Using Recycled Asphalt Shingles with Warm Mix Asphalt Technologies (2018)

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.

163 Activation Energy for Recycled Asphalt Shingle Binders and Dispersion of RAS in Asphalt Mixtures Abstract The thermal energy input required to achieve recycled asphalt shingle (RAS) asphalt binder phase change from solid to liquid is assumed to provide an indicator of how well RAS particles break down during mixing, allowing the RAS binder to become an active part of the mixture’s compos- ite binder. This study used differential scanning calorimetry (DSC) analysis to assess the phase change characteristics of the RAS particles, compared to a virgin asphalt binder. Phase changes were evident at very high temperatures, well above the normal mixing temperatures for asphalt plants. At tem- peratures up to 150°C, partial melting of RAS binder was evident. Scanning electron microscopy (SEM) was also used to assess how well RAS was dispersed in a plant-produced asphalt mixture. In this analysis, inorganic fibers in the RAS were found to provide a unique indicator of RAS dispersion. However, this technique was not practical to quantify disper- sion of RAS in an asphalt mixture or to indicate if the RAS binder is activated. Introduction One major concern impeding the use of RAS in asphalt mixtures is uncertainty about how much the RAS binder is activated during mixing and is made available to coat and bind the aggregates. A common hypothesis is that during asphalt mixture production, the RAS binder is activated by melting and then blends with the virgin binder to form a composite binder that coats the virgin and recycled aggregates. Several state agencies follow this assumption, giving full credit to RAS binder contribution in mix design and then establishing virgin binder content once the design binder content is determined (Johnson et al. 2010, Shirodkar et al. 2011, Zaumanis and Mallick 2015). However, the full activation assumption has been questioned, since binder from RAS is much stiffer than virgin binder. Thus, binder from recycled materials requires more energy to activate, which may be impossible to accomplish in a regular asphalt plant. It has been found that a very high temperature (above 200°C) is needed to drain the binder extracted and recovered from RAS (Zhao et al. 2014, Zhou et al. 2013B). Subsequent research has confirmed this doubt and has shown that partial activation—rather than com- plete activation—occurs in RAS mixes (Bowers et al. 2014B, Shirodkar et al. 2011). This issue is likely to be more signifi- cant with PC–RAS. Thermal Analysis of RAS Binder Phase Change Background DSC analysis was employed to assess the phase change characteristics of RAS particles compared to an asphalt virgin binder. DSC is a characterization technique that determines the temperature and thermal energy required to cause a phase change in a material (Shriver and Atkins 2006). The method uses a reference material (alumina) that does not undergo any phase changes at temperatures lower than 1000°C and accepts thermal energy at a constant rate throughout this temperature range. The sample and reference material are each placed in a holder with an imbedded thermocouple. The temperature is increased at a constant rate, and the tem- perature of each is measured as a function of time. As the sample undergoes a phase change, a significant portion of the thermal energy is consumed by enthalpy of the phase change. This results in a temperature lag between the sample and the reference and is used to identify the initial temperature for the phase change as well as a relative measure of the thermal energy it is consuming. If a phase change is exothermic, it gives off heat, and the temperature will increase ahead of the reference. If it is endothermic, it accepts heat and it lags behind the reference. DSC is widely used for determination of thermal transitions brought about by the first order tran- sitions, such as melting and crystallization of crystallizable species (Elseifi et al. 2010). A P P E N D I X B

164 Sample Preparation and Testing Sample preparation included sieving a sample of processed RAS into 13 particle size ranges (45, 150, 250, 300, 350, 500, 600, 710, and 850 µm; and 1, 2.23, 4.04, and 7.26 mm). The par- ticle sizes analyzed by DSC were those retained on the 45, 150, 300, 500, and 850 µm sieves. For each test, 12 mg of the sam- ple was placed on the aluminum pan. 12 mg is the maximum limit, so larger particle sizes could not be tested. It important to note that the RAS samples included mineral matter and, possibly, some very small pieces of extraneous material. The size-separated samples were individually tested in DSC with an empty aluminum blank pan sample as the reference. A sample of virgin asphalt binder was also compared. The temperature was increased from 30°C to 550°C at a rate of 10°C/min. in an inert environment (nitrogen). Testing was performed with temperatures above the maximum tempera- ture encountered in asphalt mix production (180°C) to make sure that a thermal transition was observed. Thermal Analysis of RAS Figure B-1 shows the full heat flow versus temperature plots for the tested sizes of RAS. Since the rate of tempera- ture increase was kept constant, the heat flow is the difference in energy applied to the aluminum reference standard and the tested sample. Positive flow indicates that additional heat flowed to the RAS sample relative to the reference. Peaks of the heat flow plots correspond to phase transitions. Over the temperature range from 200°C up to about 350°C, there does seem to be an ordering of the heat flow to the RAS particle sizes with less heat flow occurring for the larger sizes. All sam- ples, including the virgin binder, exhibited some sort of phase change around 370°C which is typically where the flash point of asphalt binders occurs in the presence of oxygen; however, the tests were performed in the presence of nitrogen, which restricted the decomposition of the binder. Additional high-resolution DSC tests were conducted in the lower temperature range of 30°C to 150°C. The results of DSC are shown in Figure B-2. As can be seen in this figure, at approximately 60°C, the virgin bitumen and smaller particle sizes began taking on more heat than the other samples. Above approximately 80°C, a very clear order of how these samples absorbed heat was evident, starting with the virgin bitumen absorbing the most and then in order of the particle size from smallest to largest. This is an indication that the smaller particle sizes were absorbing more heat in the form of melting transi- tions. This could lend support to requiring much finer grind sizes for processing RAS. However, visual inspections of the samples after the high- resolution DSC tests revealed that very little of the RAS par- ticles had melted during the test conducted up to 150°C. Some agglomeration was evident in the smaller particle sizes but, overall, little melting seemed to occur during DSC. Mag- nified images were taken of particle sizes prior to DSC test- ing and samples still in the aluminum pan after DSC testing Figure B-1. DSC comparison of each particle size range.

165 (Figure B-3). The magnification is constant throughout the images at 30x. The annotated micron bar is 20mm in length. Some melting and sintering connection was made between the particles. Thus, the small peaks must be related to different phase changes or decompositions not related to the melting of the RAS binder. Use of Scanning Electron Microscopy for Analysis of RAS Dispersion Background SEM is a technique that employs electron beams to produce high-resolution magnified images of small samples (Goldstein et al. 2003). The electrons interact with atoms in the sample, producing signals that contain information about the sam- ple’s surface topography and composition. SEM produces images of objects with resolution in the micrometer to nano- meter range with relatively low diffraction effects. An energy- dispersive spectroscopy (EDS) detector is used to separate the characteristic x-rays of different elements into an energy spectrum from which software is used to determine the abun- dance of specific elements in the sample (Argast and Tennis 2004). EDS can be used to analyze the chemical composition of materials and to create element composition maps. Elseifi et al. (2008) used SEM to study the surface morphology of HMA, and Yao et al. (2012) used SEM to investigate the micro structure of nanomodified asphalt binder. For this study, SEM was used to examine how well RAS par- ticles were broken up and dispersed within an asphalt mix- ture. A common technique used for SEM analysis of mixing of materials employs the addition of an artificial tracer. How- ever, the high temperature and mechanical energy occurring during mixing in an asphalt plant poses a rather rigorous environment. Only a ceramic tracer would be rugged enough to survive that environment; however, there was concern that adding a ceramic tracer could alter the dynamics of mixing. Fortunately, RAS samples contain a natural tracer material: fiberglass fibers. Nonwoven fiberglass mats serve as the skele- ton in most modern shingles. An SEM image of a RAS sample is shown in Figure B-4. The rod-shaped fiberglass fibers in the image have a unique chemical signature as measured by EDS. In Figure B-5, the silicon in the fiberglass contrasts well against the surrounding carbon-based matrix (asphalt). In cases where there is a significant amount of matrix covering the fibers, EDS mapping can be used to map their locations, as the x-rays easily penetrate through the carbon matrix. Fig- ure B-5 also illustrates that the fiberglass fibers contained in Figure B-2. High resolution DSC scans of the RAS particles in the low temperature range.

166 shingles are composed of silicon, calcium, aluminum, and oxygen. Figure B-6 illustrates that the asphalt matrix of the RAS is composed essentially of carbon, with small amounts of oxygen and sulfur. Sample Preparation and Testing of Processed RAS Sample preparation includes acquisition of a sample that will fit into the SEM chamber and coating the sample to pre- vent charge build-up on electrically insulating samples. Most electrically insulating samples are coated with a thin layer of conducting material, commonly carbon, gold, or some other metal or alloy (Egerton 2005). The choice of material for con- ductive coatings depends on the data to be acquired; carbon is most desirable if elemental analysis is a priority, while metal coatings are most effective for high-resolution electron imag- ing applications. Alternatively, an electrically insulating sample can be examined without a conductive coating in an instrument capable of low vacuum operation. RAS samples retained on fine sieves were secured with carbon tape on aluminum pucks. Next, the particles were (a) 300µm before DSC. (b) 300µm after DSC. (c) 500µm before DSC. (d) 500µm after DSC. (e) 850µm before DSC. (f) 850µm after DSC. 20mm 20mm 20mm20mm 20mm20mm Figure B-3. Images of processed RAS before and after DSC.

Figure B-4. Scanning electron micrograph using backscattered electrons of a PC–RAS stock illustrating the contrast seen between fiberglass fibers and RAS matrix. Figure B-5. Energy dispersive spectrometry of the fiber area illustrating the unique chemical signature of the fiberglass as compared to the matrix.

168 gold sputter coated to provide electrical conductivity during SEM analysis. The coated samples were mounted on alumi- num stubs, and images were taken at various magnifications with a JEOL 7000F SEM. Analysis of SEM Images of RAS Images were compared at different magnifications before and after the high-resolution DSC testing. The magnifica- tions were 50x, 250x, 500x, 1,000x, and 5,000x (Figure B-7). The morphologies were compared between these two stages. Some melting was evident in the samples after DSC testing, particularly in the smaller sizes. Overall, when compared to the pre-tested particles, the post-DSC samples were agglom- erated together. During melting, the particles fuse together, seen mostly at the 50x magnification. Surface features are more evident in the higher magnifications. In these images, it appears that the surface features of the particles are generally intact after the DSC testing to 150°C. Tracer Analysis of Asphalt Mixtures Containing RAS A plant-produced asphalt mixture containing RAS was analyzed in the SEM to assess how well the RAS binder was distributed in the mixture. The hypothesis was that well- dispersed fibers would indicate that the RAS binder had melted, allowing the fibers to be released. This involved tracking the unique chemical signature of the silicon-based fiberglass com- ponent of the RAS. For asphalt shingles containing fiberglass, the amount of fiber ranges from 2% to 15% (3M Corporation 2007, Lee 2009, National Association of Home Builders 1998). For an asphalt mixture with 5% RAS, the fiber content in the mixture could range from 0.1% to 0.75%. Procedures and Results A laboratory-compacted specimen from one of the Alabama mixtures was randomly selected for examination. This SEM was used to look for the fiberglass fibers from the shingles Figure B-6. Energy dispersive spectrometry of the matrix area illustrating the chemical signature.

169 (a) 45µm before (top row) and after (bottom row). (b) 150µm before (top row) and after (bottom row). (c) 300µm before (top row) and after (bottom row). Figure B-7. RAS binder SEM images before and after DSC. (continued on next page)

170 as an indicator of RAS presence in asphalt mixtures. In addi- tion to EDS chemical analysis, the fibers have a unique shape compared to the minerals in the aggregates containing sili- con. Thus, they can be identified both by shape and chemistry using SEM, although shape is an easier defining characteris- tic than elemental composition. Description of the steps and results from this work are as follows: 1. Small beams (approximately 12.5 mm × 6.3 mm × 125 mm), as shown in Figure B-8, were cut from the Superpave Gyra- tory Compactor specimen. The beams were coated with (d) 500µm before (top row) and after (bottom row). (e) 850µm before (top row) and after (bottom row). (f) Bitumen after DSC testing. Figure B-7. (Continued). Figure B-8. Mix sample coated with a thin gold layer.

171 a thin gold layer to make the surfaces conductive. Unfor- tunately, SEM imaging did not reveal the presence of the fibers on the surfaces (Figure B-9). 2. A few beams were then fractured in an attempt to reveal the fibers, followed by gold coating. Although a fiber can be seen in Figure B-10, it was the only one over the entire scanned area. Note that the field of view for these images is only a couple of square millimeters. It was believed that most fibers would be buried within the asphalt. 3. Beams were treated with various solvents to chemically remove the surface asphalt and reveal the fibers. Benzene, methylbenzene, carbon bisulfide, and carbon tetrachloride Figure B-9. Secondary electron image (topography), gold treated. Figure B-10. Secondary electron image fracture and gold-treated sample at two locations.

172 were used, but it was found that these materials were too dangerous for the technicians. An over-the-counter com- mercial asphalt and tar remover (Black Jack) worked well. Figure B-11 and Figure B-12 show the presence of fiber within the mixture at three different locations. Figure B-12 appears to show a piece of a cellulose fiber mat that may indicate that the RAS particle was not broken up and that the RAS was not well dispersed. Although the presence of a few fibers in a few beams cut from a Superpave Gyratory Compactor sample may indi- cate that RAS was dispersed within that sample, it was not Figure B-11. Embedded fiber in the mixture at two locations. Figure B-12. Embedded cellulose fiber in the mixture.

173 considered a feasible technique to quantify dispersion, much less validate the hypothesis that the fiber dispersion could be an indicator of RAS binder activation. Perhaps analyzing hundreds of samples that were randomly obtained from a larger quantity of mix produced with RAS could provide a better statistical measure of dispersion, but the cost of doing so would be prohibitive. Conclusions and Recommendations Based on the experimental results obtained in this study, it was concluded that: • Thermal analysis by means of high-resolution differential scanning calorimetry (DSC) in the temperature range below 150°C showed that smaller RAS particles also exhibited closer thermal behavior to the virgin binder at this temper- ature range. The smaller RAS particles increased heat flow faster than larger particles, indicating that the RAS binder of smaller particles transition from solid to liquid faster. Analysis of SEM images following DSC confirmed that RAS particles were partially but not completely melted at 150°C. • Inorganic fibers in RAS were easily identifiable both in shape and chemical composition using SEM. However, this tech- nique was found to be impractical to quantify dispersion of RAS in an asphalt mixture or to indicate if the RAS binder is melted/activated. This technique is limited by the extremely small size of the samples required to conduct the test. To quantify RAS dispersion using this technique, dozens of samples from a large quantity of mix would be required. Recently, gel permeation chromatography—a size-exclusion chemistry technique that yields molecular weight distribution of the substance based on molecular size—was introduced in the study of RAP or RAS blending (Bowers et al. 2014A). Gel permeation chromatography is capable of differentiating aged binder from virgin binder because aged binder has a higher portion of large molecules than virgin binder (Bowers et al. 2014A, Zhao et al. 2014). This technique is recommended for further research to quantify RAS activation and dispersion. Bibliography 3M Corporation. Scotchguard Algae Resistant Roofing System. 2007. http://solutions.3m.com. Argast, A., and C. F. Tennis III. A Web Resource for The Study of Alkali Feldspars and Perthitic Textures Using Light Microscopy, Scanning Electron Microscopy and Energy Dispersive X-Ray Spectroscopy. Journal of Geoscience Education. Vol. 52, No. 3, 2004, pp. 213–217. Bowers, B. F., B. Huang, X. Shu, and B. C. Miller. Investigation of Reclaimed Asphalt Pavement Blending Efficiency Through GPC and FTIR. Construction and Building Materials, Vol. 50, 2014A, pp.517–523. Bowers, B. F., J. Moore, B. Huang, and X. Shu. Blending Efficiency of Reclaimed Asphalt Pavement: An Approach Utilizing Rheologi- cal Properties and Molecular Weight Distributions. Fuel, Vol. 135, No. 1, 2014B, pp. 63–68. Egerton, R. F. Physical Principles of Electron Microscopy: An Introduction to TEM, SEM, and AEM. Springer U.S., 2005. Elseifi, M. A., I. L. Al-Qadi, S. H. Yang, and S. H. Carpenter. Validity of Asphalt Binder Film Thickness Concept in Hot-Mix Asphalt. Transportation Research Record: Journal of the Transportation Research Board, No. 2057, 2008, pp. 37–45. http://dx.doi.org/ 10.3141/2057-05. Elseifi, M., L. N. Mohammad, I. Glover, I. Negulescu, W. H. Daly, and C. Abadie. Relationship Between Molecular Compositions and Rheological Properties of Neat Asphalt Binder at Low and Inter- mediate Temperatures. Journal of Materials in Civil Engineering, Vol. 22, No. 12, 2010, pp. 1288–1294. Goldstein, J., D. E. Newbury, D. C. Joy, C. E. Lyman, P. Echlin, E. Lifshin, L. Sawyer, and J. R. Michael. Scanning Electron Microscopy and X-Ray Microanalysis. 3rd edition, Springer U.S., New York, 2003. Johnson, E., G. Johnson, S. Dai, D. Linell, J. McGraw, and M. Watson. Incorporation of Recycled Asphalt Shingles in Hot Mixed Asphalt Pavement Mixtures. Minnesota Department of Transportation, Saint Paul, 2010. Lee. R. Recycled Shingles in HMA. Presented at Southeastern Asphalt User/Producer Group Conference, Hilton Head, S.C., 2009. National Association of Home Builders. From Roofs to Roads: Recy- cling Asphalt Roofing Shingles into Paving Materials. 1998. http:// www.epa.gov. Shirodkar, P., Y. Mehta, A. Nolan, K. Sonpal, A. Norton, C. Tomlinson, E. Dubois, P. Sullivan, and R. Sauber. A Study to Determine the Degree of Partial Blending of Reclaimed Asphalt Pavement (RAP) Binder for High RAP Hot Mix Asphalt. Construction and Building Materials, Vol. 25, No. 1, 2011, pp. 150–155. Shriver, D. F., and P. W. Atkins. Shriver & Atkins Inorganic Chemistry. 4th edition, Oxford University Press, Oxford, UK, 2006. Yao, H., Z. You, L. Li, C. H. Lee, D. Wingard, Y. K. Yap, X. Shi, and S. W. Goh. Rheological Properties and Chemical Bonding of Asphalt Modified with Nanosilica. Journal of Materials in Civil Engineering, Vol. 25, No. 11, 2012, pp. 1619–1630. Zaumanis, M., and R. B. Mallick. Review of Very High-Content Reclaimed Asphalt Use in Plant Produced Pavements: State of the Art. International Journal of Pavement Engineering, Vol. 16, No. 1, 2015, pp. 39–55. Zhao, S., B. Bowers, B. Huang, and X. Shu. Characterizing Rheologi- cal Properties of Binder and Blending Efficiency of Asphalt Paving Mixtures Containing RAS Through GPC. Journal of Materials in Civil Engineering, Vol. 26, No. 5, 2014, pp. 941–946. Zhou, F., H. Li, R. Lee, T. Scullion, and G. Claros. Recycled Asphalt Shingle Binder Characterization and Blending with Virgin Binders. Transportation Research Record: Journal of the Transportation Research Board, No. 2370, 2013B, pp. 33–43. http://dx.doi.org/ 10.3141/2370-05.