 44.2 N/A 47.2

F/E particles, % (3:1 ratio) 6.8 6.0 N/A 6.8

Asphalt Binder

The single type of virgin asphalt cement was PG 76-22 which is a polymer-modified asphalt. The results for the virgin and recovered RAP asphalts are presented in Table 10. The critical high temperatures for the extracted RAP binders were obtained by testing the recovered RAP binder as if it had been RTFO aged.

20

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

TABLE 10 Critical Temperatures and Performance Grades of Virgin and Recovered RAP Binders

Aging

Property

Virgin Binder

Recovered RAP Binders

+4

-4

DG1

DG2

Original RTFO RTFO+PAV
PG

G*/sin , kPa
G*/sin , kPa G* sin , kPa BBR S, MPa BBR m-value Actual
MP1

78.9
79.2 21.1 -27.2 -24.4 78.9-24.4
76-22

---
87.4 26.0 -27.9 -25.5 87.4-25.5
82-22

---
89.0 26.5 -28.1 -20.1 89.0-20.1
88-16

---
89.0 27.6 -30.2 -23.9 89.0-23.9
88-22

---
94.2 28.5 -25.1 -18.4 94.2-18.4
94-16

It can be observed from Tables 11A-11B that blends that contain DG2 RAP have higher values of G*/sin at a given RAP content for RFTO aged samples, which indicates better resistance of the resulting binder blends to rutting. Blends containing DG2 RAP also had lower values of G*sin than the corresponding blends containing other RAP binder types (maximum 3,847 kPa at passing temperature). This is indicative of a higher fatigue cracking resistance for these binder blends. Part of the reason for these good results may be because the DG2 RAP binder was a polymer-modified binder.
Finally, the properties obtained with the BBR test (creep stiffness and creep rate) had more favorable results for blends using DG2 RAP binder. These blends had lower stiffness (maximum 151 MPa) and higher creep rate ( 0.322) at passing temperatures than the corresponding blends containing other RAP binder types. Low creep stiffness values are desired in order to minimize thermal stresses, while the m-value must be high to maximize the ability of the HMA pavement to relieve those stresses; therefore, mixtures that contain DG2 RAP have a binder blend that is more resistant to both rutting and thermal cracking.
Table 12 shows the results for the performance grades of the binder blends. The actual grade as well as the general PG designation from AASHTO MP1 is shown. The addition of 10% RAP binder did not change the performance grade of the binder blends. Increasing the RAP binder content to 20% only affected the DG2 blend by raising the high-temperature by one grade. The low temperature performance grade remained the same as the virgin asphalt binder. Finally, the use of 30% RAP binder reduced the lowtemperature grade of the -4 blend by one grade, raised the high-temperature grade of the DG1 blend by one grade, and had no further effect on the DG2 and +4 blends.
The trends of binder properties obtained with the DSR and BBR tests in the range of 0% to 30% RAP binder in the blend were analyzed for the different mixtures in this study. The rates of change of the properties in that range were calculated to assess the impact of the RAP asphalt content in the blends. These rates were computed as the change in the binder property divided by the change in RAP content, for the entire range studied.

21

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, D. Jared, and P. Wu

TABLE 11A Measured Binder Properties of Fractionated RAP Blends

Aging Original

Property G*/sin , kPa

Required Value
1.00 kPa

RTFO

G*/sin , kPa 2.20 kPa

G* sin , kPa 5,000 kPa

RTFO+PAV BBR S, MPa 300 MPa

BBR m-value 0.300

Temp. C 76 82 76 82 25 22 -12 -18 -12 -18

10% 1.290 0.762 2.695 1.570 3,427 4,915 148 273 0.332 0.276

+4 RAP 20% 1.634 0.967 2.802 1.559 3,614 5,190 157 270 0.324 0.223

30% 2.521 1.354 3.312 1.968 4,413 6,233 167 264 0.304 0.266

10% 1.578 0.884 3.018 1.760 4,370 6,227 176 287 0.304 0.261

-4 RAP 20% 1.657 0.956 3.318 1.818 4,153 5,915 179 297 0.306 0.263

30% 1.849 1.036 3.964 2.087 4,690 6,520 182 297 0.291 0.266

TABLE 11B Measured Binder Properties of DG1 and DG2 RAP Blends

Aging Original

Property G*/sin , kPa

Required Value
1.00 kPa

RTFO

G*/sin , kPa 2.20 kPa

G* sin , kPa 5,000 kPa RTFO+PAV BBR S, MPa 300 MPa
BBR m-value 0.300

Temp. C 76 82 88 76 82 88 25 22 -12 -18
-12 -18

10% 1.518 0.877
3.032 1.736
3,688 5,297 164 308 0.313 0.326

DG1 RAP 20% 1.593 0.885
3.183 1.843
3,997 5,715 164 344 0.311 0.254

30%
1.112 0.603
2.243 1.217 4,149 5,854 168 345 0.304 0.271

10% 2.039 1.150
3.488 2.042
3,030 4,374 142 317 0.334 0.281

DG2 RAP 20% 1.965 1.101
4.038 2.274
3,416 4,876 127 299 0.329 0.271

30%
2.676 0.652 4.213 2.676
3,847 5,438 151 317 0.322 0.260

22

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, D. Jared, and P. Wu

Table 12 Performance Grades of RAP Blends

RAP Source
+4
-4
DG1
DG2

% RAP blend
0 10 20 30 0 10 20 30 0 10 20 30 0 10 20 30

Performance Grade

Actual

MP 1

PG 78.9-24.4 PG 76-22

PG 78.3-26.4 PG 76-22

PG 78.5-25.0 PG 76-22

PG 80.7-22.6 PG 76-22

PG 78.9-24-4 PG 76-22

PG 79.5-22.6 PG 76-22

PG 80.1-22.8 PG 76-22

PG 81.5-19.8 PG 76-16

PG 78.9-24.4 PG 76-22

PG 79.5-23.5 PG 76-22

PG 80.1-23.5 PG 76-22

PG 82.2-22.7 PG 82-22

PG 78.9-24.4 PG 76-22

PG 81.2-25.8 PG 76-22

PG 82.6-25.1 PG 82-22

PG 83.8-24.1 PG 82-22

DSR Results Results for the rutting and fatigue parameters were analyzed for original (unaged) blends, RTFO-aged and RTFO+PAV aged blends at failing and passing temperatures. Estimated and actual critical temperatures were compared for original and aged blends. As expected, the rutting parameter G*/sin in the original blends was higher at the low temperature and increased with the addition of RAP binder because the old binder makes the resulting blends stiffer.
The results for RFTO-aged blends were similar to those of unaged blends. Lower temperature and higher RAP binder percentages increased G*/sin, but not at a high rate. This suggests that even though the rutting parameter increases with the addition of RAP for both original and RTFO-aged blends, the rates of change are so small that the rutting resistance of the mixes is not likely to be significantly improved.
For the RFTO and PAV residues, the fatigue parameter G*sin increased with the addition of RAP binder. The results were considerably higher at lower temperatures. Unlike the trends for the unaged and RTFO-aged blends, these results indicate that the fatigue resistance of the binder blends is highly influenced by temperature and RAP content. Addition of RAP binder may result in mixes more susceptible to fatigue cracking, especially at high temperatures.

BBR Results Results for creep stiffness and creep rate obtained with the bending beam rheometer were analyzed for the asphalt blends at failing and passing temperatures. Figure 9 shows that the creep stiffness of the blends increased slightly at the lower test temperature for

23

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

Creep Stiffness (MPa)

Creep Stiffness (MPa)

incremental increases of RAP binder. The results suggest that addition of RAP is not highly influential for the creep stiffness of the binder blend at the temperatures studied.

300
250 -12 C
200 -18 C
150
100 0
350 300 250 200 150 100
0

+4 RAP

-4 RAP 300

Creep Stiffness (MPa)

250

200

150

10

20

30

% RAP Blend

DG1 RAP

100 0
350

10

20

% RAP Blend

DG2 RAP

Creep Stiffness (MPa)

-12 C

300

-18 C

250

200

150

10

20

30

% RAP Blend

100 0

10

20

% RAP Blend

FIGURE 9 Creep stiffness trends for RAP blends.

-12 C -18 C
30
-12 C -18 C
30

As expected, higher RAP binder percentages generally resulted in lower creep rates. As with creep stiffness, the small rate of change suggested that increasing RAP content does not decrease the creep rate significantly, and that the thermal cracking resistance of the recycled binder blends will not be affected.

Mix Designs

Table 13 shows the optimum total asphalt content (virgin binder plus RAP binder), voids in the mineral aggregate (VMA), and voids filled with asphalt (VFA) based on mix design results for each factor level combination tested. The mix design information for all mixtures is presented in Appendix A. The VMA values were calculated using the effective specific gravity (Gse) of the aggregate blends, as specified by GDOT. It can be observed that the results for VMA and VFA do not change significantly with the addition of RAP or with the types of RAP used.
One parameter that can be useful to evaluate the impact of RAP addition on mixture performance is the ratio of RAP binder to virgin binder, shown in Table 14. Figure 10 shows that as RAP content increases, the ratio of RAP binder to virgin binder increases as well. When grouped by RAP type, mixes that contained -4 RAP had the highest ratio (up to 0.45) due to the higher asphalt content (6.2%) present in -4 RAP.
24

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

TABLE 13 Volumetric Properties of RAP Mixtures

Aggregate Source Mt. View
Lithia Springs
Camak
Ruby

% RAP
0 10 20 30 0 10 20 30 0 10 20 30 0 10 20 30

RAP Type
+4 DG2
-4 DG1
-4 DG1 +4 DG2 DG1
-4 DG2 +4 DG2 +4 DG1
-4

Total AC, %
6.2 6.1 6.8 6.3 6.2 6.0 6.4 6.0 6.9 6.4 6.9 7.3 6.2 5.9 6.6 5.8

VMA, % 18.3 18.2 19.3 18.3 18.1 17.6 18.0 17.3 19.6 18.9 19.4 19.8 18.5 18.0 18.9 17.6

VFA, % 74.6 77.2 80.3 78.9 76.5 76.8 79.4 77.1 78.8 76.4 79.0 82.8 76.7 76.4 79.8 76.5

TABLE 14 Ratios of Virgin and RAP Binder Contents for SMA Mixes

RAP Type
+4
-4
DG1
DG2

% RAP
0 10 20 30 0 10 20 30 0 10 20 30 0 10 20 30

Agg. Source
Camak Ruby Lithia Springs Mt. View Lithia Springs Mt. View Camak Ruby Mt. View Lithia Springs Ruby Camak Ruby Camak Mt. View Lithia Springs

Virgin AC, % 6.9 5.5 5.5 5.0 6.2 5.5 5.7 4.0 6.2 5.5 5.5 5.7 6.2 6.0 6.0 4.7

RAP AC, % 0.0 0.4 0.9 1.3 0.0 0.6 1.2 1.8 0.0 0.5 1.1 1.6 0.0 0.4 0.8 1.3

RAP AC/ Virgin AC
0.00 0.07 0.16 0.26 0.00 0.11 0.21 0.45 0.00 0.09 0.20 0.28 0.00 0.07 0.13 0.28

25

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

0.50 0.40 0.30

y = 0.0099x R2 = 0.8652

RAP Binder/Virgin Binder

0.20

0.10

0.00 0

10

20

30

% RAP

FIGURE 10 Ratio of RAP AC to new AC vs. RAP content.

Table 15 illustrates the savings in virgin binder content for all RAP types. These savings were calculated as a percent of reduction on virgin binder compared to the control mix. Since there are four different mixes with 0% RAP (one for each aggregate source), these percentages were computed by matching the recycled SMA mix with the control mix that contained the same aggregate source. For example, the mix that contains 20% -4 RAP and Camak aggregates was compared to the control mix that contains Camak aggregates; while the mix that contains 10% -4 RAP and Mt. View aggregates was compared to the control mix that contains Mt. View aggregates, and so on.

TABLE 15 Savings in Virgin Binder (%) for RAP Mixtures

% RAP
10 20 30

DG1 11.3 11.3 17.4

RAP Type

DG2

-4

13.0

11.3

3.2

17.4

24.2

35.5

+4

Average

11.3

11.7

11.3

10.8

19.4

24.1

At 10% and 20% RAP, the reduction in the required virgin binder is very similar (averages of 11.7% and 10.8%, respectively). Normally, the required amount of virgin binder will decrease with RAP content, so any anomaly in these results can be attributed to variability in mix design. At 30% RAP the savings increase dramatically to an average of 24.1%, which would represent an important economical benefit, since asphalt cement is the most expensive component of an HMA mix. This benefit is particularly important for mixtures containing -4 RAP, where the virgin binder required is reduced up to 35.5%. Based on the reduction in virgin binder shown in Table 15, there is a potential to save $1 - $2 for every ton of hot mix produced if higher RAP percentages may be allowed. This
26

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

could amount to as much as $10 million per year since GDOT normally places about 5 million tons of HMA annually.
Performance Tests
Moisture Susceptibility

Table 16 shows the wet (conditioned) and dry (control) strength values for the SMA mixes with recycled material. An analysis of variance (shown in Table 17) indicates that at 95% confidence level, the amount of RAP significantly influences the tensile strength (p-values < 0.001), while type of RAP has a somewhat significant effect (p-value = 0.015 for dry strength and 0.019 for wet strength). This was expected because increasing the RAP content increases the amount of old binder (binder from RAP) and makes the mixture stiffer. The analysis of variance also indicates that the interaction between aggregate source and RAP type is significant for both conditioned and unconditioned tensile strengths (p-values < 0.001).

RAP Type +4 -4 DG1 DG2

TABLE 16 Tensile Strength of SMA Mixtures

% RAP
0 10
20
30
0
10 20 30 0
10
20 30 0 10 20
30

Aggregate Source
Camak Ruby Lithia Springs Mt. View Lithia Springs Mt. View Camak Ruby Mt. View Lithia Springs Ruby Camak Ruby Camak Mt. View Lithia Springs

Control Tensile Strength
(psi) 87.7 103.6
87.0
105.7
89.4
87.2 98.5 144.7 71.6
83.6
83.1 124.8 69.7 85.1 101.9
94.8

Conditioned Tensile
Strength (psi)
78.7 92.3
90.8
99.8
70.9
88.2 85.7 141.7 71.6
79.7
83.6 118.1 70.2 72.3 91.9
95.8

TSR, %
89.7 89.1 104.4
94.4 79.3
101.1 87.0 97.9 100.0 95.3
100.6 94.6 100.7 85.0 90.2 101.1

27

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

TABLE 17 Analysis of Variance for Tensile Strengths

Factor
Agg. Source RAP Content
RAP Type Agg. Source x
RAP Type

Unconditioned

F-statistic

p-value

2.24

0.099

18.21

0.000

3.96

0.015

16.33

0.000

Conditioned

F-statistic

p-value

3.18

0.035

33.61

0.000

3.73

0.019

33.90

0.000

All mixtures contained between 0.8 and 1.0 percent hydrated lime by total weight of aggregate as an anti-stripping agent. The rate of lime treatment varied proportionally with the amount of virgin aggregate to be treated. A minimum TSR of 0.80 is generally required by GDOT for SMA mixtures. However, a TSR of 0.7 may be acceptable so long as all individual test values exceed 100 psi. All mixtures met or exceeded the 0.8 minimum retained strength, except the virgin aggregate blend from Lithia Springs which only marginally failed.
Figure 11 shows that tensile strength increased as the percentage of RAP increased. This is not surprising since recycled mixtures contain RAP binder that would be expected to have an effect on the mixture stiffness. This can be attributed to the higher old to new binder ratio for mixtures with higher RAP contents.

Tensile Strength, psi

140 120

y = 1.1648x + 77.427 R2 = 0.8746

100 80

y = 1.2789x + 70.266 R2 = 0.8956

60 40 20
0 0

10

20

% RAP

Unconditioned Conditioned Unconditioned Conditioned
30

FIGURE 11 Effect of RAP percentage on tensile strength.

28

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

Rutting Susceptibility

An analysis of variance (Table 18) indicates that aggregate source, RAP type and the interaction between RAP proportion and RAP type were significant factors, while RAP proportion alone did not have an effect on rutting susceptibility (p-value = 0.720).
Rutting susceptibility test results using the APA are shown in Table 19. Table 19 shows that the rut depths for different RAP contents ranged from 1.7 mm to 5.4 mm. Only blends with Ruby aggregate failed to meet the maximum 5 mm criteria specified by GDOT. The fact that mixes with these RAP types only affected rutting performance at a particular RAP content can be attributed to test variability.

TABLE 18 Analysis of Variance for Rut Depths

Factor Agg. Source RAP Content RAP Type RAP Proportion x RAP Type

F-statistic 14.36 0.45 2.86 15.24

p-value 0.000 0.720 0.038 0.000

TABLE 19 Rutting Susceptibility Results for RAP Mixtures

Aggregate Source Mt. View
Lithia Springs
Camak
Ruby

RAP Type DG1
-4 DG2 +4
-4 DG1 +4 DG2 +4 DG2
-4 DG1 DG2 +4 DG1
-4

% RAP
0 10 20 30 0 10 20 30 0 10 20 30 0 10 20 30

Rut depth, mm 3.11 3.23 4.41 1.70 2.37 2.00 2.44 4.50 3.67 1.96 1.48 3.25 3.58 5.16 5.38 3.85

Indirect Tensile Creep Compliance

The creep compliance test is used to evaluate thermal cracking resistance of the mixtures. Figure 12 shows the creep compliance results at 50 seconds. It is clear that, as expected, creep compliance increases with temperature. However, there is not a clear relationship

29

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

between creep compliance and RAP content in these data. The addition of RAP does not clearly change the stiffness of the mix, as characterized by the IDT creep test, suggesting that the low temperature performance would not be as significantly affected by the higher RAP content as one might think.
The other parameter that measures thermal cracking resistance is the m-value. The creep compliance master curves from Figure 12 were used to calculate the m-value for each mix (Table 20). The m-value is obtained by fitting a power law through the master compliance curve obtained from the indirect tensile creep tests. Mixtures with higher mvalues tend to have greater resistance to thermal cracking.
Table 20 shows that the m-value is not influenced by RAP proportion. This was verified by an ANOVA which shows a p-value of 0.552 for the RAP proportion factor Table 21. However, the ANOVA shows that the m-value is significantly affected by the aggregate source and to some extent by RAP type. The influence of aggregate source is somewhat unexpected because thermal cracking resistance is not really defined by the aggregate properties, and is instead generally believed to be dictated by the binder properties. This may be attributed to the difference in mean old to new binder ratios among mixtures with different aggregate sources. The poor correlation between RAP proportion and m-value leads one to conclude that the addition of RAP would not affect thermal cracking potential significantly.

Creep Compliance, 1/psi

+4 RAP

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+00

Control 10% RAP 20% RAP 30% RAP
1.0E+02 1.0E+04

1.0E+06

1.0E+08

1.0E+10

Reduced Time, s

Regular RAP

1.0E+00 1.0E-01 1.0E-02 1.0E-03

Control 10% RAP 20% RAP 30% RAP

1.0E-04

1.0E-05

1.0E-06

1.0E-07 1.0E+00

1.0E+02

1.0E+04

1.0E+06

1.0E+08

1.0E+10

Reduced Time, s

Creep Compliance, 1/psi

Creep Compliance, 1/psi

-4 RAP

1.0E+00 1.0E-01 1.0E-02 1.0E-03

Control 10% RAP 20% RAP 30% RAP

1.0E-04

1.0E-05

1.0E-06

1.0E-07 1.0E+00

1.0E+02

1.0E+04

1.0E+06

1.0E+08

1.0E+10

Reduced Time, s

SMA RAP

1.0E+00 1.0E-01 1.0E-02 1.0E-03

Control 10% RAP 20% RAP 30% RAP

1.0E-04

1.0E-05

1.0E-06

1.0E-07 1.0E+00

1.0E+02

1.0E+04

1.0E+06

1.0E+08

1.0E+10

Reduced Time, s

Creep Compliance, 1/psi

FIGURE 12 Creep compliance master curves for RAP mixtures.

30

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

TABLE 20 m-values for SMA Mixes

RAP Type +4 -4 DG1 DG2

% RAP 0 10 20 30 0 10 20 30 0 10 20 30 0 10 20 30

Agg. Source Camak Ruby
Lithia Springs Mt. View
Lithia Springs Mt. View Camak Ruby Mt. View
Lithia Springs Ruby Camak Ruby Camak
Mt. View Lithia Springs

m-value 0.440 0.548 0.484 0.372 0.339 0.410 0.518 0.331 0.574 0.320 0.630 0.687 0.598 0.599 0.386 0.432

TABLE 21 Analysis of Variance for m-value

Factor Agg. Source RAP Proportion RAP Type

F-statistic 5.16 0.71 3.49

p-value 0.004 0.552 0.025

Flexural Beam Fatigue

For this procedure, beams were prepared and tested at two strain levels to simulate different pavement structures. The high strain level (800 ) simulates a thin pavement with weak structure or poor subgrade, while the low strain level (400 ) simulates a thicker pavement with adequate subgrade. Because low strain beams typically do not reach the termination stiffness in a reasonable amount of time, it was necessary to establish a cut off point of 1,000,000 cycles, which allowed the test to be completed in a maximum time of nearly 28 hours.

High Strain Results Table 22 shows the results for the high strain beams. An analysis of variance indicated that the number of cycles to failure is affected by the aggregate source and RAP content (p-values < 0.001) and the interaction of the two (p-value = 0.023).

31

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

TABLE 22 Test Results for High Strain Beams (800 )

Aggregate Source

% RAP

0

Mt. View

10 20

30

0

Lithia

10

Springs

20

30

0

Camak

10 20

30

0

Ruby

10 20

30

RAP Type
DG1 -4
DG2 +4 -4 DG1 +4 DG2 +4 DG2 -4 DG1 DG2 +4 DG1 -4

Cycles to Failure
45,800 50,143 31,013 19,880 58,703 44,877 57,940 16,753 92,070 40,947 71,070 74,760 72,680 40,933 21,123 4,273

Initial Stiffness (MPa)
2,952 3,169 4,325 4,315 3,090 3,411 3,179 4,949 3,220 3,221 3,380 3,433 3,028 3,141 3,348 4,798

Final Stiffness (MPa)
1,473 1,577 2,148 2,150 1,532 1,690 1,583 2,469 1,607 1,600 1,677 1,710 1,510 1,566 1,671 2,382

Initial Dissipated Energy (kPa)
0.985 1.069 1.105 0.964 1.039 0.937 0.965 0.662 0.985 1.055 1.028 0.717 0.862 1.034 0.894 0.542

Figure 13 shows that as the amount of RAP increases, the number of cycles to failure decreases. As shown in Figure 13, the fatigue life of the mixes is reduced as soon as RAP is added. This was expected because as the old to new binder ratio increases mixes become stiffer and tend to fail sooner in a constant strain test. However, the performance of the 30% RAP mixtures in fatigue may be improved by reducing the PG binder grade to compensate for the stiffer composite binder.

80,000

70,000

60,000

Cycles to Failure

50,000

40,000

30,000

20,000 10,000

y = -1141.3x + 63555 R2 = 0.8695

0

0

10

20

30

% RAP

FIGURE 13 Number of cycles to failure for RAP mixtures (800). 32

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

As mentioned earlier, in a controlled-strain test, stiffer mixes are expected to fail earlier (lower number of cycles to failure). The results shown in Figure 14 followed the expected trend, and shows that the higher initial stiffnesses are related to higher RAP contents and higher old to new asphalt ratios.
Figure 15 shows the effect of aggregate source on fatigue life. Along with Table 22, Figure 15 indicates that Camak mixtures reached a higher number of cycles to failure than the other aggregate sources. This was somewhat surprising because Camak blends did not have aggregate properties that would seem to significantly improve the fatigue resistance of the mixture. However, they had a low old to new binder ratio (an average of 0.12) which would be expected to result in higher fatigue life in HMA mixtures. Table 23 shows the difference of mean cycles to failure (Nf) for the interaction between aggregate source and RAP content. It is important to note that the number of cycles to failure only changes significantly with RAP content for Ruby mixtures, and only at the 30% RAP level. (Significant differences at 95% confidence level are in bold.) This was expected because these mixtures have the highest old to new binder ratio (0.45). A comparison of the differences in mean values for initial dissipated energy (Table 24) also shows that there was not a significant difference until 30% RAP was used.

5,500 5,000 4,500

0% 10% 20% 30%

Initial Stiffness (MPa)

4,000

3,500

3,000 2,500

y = -535.04Ln(x) + 9174.5 R2 = 0.4454

2,000 0

20,000

40,000

60,000

80,000

Cycles to Failure

100,000

120,000

FIGURE 14 Relationship between initial stiffness and number of cycles to failure (800 ).

33

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

100,000

80,000

Cycles to Failure

60,000

40,000

20,000

0 Camak

Lithia Springs

Mt. View

Aggregate Source

Ruby

FIGURE 15 Effect of aggregate source on Nf (800).

TABLE 23 Nf Comparisons for RAP Content Agg. Source Interaction (800 )

% RAP
0 10 0 20 0 30

Mt. View 4,343 -14,787 -25,920

Difference of Means (Cycles) Lithia Spr. Camak

-13,827

-51,123

-763

-21,000

-41,950

-17,310

Ruby -31,747 -51,557 -68,407

10 20 10 30

-19,130 -30,263

13,063 -28,123

30,123 33,813

-19,810 -36,660

20 30

-11,133

-41,187

3,690

-16,850

TABLE 24 Initial Dissipated Energy Comparisons for RAP Contents (800)

% RAP
0 10 0 20 0 30

Difference of Means (kPa)
0.088 0.063 -0.282

Are differences significant at 95% confidence level?
No No Yes

10 20

-0.026

No

10 30

-0.371

Yes

20 30

-0.345

Yes

34

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

Fatigue cracking is a distress that typically initiates at the bottom of the HMA layer where the tensile stress is the highest then propagates to the surface as one or more longitudinal cracks. Since SMA mixes are typically placed at or near the surface of the pavement, fatigue cracking may not be a major concern.

Low Strain Results The results for the low strain beams are shown in Table 25. Most tests were stopped at the cutoff point of 1,000,000 cycles, and only a few specimens reached 50% of the initial stiffness before that. The number of cycles to failure shown in Table 25 corresponds to the extrapolated value from the best fit curve of the Nf vs stiffness plot at 50% of the initial stiffness. Because data extrapolation can be a source of error, the percent drop in initial stiffness at the end of the test (1,000,000 cycles maximum) was also included in the results. Low percentages indicate that the specimen had experienced less damage at the end of the test and it was likely to withstand a higher number of cycles before reaching the failure point. As shown in Figure 16, there is a good correlation between the percent drop in stiffness and estimated Nf (R2 = 0.66).

TABLE 25 Average Test Results for Low Strain Beams (400 )

Aggregate Source

% RAP

Mt. View

0

10

20

30

Lithia

0

Springs

10

20

30

Camak

0

10

20

30

Ruby

0

10

20

30

RAP Type
DG1 -4
DG2 +4 -4 DG1 +4 DG2 +4 DG2 -4 DG1 DG2 +4 DG1 -4

Cycles to Failure*
3,601,543 5,153,700 2,338,070 1,332,883 5,526,153 3,686,010 4,791,923 1,532,050 4,353,263 4,686,913 2,689,257 3,857,159 5,405,080 4,552,187 4,460,207 759,820

Initial Stiffness (MPa)
3,369 3,450 3,485 4,715 3,653 3,845 3,547 4,166 3,187 3,571 3,139 3,285 3,502 3,57 0 3,395 4,621

% Drop in
Stiffness**
38.4 37.6 35.9 46.6 36.5 39.6 38.6 45.2 38.3 34.7 41.8 33.3 38.6 39.1 39.3 49.9

Initial Dissipated
Energy (kPa) 0.271 0.267 0.279 0.199 0.290 0.292 0.279 0.485 0.267 0.438 0.245 0.251 0.223 0.294 0.274 0.317

* Extrapolated values for specimens that did not reach 50% of the initial stiffness after 1,000,000 cycles. ** Measured at 1,000,000 cycles.

35

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

60

0%

10%

20%

50

30%

% Drop in Stiffness

40

30 y = -5.6204Ln(x) + 123.23 R2 = 0.6636

20 0.0.E+00

2.0.E+06

4.0.E+06

6.0.E+06

Cycles to Failure

8.0.E+06

1.0.E+07

FIGURE 16 Relationship between drop in initial stiffness at 1,000,000 cycles and estimated Nf (400 ).

An analysis of variance indicated the main factor that influenced the number of cycles to failure was RAP content (p-value = 0.002). Again, this is related to the old to new binder ratio that increases with RAP content. Fatigue life is more affected at high ratios (over 0.3), where the number of cycles to failure was reduced by up to 2.8 million cycles (Table 26). Similar trends were obtained when using the percent drop in initial stiffness instead of Nf.

Table 26 Nf and Percent Drop Comparisons for RAP Contents (400 )

% RAP
0 10 0 20 0 30

Difference of Means (Cycles)
-201,808 -1,151,646 -2,851,032

Difference of Means (%)
-0.22 0.94 5.75

Are differences significant at 95% confidence level?
No No Yes

10 20

-949,838

1.17

No

10 30

-2,649,224

5.98

Yes

20 30

-1,699,386

4.81

No

Figure 17 and Table 27 show that mixtures that contained Camak aggregates had the lower stiffness among aggregate sources. As mentioned earlier, these mixtures have more virgin binder content which is likely the cause of this result. Table 27 shows the difference of means for the initial stiffness for the interaction between RAP content and aggregate source.

36

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

6,000 5,500 5,000 4,500

0% 10% 20% 30%

Initial Stiffness (MPa)

4,000

3,500

3,000 2,500

y = -366.78Ln(x) + 9115 R2 = 0.3238

2,000 0.E+00

2.E+06

4.E+06

6.E+06

Cycles to Failure

8.E+06

1.E+07

FIGURE 17 Relationship between initial stiffness and Nf (400 ).

TABLE 27 Initial Stiffness Comparisons for RAP Content Aggregate Source

Interaction (400 )

% RAP

Difference of Means (MPa)

Mt. View Lithia Spr. Camak

Ruby

0 10

81

192

383

68

0 20

116

-106

-48

-107

0 30

1,346

514

98

1,119

10 20

35

-298

10 30

1,265

321

-432

-175

-286

1,050

20 30

1,230

620

146

1,226

Significant differences at 5% confidence level are in bold.

It was found that specimens that contained Lithia Springs and Camak aggregates (both with old to new asphalt ratios = 0.13, lowest among aggregate sources) did not change their initial stiffness significantly with an increase in RAP content. Table 27 also shows that the difference in mean values was not significant until the high RAP proportion of 30% was used. For low strain specimens, an analysis of variance indicated that the dissipated energy was not influenced by any of the main factors or interaction terms at the 95% significance level.

37

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

High Strain and Low Strain Comparison As expected, the number of cycles to failure was higher for low strain samples than for high strain samples (in the order of millions of cycles higher, as shown in Figure 18).

1000 800

0% 10% 20% 30%

Strain Level,

600

400

200 10,000

100,000

1,000,000

Cycles to Failure

10,000,000

FIGURE 18 Number of cycles to failure for high and low strain levels.

Mixtures with no RAP showed the best performance (Nf up to 2.5 times higher than recycled mixtures, as seen in Table 28) because they only contain virgin binder that is less stiff. As the RAP content is increased and more old binder goes into the mix, the fatigue life of the specimens is significantly reduced for both strain levels. This is associated with an increase in initial stiffness that causes earlier failure in controlledstrain specimens. However, this increase in stiffness may not affect performance when the mixes are placed near the top of the pavement, since fatigue cracking generally originates at the bottom of the HMA layer.

TABLE 28 Fatigue Life Comparisons for Strain Levels

% RAP
0 10 20 30

Old binder/ New binder
0.00 0.09 0.17 0.32

Nf (400 )
67,313 44,225 45,287 28,917

Nf (800 )
4,721,510 4,519,703 3,569,864 1,870,478

Summary

The results of this study have shown that adding RAP to an SMA mix has an impact in the sense that a portion of the total binder content corresponds to old (aged) binder. As
38

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

the RAP content increases, this portion of binder increases as well. The main implication is that the stiffness of the resulting asphalt blend is higher than that of the virgin binder, and therefore mixture stiffness increases with RAP content. For RAPs with high asphalt content, such as the fine-graded portion of screened RAP, the old to new binder ratio is higher and the effect is greater. The magnitude of the effect also depends on the properties of the RAP binder compared to the virgin binder.
For the performance tests conducted in this study, the increase in stiffness caused by higher old to new asphalt ratios did not have a significant effect in most cases. Figure 19 shows that there is a poor correlation between old to new binder ratio and TSR (R2 = 0.08). However, all mixtures were above the minimum requirement and the TSR values increased slightly as RAP content increased; therefore moisture susceptibility was not adversely affected by recycled SMA mixes.
1.10

TSR

1.00

y = 0.1504x + 0.9229

0.90

R2 = 0.0768

0.80

GDOT Minumum

0.70

Required

0.60

0.00

0.10

0.20

0.30

0.40

0.50

RAP Binder/Virgin Binder

FIGURE 19 Effect of RAP binder on TSR.

Figure 20 shows that rutting performance was not correlated to the old to new binder ratio either (R2 = 0.01). Still, most mixtures had average rut depths below the maximum specified by GDOT. Two recycled mixtures had results that exceeded the design criteria but by no more than 0.4 mm, which could be attributed to test variability. In general, mixtures had good rutting performance that was not significantly changed by the presence of old binder or increased RAP.
Susceptibility of the mixes to thermal cracking was poorly correlated to the old to new binder ratio. This could be due to the fact that even though the old binder content increases, there was not a significant change in the combined binder blend properties that control thermal cracking (creep stiffness and creep rate, shown in Figure 21).

39

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

7.0
GDOT Maximum Limit 6.0

Rut Depth, mm

5.0

4.0

3.0 y = 0.878x + 3.1287 R2 = 0.0087
2.0

1.0

0.00

0.10

0.20

0.30

0.40

0.50

RAP Binder/Virgin Binder

FIGURE 20 Effect of RAP binder on rutting performance.

Creep Stiffness (MPa) Creep Rate

400 300 200 100
0 0

-12C -18C 0.40
y = 1.45x + 269.25 R2 = 0.3239 0.36
0.32

-12C -18C
y = -0.0009x + 0.2823 R2 = 0.3496

y = 0.6525x + 146.78 R2 = 0.2363

10

20

30

% RAP Blend

0.28
0.24
0.20 0

y = -0.0006x + 0.3258 R2 = 0.31

10

20

30

% RAP Blend

FIGURE 21 Effect of RAP on low temperature binder properties.

One result that was affected by the increase in the old to new binder ratio was the fatigue life of the mixes. Figure 22 shows that adding more RAP binder to the mixture produces lower number of cycles to failure for specimens tested in controlled-strain mode. This occurs because the stiffness of the mix increases with higher old to new asphalt ratios.
It was observed that the fatigue life of the mixes was significantly reduced at high strain levels. However, recycled mixes are stiffer and will have less strain. Fatigue life may not be as affected unless the mixes are used in thin pavements. Also, SMA mixes are likely to be used as a surface layer, and because fatigue cracking generally originates at the bottom of the HMA pavement layers, it may not be a concern for this type of mixture.

40

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

100,000

10,000,000

Cycles to Failure (800 ) Cycles to Failure (400 )

10,000

y = 72064e-4.4141x R2 = 0.5453

1,000,000

y = 5E+06e-3.6289x R2 = 0.6491

1,000 0.00

0.10

0.20

0.30

0.40

RAP Binder/Virgin Binder

0.50

100,000

0.00

0.10

0.20

0.30

0.40

0.50

RAP Binder/Virgin Binder

FIGURE 22 Effect of RAP binder on fatigue life.

It is important to mention that the type of RAP used in recycled mixes can also have an important influence in performance. When the fine-graded portion of the RAP was used, the amount of old binder was increased because this portion of the RAP typically has a higher asphalt content than dense-graded or coarse graded RAP. As already discussed, this results in higher stiffness of the mix. It is expected that mixes that contain fine-graded RAP will have good resistance to moisture susceptibility and permanent deformation, but low fatigue life. Thermal cracking may not be a major concern for the reasons discussed above. One benefit of using fine-graded RAP is that the virgin binder requirement can be considerably reduced, lowering the cost of the mix.
Replacing a percentage of the No. 7 stone with coarse-graded aggregate did not affect the performance of the recycled mixes significantly. The low asphalt content characteristic of coarse-graded RAP produces asphalt blends with a low old to new asphalt ratio that did not increase the stiffness of the mixes in a way that would adversely influence performance. Moisture susceptibility, permanent deformation and thermal cracking were not a concern for mixtures containing coarse-graded RAP. Fatigue life may be reduced depending on the RAP content used. The advantage of replacing virgin material with RAP aggregate is that recycled aggregates could be used if quarries were faced with a critical supply shortage of No. 7 stone due to its high demand, and still obtain a mix with characteristics similar to those of a virgin SMA mix.
The feasibility of using RAP from reclaimed SMA could not be evaluated conclusively. The SMA RAP received for this project did not match some of the characteristics of an SMA mix (asphalt content, percent passing the No. 4 sieve). This has been partially attributed to circumstances that occurred during the milling process and the fact that the material was crushed to have 100 percent passing the 12.5 mm sieve. The resulting RAP was more similar to a dense-graded mix with low asphalt content. No general conclusions can be made unless it is assured that the same conditions would be repeated as part of a standard procedure for this type of RAP.

41

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared
CONCLUSIONS AND RECOMMENDATIONS
The following general conclusions and recommendations are made based on test results from this research:
Tests on the aggregate properties of the combined blends indicated that addition of RAP generally improved LA abrasion and F/E particle content depending on the properties of the RAP aggregates, but the change was not significant up to 30% RAP.
Use of RAP changed the engineering properties of the resulting binder blends due to the increased old to new binder ratio. The stiffness of the binder blend (G*/sin, G*sin and creep stiffness) increases with RAP content, particularly increasing the fatigue cracking potential. To counter this effect, a lower PG binder grade may need to be used with high RAP proportions.
The volumetric properties of the mix (air voids, VMA and VFA) were met with all of the RAP stockpiles and various RAP contents.
RAP content influenced only the tensile strength and fatigue life (Nf) of the mixtures. Increasing RAP content resulted in higher tensile strengths (conditioned and unconditioned) and lower number of cycles to failure. It can be concluded that the addition of RAP may be beneficial for resistance to moisture damage, and adversely affects only the fatigue performance of the mixtures, especially at high strain levels.
Separating the RAP into fine and coarse-graded fractions produced two stockpiles with different properties. The fine-graded portion had a high asphalt content and therefore, produced mixes with high old to new binder ratios. Additionally, finegraded RAP contains more material passing the No. 200 sieve, which must be accounted for during mix design. The coarse-graded portion had lower asphalt content, which indicates that mixtures containing this material will have a lower amount of old binder and less increase in stiffness.
Use of fine-graded RAP reduced the virgin binder requirements due its high asphalt content, which translates into increased economic benefits. However, mixes that contain fine-graded graded RAP are stiffer because they have higher old to new binder ratios and are more susceptible to fatigue cracking. Use of finegraded RAP may require the use of binder blending charts or a reduction in PG binder grade.
Use of coarse-graded RAP allowed reducing the No. 7 stone proportion without affecting the performance of the mixes. This may be beneficial in case quarries are faced with a shortage due to the high demand of this material.
It is uncertain whether SMA pavement material can be successfully recycled back into an SMA mixture at high RAP proportions. Because of crushing and possible contamination with other pavement layers during milling, the reclaimed asphalt used in this study had a mix gradation that resembled a dense-graded mix and it had low asphalt content.
42

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared
Because fatigue cracking is the main concern for recycled mixtures and this distress originates at the bottom of the HMA layer, it is recommended that SMA mixes containing RAP be used primarily in the top layers of the pavement.
Adding RAP up to 30% had little effect on the low temperature PG properties. The low temperature grade of the combined binder blends was raised by one grade on only one of the cases. This may indicate that the grade of virgin binder (especially the low temperature grade) does not have to be adjusted to provide the desired properties.
It is recommended that GDOT specifications be modified to allow up to 20% RAP in SMA mixtures with no change in virgin binder grade. Mixtures will still need to meet the same gradation, volumetric, and performance criteria as virgin mixtures.
RAP proportions higher than 20% may be allowed, but the virgin binder grade may need to be reduced to improve fatigue performance properties. To maintain a similar dosage rate of polymer modification, it would be necessary to reduce both the high and low grades. (For example, reduce the grade from PG 76-22 to PG 7028.)
ACKNOWLEDGEMENTS
The authors wish to thank the Georgia Department of Transportation for its support in sponsoring this study. Thanks are also extended to Dr. Saeed Maghsoodloo, Statistician, for conducting the factorial design experiment and the statistical analysis of test results.
REFERENCES
1. Beyond Roads, Questions and Answers. Asphalt Education Partnership. http://www.beyondroads.com/index.cfm?fuseaction=page&filename=asphaltQan dA.html. Accessed March 15th, 2006.
2. Summary of Georgia's Experience with Stone Matrix Asphalt Mixes. Georgia Department of Transportation. http://www.dot.state.ga.us/dot/construction/materials-research/badmin/research/onlinereports%5Cr-SMA2002.pdf. Accessed February 21st, 2006.
3. Brown, E.R., R.B. Mallick, J. E. Haddock and J. Bukowski. Performance of Stone Matrix Asphalt (SMA) Mixtures in the United States. National Center for Asphalt Technology, Report No. 97-01, Jan. 1997.
4. Watson, D. E. Updated Review of Stone Matrix Asphalt and Superpave Projects. In Transportation Research Record 1832, TRB, National Research Council, Washington, D.C., 2003, pp. 217-223.
5. Brown, E.R., R.B. Mallick, J. E. Haddock and T.A. Lynn. Development of a Mixture Design Procedure for Stone Matrix Asphalt (SMA). National Center for Asphalt Technology, Report No. 97-03, March 1997.
43

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared
6. McDaniel, R.S., H. Soleymani, R.M. Anderson, P. Turner, and R. Peterson. Recommended Use of Reclaimed Asphalt Pavement in the Superpave Mix Design Method. Contractor's Final Report. NCHRP Web Document 30 (Project D9-12). 2000.
7. Huang, B., G. Li, D. Vukosavljevic, X. Shu, and B.K. Egan. Laboratory Investigation of Mixing Hot-Mix Asphalt with Reclaimed Asphalt Pavement. In Transportation Research Record 1929, TRB, National Research Council, Washington, D.C., 2005, pp. 37-45.
8. Kandhal, P.S. and K.Y. Foo. Designing Recycled Hot Mix Asphalt Mixtures Using Superpave Technology. National Center for Asphalt Technology, Report No. 9605, Jan. 1997.
9. Bukowski, J. R., Guidelines for the Design of Superpave Mixtures Containing Reclaimed Asphalt Pavement (RAP), Memorandum, ETG Meeting, FHWA Superpave Mixtures Expert Task Group, San Antonio, Texas, March, 1997.
10. Li X., T.R. Clyne, and M.O. Marasteanu. Recycyled Asphalt Pavement (RAP) Effects on Binder and Mixture Quality. Minnesota Department of Transportation. July, 2004.
11. Huang, B., W.R. Kingery, and Z. Zhang. Laboratory Study of Fatigue Characteristics of HMA Mixtures Containing RAP. Presented at the International Symposium on Design and Construction of Long Lasting Asphalt Pavements, Auburn, AL. June, 2004.
12. McGennis, R.B., S. Shuler, and H.U. Bahia. Background of Superpave Asphalt Binder Test Methods. FHWA, Report No. FHWA-SA-94-069, July 1994.
13. Roberts, F.L., P.S. Kandhal, E.R. Brown, D.Y. Lee, T.W. Kennedy. Hot Mix Asphalt Materials, Mixture Design and Construction. National Asphalt Pavement Association Research and Education Foundation, 1996.
44

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared APPENDIX A
Laboratory Mix Designs
45

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

A.1 Mix Designs for Mt. View Mixtures

Mt. View 0% RAP

Sieve size Proportions
1" 3/4" 1/2" 3/8" #4 #8 #16 #30 #50 #100 #200

007 68.0% 100.0 100.0 97.0 48.0
3.0 3.0 2.0 2.0 2.0 1.0 1.0

Aggregate Components

089

M10

Marble dust

12.0%

13.0%

6.0%

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

22.0

99.0

100.0

4.0

83.0

100.0

2.0

66.0

100.0

2.0

53.0

100.0

1.0

37.0

100.0

1.0

18.0

98.0

1.0

6.0

90.0

Lime 1.0% 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

Series 1

% AC 6.0

VMA 18.3

VFA 74.6

Mt. View 10% -4 RAP

Sieve size
Proportions 1" 3/4" 1/2" 3/8" #4 #8 #16 #30 #50
#100 #200

007
77.0% 100.0 100.0 97.0 48.0 3.0 3.0 2.0 2.0 2.0 1.0 1.0

089
0.0% 100.0 100.0 100.0 100.0 22.0 4.0 2.0 2.0 1.0 1.0 1.0

Aggregate Components

M10

Marble dust

7.0%

5.0%

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

99.0

100.0

83.0

100.0

66.0

100.0

53.0

100.0

37.0

100.0

18.0

98.0

6.0

90.0

Lime
1.0% 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

RAP
10.0% 100.0 100.0 100.0 100.0 100.0 81.0 65.0 53.0 40.0 25.0 15.0

Series 1 2 3

% AC 4.5 5.0 5.5

VMA 17.3 17.1 18.2

VFA 66.8 74.7 77.2

Blend 100% 100.0 100.0 98.0 64.6 24.6 20.3 17.2 15.5 13.3 10.0
8.0
Blend
100% 100.0 100.0 98.0 64.6 24.6 20.3 17.2 15.5 13.3 10.0
8.0

46

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

Air Voids

6.0 5.5 5.0

4.5

4.0

3.5

3.0

4.0

4.5

5.0

5.5

6.0

%AC

VM A

18.4

18.0

17.6

17.2

16.8

4.0

4.5

5.0

5.5

6.0

%AC

VFA

80.0

76.0

72.0

68.0

64.0

4.0

4.5

5.0

5.5

6.0

%AC

Mt. View 20% DG2 RAP

Sieve size
Proportions 1" 3/4" 1/2" 3/8" #4 #8 #16 #30 #50
#100 #200

007
67.0% 100.0 100.0 97.0 48.0 3.0 3.0 2.0 2.0 2.0 1.0 1.0

089
7.0% 100.0 100.0 100.0 100.0 22.0 4.0 2.0 2.0 1.0 1.0 1.0

Aggregate Components

M10

Marble dust

0.0%

4.8%

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

99.0

100.0

83.0

100.0

66.0

100.0

53.0

100.0

37.0

100.0

18.0

98.0

6.0

90.0

Series 1 2 3

% AC 5.5 6.0 6.5

VMA 20.4 19.3 19.7

VFA 69.9 80.3 84.1

Lime
0.9% 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

RAP
20.3% 100.0 100.0 100.0 95.0 77.0 61.0 50.0 42.0 32.0 20.0 12.0

Blend
100% 100.0 100.0 98.0 64.1 24.9 20.4 17.3 15.7 13.6 10.4
8.4

47

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

Air Voids

7.0

6.0

5.0

4.0

3.0

2.0

5.0

5.5

6.0

6.5

7.0

%AC

VM A

20.8

20.4

20.0

19.6

19.2

18.8

5.0

5.5

6.0

6.5

7.0

%AC

VFA

90.0

86.0

82.0

78.0

74.0

70.0

66.0

5.0

5.5

6.0

6.5

7.0

%AC

Mt. View 30% +4 RAP

Sieve size
Proportions 1" 3/4" 1/2" 3/8" #4 #8 #16 #30 #50
#100 #200

007
58.7% 100.0 100.0 97.0 48.0 3.0 3.0 2.0 2.0 2.0 1.0 1.0

089
0.0% 100.0 100.0 100.0 100.0 22.0 4.0 2.0 2.0 1.0 1.0 1.0

Aggregate Components

M10

Marble dust

5%

5%

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

99.0

100.0

83.0

100.0

66.0

100.0

53.0

100.0

37.0

100.0

18.0

98.0

6.0

90.0

Series 1 2 3

% AC 5.0 5.5 6.0

VMA 18.3 18.9 19.9

VFA 78.9 81.8 82.8

Lime
0.8% 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

RAP
30.5% 100.0 99.0 96.0 84.0 37.0 25.0 21.0 18.0 15.0 10.0 6.2

Blend
100% 100.0 99.7 97.0 64.6 23.8 19.3 16.7 15.1 13.4 10.2
8.1

48

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

Air Voids

3.9 3.8 3.7

3.6 3.5

3.4

3.3

4.5

5.0

5.5

6.0

6.5

%AC

VM A

20.0

19.6

19.2

18.8

18.4

18.0

4.5

5.0

5.5

6.0

6.5

%AC

VFA

86.0

84.0

82.0

80.0

78.0

76.0

4.5

5.0

5.5

6.0

6.5

%AC

A.2 Mix Designs for Lithia Springs Mixtures

Lithia Springs 0% RAP

Aggregate Components

Sieve size

007

089

810

Marble dust

Proportions 67.0%

13.0%

13.0%

6.0%

1"

100.0

100.0

100.0

100.0

3/4"

100.0

100.0

100.0

100.0

1/2"

85.0

100.0

100.0

100.0

3/8"

50.0

100.0

100.0

100.0

#4

6.0

30.0

84.0

100.0

#8

2.0

2.0

62.0

100.0

#16

1.0

2.0

50.0

100.0

#30

1.0

1.0

41.0

100.0

#50

1.0

1.0

28.0

100.0

#100

1.0

1.0

21.0

98.0

#200

1.0

1.0

10.0

90.0

Series 1

% AC 6.0

VMA 17.8

VFA 75.8

Lime 1.0% 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

Blend 100% 100.0 100.0 90.0 66.5 25.8 16.7 14.4 13.1 11.4 10.4
8.5

49

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

Lithia Springs 10% DG1 RAP

Sieve size
Proportions 1" 3/4" 1/2" 3/8" #4 #8 #16 #30 #50
#100 #200

007 71.9% 100.0 100.0 85.0 50.0
6.0 2.0 1.0 1.0 1.0 1.0 1.0

089 0.0% 100.0 100.0 100.0 100.0 30.0 2.0 2.0 1.0 1.0 1.0 1.0

Aggregate Components

Marble

810

dust

12.6%

4.5%

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

84.0

100.0

62.0

100.0

50.0

100.0

41.0

100.0

28.0

100.0

21.0

98.0

10.0

90.0

Series 1 2 3

% AC 5.0 5.5 6.0

VMA 17.2 17.6 18.5

VFA 72.0 76.8 78.9

Lime 1.0% 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

RAP 10.0% 100.0 100.0 99.0 93.0 73.0 58.0 47.0 38.0 29.0 19.0 11.2

Blend 100% 100.0 100.0 89.1 63.4 27.7 20.6 17.2 15.2 12.6 10.7
8.1

Air Voids

5.0

4.7

4.4

4.1

3.8

3.5

4.5

5.0

5.5

6.0

6.5

%AC

VM A

18.8

18.4

18.0

17.6

17.2

16.8

4.5

5.0

5.5

6.0

6.5

%AC

VFA

82.0

80.0

78.0

76.0

74.0

72.0

70.0

4.5

5.0

5.5

6.0

6.5

%AC

50

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

Lithia Springs 20% +4 RAP

Sieve size
Proportions 1" 3/4" 1/2" 3/8" #4 #8 #16 #30 #50
#100 #200

007 61.8% 100.0 100.0 85.0 50.0
6.0 2.0 1.0 1.0 1.0 1.0 1.0

089 0.0% 100.0 100.0 100.0 100.0 30.0 2.0 2.0 1.0 1.0 1.0 1.0

Aggregate Components

Marble

810

dust

12.0%

5.0%

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

84.0

100.0

62.0

100.0

50.0

100.0

41.0

100.0

28.0

100.0

21.0

98.0

10.0

90.0

Series 1 2 3

% AC 5.0 5.5 6.0

VMA 17.7 18.0 18.4

VFA 74.4 79.4 83.8

Lime 0.9% 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

RAP 20.3% 100.0 99.0 96.0 84.0 37.0 25.0 21.0 18.0 15.0 10.0
6.2

Blend 100% 100.0 99.8 89.9 65.9 27.2 19.7 16.8 15.1 12.9 11.0
8.5

Air Voids

5.0 4.5 4.0

3.5 3.0

2.5

2.0

4.5

5.0

5.5

6.0

6.5

%AC

VM A

18.6

18.4

18.2

18.0

17.8

17.6

17.4

17.2

4.5

5.0

5.5

6.0

6.5

%AC

VFA

90.0

86.0

82.0

78.0

74.0

70.0

66.0

4.5

5.0

5.5

6.0

6.5

%AC

51

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

Lithia Springs 30% DG2 RAP

Sieve size
Proportions 1" 3/4" 1/2" 3/8" #4 #8 #16 #30 #50
#100 #200

007 65.7% 100.0 100.0 85.0 50.0
6.0 2.0 1.0 1.0 1.0 1.0 1.0

089 0.0% 100.0 100.0 100.0 100.0 30.0 2.0 2.0 1.0 1.0 1.0 1.0

Aggregate Components

Marble

810

dust

0.0%

3.0%

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

84.0

100.0

62.0

100.0

50.0

100.0

41.0

100.0

28.0

100.0

21.0

98.0

10.0

90.0

Series 1 2 3

% AC 4.5 5.0 5.5

VMA 17.3 17.5 18.4

VFA 74.9 80.8 82.2

Lime 0.8% 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

RAP 30.5% 100.0 100.0 100.0 95.0 77.0 61.0 50.0 42.0 32.0 20.0 12.0

Blend 100% 100.0 100.0 90.1 65.6 31.2 23.7 19.7 17.3 14.2 10.5
7.8

Air Voids

4.5

4.0

3.5

3.0

2.5

4.0

4.5

5.0

5.5

6.0

%AC

VM A

18.8

18.4

18.0

17.6

17.2

16.8

4.0

4.5

5.0

5.5

6.0

%AC

VFA

88.0

84.0

80.0

76.0

72.0

4.0

4.5

5.0

5.5

6.0

%AC

52

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

A.3 Mix Designs for Camak Mixtures

Camak 0% RAP

Aggregate Components

Sieve size

007

M10

Marble dust

Proportions 83.0%

10.0%

6.0%

1"

100.0

100.0

100.0

3/4"

100.0

100.0

100.0

1/2"

94.0

100.0

100.0

3/8"

46.0

100.0

100.0

#4

2.0

98.0

100.0

#8

1.0

82.0

100.0

#16

1.0

62.0

100.0

#30

1.0

50.0

100.0

#50

1.0

36.0

100.0

#100

1.0

25.0

98.0

#200

1.0

12.0

90.0

Lime 1.0% 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

Blend 100% 100.0 100.0 95.0 55.2 18.5 16.0 14.0 12.8 11.4 10.2
8.4

Series 1

% AC 6.0

VMA 21.1

VFA 73.3

Camak 10% DG2 RAP

Sieve size
Proportions 1" 3/4" 1/2" 3/8" #4 #8 #16 #30 #50
#100 #200

007 73.0% 100.0 100.0 94.0 46.0
2.0 1.0 1.0 1.0 1.0 1.0 1.0

Aggregate Components

Marble

M10

dust

Lime

11.0%

4.9%

0.9%

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

98.0

100.0

100.0

82.0

100.0

100.0

62.0

100.0

100.0

50.0

100.0

100.0

36.0

100.0

100.0

25.0

98.0

100.0

12.0

90.0

100.0

RAP 10.2% 100.0 100.0 100.0 95.0 77.0 61.0 50.0 42.0 32.0 20.0 12.0

Blend 100% 100.0 100.0 95.6 60.1 25.9 21.8 18.5 16.3 13.8 11.2
8.6

Series 1 2 3

% AC 5.5 6.0 6.5

VMA 18.4 18.9 19.2

VFA 72.5 76.4 81.1

53

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

Air Voids

5.5

5.0

4.5

4.0

3.5

3.0

5.0

5.5

6.0

6.5

7.0

%AC

VM A

19.4

19.2

19.0

18.8

18.6

18.4

18.2

5.0

5.5

6.0

6.5

7.0

%AC

VFA

86.0

82.0

78.0

74.0

70.0

5.0

5.5

6.0

6.5

7.0

%AC

Camak 20% -4 RAP

Sieve size
Proportions 1" 3/4" 1/2" 3/8" #4 #8 #16 #30 #50
#100 #200
Series 1 2 3

007 75.1% 100.0 100.0 94.0 46.0
2.0 1.0 1.0 1.0 1.0 1.0 1.0
% AC 5.0 5.5 6.0

Aggregate Components

Marble

M10

dust

Lime

0.0%

4.0%

0.9%

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

98.0

100.0

100.0

82.0

100.0

100.0

62.0

100.0

100.0

50.0

100.0

100.0

36.0

100.0

100.0

25.0

98.0

100.0

12.0

90.0

100.0

VMA 18.7 19.2 19.8

VFA 74.2 77.8 81.1

RAP 20.0% 100.0 100.0 100.0 100.0 100.0 81.0 65.0 53.0 40.0 25.0 15.0

Blend 100% 100.0 100.0 95.5 59.4 26.4 21.9 18.7 16.3 13.7 10.6
8.3

54

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

Air Voids

5.0

4.5

4.0

3.5

3.0

4.5

5.0

5.5

6.0

6.5

%AC

VM A

20.0

19.6

19.2

18.8

18.4

18.0

4.5

5.0

5.5

6.0

6.5

%AC

VFA

86.0

82.0

78.0

74.0

70.0

4.5

5.0

5.5

6.0

6.5

%AC

Camak 30% DG1 RAP

Sieve size
Proportions 1" 3/4" 1/2" 3/8" #4 #8 #16 #30 #50
#100 #200

007 65.6% 100.0 100.0 94.0 46.0
2.0 1.0 1.0 1.0 1.0 1.0 1.0

Aggregate Components

Marble

M10

dust

Lime

0.0%

3.5%

0.8%

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

98.0

100.0

100.0

82.0

100.0

100.0

62.0

100.0

100.0

50.0

100.0

100.0

36.0

100.0

100.0

25.0

98.0

100.0

12.0

90.0

100.0

Series 1 2 3

% AC 5.5 6.0 6.5

VMA 19.6 20.1 20.6

VFA 81.3 85.1 88.0

RAP 30.1% 100.0 100.0 99.0 93.0 73.0 58.0 47.0 38.0 29.0 19.0 11.2

Blend 100% 100.0 100.0 95.8 62.5 27.6 22.4 19.1 16.4 13.7 10.6
8.0

55

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

Air Voids

4.0

3.6

3.2

2.8

2.4

2.0

5.0

5.5

6.0

6.5

7.0

%AC

VM A

20.8

20.4

20.0

19.6

19.2

5.0

5.5

6.0

6.5

7.0

%AC

VFA

90.0

88.0

86.0

84.0

82.0

80.0

78.0

5.0

5.5

6.0

6.5

7.0

%AC

A.4 Mix Designs for Ruby Mixtures

Ruby 0% RAP

Aggregate Components

Sieve size

007

M10

Marble dust

Proportions 77.0%

18.0%

4.0%

1"

100.0

100.0

100.0

3/4"

100.0

100.0

100.0

1/2"

96.0

100.0

100.0

3/8"

55.0

100.0

100.0

#4

2.0

99.0

100.0

#8

1.0

82.0

100.0

#16

1.0

62.0

100.0

#30

1.0

49.0

100.0

#50

1.0

37.0

100.0

#100

1.0

27.0

98.0

#200

1.0

18.0

90.0

Series 1 2

% AC 6.5 7.0

VMA 19.4 18.8

VFA 76.7 86.5

Lime 1.0% 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

Blend 100% 100.0 100.0 96.9 65.4 24.4 20.5 16.9 14.6 12.4 10.6
8.6

56

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

Ruby 10% +4 RAP

Sieve size
Proportions 1" 3/4" 1/2" 3/8" #4 #8 #16 #30 #50
#100 #200
Series 1 2 3

007 69.9% 100.0 100.0 96.0 55.0
2.0 1.0 1.0 1.0 1.0 1.0 1.0
% AC 5.0 5.5 6.0

Aggregate Components

Marble

M10

dust

Lime

15.0%

4.0%

0.9%

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

99.0

100.0

100.0

82.0

100.0

100.0

62.0

100.0

100.0

49.0

100.0

100.0

37.0

100.0

100.0

27.0

98.0

100.0

18.0

90.0

100.0

VMA 18.1 18.0 18.5

VFA 68.9 76.4 80.2

RAP 10.2% 100.0 99.0 96.0 84.0 37.0 25.0 21.0 18.0 15.0 10.0
6.2

Blend 100% 100.0 99.9 96.8 66.9 24.9 20.4 17.0 14.8 12.7 10.6
8.5

Air Voids

6.0

5.5

5.0

4.5

4.0

3.5

3.0

4.5

5.0

5.5

6.0

6.5

%AC

VM A

18.6

18.4

18.2

18.0

17.8

17.6

4.5

5.0

5.5

6.0

6.5

%AC

VFA

82.0

78.0

74.0

70.0

66.0

4.5

5.0

5.5

6.0

6.5

%AC

57

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

Ruby 20% DG1 RAP

Sieve size
Proportions 1" 3/4" 1/2" 3/8" #4 #8 #16 #30 #50
#100 #200
Series 1 2 3

007 70.0% 100.0 100.0 96.0 55.0
2.0 1.0 1.0 1.0 1.0 1.0 1.0
% AC 5.0 5.5 6.0

Aggregate Components

Marble

M10

dust

Lime

5.0%

4.0%

0.9%

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

99.0

100.0

100.0

82.0

100.0

100.0

62.0

100.0

100.0

49.0

100.0

100.0

37.0

100.0

100.0

27.0

98.0

100.0

18.0

90.0

100.0

VMA 18.4 18.9 19.7

VFA 75.6 79.8 81.9

RAP 20.1% 100.0 100.0 99.0 93.0 73.0 58.0 47.0 38.0 29.0 19.0 11.2

Blend 100% 100.0 100.0 97.0 67.1 25.9 21.4 18.1 15.7 13.3 10.7
8.4

Air Voids

5.0

4.5

4.0

3.5

3.0

4.5

5.0

5.5

6.0

6.5

%AC

VM A

20.0

19.6

19.2

18.8

18.4

18.0

17.6

4.5

5.0

5.5

6.0

6.5

%AC

VFA

86.0

82.0

78.0

74.0

70.0

4.5

5.0

5.5

6.0

6.5

%AC

58

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

Ruby 30% -4 RAP

Sieve size
Proportions 1" 3/4" 1/2" 3/8" #4 #8 #16 #30 #50
#100 #200
Series 1 2 3

007 66.7% 100.0 100.0 96.0 55.0
2.0 1.0 1.0 1.0 1.0 1.0 1.0
% AC 4.0 4.5 5.0

Aggregate Components

Marble

M10

dust

Lime

0.0%

2.5%

0.9%

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

99.0

100.0

100.0

82.0

100.0

100.0

62.0

100.0

100.0

49.0

100.0

100.0

37.0

100.0

100.0

27.0

98.0

100.0

18.0

90.0

100.0

VMA 17.6 17.5 18.4

VFA 76.5 83.9 85.4

RAP 29.9% 100.0 100.0 100.0 100.0 100.0 81.0 65.0 53.0 40.0 25.0 15.0

Blend 100% 100.0 100.0 97.3 70.0 34.6 28.3 23.5 19.9 16.0 11.5
8.3

Air Voids

4.5

4.0

3.5

3.0

2.5

2.0

3.5

4.0

4.5

5.0

5.5

%AC

VM A

18.8

18.4

18.0

17.6

17.2

16.8

3.5

4.0

4.5

5.0

5.5

%AC

VFA

90.0

86.0

82.0

78.0

74.0

3.5

4.0

4.5

5.0

5.5

%AC

59

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared APPENDIX B
Individual Test Results
60

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

Table B.1. Results from Moisture Susceptibility Test.

Agg. Source

% RAP RAP Type

% Air Voids

Wet Strength,
psi

Dry Strength,
psi

TSR

Mt. View

0

DG1

7.2

69.90

77.86

0.90

Mt. View

0

DG1

7.2

77.09

64.94

1.19

Mt. View

0

DG1

7.1

67.67

71.94

0.94

Mt. View

10

-4

6.7

90.91

100.65

0.90

Mt. View

10

-4

6.9

86.45

83.21

1.04

Mt. View

10

-4

7.1

87.09

77.79

1.12

Mt. View

20

DG2

7.3

95.75

107.80

1.12

Mt. View

20

DG2

9.6

84.03

100.60

0.89

Mt. View

20

DG2

6.6

95.87

97.30

0.84

Mt. View

30

+4

6.7

101.99

107.84

0.95

Mt. View

30

+4

6.3

110.14

105.62

1.04

Mt. View

30

+4

7.3

87.28

103.77

0.84

Lithia Spr.

0

-4

7.6

67.55

91.80

0.74

Lithia Spr.

0

-4

7.9

74.36

97.08

0.77

Lithia Spr.

0

-4

7.5

70.79

79.26

0.89

Lithia Spr.

10

DG1

7.1

84.93

85.56

0.99

Lithia Spr.

10

DG1

6.8

73.78

79.96

0.92

Lithia Spr.

10

DG1

7.4

80.41

85.18

0.94

Lithia Spr.

20

+4

7.4

88.87

95.68

0.93

Lithia Spr.

20

+4

7.0

98.74

77.35

1.28

Lithia Spr.

20

+4

7.4

84.67

88.04

0.96

Lithia Spr.

30

DG2

7.3

98.99

98.55

1.00

Lithia Spr.

30

DG2

6.8

92.37

85.94

1.07

Lithia Spr.

30

DG2

6.5

95.94

99.76

0.96

Camak

0

+4

7.5

77.70

87.54

0.96

Camak

0

+4

8.0

78.80

90.34

0.89

Camak

0

+4

7.5

79.50

85.18

0.87

Camak

10

DG2

6.6

76.90

100.39

0.77

Camak

10

DG2

6.9

68.05

79.26

0.86

Camak

10

DG2

7.2

72.00

75.69

0.95

Camak

20

-4

6.8

96.58

108.54

0.89

Camak

20

-4

6.7

76.52

100.20

0.76

Camak

20

-4

7.8

84.03

86.64

0.97

Camak

30

DG1

6.4

127.39

136.81

0.93

Camak

30

DG1

7.1

114.78

124.08

0.93

Camak

30

DG1

6.2

112.05

113.51

0.99

Ruby

0

DG2

7.0

72.10

70.00

1.03

Ruby

0

DG2

6.0

71.60

70.00

1.02

Ruby

0

DG2

6.7

67.00

69.10

0.97

Ruby

10

+4

6.7

93.65

102.24

0.92

Ruby

10

+4

6.7

89.89

102.56

0.88

Ruby

10

+4

6.9

93.39

106.12

0.88

Ruby

20

DG1

6.1

81.42

84.54

0.96

Ruby

20

DG1

6.4

76.78

85.63

0.90

Ruby

20

DG1

6.1

92.69

79.13

1.17

Ruby

30

-4

6.1

154.89

168.96

0.92

Ruby

30

-4

6.9

134.77

139.99

0.96

Ruby

30

-4

6.3

135.54

125.16

1.08

61

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

Table B.2. Results from Rutting Susceptibility Test.

Agg. Source

% RAP

RAP Type

Mt. View

0

DG1

Mt. View

0

DG1

Mt. View

0

DG1

Mt. View

0

DG1

Mt. View

0

DG1

Mt. View

0

DG1

Mt. View

10

-4

Mt. View

10

-4

Mt. View

10

-4

Mt. View

10

-4

Mt. View

10

-4

Mt. View

10

-4

Mt. View

20

DG2

Mt. View

20

DG2

Mt. View

20

DG2

Mt. View

20

DG2

Mt. View

20

DG2

Mt. View

20

DG2

Mt. View

30

+4

Mt. View

30

+4

Mt. View

30

+4

Mt. View

30

+4

Mt. View

30

+4

Mt. View

30

+4

Lithia Spr.

0

-4

Lithia Spr.

0

-4

Lithia Spr.

0

-4

Lithia Spr.

0

-4

Lithia Spr.

0

-4

Lithia Spr.

0

-4

Lithia Spr.

10

DG1

Lithia Spr.

10

DG1

Lithia Spr.

10

DG1

Lithia Spr.

10

DG1

Lithia Spr.

10

DG1

Lithia Spr.

10

DG1

Lithia Spr.

20

+4

Lithia Spr.

20

+4

Lithia Spr.

20

+4

Lithia Spr.

20

+4

Lithia Spr.

20

+4

Lithia Spr.

20

+4

Lithia Spr.

30

DG2

Lithia Spr.

30

DG2

Lithia Spr.

30

DG2

Lithia Spr.

30

DG2

Lithia Spr.

30

DG2

Lithia Spr.

30

DG2

% Air Voids 5.4 5.0 5.2 5.3 5.0 5.1 5.0 5.1 4.8 4.9 5.1 5.1 5.1 4.7 4.9 4.8 4.4 5.0 5.2 4.9 4.9 4.9 4.9 4.8 4.7 4.2 4.3 4.7 4.7 4.3 4.6 4.7 4.8 4.6 4.7 4.7 4.7 4.8 4.7 4.9 4.4 4.2 5.6 5.2 5.8 5.3 5.6 5.4

Rut Depth, mm 3.13 2.94 1.76 3.39 0.95 6.50 4.42 3.18 1.40 3.11 5.86 1.41 4.23 6.01 3.34 2.05 8.33 2.50 1.25 2.35 1.89 1.23 2.49 0.99 3.14 1.86 1.89 3.59 1.98 1.76 2.73 2.37 1.31 2.15 1.25 2.19 2.69 2.48 1.77 0.93 3.33 3.45 2.77 5.75 5.15 2.88 5.69 4.76

62

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

Table B.2 (cont.). Results from Rutting Susceptibility Test.

Agg. Source

% RAP

RAP Type

% Air Voids

Camak

0

+4

4.9

Camak

0

+4

4.8

Camak

0

+4

4.7

Camak

0

+4

4.5

Camak

0

+4

4.2

Camak

0

+4

4.3

Camak

10

DG2

4.7

Camak

10

DG2

4.6

Camak

10

DG2

4.4

Camak

10

DG2

4.9

Camak

10

DG2

4.5

Camak

10

DG2

4.6

Camak

20

-4

4.3

Camak

20

-4

4.4

Camak

20

-4

4.1

Camak

20

-4

4.8

Camak

20

-4

4.2

Camak

20

-4

5.2

Camak

30

DG1

4.7

Camak

30

DG1

4.8

Camak

30

DG1

4.7

Camak

30

DG1

5.0

Camak

30

DG1

4.3

Camak

30

DG1

4.1

Ruby

0

DG2

5.1

Ruby

0

DG2

5.4

Ruby

0

DG2

4.5

Ruby

0

DG2

4.2

Ruby

0

DG2

4.8

Ruby

0

DG2

4.4

Ruby

10

+4

5.4

Ruby

10

+4

5.2

Ruby

10

+4

5.3

Ruby

10

+4

5.6

Ruby

10

+4

5.0

Ruby

10

+4

5.0

Ruby

20

DG1

4.8

Ruby

20

DG1

4.7

Ruby

20

DG1

4.8

Ruby

20

DG1

5.0

Ruby

20

DG1

5.2

Ruby

20

DG1

4.5

Ruby

30

-4

5.4

Ruby

30

-4

5.2

Ruby

30

-4

5.8

Ruby

30

-4

5.2

Ruby

30

-4

5.3

Ruby

30

-4

4.4

Rut Depth, mm 6.83 4.27 2.39 4.42 3.45 0.68 1.62 2.14 1.97 1.58 2.56 1.89 1.40 1.83 1.93 0.95 1.87 0.91 4.04 2.95 2.53 4.20 2.75 3.04 5.89 1.73 3.85 4.05 2.70 3.30 4.40 5.60 6.17 5.83 5.03 3.92 5.74 4.56 5.99 6.47 5.84 3.67 4.95 3.08 4.03 4.84 3.25 2.97

63

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, D. Jared, and P. Wu

TABLE B.3. Results from Creep Compliance Test

Agg. Source
Mt. View Mt. View

% RAP
0 10

RAP Type
DG 1 -4

Creep Compliance @ 50 sec, 1/psi

-20C

-10C

0C

8.35E-07 1.13E-06 2.68E-06

6.46 E-07 9.91E-07 1.87E-06

Mt. View

20

Mt. View

30

Lithia Springs

0

Lithia Springs

10

Lithia Springs

20

Lithia Springs

30

Camak

0

Camak

10

DG 2
+4 -4 DG 1 +4 DG 2 +4
DG 2

5.89 E-07
9.63 E-07 4.58 E-07 5.00 E-07 7.38 E-07 1.42E-06 1.15 E-06
1.69 E-06

1.55E-06
2.59E-06 9.69E-07 1.03E-06 1.13E-06 2.38E-06 2.70E-06
1.60E-06

3.37E-06
3.12E-06 2.14E-06 2.32E-06 1.88E-06 6.63E-06 7.89E-06
2.60E-06

Camak

20

-4

9.82 E-07 1.50E-06 3.35E-06

Camak

30

DG 1

1.61 E-06 1.85E-06 6.03E-06

Ruby

0

DG 2

1.83 E-06 3.37E-06 1.03E-05

Ruby

10

+4

1.61 E-06 2.40E-06 4.71E-06

Ruby

20

DG 1

1.48 E-06 2.51E-06 6.77E-06

Ruby

30

-4

1.49E-06 2.70E-06 4.15E-06

Log aT

-10C

0C

-0.87

-1.75

-0.82

-1.64

-1.03
-0.78 -1.15 -1.27 -0.70 -1.34 -1.14

-2.07
-1.55 -2.29 -2.54 -1.40 -2.68 -2.27

-0.31

-0.62

-0.83

-1.66

-0.86

-1.71

-0.92

-1.84

-0.61

-1.21

-0.85

-1.70

-0.79

-1.58

m-value
0.574 0.410
0.386 0.372 0.339 0.320 0.484 0.432 0.440 0.599 0.518 0.687 0.598 0.548 0.630 0.331

64

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, D. Jared, and P. Wu

Table B.4. Results from Fatigue Test (400 ).

Agg. Source
Mt. View Mt. View Mt. View Mt. View Mt. View Mt. View Mt. View Mt. View Mt. View Mt. View Mt. View Mt. View Lithia Spr. Lithia Spr. Lithia Spr. Lithia Spr. Lithia Spr. Lithia Spr. Lithia Spr. Lithia Spr. Lithia Spr. Lithia Spr. Lithia Spr. Lithia Spr.
Camak Camak Camak Camak Camak Camak Camak Camak Camak Camak Camak Camak Ruby Ruby Ruby Ruby Ruby Ruby Ruby Ruby Ruby Ruby Ruby Ruby

% RAP
0 0 0 10 10 10 20 20 20 30 30 30 0 0 0 10 10 10 20 20 20 30 30 30 0 0 0 10 10 10 20 20 20 30 30 30 0 0 0 10 10 10 20 20 20 30 30 30

RAP Type
DG1 DG1 DG1
-4 -4 -4 DG2 DG2 DG2 +4 +4 +4 -4 -4 -4 DG1 DG1 DG1 +4 +4 +4 DG2 DG2 DG2 +4 +4 +4 DG2 DG2 DG2 -4 -4 -4 DG1 DG1 DG1 DG2 DG2 DG2 +4 +4 +4 DG1 DG1 DG1 -4 -4 -4

% Air Voids
6.8 6.5 6.9 6.6 6.3 6.3 5.3 5.5 5.7 7.4 6.9 6.5 6.4 5.9 5.4 5.4 5.6 6.0 6.1 6.1 6.7 5.8 5.2 5.1 5.8 6.0 6.4 6.0 5.8 5.3 5.2 7.0 6.8 5.7 5.0 5.1 6.7 6.2 6.2 5.6 6.3 5.4 6.2 5.7 5.5 6.9 6.4 6.2

65

Nf
1,202,320 6,365,620 3,236,690 5,319,730 4,017,630 6,123,740 1,755,400 2,890,920 2,367,890 2,450,210 342,580 1,205,860 6,098,480 4,917,110 5,562,870 5,434,130 3,055,130 2,568,770 6,648,340 2,593,300 5,134,130 2,274,170 2,053,980 268,000 3,062,380 3,576,160 6,421,250 3,149,900 4,033,030 6,877,810 3,163,080 4,225,320 679,370 3,771,280 4,144,106 3,656,092 5,528,870 4,984,710 5,701,660 5,364,710 2,020,360 6,271,490 8,368,160 3,557,870 1,454,590 543,000 1,091,240 645,220

Initial Stiffness,
MPa 3,102 3,374 3,632 3,194 3,236 3,921 3,660 3,366 3,430 4,314 5,572 4,260 3,564 3,738 3,656 3,972 3,896 3,667 3,472 3,674 3,494 4,160 4,126 4,213 3,480 3,252 2,830 3,605 3,463 3,644 3,542 3,077 2,798 2,858 3,267 3,730 3,636 3,568 3,302 3,726 3,466 3,519 3,074 3,488 3,623 4,411 4,863 4,588

Diss. Energy (kPa) 0.247 0.277 0.288 0.246 0.254 0.302 0.272 0.277 0.288 0.197 0.197 0.203 0.277 0.298 0.295 0.298 0.299 0.278 0.278 0.285 0.275 0.296 0.265 0.894 0.277 0.273 0.251 0.513 0.513 0.288 0.265 0.247 0.224 0.238 0.258 0.258 0.221 0.224 0.224 0.311 0.296 0.275 0.257 0.289 0.275 0.308 0.321 0.321

Watson, D. E., A. Vargas-Nordcbeck, J. R. Moore, P. Wu, and D. Jared

Table B.5. Results from Fatigue Test (800 ).

Agg. Source

% RAP

RAP Type

% Air Voids

Nf

Mt. View

0

DG1

6.7

78,940

Mt. View

0

DG1

6.1

35,440

Mt. View

0

DG1

6.6

23,020

Mt. View

10

-4

6.7

62,700

Mt. View

10

-4

5.2

57,900

Mt. View

10

-4

6.0

29,830

Mt. View

20

DG2

5.0

37,600

Mt. View

20

DG2

6.3

37,420

Mt. View

20

DG2

5.4

18,020

Mt. View

30

+4

6.9

22,860

Mt. View

30

+4

7.2

20,520

Mt. View

30

+4

6.7

16,260

Lithia Spr.

0

-4

7.0

49,280

Lithia Spr.

0

-4

5.9

73,420

Lithia Spr.

0

-4

7.0

53,410

Lithia Spr.

10

DG1

5.9

44,850

Lithia Spr.

10

DG1

5.6

32,990

Lithia Spr.

10

DG1

5.5

56,790

Lithia Spr.

20

+4

6.3

43,640

Lithia Spr.

20

+4

6.4

80,150

Lithia Spr.

20

+4

6.1

50,030

Lithia Spr.

30

DG2

5.2

25,850

Lithia Spr.

30

DG2

5.4

11,150

Lithia Spr.

30

DG2

5.0

13,260

Camak

0

+4

5.7

109,140

Camak

0

+4

5.0

109,470

Camak

0

+4

5.9

57,600

Camak

10

DG2

5.5

58,090

Camak

10

DG2

6.5

21,880

Camak

10

DG2

5.6

42,870

Camak

20

-4

5.0

82,650

Camak

20

-4

6.9

33,500

Camak

20

-4

6.1

97,060

Camak

30

DG1

5.0

93,550

Camak

30

DG1

5.0

65,730

Camak

30

DG1

6.1

65,000

Ruby

0

DG2

6.8

78,700

Ruby

0

DG2

5.2

50,660

Ruby

0

DG2

6.5

88,680

Ruby

10

+4

6.8

59,420

Ruby

10

+4

7.0

26,920

Ruby

10

+4

5.6

36,460

Ruby

20

DG1

6.4

37,170

Ruby

20

DG1

5.8

11,390

Ruby

20

DG1

5.6

14,810

Ruby

30

-4

6.6

5,180

Ruby

30

-4

6.7

4,340

Ruby

30

-4

6.7

3,300

Ini. Stiffness, MPa 2,929 3,030 2,898 3,156 3,599 2,752 4,480 4,277 4,219 4,372 4,331 4,243 3,003 3,325 2,941 3,506 3,480 3,248 3,226 3,064 3,246 5,029 4,569 5,250 3,323 3,273 3,065 3,265 3,277 3,121 3,556 3,253 3,331 3,679 3,578 3,043 3,096 3,133 2,854 3,192 2,969 3,262 2,978 3,550 3,515 4,933 4,709 4,753

Diss. Energy (kPa) 0.989 1.004 0.963 0.995 0.900 0.916 0.898 0.910 0.873 0.701 0.670 0.698 1.017 1.119 0.980 1.095 1.087 1.024 1.041 0.997 1.045 0.526 0.586 0.513 0.894 0.867 0.801 1.075 1.025 1.001 1.084 1.012 0.799 1.128 0.799 1.007 0.896 0.885 0.806 1.065 1.013 1.087 0.978 1.199 1.138 0.722 0.638 0.627

66