Laboratory Study for Comparing Rutting Performance of Limestone and Basalt Superpave Asphalt Mixtures

Laboratory Study for Comparing Rutting Performance of Limestone and Ghazi G. Al-Khateeb 1 ; Taisir S. Khedaywi 2 ; Turki I. Al-Suleiman Obaidat 3 ; and Ahmad Mirwais Najib 4 Abstract: The primary objective of this research effort was to conduct a rutting performance based comparison between limestone and basalt Superpave asphalt mixtures using dynamic creep rutting tests. Two sets of mixtures were prepared using limestone and basalt aggregate, mixed with one asphalt binder having a Superpave performance grade of PG 64-10. To overcome the stripping potential of the Superpave basalt asphalt mixtures, 1% by total weight of the basalt aggregate was replaced by hydrated lime for the filler portion of the aggregate. Rutting was evaluated at four different temperatures (40, 50, 60, and 65 C) and one loading frequency of 8 Hz. Rutting test results indicated that the basalt Superpave asphalt mixtures exhibited superior performance relative to the limestone Superpave asphalt mixtures. The difference in the rut depth at 19,200 loading cycles between the limestone and basalt asphalt mixtures was statistically significant at levels of α ¼ 1, 5, 1, and 0.5% for the temperatures 40, 50, 60, and 65 C, respectively. The difference in the rut depth at 200,000 loading cycles between the two asphalt mixtures was statistically significant at levels of α ¼ 1, 5, 0.1, and 0.1% for the temperatures 40, 50, 60, and 65 C, respectively. In addition, the difference in the number of loading cycles to rutting failure between limestone and basalt asphalt mixtures was also statistically significant at a level of α ¼ 0.1% for all temperatures. DOI: 10.1061/(ASCE)MT.1943-5533.0000519. 2013 American Society of Civil Engineers. CE Database subject headings: Limestone; Creep; Asphalts; Mixtures; Laboratory tests. Author keywords: Basalt; Limestone; Superpave; Performance; Dynamic creep test; Rutting; Asphalt mixture. Introduction Rutting is considered to be one of the primary types of distresses affecting design life of hot-mix asphalt (HMA) pavements worldwide. In Jordan, limestone aggregate is the most common aggregate type. It is used in the construction of roads, and the Marshall mix design method is the only procedure that has been commonly used in the design of asphalt mixtures. Searching for other aggregate types that are available, including basalt aggregate, is becoming vital in Jordan. Research studies conducted around the world have proven that aggregate consensus properties play an important role in controlling rutting. In addition, aggregate type and gradation are among the major factors that affect rutting occurrence in asphalt pavements. Several researchers have used the basalt aggregate in asphalt mixtures for different purposes. Kandhal and Cooley (2002) 1 Associate Professor of Civil Engineering and Research Scientist, Vice Dean of Engineering, Dept. of Civil Engineering, Jordan Univ. of Science and Technology, P.O. Box 3030, Irbid 22110, Jordan (corresponding author). E-mail: ggalkhateeb@just.edu.jo 2 Professor of Civil Engineering, Dept. of Civil Engineering, Jordan Univ. of Science and Technology, P.O. Box 3030, Irbid 22110, Jordan. E-mail: khedaywi@just.edu.jo 3 Professor of Civil Engineering, Dept. of Civil Engineering, Jordan Univ. of Science and Technology, P.O. Box 3030, Irbid 22110, Jordan. E-mail: turk957@just.edu.jo 4 Graduate Research Assistant, Dept. of Civil Engineering, Jordan Univ. of Science and Technology, P.O. Box 3030, Irbid 22110, Jordan. E-mail: mirwais_najibkk21@yahoo.com Note. This manuscript was submitted on August 2, 2011; approved on March 16, 2012; published online on March 21, 2012. Discussion period open until June 1, 2013; separate discussions must be submitted for individual papers. This paper is part of the Journal of Materials in Civil Engineering, Vol. 25, No. 1, January 1, 2013. ASCE, ISSN 0899-1561/ 2013/1-21-29/$25.00. and Buchanan (2000) found through mix design results that, on average, limestone produced mixes with lower voids in mineral aggregate (VMA) than those prepared with basalt aggregate (attributable to the crushing of limestone aggregate during compaction). Asi (2007) concluded that the use of basalt is preferred over limestone to improve the skid resistance of asphalt pavements. Al-Shweily (2002) in a comparison study indicated that the creep deformation of control mixes prepared using basalt aggregate was less than that for those prepared using limestone aggregate. Furthermore, Asi et al. (2009) evaluated the use of basalt in asphalt concrete mixes and found that basalt coarse aggregate, limestone fine aggregate and mineral filler, and 1% hydrated lime by total weight of aggregate was the optimal mix among the other mixes used in their study. Several researchers have used hydrated lime in asphalt mixtures, particularly when basalt aggregate was used to reduce stripping and moisture sensitivity. For instance, Little and Epps (2001) studied the benefits of hydrated lime in hot-mix asphalt (HMA). They concluded that lime has the ability to improve the resistance of HMA mixtures to moisture damage, reduce oxidative aging, improve mechanical properties, and improve resistance to fatigue and rutting. Consequently, the addition of lime led to observed improvements in the field performance of lime-treated HMA pavements. Life cycle cost analyses in their study also showed that using lime resulted in approximate savings of $20/ton of HMA mix, whereas field performance data exhibited an increase of 38% in the expected pavement life. In addition, lime reduced stripping and acted as a mineral filler to stiffen the asphalt binder and HMA, which reduced rutting. Sebaaly (2007) investigated the effect of lime and liquid additives on the moisture damage of HMA mixtures. He concluded that the lime-treated HMA mixtures improved the rutting resistance of the undamaged HMA pavements for both the Nevada and California mixtures. McCann and Sebaaly (2003) also evaluated the mechanical properties of lime-treated HMA mixtures before JOURNAL OF MATERIALS IN CIVIL ENGINEERING ASCE / JANUARY 2013 / 21

and after multiple cycles of freeze thaw moisture conditioning. The mechanical tests measured the resilient modulus, tensile strength, and simple shear. In addition, the three test procedures were compared for evaluating the moisture sensitivity of HMA mixtures. They concluded that with the addition of lime and after multiple cycles of freeze thaw moisture conditioning, all asphalt mixtures demonstrated an enhanced ability to retain the original measured properties. In this study, two types of aggregates, limestone and basalt, were used to produce Superpave asphalt mixtures. The two types of asphalt mixtures were evaluated for rutting performance using dynamic creep tests to compare basalt mixtures with limestone mixtures. Basalt is a volcanic material, which is derived from the magma. Basalt can be used in many industrial applications and as construction materials. In Jordan, basalt is spread over different localities (Natural Resources Authority 2007). Objectives The primary objectives of this study are the following: 1. Utilize the Superpave system to design limestone and basalt Superpave asphalt mixtures, 2. Evaluate limestone and basalt Superpave asphalt mixtures for rutting performance using the dynamic creep tests, and 3. Compare basalt asphalt mixtures with limestone asphalt mixtures in terms of rutting performance. Methodology Fig. 1 shows a schematic representation of the experimental plan for this study. The experimental plan was conducted in four primary stages (Najib 2010), as shown next. Stage 1 included collection, physical evaluation, and characterization of limestone from Al-Huson, basalt from Al-Hallabat, and asphalt binder PG 64-10 from Jordan Petroleum Refinery (JPR). Results of this stage are shown in Tables 1 and 2. Stage 2 covered the use of the Superpave mix design method to design limestone and basalt Superpave asphalt mixtures. Stage 3 covered the preparation of specimens for limestone and basalt asphalt mixtures using the Superpave gyratory compactor (SGC). Stage 4 included testing limestone and basalt Superpave asphalt mixtures for rutting performance, using dynamic creep tests to compare the results. Evaluation and testing of limestone and basalt aggregates were performed using traditional and Superpave test methods. The traditional aggregate tests to measure the source properties included the following: specific gravity and absorption of coarse aggregate, specific gravity and absorption of fine aggregate, and Los Angeles (LA) abrasion. The Superpave aggregate tests to measure consensus properties included the following: flat and elongated (F&E) particles, coarse aggregate angularity (CAA), fine aggregate angularity (FAA), and sand equivalent (SE). Table 1 shows the results of the source and consensus aggregate properties for limestone and basalt aggregates. Evaluation of the asphalt binder used in this study covered three stages. Stage 1 evaluated the asphalt binder using the traditional asphalt binder tests, including penetration, ductility, softening point, flash and fire points, and specific gravity. Stage 2 tested the asphalt binder using the Superpave asphalt binder tests, including the rotational viscosity (RV), dynamic shear rheometer (DSR), rolling thin-film oven (RTFO), pressure aging vessel (PAV), and bending beam rheometer (BBR) tests. Stage 3 classified the asphalt binder according to the Superpave asphalt binder performance grading system. A series of tests were performed in this stage, including the DSR test for fresh and RTFO asphalt binder at high temperatures, RTFO test for fresh asphalt binder, PAV test for RTFO asphalt binder, DSR test for PAV asphalt binder at intermediate temperatures, and BBR test for PAV asphalt binder at low temperatures. The performance grade for the asphalt binder used in this study was PG 64-10 according to the Superpave system. Table 2 presents the results of the traditional and Superpave tests for this asphalt binder. Collection, Physical Evaluation and Characterization of Asphalt Binder, Limestone, and Basalt Design of Asphalt Mixtures Using Superpave Mix Design Procedure Testing Limestone and Basalt Superpave Asphalt Mixtures For Rutting Performance Performance Rutting-Based Comparison Between Limestone and Fig. 1. Schematic representation of the experimental plan 22 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING ASCE / JANUARY 2013

Table 1. Properties, Criteria, and Tests of the Limestone and Basalt Aggregates Test Coarse aggregate Limestone Fine aggregate Coarse aggregate Basalt Fine aggregate Criteria Standard test method G sb-bulk 2.605 2.507 2.677 2.789 NA AASHTO T 84/85 (AASHTO 2002a) G sb-ssd 2.699 2.549 2.737 2.858 AASHTO T 84/85 (AASHTO 2002a) G sa 2.876 2.712 2.847 2.996 AASHTO T 84/85 (AASHTO 2002a) Absorption (%) 3.624 4.700 2.240 2.461 AASHTO T 84/85 (AASHTO 2002a) Flat and elongated (F&E) particles 0 NA 0 NA 10 (max) ASTM D4791 (ASTM 2010) Coarse aggregate angularity (CAA) 91=97 NA 95=97 NA 95=90 AASHTO TP 61 (AASHTO 2002b) Fine aggregate angularity (FAA) NA 45 NA 46 45 (min) AASHTO T 304 (AASHTO 2011a) Sand equivalent (SE) (%) NA 58 NA 84 45 (min) AASHTO T 176 (AASHTO 2008) Los Angeles (LA) abrasion mass loss (%) 27 NA 27 N.A 45 (max) AASHTO T 96 (AASHTO 2002c) Note: G sb-bulk = bulk specific gravity, G sb-ssd = saturated surface-dry bulk specific gravity, and G sa = apparent specific gravity. Table 2. Test Properties of Asphalt Binder Test Standard test method Test result Rotational viscosity at 135 C AASHTO T 316 650 (mpa.s ¼ cp) (AASHTO 2011b) Rotational viscosity at 160 C AASHTO T 316 220 (mpa.s ¼ cp) (AASHTO 2011b) Penetration at 25 C (mm) ASTM D5 (ASTM 2006) 64 Specific gravity at 25 C ASTM D70 (ASTM 2009) 1.02 Ductility at 25 C (cm) ASTM D113 (ASTM 2007) 100þ Flash point ( C) ASTM D92 (ASTM 2012) 318 Fire point ( C) ASTM D92 (ASTM 2012) 322 Softening point ( C) ASTM D2398 (ASTM 1982) 59 The design of limestone and basalt asphalt mixtures using the Superpave mix design method was conducted in this study. Details of the design process are not shown in this paper because they are beyond the paper s scope. However, the Superpave mix design procedure described in the Asphalt Institute (AI) SP-2 manual (1996) was followed in the design process. Table 3 presents the design asphalt binder content and the volumetric properties for both limestone and basalt asphalt mixtures designed using the Superpave mix design method. The volumetric properties include the voids in mineral aggregate (VMA), voids filled with asphalt (VFA), dust proportion (DP), percentage of theoretical maximum specific gravity (G mm ) at the initial number of gyrations (%G mm @N initial ), percentage of theoretical maximum specific gravity (G mm ) at the design number of gyrations (%G mm @N design ), and percentage of Table 3. Design Asphalt Binder Content and Volumetric Properties for Limestone and Basalt Asphalt Mixtures Limestone mixtures Basalt mixtures Superpave criteria Property Design asphalt binder 5.1 4.8 content (%) Design air voids (%) VMA (%) 14.2 1 1 VFA (%) 71.6 65.0 65 75 DP 0.7 1 0.6 1.2 %G mm @N initial 88.8 87.4 89 %G mm @N design 96.0 96.0 96 %G mm @N max 96.4 97.1 98 Note: DP = Dust proportion = % filler passing No. 200 / effective asphalt binder content. theoretical maximum specific gravity (G mm ) at the maximum number of gyrations (%G mm @N max ). Materials Limestone aggregate was a crushed limestone acquired from a quarry in the Shatana area in Al-Huson town, located in northern Jordan. Basalt aggregate was acquired from quarries in Al-Hallabat area in Al-Azraq city, located in southeastern Jordan. The same aggregate gradation was used in this study for limestone and basalt for comparison purposes, as shown in Table 4. Fig. 2 presents the 0.45 power chart for this gradation. The aggregate gradation was selected based on the aggregate stockpiles and their sieve analysis obtained from the quarries and Superpave specifications for aggregate gradations with a nominal maximum aggregate size (NMAS) of 12.5 mm. The aggregate gradation used in this study passed between the control points and below the restricted zone (BRZ). The source properties and consensus properties of both limestone and basalt aggregates used in this study were measured based on laboratory testing and evaluation for these two aggregates. Table 1 shows the results. Hydrated lime was added to the basalt aggregate at 1.0% by total weight of aggregate to replace the filler portion. The addition of hydrated lime for basalt asphalt mixtures aimed at preventing stripping and improving the asphalt coating and bond between the basalt aggregate and the asphalt binder. The asphalt binder used in this study was obtained from Jordan Petroleum Refinery (JPR) in Zarqa city of Jordan, with a penetration grade of 60=70 and a Superpave performance grade of PG 64-10. The physical and mechanical properties for this asphalt binder were measured based on laboratory testing for this material Table 4. Aggregate Gradation Used in this Study Sieve size (mm) Passing (%) 19.0 100 12.5 91 9.5 70 4.75 50 2.36 35 1.18 22 0.600 15 0.300 12 0.150 5 75 3 JOURNAL OF MATERIALS IN CIVIL ENGINEERING ASCE / JANUARY 2013 / 23

% Passing 10 8 6 4 Target Blend Control Points Restricted Zone 2 75 0.300 0.600 1.18 2.36 4.75 9.5 Sieve Size (mm) 12.5 19.0 Fig. 2. 0.45 power chart for aggregate gradation used in this study Fig. 4. Universal testing machine using traditional and Superpave asphalt binder tests. Table 2 presents the results. Rutting Performance Testing Preparation of Superpave Specimens for Rutting Tests Specimens for rutting performance testing were prepared using the Superpave gyratory compactor (SGC). After determining the design asphalt binder content and measuring G mm at the design asphalt binder content, trial specimens for both limestone and basalt Superpave asphalt mixtures were prepared to achieve 4% air voids. The specimens (Fig. 3) were compacted using the height mode of SGC compaction. The height was fixed at 115 mm. Dynamic Creep Tests After preparing all rutting specimens using the SGC, the rutting performance test was conducted using the Universal Testing Machine (UTM) that is available in the authors laboratory (Fig. 4) based on the following steps: 1. The dynamic creep test set-up was mounted in the UTM (Fig. 5). 2. Prior to the test, specimens were conditioned at the required test temperatures for a sufficient length of time, and linear Fig. 3. SGC specimens prepared for rutting testing Fig. 5. Dynamic creep setup mounted in the universal testing machine variable differential transformers (LVDTs) were mounted on the specimen before testing. 3. Using the software templates for rutting performance testing, required settings and parameters of the UTM were entered. 4. The temperature of the chamber was controlled at the target test temperature using the UTM machine temperature controller. 5. An axial load of 2.5 kn, loading frequency of 8 Hz, and four different temperatures (40, 50, 60, and 65 C) were used. Three replicates were tested at each condition. 6. The data acquisition system (DAS) of the UTM during the test collected the test data from the different channels. Fig. 6 illustrates typical rutting test data. 7. In dynamic creep testing, asphalt mixtures typically experience three stages of rutting, as schematically illustrated in Fig. 6. 8. The dynamic creep test was conducted until the specimen reached failure (tertiary stage). The primary stage, also known as the initial stage, is the first stage that identifies the initial effect of loading cycle on the specimen, and in which the compacted asphalt mixture experiences additional compaction because of the vertical loading applied axially on the specimen. In this stage, the slope is high, which represents a rapid increase in the rut depth, whereas in the second stage, the actual rut takes place in the asphalt mixture because of the effect of 24 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING ASCE / JANUARY 2013

Permanent Deformation (mm) 20 15 10 5 Primary Stage 0 0 5000 10000 15000 20000 25000 Load Cycles dynamic loading on the specimen. The final stage, known as the tertiary stage, is the last stage of the rutting test. At this stage, the asphalt mixture starts to exhibit a noticeable damage and flow attributable to the effects of rutting, as shown in Fig. 7. Reduction of Test Data Secondary Stage Tertiary Stage Fig. 6. Three stages of rutting This section concerns the methodology used to reduce the test data obtained from the unconfined dynamic creep tests for rutting performance and the analysis methods used to present the final results for rutting for both types of Superpave asphalt mixtures (the limestone and the basalt aggregate asphalt mixtures). After conducting the tests and calculating the averages of three replicates for each condition, rut depth (RD) was plotted versus the number of loading cycles (ND) (Fig. 8), and models were developed using statistical regression analysis techniques (linear or non-linear). Based on the results obtained from each condition for both limestone and basalt aggregate asphalt mixtures using the dynamic creep tests, several plots were prepared in a manner intended to show the differences between the behavior of both materials at different conditions. The failure point for rutting (the number of loading cycles to rutting failure) was identified based on the criterion and method described in Al-Khateeb and Basheer (2009). In this method, the onset of the secondary and tertiary stages may be obtained graphically by extending tangents of the nearly linear segments around the transition point. The corresponding N in which the two tangents intersect can then be determined on the x-axis. Test Results and Analysis Three parameters were obtained from the dynamic creep tests to compare the two types of asphalt mixtures: the rut depth (RD) at 19,200 loading cycles, the rut depth (RD) at 200,000 loading cycles, and the number of loading cycles to rutting failure as described previously. The 19,200 loading cycles was selected because it was used in the literature for the Hamburg wheel tracking device (HWTD) rutting test for different asphalt mixtures and found to be a reasonable number of loading cycles to differentiate between the rutting performance of asphalt mixtures. Table 5 presents the rut depths for both limestone and basalt Superpave asphalt mixtures at 19,200 and 200,000 cycles at the four temperatures. The reduction in the rut depth at 19,200 cycles between limestone and basalt asphalt mixtures at the four test temperatures was determined, as shown in Table 5. It ranged between 50 and 75%, which is considered a dramatic decrease in rutting when the basalt aggregate is used. Fig. 9 illustrates in a histogram the difference in behavior of both limestone and basalt Superpave asphalt mixtures at 19,200 cycles. In contrast, obvious longer rutting life was obtained for basalt asphalt mixtures over limestone asphalt mixtures, and that was clear from the low values of rut depths at 200,000 cycles for basalt mixtures compared with limestone mixtures, as shown in Table 5. This indicates a reduction in the rut depth at 200,000 cycles between limestone and basalt asphalt mixtures of 78, 57, 88, and 92% at the temperatures 40, 50, 60, and 65 C, respectively. Statistical hypothesis testing was also conducted using the t-test to check if the differences in the rut depths between the basalt asphalt mixtures and the limestone asphalt mixtures were significant. Table 5 summarizes the results of the t-test hypothesis 1 1 8.0 6.0 0 10,000 20,000 30,000 40,000 50,000 60,000 70,000 Fig. 8. Sample of rutting performance test results Fig. 7. Specimen before and after unconfined dynamic creep rutting test JOURNAL OF MATERIALS IN CIVIL ENGINEERING ASCE / JANUARY 2013 / 25

Table 5. Rut Depth of Limestone and Basalt Asphalt Mixtures Rut depth (mm) at 19,200 cycles Temperature ( C) Limestone mixtures Basalt mixtures Reduction in RD (%) t exp t stat ðdf; αþ 40 0.8 0.2 75 4.90 t ð4; 1Þ ¼ 4.6 50 1.6 0.8 50 3.64 t ð4; 5Þ ¼ 2.78 60 4.6 1.3 72 5.07 t ð4; 1Þ ¼ 4.60 65 6.5 1.6 75 6.09 t ð4; 05Þ ¼ 5.60 Rut depth (mm) at 200,000 Cycles Temperature ( C) Limestone mixtures Basalt mixtures Reduction in RD (%) t exp t statistic ðdf; αþ 40 1.6 0.4 78 5.18 t ð4; 1Þ ¼ 4.60 50 4.2 1.8 57 3.20 t ð4; 5Þ ¼ 2.78 60 21.7 2.6 88 14.84 t ð4; 01Þ ¼ 8.61 65 41.8 3.2 92 15.14 t ð4; 01Þ ¼ 8.61 Note: t exp = calculated t value from experiment, t-stat (df, a) = statistical t value at df degrees of freedom and α level. 7.0 6.0 5.0 3.0 1.0 Basalt Superpave Aspahlt Mixtures Limestone Superpave Aspahlt Mixtures 40 50 60 65 Temperature ( C) Fig. 9. Rut depth at 19,200 cycles for limestone and basalt Superpave asphalt mixtures at four different temperatures for the rut depths at 19,200 and 200,000 cycles. The difference in the rut depth at 19,200 cycles between the two mixtures was statistically significant at levels of α ¼ 1, 5, 1, and 0.5% for the temperatures 40, 50, 60, and 65 C, respectively. In contrast, the difference in the rut depth at 200,000 cycles between the two mixtures was statistically significant at levels of α ¼ 1, 5, 0.1, and 0.1% for the temperatures 40, 50, 60, and 65 C, respectively. The relationship between rut depth and number of loading cycles was also plotted at the four test temperatures for both limestone and basalt asphalt mixtures, as shown in Figs. 10(a and b). The basalt asphalt mixtures were superior relative to the limestone asphalt mixtures, particularly at the temperatures 60 and 65 C. For limestone asphalt mixtures, rutting failure took place for these two temperatures (60 and 65 C) at 48,000 and 37,000 loading cycles, respectively. In contrast, for basalt asphalt mixtures at the same temperatures (60 and 65 C), failure occurred at 167,800 and 147,400 loading cycles, respectively. This difference between the two mixtures was statistically significant at a level of α ¼ 0.1% for all temperatures (40, 50, 60, and 65 C). Table 6 summarizes the results for the statistical hypothesis testing using the t-test. Table 6 presents the results for the number of loading cycles to rutting failure for both asphalt mixtures. The increase in the number of loading cycles to failure between limestone and basalt asphalt mixtures was 106, 104, 250, and 298% at the temperatures 40, 50, 60, and 65 C, respectively. Statistical models were developed between rut depth (RD) and number of loading cycles (ND) at the four temperatures using regression analysis techniques. The secondary stage was used in 1 1 8.0 6.0 0 (a) 3.5 3.0 2.5 1.5 1.0 0.5 40C 50C 60C 65C 40C 50C 60C 65C 50,000 100,000 150,000 200,000 0 50,000 100,000 150,000 200,000 (b) Fig. 10. Rut depth of limestone and basalt Superpave asphalt mixtures at four different temperatures: (a) limestone asphalt mixtures; (b) basalt asphalt mixtures this analysis because it represents the actual rutting that takes place for asphalt mixtures. The other two stages of rutting are used to determine the transition points between the primary and secondary stages and between the secondary and tertiary stages, in addition to the failure point when flow or damage of the asphalt mixture occurs. Tables 7 and 8 summarize the models developed for limestone and basalt asphalt mixtures, respectively. The linear model was the best type to fit the rutting results in the secondary stage. The coefficient of simple determination (r 2 ) for these linear models was high and ranged from 0.86 0.95 and 0.83 0.99 for basalt and limestone mixtures, respectively. By plotting the rut depth versus the test temperature in a scatter plot, and fitting the best-fit model for both sets of data (Fig. 11), the 26 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING ASCE / JANUARY 2013

Table 6. Number of Loading Cycles to Rutting Failure of Limestone and Basalt Asphalt Mixtures Number of loading cycles to rutting failure Temperature ( C) Limestone mixtures Basalt mixtures Increase in ND RF (%) t exp t statistic ðdf; αþ 40 540,000 1,110,650 106 8.84 t ð4; 01Þ ¼ 8.61 50 250,000 510,450 104 8.76 t ð4; 01Þ ¼ 8.61 60 48,000 167,800 250 9.28 t ð4; 01Þ ¼ 8.61 65 37,000 147,400 298 9.71 t ð4; 01Þ ¼ 8.61 Note: ND RF = number of loading cycles to rutting failure. Table 7. Linear Regression Models for Limestone Superpave Asphalt Mixtures Temperature ( C) Model r 2 40 RD ¼ 5E-06 ðndþþ0.8243 0.83 50 RD ¼ 2E-05 ðndþþ1.3778 0.96 60 RD ¼ 001 ðndþþ1.7004 0.99 65 RD ¼ 002 ðndþþ1.7934 0.99 Note: RD = rut depth (mm), and ND = number of loading cycles. The developed models indicate that basalt asphalt mixtures do not accumulate rutting (permanent deformation) fast with the increase in temperature; the behavior is linear. In contrast, limestone asphalt mixtures do rut faster with the increase in temperature in an exponential form. Wherever there is a dramatic change in temperature, basalt mixtures rather than limestone mixtures are recommended to prevent a rapid rate of change (increase) in rutting occurrence for asphalt pavements. Figs. 12 15 describe a pair comparison in rutting behavior between limestone asphalt mixtures and basalt asphalt Table 8. Linear Regression Models for Temperature ( C) Model r 2 40 RD ¼ 1E-06 ðndþþ0.1577 0.86 50 RD ¼ 6E-06 ðndþþ0.7734 0.91 60 RD ¼ 7E-06 ðndþþ1.1652 0.91 65 RD ¼ 9E-06 ðndþþ1.4182 0.95 1.6 Note: RD = rut depth (mm), and ND = number of loading cycles. Rut Depth at 19,200 Cycle (mm) 7 6 5 4 3 2 1 0 Limestone Mixtures Basalt Mixtures RD = 238e 865T r 2 = 0.992 RD = 553T - 1.9949 r 2 = 0.9985 30 35 40 45 50 55 60 65 70 Test Temperature ( C) Fig. 11. Rut depth versus temperature for limestone and basalt Superpave asphalt mixtures authors found that the relationship between rut depth and temperature is exponential for limestone asphalt mixtures and linear for basalt asphalt mixtures. The coefficients of simple determination (r 2 ) for these two models are 0.992 and 0.999, respectively. With the increase in temperature, rutting always increased. However, the type of this relationship and the rate of increase or change in rutting are crucial for the different types of asphalt mixtures. In other words, the occurrence of rutting is normal at high temperatures, but the rate of change in rutting is important to determine the rutting life and performance of asphalt mixtures. Therefore, it was necessary to investigate the major difference in the rate of change in rutting with temperature between limestone and basalt asphalt mixtures. 1.2 0.8 0.4 0 50,000 100,000 150,000 200,000 Fig. 12. Rutting behavior of limestone and basalt Superpave asphalt mixtures at 40 C 4.5 3.5 3.0 2.5 1.5 1.0 0.5 0 50,000 100,000 150,000 200,000 Fig. 13. Rutting behavior of limestone and basalt Superpave asphalt mixtures at 50 C JOURNAL OF MATERIALS IN CIVIL ENGINEERING ASCE / JANUARY 2013 / 27

1 1 8.0 6.0 0 50,000 100,000 150,000 200,000 Fig. 14. Rutting behavior of limestone and basalt Superpave asphalt mixtures at 60 C 1 1 8.0 6.0 0 50,000 100,000 150,000 200,000 Fig. 15. Rutting behavior of limestone and basalt Superpave asphalt mixtures at 65 C mixtures at four different temperatures: 40, 50, 60, and 65 C, respectively. Finally, statistical analysis for the rutting test results was conducted using the analysis of variance (ANOVA) method. Based upon the number of variables (type of aggregate and temperature), two-staged nested design was used for rutting test results. Table 9 presents a brief summary of this design. Both the type of aggregate and the test temperature exerted a statistically significant effect on the rutting behavior of asphalt mixtures, which was obvious from the very high value of F o over the value of F critical. Conclusions Based on the analysis and results of this study, the following conclusions were drawn: 1. Basalt asphalt mixtures are superior relative to limestone asphalt mixtures in terms of rutting performance. This was obvious when the rut depths for basalt mixtures were compared with those for limestone mixtures at the four test temperatures and 19,200 loading cycles. The difference in the rut depth at 19,200 cycles between limestone and basalt asphalt mixtures was statistically significant at levels of α ¼ 1, 5, 1, and 0.5% for the temperatures 40, 50, 60, and 65 C, respectively. The reduction in the rut depth at 19,200 loading cycles between limestone and basalt asphalt mixtures was significant and ranged from 50 75%. 2. The difference in the rut depth at 200,000 loading cycles between limestone and basalt asphalt mixtures was statistically significant at levels of α ¼ 1, 5, 0.1, and 0.1% for the temperatures 40, 50, 60, and 65 C, respectively. 3. The reduction in the rut depth at 200,000 cycles between limestone and basalt asphalt mixtures was 78, 57, 88, and 92% at the temperatures 40, 50, 60, and 65 C, respectively. 4. The difference in the number of loading cycles to rutting failure based on the tertiary stage transition point criterion between limestone and basalt asphalt mixtures was statistically significant at level of α ¼ 0.1% for all temperatures (40, 50, 60, and 65 C). 5. The relationship between rut depth and number of loading cycles during the secondary stage of rutting was linear for both limestone and basalt asphalt mixtures, with a high coefficient of simple determination that ranged from 0.83 0.99 and 0.86 0.95 for limestone and basalt mixtures, respectively. 6. At 8 Hz loading frequency, the relationship between temperature and rutting life was highly significant (r 2 ¼ 0.992 and 0.999 for limestone and basalt mixtures, respectively). 7. The rate of change (increase) of rutting with temperature was significantly different between limestone asphalt mixtures and basalt asphalt mixtures. For basalt asphalt mixtures, the rutting behavior with temperature was linear; in contrast, it was exponential for limestone asphalt mixtures. Consequently, limestone asphalt mixtures may be risky to use for highways in areas where dramatic changes in temperature occur during summer. In contrast, basalt asphalt mixtures may be more reliable in this scenario. 8. As the temperature increased, differences between limestone and basalt asphalt mixtures in terms of rutting performance were significantly higher. 9. Basalt seemed to be able to solve rutting problems for asphalt mixtures in Jordan. Asphalt pavements of major highways in Jordan experience medium to high severity rutting (permanent deformation), particularly those highways that carry heavy vehicular loadings and connect Jordan to neighboring countries such as Iraq, Saudi Arabia, and Syria. Table 9. Summary of ANOVA Statistical Analysis for Rutting Test Results Source Degrees of freedom Sum of squares (SS) Mean of squares (MS) F 0 F critical Results Sum of square A 1 34.56 34.56 3,206.18 4.49 Significant Sum of square BðAÞ 6 66.59 11.10 1,029.58 2.74 Significant Sum of square Error 16 0.17 1 Sum of square Total 23 101.32 4.41 Note: A = type of aggregate, and B = temperature. 28 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING ASCE / JANUARY 2013

Acknowledgments The authors of this paper are grateful to the Scientific Research Fund (SRF) of the Ministry of Higher Education and Scientific Research in Jordan for their financial support. This paper is part of a research project funded financially from the SRF. References AASHTO. (2002a). Specific gravity and absorption of fine aggregate. T 84/84, Washington, DC. AASHTO. (2002b). Standard method of test for determining the percentage of fracture in coarse aggregate. TP 61, Washington, DC. AASHTO. (2002c). Standard method of test for resistance to degradation of small-size coarse aggregate by abrasion and impact in the Los Angeles machine. T 96, Washington, DC. AASHTO. (2008). Standard method of test for plastic fines in graded aggregates and soils by use of the sand equivalent test. T 176, Washington, DC. AASHTO. (2011a). Standard method of test for uncompacted void content of fine aggregate. T 304, Washington, DC. AASHTO. (2011b). Standard method of test for viscosity determination of asphalt binder using rotational viscometer. T 316, Washington, DC. Al-Khateeb, G., and Basheer, I. (2009). A three-stage rutting model utilizing rutting performance data from the Hamburg wheel-tracking device (WTD). Road Transport Res. J., 18(3), 32 45. Al-Shweily, H. (2002). Effect of bituminous mixtures stripping on creep behavior. Master thesis, Jordan Univ. for Science and Technology, Irbid, Jordan. Asi, I. (2007). Evaluating skid resistance of different asphalt concrete mixes. Build. Environ. J., 42(1), 325 329. Asi, I., Shalabi, F., and Jamil, N. (2009). Use of basalt in asphalt concrete mixes. Constr. Build. Mater. J., 23(1), 498 506. Asphalt Institute (AI). (1996). Asphalt Institute mix design program. http://www.asphaltinstitute.org/mix_design_sw/sw_2_manual_9_03.pdf (Oct. 15, 2012). ASTM. (1982). Method for softening point of bitumen in ethylene glycol (ring-and-ball). D2398, West Conshohocken, PA. ASTM. (2006). Standard test method for penetration of bituminous materials. D5, West Conshohocken, PA. ASTM. (2007). Standard test method for ductility of bituminous materials. D113, West Conshohocken, PA. ASTM. (2009). Standard test method for density of semi-solid bituminous materials (pycnometer method). D70, West Conshohocken, PA. ASTM. (2010). Standard test method for flat particles, elongated particles, or flat and elongated particles in coarse aggregate. D4791, West Conshohocken, PA. ASTM. (2012). Standard test method for flash and fire points by Cleveland open cup tester. D92, West Conshohocken, PA. Buchanan, M. (2000). Evaluation of the effect of flat and elongated particles on the performance of hot mix asphalt mixtures. Rep. No. 2000 03, National Center for Asphalt Technology, Auburn Univ., Auburn, AL. Kandhal, P., and Cooley, A. (2002). Coarse versus fine-graded Superpave mixtures: Comparative evaluation of resistance to rutting. Rep. No. 2002 02, National Center for Asphalt Technology, Auburn Univ., Auburn, Alabama. Little, D. N., and Epps, J. A. (2001) and updated by Sebaaly, P. E. (2006). The benefits of hydrated lime in hot-mix asphalt. Rep. Prepared for the National Lime Association, Arlington, VA. McCann, M., and Sebaaly, P. E. (2003). Evaluation of moisture sensitivity and performance of lime in hot-mix asphalt. Transportation Research Record 1832, Transportation Research Board, Washington, DC, 9 16. Najib, A. (2010). Performance based comparison between basalt asphalt mixtures and limestone asphalt mixtures. M.Sc. Dissertation, Dept. of Civil Engineering, Jordan Univ. of Science and Technology. Natural Resources Authority. (2007). http://www.nra.gov.jo. Sebaaly, P. E. (2007). Comparison of lime and liquid additives on the moisture damage of HMA mixtures. Rep. Prepared for the National Lime Association, Arlington, VA. JOURNAL OF MATERIALS IN CIVIL ENGINEERING ASCE / JANUARY 2013 / 29

Copyright of Journal of Materials in Civil Engineering is the property of American Society of Civil Engineers and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.