Annotated Literature Review for NCHRP Report 640 (2009)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

201 1.50 Watson, D. E., K. A. Moore, K. Williams and L. A. Cooley, Jr. “Refinement of New Generation Open-Graded Friction Course Mix Design.” Transportation Research Record No: 1832. Transportation Research Board. National Research Council. Washington, D.C. 2003. 1.50.1 General Starting with a discussion on the mix design developed earlier by NCAT researchers, Watson et al points out certain concerns and the need for evaluation of some test methods and specifications for the design of PFCs. Watson et al then describes the results of a laboratory study carried out with the following objectives: 1.) Determine if the Ndesign of 50 for Open Graded Friction Course is appropriate or not; 2.) How can the method for determination of air voids in OGFC mixes be improved; 3.) What are the implications of using samples compacted with the Superpave gyratory compactor (SGC) (as opposed to Marshall hammer) for Cantabro test; and 4.) Is there any advantage of changing the sieve size in the basket used for evaluation of draindown in OGFC mixes during mix design? Watson et al indicate that the study consisted of preparation of samples of OGFC with three different types of aggregates and three different types of asphalt binder, and one type of fiber. Samples were compacted at different gyrations with the SGC, as well as with the Marshall hammer, with two different compaction levels. The air voids were determined using dimensional analysis as well as the CoreLok method. Voids in coarse aggregate for the mix were compared to voids in coarse aggregate for dry-rodded condition. Breakdown of aggregates were also evaluated. Conditioned and unconditioned samples were subjected to the Cantabro abrasion loss testing. Finally, baskets with two different opening sizes were used for determination of draindown of the different mixes. Based on the results, Watson et al concluded that the originally proposed Ndesign of 50 seemed to be appropriate, that future work needed to be done for better prediction of air voids in OGFC mixes, breakdown of aggregates was significantly higher in samples compacted with the Marshall hammer, the Cantabro abrasion loss criteria should be changed to accommodate the reduced stone loss for the samples compacted with the SGC, and that a reduced draindown basket opening size could be used for the draindown testing baskets. 1.50.2 Benefits of Permeable Asphalt Mixtures In their opening paragraph, Watson et al indicate that OGFC pavements have high air void contents that allow water to be removed from the pavement through the asphalt layer. They mention that this removal of water can help to minimize splash and spray, increase visibility during rain, improve friction resistance and reduce noise levels. Watson et al also mention that OGFCs have primarily been used in the United States (U.S.) for friction resistance.

202 1.50.3 Materials and Design Watson et al indicate that three different types of aggregates and three different types of asphalt binder were used in this study. Samples were compacted with the SGC as well as the Marshall hammer. Table 101 shows the matrix of materials used. Note that the aggregates were blended to meet the medium gradation requirement. The master gradation shown in Table 101 was developed from recent experiences in Georgia – the fine gradation represents mixes originally used (for 9.5 mm NMAS OGFC mixes) in Georgia, whereas the medium and the coarse gradations represent typical porous asphalt mixes used in Europe. Table 101: Materials and Mixes Used in This Study Material/property Type Aggregate Granite, crushed gravel, and traprock Aggregate Properties The traprock is the heaviest aggregate used, followed by granite and crushed gravel respectively. In terms of LA Abrasion loss, the traprock is the toughest, followed by crushed gravel and granite. Aggregate gradation, percent passing Sieve Size Master Gradation Fine Medium Coarse 19 mm 100 100 100 100 12.5 mm 80-100 100 90 80 9.5 mm 35-60 90 47 35 4.75 mm 10-25 25 17 10 2.36 mm 5-10 10 7 5 0.075 mm 2-4 3 3 3 Asphalt Binder PG 67-22, PG 76-22 (SBS modified) and PG 76-34 (rubber modified; chemically modified within the refining process). Rubber modification process Chemically modified crumb rubber asphalt (CMCRA) was prepared by using an asphalt of PG 55-34 using ten percent chemically modified crumb rubber by weight of asphalt (minus 180µm – No. 80 – mesh rubber), followed by the addition of 1.7% Styrene Butadiene Styrene (SBS) in continuous stirring mode at an elevated temperature of 170-175ºC. After the complete dispersion of the polymeric material a double action activator/linking agent (0.6%) was added to the mixture of asphalt, crumb rubber and polymer. The CMCRA obtained after modification meets requirements for PG 76-34 with a solubility of 98 percent. Stabilizer Fiber, 0.4% of the total mix weight. Mix With fiber - PG 67-22, PG 76-22, PG 76-34 Without fiber - PG 67-22, PG 76-22 The test methods that were conducted are shown in Table 102.

203 Table 102: Tests for Different Properties Property Standard Tests and Procedures Used in this Study Compaction Marshall – 25 and 50 blows; SGC – 30, 45 and 60 gyrations Aggregate Breakdown Extraction in ignition furnace (AASHTO TP 53-97, without calibration) and washed sieve analysis (AASHTO T 11-91). Stone-on- Stone Contact Voids in coarse aggregate in dry rodded condition (VCADRC) (AASHTO T 19) and voids in coarse aggregate in mix (VCAmix). Air Voids Dimensional volume: CoreLok – using double bags. Draindown NCAT draindown test; draindown was determined for three binder grades – PG 67-22, PG 76-22 and PG 76-34 and three aggregate types. Abrasion/ Stone Loss Cantabro tests (European standard: prEN 12697-17); In this study, samples were compacted using both Marshall hammer and the Superpave gyratory compactor (SGC). The SGC samples were compacted to 50 gyrations and the Marshall samples were compacted to 50 blows per face. Testing was completed using the medium gradation and a PG 67-22 binder and then repeated with a PG 76-22 polymer modified binder. Six samples (rather than ten) were made using each aggregate, the medium gradation, and two asphalt contents. Three of these samples were tested after cooling and are considered unconditioned samples. Three samples were conditioned by placing them in a forced draft oven at 64oC for seven days. The conditioned samples were allowed to cool at room temperature (25°C) a minimum of four hours before testing. Draindown With draindown baskets, AASHTO T305-97, with two modifications: 1. Two different basket sieve sizes, and 2. Weighing the basket after the mix was emptied out of the basket. Regarding compaction characteristics, Watson et al indicate that the PG 76-22 SBS modified asphalt and fiber stabilizer were used to prepare three replicates each for the five levels of compaction. A 6.0 percent binder content was used for all mixes because it was estimated to be near the optimum level needed for these mixes based on previous experience. The compaction slope (gyration versus sample height) for these mixes indicated that the density of these mixtures continued to increase even after 60 gyrations. Watson et al mention that when the change in height per gyration was plotted the resistance to compaction of all three aggregates was seen to be virtually identical. The results showed that most of the density is obtained within the first 20 gyrations. The authors concluded that stone-on-stone contact was beginning to occur at this point and that the increase in density beyond that point was due to aggregate breakdown from excessive compactive effort. Watson et al made a comparison between the number of gyrations with the SGC and the Marshall hammer. It was found that there was not a close correlation with any of the gyratory compaction levels to the 25 blow Marshall method. However, the 50 blow Marshall compaction effort matched very closely to the gyratory compaction effort at 50 gyrations. Watson et al showed that the SGC results for the traprock and granite mixes

204 are equivalent to that of the 50 blow Marshall within a range of 45 to 53 gyrations. The gravel mix had a higher density with the 50 blow Marshall hammer than with the gyratory compactor. To evaluate aggregate breakdown, Watson et al indicate that samples for each of the five compaction levels were tested for gradation and compared to the actual gradation blend. The results showed that aggregate breakdown (difference in gradation before and after compaction) ranged from zero on the 0.075mm (No. 200) sieve to about ten percent on the 4.75mm (No. 4) sieve. The authors mention that a comparison of breakdown for the granite mixes showed that the SGC breakdown was not directly dependent on the number of gyrations and was not as great as the breakdown caused by the Marshall method. They mention that even as little as 25 blows with the Marshall hammer resulted in greater breakdown than 60 gyrations with the SGC. The authors concluded that the dynamic impact of the Marshall hammer provided more breakdowns of aggregate particles than the kneading action of the SGC. It was also seen that the breakdown from 60 gyrations of the SGC was practically the same as the breakdown from 30 gyrations. Based on this data, Watson et al concluded that the 50 gyration compactive effort previously recommended by NCAT is an acceptable value for design compactive effort. Watson et al also looked at stone-on-stone contact of different mixes to evaluate the Ndesign criterion. Results from tests using the traprock aggregate indicated that VCAMIX was slightly higher than VCADRC at the 30 gyration level. This indicates that 30 gyrations were insufficient to compact the specimens to the point of stone-on-stone contact. At both 45 and 60 gyrations, all samples met stone-on-stone criteria. Based on this observation, the authors conclude that a design compactive effort of 50 gyrations is acceptable. Watson el at indicate that the determination of mixture density and air voids is an inherent problem in the traditional mix design system for OGFC. In this study, air voids were determined based on three methods of determining bulk specific gravity: dimensional analysis, use of the vacuum sealing method (CoreLok), and the use of the Corelok to determine the effective air void content. Bulk specific gravity of the OGFC samples was also obtained using the CoreLok device, with one modification. The use of a single bag frequently resulted in punctures (when testing SGC samples) that would allow air back into the vacuum-sealed bag. Therefore, the double bag method was used on all gyratory samples to prevent bag punctures. A double bag correction factor was used. The standard single bag method was performed on all Marshall samples and bag punctures were not a problem. Watson et al conducted Cantabro Abrasion loss tests to determine whether samples compacted with the SGC could be utilized in this test. From a comparison of a sample before and after testing, the authors comment that the test is a severe test of mixture durability. Test results on Marshall compacted samples for each of the aggregate sources indicated that none of the samples would meet the maximum requirements. The failing results applied to both the unconditioned and conditioned samples in which 6.0 percent asphalt binder had been added. Only granite mixes met maximum loss requirements only when a polymer-modified asphalt was used.

205 Watson et al mention that with Superpave gyratory compacted samples, all mixtures except traprock and gravel with the PG 64-22 binder met the requirements. All the gyratory samples with polymer-modified binder met the maximum stone loss requirements of 20 percent for unconditioned specimens and 30 percent for conditioned samples. They indicate that the results show that 6.0 percent asphalt binder content is insufficient for the traprock and gravel mixtures in order to provide the durability needed for long-term performance. When the asphalt content was increased to 6.5 percent and a polymer-modified binder was used, all mix types met the Cantabro requirements for 50 blow Marshall specimens. Watson et al indicates that the results emphasize the important contribution of polymer- modified asphalt in providing resistance to stone loss and improving mixture durability. As an example, they mention that there was as much as 55 percent difference in the percent of stone loss for the traprock aggregate when comparing results of unmodified asphalt to that of the polymer modified binder. Watson et al mention that stone loss by the Cantabro procedure is almost always lower for the SGC samples compacted to 50 gyrations than the Marshall specimens compacted to 50 blows. They comment that for this reason, the criteria developed in Europe for use with Marshall design procedures may not be applicable to the Superpave compaction method used in the U.S., and that the criteria may need to be changed. They mention that the data showed that when stone loss from Marshall samples is compared to stone loss from SGC samples, a corresponding limit for unconditioned samples should be 15 percent maximum loss and that the stone loss for conditioned samples should be near 20 percent maximum. Watson et al conducted draindown tests with different mixes, using a variation of the standard basket drainage test procedure. They used baskets with 4.75 mm as well as 2.36 mm openings, and they weighed the baskets at the end of the draindown test. The smaller sieve was used because the authors felt that there may be some intermediate aggregate particles that pass through the 4.75mm (No. 4) mesh that would adversely affect test results for binder draindown, by showing an artificially high value for binder draindown. The second modification was performed to determine the asphalt material that remained on the basket and in effect could be considered as part of the binder draindown similar to the Schellenberger drainage test. To evaluate the draindown results an analysis of variance (ANOVA), was conducted. The factors considered were 2 aggregate types, 3 gradations, 3 binder types, 2 fiber conditions, and 4 asphalt contents. Based on previous experience, only traprock and granite were considered as aggregate type factors in this analysis. A regression analysis was also performed which indicated a very strong correlation between the two basket types and two methods used to calculate draindown. An analysis of test results using the standard AASHTO procedure was compared to test results using the modified method where the amount of binder retained on the 4.75mm (No. 4) wire mesh was also considered as draindown. The regression analysis had a coefficient of determination (R2) of 98.5 percent for this comparison. The authors comment that this indicates that considering the amount of binder retained on the 4.75mm (No. 4) basket may not be a significantly better method of determining draindown than