J. Cent. South Univ. Technol. (2011) 18: 1316-1320

DOI: 10.1007/s11771-011-0839-7

Evaluation of reclaimed asphalt pavement binder stiffness without extraction and recovery

MA Tao(马涛)^{1, 2}, HUANG Xiao-ming(黄晓明)¹, U. B. Hussain²

1. School of Transportation, Southeast University, Nanjing 210096, China;

2. Department of Civil and Environmental Engineering, University of Wisconsin-Madison, WI 53706, USA

? Central South University Press and Springer-Verlag Berlin Heidelberg 2011

Abstract: A new testing procedure to estimate the low-temperature stiffness of the reclaimed asphalt pavement (RAP) binder was developed. In the testing procedure, the Superpave^TM Bending Beam Rheometer (BBR) with special modifications and binder blending charts by Asphalt Institute were utilized. Modifications involved the development of a new kind of sample mold and different testing parameters were made to BBR testing procedure to capture the rheological properties of bitumen mortars produced by mixing fresh binder with fine RAP materials or RAP aggregate. The stiffness relationship between binder and bitumen mortar was established based on the BBR test results. The blended binder stiffness in bitumen RAP mortar was estimated from the RAP mortar stiffness based on the binder-mortar relationship. And finally, the RAP binder stiffness was estimated from the blended binder and fresh binder stiffness based on the blending charts by Asphalt Institute. The results indicate that the new procedure can capture the rheological properties of bitumen mortar and can be used to estimate the low temperature stiffness of RAP binder without binder extraction and/or any chemical treatments.

Key words: aged binder; reclaimed asphalt pavement; mortar; stiffness; Bending Beam Rheometer; blending chart

1 Introduction

Because of the increase of construction costs along with the rise in environmental awareness, the use of reclaimed asphalt pavement (RAP) is becoming more and more a common practice around the world [1]. However, there is still no commonly accepted design procedure for hot mixture asphalt (HMA) containing RAP [2]. Therefore, the overall objective of this research is to develop testing and analysis procedures that can be effectively used to evaluate RAP materials and optimize the performance of HMA mixtures containing RAP materials.

One of the main barriers to a wide RAP usage is the difficulty of evaluating the properties of the binder in an aged pavement. The current method of testing the properties of aged binder in the RAP is done by the extraction and recovery of asphalt binder with solvents method [3-4]. However, research studies [5-6] have consistently shown that this method is not accurate and property alteration of bitumen binder during solvent extraction is the most concern. Therefore, a new testing procedure to estimate the properties of bitumen binder more precisely should be developed.

In general, the inclusion of RAP materials in the HMA design can improve its resistance to rutting while it may significantly jeopardize its resistance to thermal cracking [7]. Therefore, as the first critical step of the whole research effort, this work focuses on testing the procedure development to estimate the low-temperature rheological properties of bitumen-RAP mortar and to facilitate the low-temperature rheological properties evaluation of RAP binder.

2 Hypothesis and research methodology

Based on the previous study and trial tests, the following hypothesis can be summarized.

According to Bonnaure Model [8], Eq.(1) depicts the relationship between binder and mixture stiffness:

                                           (1)

where α and β are parameters calculated from the volume fractions of binder and aggregate in mixture; S_b is the stiffness modulus of bitumen; S_m is the stiffness modulus of mixture.

Fractionating the RAP material at the 8^# sieve and using the passing 8^# fraction for evaluating the aged binder properties provide a convenient and accurate asphalt blending analysis. This also allows a more accurate representation of the blending that occurs between aged and virgin binders [9]. Therefore, based on Eq.(1), the blending binder properties in bitumen-RAP mortar consisting of passing 8^# RAP and certain amounts of fresh binder can be evaluated from bitumen-RAP mortar properties.

Meanwhile, the properties of aged binder on the RAP aggregates determine the properties of virgin asphalt that needs to be added to new mixtures containing RAP [10]. Based on the requirements of binder property in new HMA mixture, the property of virgin asphalt and the content limit of RAP used in the mixture could be estimated. The blending charts by Asphalt Institute [11] can be used to estimate RAP binder properties from the blended binder and fresh binder properties and can be represented by the following equation:

              (2)

where S and v are the stiffness and the content of bitumen binder and the subscripts “blended”, “aged” and “fresh” represent blended, aged and fresh bitumen binder.

The procedure using existing Superpave Binder equipment will eliminate the need for the asphalt extraction and recovery. The bending beam rheometer (BBR) is the Superpave^TM standard method for determining binder low-temperature properties. A modified version of this test is used in this work to obtain the bitumen-RAP mortar properties. RAP materials passing sieve 8^# are blended with fresh binder to generate voidless bitumen-RAP mortar samples, which are then tested using the BBR.

3 Materials

One binder and one RAP source were used in this work. The asphalt binder used was a PG 64-22 from Flint Hills, commonly used in Wisconsin, and the RAP was a 5/8″ RAP, supplied by Pitlick & Wick. This RAP was recovered from a county trunk highway in the State of Wisconsin, USA.

In order to prepare the bitumen-RAP mortars to be tested, the RAP material was sieved on the 8^# sieve. The asphalt content was determined by burning it in an ignition oven using standard AASHTO protocol. The asphalt content was found to be 5% (mass fraction).

4 Experimental design

4.1 Mortar sample preparation

Bitumen-RAP mortar was prepared by mixing RAP material with 15% (mass fraction) of fresh PG 64-22 binder. Three kinds of RAP materials were used as shown in Fig.1.

1) One is the material supplied by Pitlick and Wick, sieved at 8^# sieve, hereinafter referred to as “Natural RAP”.

2) The second one, called “Artificial RAP” consists of the recovered RAP aggregates component of the same material of the natural one (therefore with the same size distribution), mixed with 5% (mass fraction) aged binder, obtained by processing the PG 64-22 to two full cycles of PAV aging, in order to simulate the aging of an in-service pavement at the end of its service life.

3) The third type of RAP material is, in reality, a control mixture, obtained by mixing the same RAP aggregates used for the artificial RAP with 5% (mass fraction) fresh PG 64-22 binder and referred to as “Fresh RAP”.

Thus, three kinds of mortars were prepared as “Natural mortar”, “Artificial mortar” and “Fresh mortar”.

4.2 BBR modification

In order to test the mortar using the BBR, the standard BBR procedure had to be modified. The modified mold produced samples with end cross- sectional dimensions of 12.7 mm (b)×10 mm (h) instead of 12.7 mm (b)×6.35 mm (h). The new thickness is more than four times the maximum aggregate size (passing 8^#≤2.36 mm) of the SRAP, and thus it is believed that there is no interference of the maximum size with response measured. Based on the elementary bending theory [12], the shear effect, due to the increased thickness, contributes to only 2% of the center deflection, which is deemed acceptable. Based on many trials, it is found that the plastic strips commonly used in the BBR should be replaced by Teflon tape. Also, the use of screws for assembly of the end pieces, instead of O-rings, is necessary to allow the pressure to be applied during molding, while keeping the thickness of specimen constant. This modification eases the molding, trimming, and de-molding processes. Figure 2 shows a picture of the modified BBR molds and mortar samples.

Fig.1 Preparation of mortar samples

Fig.2 BBR molds and samples for bitumen-RAP mortar:     (a) Side bar covered by Teflon; (b) Assembled mold by end screws; (c) Mortar samples

After the initial testing, testing temperature, 0 °C, was selected with a testing load of 4 000 mN. These parameters allowed us to obtain acceptable values of deflection of mortar specimens, which leads to reasonable stiffness of mortar with passing 8^# SRAP in BBR.

4.3 Testing and correlation between binder and mortar

Regular BBR tests on fresh and aged pure binders were performed at the regular superpave grade temperature: for the PG 64-22 binder, the temperature chosen was -12 °C. The modified BBR tests on different mortar samples were performed at 0 °C.

In order to correlate the resulting data of binder with those acquired for the mortar (at 0 °C), an estimation of the binder stiffness at 0 °C was necessary. Since the stiffness of the samples at this temperature would be too low to allow for convenient deflections, a further set of tests was run at -18 °C and the results were used to build a master curve for each sample, by finding the adequate shift factor, according to the time-temperature superposition principle [13-14]. Also, the values were used to estimate the temperature sensitivity and estimate a temperature correction factor. Assuming the shift factors-temperature relationship to be linear, the correction factor for 0 °C was found and the stiffness values were estimated and used in the analysis.

As shown in Fig.1, the binder in artificial mortar was blended binder consisting of fresh binder and “2PAV” aged binder. Therefore, an estimation of the blended binder stiffness was needed to be correlated with the artificial mortar stiffness. Since the fresh binder and “2PAV” binder were already tested, the stiffness of the blended binder can be estimated based on Eq.(2).

The stiffness values for binder (fresh and aged) were correlated to values of mortars (fresh and artificial) by plotting the two sets at equal loading time and temperature. The plots indicated a very good linear correlation. The constants of the linear correlations defined the binder to obey mortar conversion equation.  The same equation was used to estimate the stiffness of the binder in the natural RAP, by assuming that it remains valid, and the stiffness value for the mortar is known, which can lead to a good estimation of the binder in the mortar.

5 Results and discussion

Data from BBR tests on mortars at 0 °C are reported in Table 1. Data from BBR tests on fresh binder and “2PAV” binder at -12 °C and -18 °C are also reported in Table 1. Stiffness at 0 °C for fresh and “2PAV” binder are calculated based on stiffness at -12 °C and -18 °C, respectively, and reported in Table 1. The stiffness of blended binder in artificial mortar is calculated based on the stiffness of fresh and “2PAV” binder according to Eq.(2) and also reported in Table 1.

The stiffness values (Table 1) for binder (fresh and blended) and mortar (fresh and artificial) are plotted in Fig.3(a), showing an extremely accurate linear correlation. As shown in Fig.3(a), the two fitting lines and the correlation equations, for the fresh and the artificial mortars, are very similar, proving the validity of the correlation between binder and mortar stiffness. This also proves that mortar property solely depends on the binder property since the volume and property of aggregate are the same for different mortars with different binders.

Therefore, the parameters of the fitting equation relating binder with mortar stiffness derived from the binder (both fresh and blended) and mortar (both fresh and artificial) correlations, shown in Fig.3(b), are applied to the natural mortar to back-calculate the stiffness of the blended binder (aged and fresh binder) in the natural mortar.

Table 1 Stiffness of mortars and binders (MPa)

Fig.3 Correlation between mortar and binder stiffness:       (a) Correlation; (b) Fitting

A blending chart (Fig.4), following the assumption of linear relationship between the logarithm of the stiffness and the content of RAP bitumen (Eq.(2)) is then used to estimate the stiffness of the aged binder in natural RAP material based on the stiffness of fresh and blended binder. Stiffness values at 60 s for different binders (fresh, 2PAV aged and RAP) and mortars (fresh, artificial and natural) are summarized in Fig.5.

The analysis shows that the stiffness of natural RAP mortar is slightly higher than that of artificial RAP mortar, which is reasonable, considering that the RAP material extracted from Wisconsin is more than ten years old. Moreover, the estimation of the stiffness of the binder in the natural RAP gives a higher value than the measured stiffness of the “2PAV” binder, complying with the results of the BBR tests. This proves the validity of the testing procedure for the bitumen-RAP mortar and the estimating procedure for the aged binder in RAP.

Fig.4 Blending chart to estimate stiffness of blended binder in natural RAP

Fig.5 Stiffness at 60 s for different binders and mortars:      (a) Binders; (b) Mortars

6 Conclusions

1) The protocol requires mixing fine RAP portion (passing 8^# sieve) with 15% (mass fraction) fresh binder to get mortar samples and uses the existing BBR configuration with some modifications to allow for the testing of RAP mortars for low temperature stiffness.

2) Simple binder-mortar relationships, as well as widely used linear blending charts, are found sufficient to get very good estimations of the binder properties in the RAP. Based on a linear relationship between the mortar and binder stiffness, and the concept of blending charts, the stiffness of RAP binder is estimated.

3) The RAP materials used in the experimental study are both extracted from real reclaimed pavement and prepared in the laboratory, by aging binder through repeated PAV cycles. Thus, the procedure is validated using artificially aged binder mixed with aggregates from natural RAP. Although the data are limited, the modification of the BBR appears to be simple and provides repeatable data. And the developed test procedure is proved to be reliable and relatively straight forward.

References

[1] JUNAN S, SERJI A. Effects of rejuvenator on performance-based properties of rejuvenated asphalt binder and mixtures [J]. Construction and Building Materials, 2007, 21(3): 958-964.

[2] KANDHAL P S, RAO S S. Performance of recycled hot mix asphalt mixtures [R]. Auburn: National Center for Asphalt Technology, 1995.

[3] AASHTO. Standard method of test for quantitative extraction of asphalt binder from hot mix asphalt (HMA) [S].

[4] ASTM D 2172. Standard test methods for quantitative extraction of bitumen from bituminous paving mixtures [S].

[5] STROUP-GARDINER M, NELSON J W. Use of normal propyl bromide solvents for extraction and recovery of asphalt cement [M]. Auburn: Auburn University Press, 2000: 75-82.

[6] MA Tao, ZHANG D Y. Research on influence and modification of extraction and recovery experiments for SBS modified asphalt [J]. Journal of Southeast University, 2008, 40(5): 511-523.

[7] BROWN E R. Evaluation of properties of recycled asphalt concrete hot mix [R]. Vicksburg: US Army Engineer Waterways Experiment Station, 1984.

[8] MEYERS F, TESSIER G R. Study of hot mix recycling of asphalt pavements [R]. Ottawa: Roads and Transportation Association of Canada, 1983.

[9] McDANIEL R S, SOLEYMANI H. Recommended use of reclaimed asphalt pavement in the SuperPave? mixture design method [R]. Washington: National Research Council, 2000.

[10] STROUP-GARDINER M, WAGNER C. Use of reclaimed asphalt pavement in Superpave hot-mix asphalt applications [J]. Journal of the Transportation Research Board, 1999, 1681: 1-9.

[11] MA Tao, HUANG Xiao-ming. Recycling law of aged asphalt based on composite theory of material [J]. Journal of Southeast University, 2008, 38(3): 520-524. (in Chinese)

[12] SOLEYMANI H R, BAHIA H U. Blending charts based on performance-graded asphalt binder specification [J]. Journal of the Transportation Research Board, 1999, 1661: 7-14.

[13] AL-QADI IL, LOULIZ A. Modification of bending beam rheometer specimen for low-temperature evaluation of bituminous crack sealants [J]. Journal of the Transportation Research Board, 2005, 2238: 1-26.

[14] WANG Ji, PANG Ling, WU Shao-peng, LIU Quan-tao, CHEN Zheng. Effect of ageing on fatigue properties of asphalt [J]. Journal of Central South University of Technology, 2008, 15(S1): 111-114.

(Edited by YANG Bing)

Foundation item: Project(200831800044) supported by the Ministry of Communication of China; Project(50878054) supported by the National Natural Science Foundation of China; Project(06Y31) supported by the Department of Communication of Zhejiang Province, China

Received date: 2010-05-25; Accepted date: 2010-08-09

Corresponding author: MA Tao, PhD; Tel: +86-25-83791654; E-mail: mataoseu@163.com