e SPR 617 SEPTEMBER 2014 Effects of Deicing/Anti-Icing Chemicals (DIAICs) on Rubberized Asphalt Pavements Arizona Department of Transportation Research Center Effects of Deicing/Anti‐Icing Chemicals (DIAICs) on Rubberized Asphalt Pavements SPR‐617 September 2014 Rita Leahy Shih‐Shien Sam Yang Adriana Vargas Peter Schmalzer Nichols Consulting Engineers, Chtd. 1885 S. Arlington Ave. Reno, NV 89509 Arizona Department of Transportation 206 S. 17th Avenue Phoenix, AZ 85007 in cooperation with US Department of Transportation Federal Highway Administration This report was funded in part through grants from the Federal Highway Administration, U.S. Department of Transportation. The contents of this report reflect the views of the authors, who are responsible for the facts and the accuracy of the data, and for the use or adaptation of previously published material, presented herein. The contents do not necessarily reflect the official views or policies of the Arizona Department of Transportation or the Federal Highway Administration, U.S. Department of Transportation. This report does not constitute a standard, specification, or regulation. Trade or manufacturers’ names that may appear herein are cited only because they are considered essential to the objectives of the report. The U.S. government and the State of Arizona do not endorse products or manufacturers. Technical Report Documentation Page 1. Report No. 2. Government Accession No. 3. Recipient's Catalog No. FHWA‐AZ‐14‐617 4. Title and Subtitle 5. Report Date Effects of Deicing/Anti‐Icing Chemicals (DIAICs) on Rubberized Asphalt Pavements September 2014 7. Author(s) 8. Performing Organization Report No. 6. Performing Organization Code Rita Leahy, Shih‐Shien Sam Yang, Adriana Vargas, and Peter Schmalzer 9. Performing Organization Name and Address 10. Work Unit No. Nichols Consulting Engineers, Chtd. 1885 S. Arlington Avenue, Suite 111 Reno, Nevada 89509 11. Contract or Grant No. 12. Sponsoring Agency Name and Address 13.Type of Report & Period Covered Arizona Department of Transportation 206 S. 17th Avenue Phoenix, Arizona 85007 FINAL REPORT SPR‐000 1(169) 617 14. Sponsoring Agency Code 15. Supplementary Notes Prepared in cooperation with the US Department of Transportation, Federal Highway Administration 16. Abstract The objective of this study was to evaluate the effect of typical chemical winter maintenance practices on Arizona Department of Transportation (ADOT) pavements. A review of previous studies on the effect of deicing/anti‐icing chemicals (DIAICs) did not yield definitive recommendations, especially for DIAICs typically used by ADOT. Researchers conducted a laboratory study evaluating the effects of magnesium chloride, potassium chloride, sodium chloride, and distilled water on eight different open‐graded rubber‐modified asphalt concrete mixes using the boiling test (ASTM D3625). All experimental factors were found to be statistically significant, and the researchers provide recommendations on which DIAICs should be used for different binder and aggregate types. 17. Key Words 18. Distribution Statement Deicing, anti‐icing, rubberized asphalt, salt, magnesium chloride, potassium chloride, sodium chloride, DIAIC, binder, aggregate Document available to the US public through the National Technical Information Service, Springfield, Virginia, 22161 19. Security Classification 20. Security Classification 21. No. of Pages Unclassified Unclassified 54 22. Price 23. Registrant's Seal SI* (MODERN METRIC) CONVERSION FACTORS APPROXIMATE CONVERSIONS TO SI UNITS Symbol When You Know Multiply By LENGTH in ft yd mi inches feet yards miles in 2 ft 2 yd ac 2 mi 2 square square square acres square fl oz gal 3 ft 3 yd fluid ounces gallons cubic feet cubic yards oz lb T ounces pounds short tons (2000 lb) o Fahrenheit fc fl foot-candles foot-Lamberts lbf 2 lbf/in poundforce poundforce per square inch Symbol When You Know mm m m km millimeters meters meters kilometers 25.4 0.305 0.914 1.61 To Find Symbol millimeters meters meters kilometers mm m m km square millimeters square meters square meters hectares square kilometers mm 2 m 2 m ha 2 km milliliters liters cubic meters cubic meters 3 shown in m mL L 3 m 3 m grams kilograms megagrams (or "metric ton") g kg Mg (or "t") AREA inches feet yard 645.2 0.093 0.836 0.405 2.59 miles 2 VOLUME 29.57 3.785 0.028 0.765 NOTE: volumes greater than 1000 L shall be MASS 28.35 0.454 0.907 TEMPERATURE (exact degrees) F 5 (F-32)/9 or (F-32)/1.8 Celsius o lux 2 candela/m lx 2 cd/m C ILLUMINATION 10.76 3.426 FORCE and PRESSURE or STRESS 4.45 6.89 newtons kilopascals N kPa APPROXIMATE CONVERSIONS FROM SI UNITS Multiply By LENGTH 0.039 3.28 1.09 0.621 To Find Symbol inches feet yards miles in ft yd mi AREA 2 mm 2 m 2 m ha 2 km square millimeters square meters square meters hectares square kilometers 0.0016 10.764 1.195 2.47 0.386 square square square acres square inches feet yards miles 2 in 2 ft 2 yd ac 2 mi VOLUME mL L 3 m m3 milliliters liters cubic meters cubic meters 0.034 0.264 35.314 1.307 g kg Mg (or "t") grams kilograms megagrams (or "metric ton") o Celsius fluid ounces gallons cubic feet cubic yards fl oz gal 3 ft yd3 ounces pounds short tons (2000 lb) oz lb T MASS 0.035 2.202 1.103 TEMPERATURE (exact degrees) C 1.8C+32 Fahrenheit o foot-candles foot-Lamberts fc fl F ILLUMINATION lx 2 cd/m lux 2 candela/m N kPa newtons kilopascals 0.0929 0.2919 FORCE and PRESSURE or STRESS 0.225 0.145 poundforce poundforce per square inch lbf 2 lbf/in *SI is the symbol for the International System of Units. Appropriate rounding should be made to comply with Section 4 of ASTM E380. (Revised March 2003) Contents Executive Summary ............................................................................................................................. 1 Chapter 1. Introduction ........................................................................................................................ 3 Problem............................................................................................................................................ 3 Objectives ........................................................................................................................................ 3 Chapter 2. Literature Review................................................................................................................ 5 Deicing and Anti‐Icing Chemicals ..................................................................................................... 5 Recent Studies of Deicer Damage to HMA Pavements ................................................................... 5 Laboratory Investigations .......................................................................................................... 5 Field Study ............................................................................................................................... 11 Possible Mechanisms for DIAIC‐Induced Damage ......................................................................... 11 Factors Affecting Deicer‐Induced Damage in HMA Pavements .................................................... 12 Test Methods for Evaluating Resistance to Deicer‐Induced Damage ........................................... 13 Identifying Deicer‐Induced Damage in HMA ................................................................................. 15 Summary ........................................................................................................................................ 16 Chapter 3. Laboratory Testing ............................................................................................................ 19 Experimental Design ...................................................................................................................... 19 Materials ........................................................................................................................................ 19 Aggregate Sources ................................................................................................................... 19 Asphalt Binder and Crumb Rubber.......................................................................................... 21 Mix Design ............................................................................................................................... 23 DIAIC Solutions ........................................................................................................................ 23 Test Procedure ........................................................................................................................ 23 Results ............................................................................................................................................ 24 Chapter 4. Summary and Recommendations...................................................................................... 33 Summary ........................................................................................................................................ 33 Recommendations ......................................................................................................................... 34 References ......................................................................................................................................... 37 Appendix: Boiling Test Results ........................................................................................................... 39 v List of Figures Figure 1. Effect of Different DIAIC Solutions on Indirect Tensile Strength (Farha et al. 2002, Figure 6) ................................................................................................................. 7 Figure 2. Percentage of Material Stripped by Aggregate Source................................................................ 25 Figure 3. Percentage of Material Stripped by Binder Source ..................................................................... 26 Figure 4. Percentage of Material Stripped by Type of DIAIC ...................................................................... 27 Figure 5. Interaction Between Aggregate Source and Binder Blend .......................................................... 28 Figure 6. Interaction Between Aggregate Source and Type of DIAIC Solution ........................................... 28 Figure 7. Interaction Between Binder Blend and Type of DIAIC Solution ................................................... 29 Figure 8. Percentage of Material Stripped by DIAIC Type (CM‐0343 and CM‐0323 Mixes) ....................... 30 Figure 9. Percentage of Material Stripped by DIAIC Type (CM‐2176 and CM‐2058 Mixes) ....................... 31 vi List of Tables Table 1. Experimental Matrix ...................................................................................................................... 20 Table 2. Aggregate Properties..................................................................................................................... 21 Table 3. Asphalt‐Rubber Binder Properties ................................................................................................ 21 Table 4. Crumb Rubber Gradation .............................................................................................................. 22 Table 5. Physical Properties of Asphalt‐Rubber Binder (Blend 1) ............................................................... 22 Table 6. Physical Properties of Asphalt‐Rubber Binder (Blend 2) ............................................................... 23 Table 7. Mix Properties ............................................................................................................................... 23 Table 8. Analysis of Variance of Boiling Test Results .................................................................................. 24 Table 9. Level of Stripping Caused by DIAIC Solutions vs. Water ............................................................... 32 Table 10. Best and Worst Combinations of Mix Types and DIAICs ............................................................. 34 Table 11. Proposed Experiment Design for Testing Binder–DIAIC Interaction ........................................... 35 vii List of Abbreviations, Acronyms, and Symbols AAPTP Airfield Asphalt Pavement Technology Program AASHTO American Association of State Highway and Transportation Officials ADOT Arizona Department of Transportation ASTM American Society for Testing and Materials CaCl2 calcium chloride DIAIC deicing/anti‐icing chemical DOT department of transportation FTIR Fourier transform infrared HMA hot‐mix asphalt IDT indirect tension IT immersion tension ITS indirect tensile strength KCl potassium chloride MgCl2 magnesium chloride NaCl sodium chloride NMR nuclear magnetic resonance NMRI nuclear magnetic resonance imaging OGFC open‐graded friction course PAH polycyclic aromatic hydrocarbon PG performance‐graded TSR tensile strength ratio viii EXECUTIVE SUMMARY The Arizona Department of Transportation (ADOT) makes significant use of open‐graded rubber‐ modified asphalt concrete for roadway surfacing. There is concern that this type of surface may be more vulnerable than conventional dense‐graded asphalt concrete to damage caused by the chemical deicing techniques ADOT uses to fight winter storms. Previous studies on the effects of deicing/anti‐icing chemicals (DIAICs) on pavements have yielded varying conclusions. Most studies have focused on airfields, where urea, formates, or acetates are primarily used, as opposed to the chloride DIAICs predominantly used on highways. There is general agreement that these airfield DIAICs accelerate stripping (loss of adhesion between asphalt binder and aggregate) to some degree. However, the damage mechanism is unclear, as is the degree to which the effect is significant under real‐world conditions. For this project, researchers conducted a laboratory study to investigate the effect of DIAICs commonly used by ADOT on typical open‐graded rubber‐modified asphalt concrete mixes. Asphalt binder and crumb rubber were obtained from two ADOT suppliers, and aggregate was obtained from four ADOT‐ approved sources. These components were used to produce eight asphalt concrete mixes in accordance with Section 414 of the ADOT Standard Specifications, “Asphaltic Concrete Friction Course (Asphalt‐ Rubber).” The mixes were evaluated for stripping potential using the boiling test (ASTM D3625) after soaking in solutions of magnesium chloride, sodium chloride, potassium chloride, and distilled water. None of the mixes performed significantly worse after soaking in magnesium chloride than after soaking in distilled water. The effects of potassium chloride and sodium chloride varied with aggregate and binder source. Based on the results of the laboratory study, the researchers recommend magnesium chloride as the DIAIC with the least potential to damage open‐graded rubber‐modified asphalt concrete pavements. If DIAIC‐induced damage is a concern, they recommend using stiffer binders, which appear to help mitigate the damage. Of the four aggregate sources tested, CM‐2176 aggregate is recommended for rubberized pavements where DIAICs will be used. 1 2 CHAPTER 1. INTRODUCTION PROBLEM Recent observations show that the winter maintenance techniques that the Arizona Department of Transportation (ADOT) currently uses may adversely affect pavement performance. ADOT’s standard approaches to clearing winter precipitation are to apply deicing/anti‐icing chemicals (DIAICs) and to mechanically remove accumulated snow and ice. In Arizona, significant portions of the roadways are surfaced with open‐graded rubber‐modified asphalt concrete. This type of pavement may be more vulnerable than conventional dense‐graded pavement to damage caused by chemical deicing techniques. ADOT currently selects winter maintenance treatment strategies based on pavement temperature and pavement surface conditions, following these guidelines:  If the pavement temperature is steadily below 15° F and falling, accumulated snow and ice should be plowed and/or abrasives should be applied to the road.  If the pavement temperature is above 15° F and the pavement surface is icy, accumulated snow and ice should be plowed and/or deicing should be initiated.  If the pavement temperature is above 15° F and the pavement surface is not yet icy but precipitation is anticipated, anti‐icing should be initiated. Chemical treatment strategies involve applying deicing or anti‐icing chemicals to the pavement surface. Anti‐icing agents are used to prevent snow and ice from bonding with the pavement surface through the timely application of a chemical freezing point depressant, while deicing agents are used to destroy an existing bond between snow and ice and the pavement surface. The DIAICs that ADOT most commonly uses are magnesium chloride (MgCl2), sodium chloride (NaCl), calcium chloride (CaCl2), and calcium magnesium acetate. It has been hypothesized that the use of these chemicals exacerbates the effects of freeze‐thaw cycles on pavements. There is also concern that the chemicals may adversely affect the component materials of asphalt concrete (asphalt, rubber, and aggregate). Additionally, the chemicals may accelerate the oxidation process or cause other chemical or physical changes that may reduce the overall serviceability of the pavement. OBJECTIVES The effect of deicing/anti‐icing chemicals on open‐graded rubberized asphalt pavements needs to be fully understood so that pavement damage related to these chemicals can be minimized. Accordingly, the objectives of this study were to (1) use laboratory experiments to evaluate the effect of DIAICs on open‐graded rubberized pavements, and (2) select the best combinations of DIAICs, aggregates, and binders for open‐graded rubberized pavements. 3 4 CHAPTER 2. LITERATURE REVIEW This chapter summarizes existing research on the effect of deicing and anti‐icing chemicals on hot‐mix asphalt (HMA) pavements. It includes recent studies on DIAIC‐induced damage in HMA pavements; possible mechanisms for this damage and factors affecting its extent; test methods for evaluating pavement susceptibility to DIAIC‐induced damage; and methods for identifying when pavement damage has been caused by DIAICs. This chapter concludes with a short summary of the chemical composition of asphalt. DEICING AND ANTI‐ICING CHEMICALS Applying DIAIC compounds to roadways is an essential winter maintenance strategy in areas where ice and snow are common. These chemicals help vehicles maintain traction by melting ice on the pavement surface. Unfortunately, common deicing salts such as sodium chloride and calcium chloride can cause corrosion to vehicles and infrastructure. Other deicing chemicals have been developed that perform effectively without significantly accelerating the corrosion of metals and alloys. These noncorrosive DIAIC compounds include ethylene glycol, propylene glycol, urea, potassium acetate, sodium acetate, sodium formate, and calcium magnesium acetate. Noncorrosive DIAICs may be applied to the pavement in solid form (pellets or granules), or in liquid form as solutions. For many years, it has been assumed that these DIAICs do little or no damage to either portland cement pavements or HMA pavements. However, recent evidence suggests that these DIAICs can damage both materials. RECENT STUDIES OF DEICER DAMAGE TO HMA PAVEMENTS Studies on DIAIC‐induced damage to HMA can be grouped into two categories: laboratory investigations and field testing. In the laboratory investigations described below, the mechanical and physical properties of asphalt mixture samples conditioned with DIAIC solutions were compared with the properties of mixtures conditioned with dry or distilled water. The field studies were conducted by comparing the friction resistance of DIAIC‐treated pavement sections and sections not treated with DIAICs. Laboratory Investigations Canadian Airfields Study Researchers from Transport Canada and Carleton University (Ottawa, Ontario) conducted a study on the effects of various DIAICs on asphalt concrete pavements at Canadian airfields (Hassan et al. 2001). This study was initiated not because DIAIC‐induced damage was believed to be a problem at Canadian airfields, but because airports in Canada had recently switched from urea‐based pavement deicers to newer compounds such as sodium formate, potassium acetate, and sodium acetate, and there was concern that these new DIAICs might damage HMA pavements. 5 This research was conducted in three phases. In the first phase, two types of aggregate representative of those commonly used in HMA mixes for Canadian airfield pavements were subjected to freeze‐thaw cycles after being immersed in urea, potassium acetate, and sodium formate at various concentrations, as well as distilled water as a control. All of the samples that had been soaked in the DIAICs showed significantly greater weight loss after freeze‐thaw cycling than the samples that had been soaked in distilled water, with those soaked in urea exhibiting the greatest weight loss. The quartzite aggregate had greater weight loss than the limestone aggregate. Weight loss was greatest at a 1 to 2 percent concentration of the DIAICs. In the second phase of the research, HMA cores were soaked in solutions of urea, potassium acetate, sodium acetate, and sodium formate at the most damaging concentrations identified in Phase 1, and then subjected to freeze‐thaw cycles. A control group of cores was soaked in distilled water prior to freeze‐thaw cycling, and a second control group was not conditioned at all. The cores were tested to determine change in weight, indirect tensile strength (ITS), elastic modulus, penetration of extracted binder, and gradation of extracted aggregate. All of the specimens gained weight during freeze‐thaw testing, and the rate of weight gain decreased as the number of freeze‐thaw cycles increased. The first indications of damage to the specimens occurred after 15 cycles. ITS was significantly lower for all of the conditioned samples relative to the unconditioned samples. After 25 freeze‐thaw cycles, the samples soaked in distilled water had the lowest ITS values, while the samples soaked in DIAICs had similar and slightly higher ITS values. After 50 freeze‐thaw cycles, the ITS of the samples soaked in distilled water had not decreased further, whereas the ITS of the samples soaked in DIAICs had decreased substantially. The samples soaked in urea had the lowest ITS values after 50 cycles by a significant amount, whereas the ITS values of the samples soaked in potassium acetate and sodium formate were slightly lower than those soaked in distilled water. The ITS values of the samples soaked in sodium acetate were slightly higher than those soaked in distilled water. Phase 3 was similar to Phase 2, except that only 15 freeze‐thaw cycles were conducted, followed by 40 wet‐dry cycles at 104° F. As shown in Figure 1, ITS decreased after the 15 freeze‐thaw cycles, but recovered after the wet‐dry cycles. The trend is less clear for elastic modulus. While the modulus of all the specimens decreased following freeze‐thaw cycling, it recovered substantially in some cases. Penetration of extracted binder increased for some specimens following freeze‐thaw cycling. In the cases where penetration increased, it recovered fully after the wet‐dry cycles. No significant change in the gradation of extracted aggregate was observed. 6 DS = dry specimen; DW = distille ed water; PA = potassium aacetate; SF = so odium formatte; UR = urea;; SA = sodium m acetate. Figure 1. Effect of Different D DIAIC Solutions o on Indirect Tensile Strenggth (Farha et al. 2002, Figu ure 6) Several faactors limit this study’s app plicability to Arizona A roadw ways. First, th he study did n not include th he DIAICs commonly used d in Arizona, especially e chlo orides. In adddition, the meechanical testt results show wed d inducced by the DIA AIC solutions was similar t o moisture‐in nduced damaage, and it is w well that the damage known that moisture damage d in HM MA is stronglyy related to booth binder so ource and agggregate type (Thelen 19 958, Majidzad deh and Brovvold 1968, Kim m et al. 1985,, Curtis et al. 1991, Tunnicliff and Root 1982). However, this study included a veryy limited rangge of aggregatte and binderr types. The m material selecctions are poorlyy documented in the report, but they are presumabbly most applicable to airfieeld pavements in cold climaates. Therefore, it is possib ble that the re esults of this study would be different iif a broader range of materiaals had been used. As discussed below,, it is also posssible that thee temperatures used in thee 7 study’s high‐temperature cycling were not high enough to produce the reactions needed to cause DIAIC‐ induced damage. Helsinki University of Technology Study Researchers at the Helsinki University of Technology in Finland conducted a study (Alatyppö and Valtonen 2007) similar in concept to the Canadian airfields study. However, the study was conducted reactively, to investigate the potential role of new DIAICs in observed airfield pavement failures, rather than proactively. The failures occurred on recently overlaid pavements where the original surface layer had been subjected to acetates and formates as part of routine deicing procedures. Observed surface distresses included blistering, raveling, and staining. Coring of failure areas revealed that the lower layers of HMA had become very soft and odoriferous. The study included laboratory testing of mixture components (binder and aggregate), laboratory‐mixed HMA specimens, and cores taken from in‐service pavements and specially constructed test sections. Key findings included:  Samples of binder boiled in DIAIC solutions did not appear to undergo chemical changes (as determined by gas chromatography and pH measurements). Specimen weight did increase, but the investigators could not determine the reason.  Surface tension was significantly higher between distilled water and binder than between DIAIC solutions and binder. An emulsion consisting of finely dispersed formate–water solution in binder could be created easily in the lab. Distilled water could not be as easily or as finely mixed into the same binder. The investigators concluded that DIAIC solutions facilitate the penetration of water into binder.  Neutralization was observed between DIAIC solutions and felsic aggregates. Limited aggregate soundness testing indicated that acetic aggregates degrade more when soaked in DIAIC solutions than basic aggregates do.  Samples of binder‐coated aggregates boiled in DIAIC solutions showed significant stripping. Formate solutions showed greater stripping than acetate solutions. The softer binder performed worse than the harder binder. The rubber‐modified binder exhibited little stripping. Gas chromatography detected the presence of polycyclic aromatic hydrocarbons (PAHs), which were interpreted as potential binder decomposition products. PAHs were not detected when the binder alone was boiled in DIAICs, suggesting that a complex chemical interaction involving the binder, aggregate, and DIAIC was occurring at high temperatures.  Testing of laboratory‐compacted mixes and field cores subjected to freeze‐thaw cycling showed little difference among ITS values for specimens soaked in water and DIAICs.  Core samples from a failed runway showed that the bottom of the HMA pavement was extremely soft and odoriferous. Samples had low concentrations of DIAICs and high pH values. One sample had high PAH content. In one experiment, 10 test sections were constructed consisting of an HMA overlay over existing pavement treated with potassium formate. Experimental variables included binder, aggregate type, and mineral filler. Minor damage ascribed to the potassium formate was observed after 1 year. 8 A second experiment consisting of 12 test sections was constructed at a different location. Little damage was observed after 3 years. Some damage was observed after 4 years, and the only section without damage had an alkaline aggregate and a stiff binder with a gilsonite modifier. Cores from damaged areas bled a very soft, sticky, bitumen‐like fluid. To manage the effect of DIAICs on HMA pavement, the investigators suggested: (1) using low air void contents; (2) using high‐viscosity or polymer‐modified binders; (3) using alkaline aggregates or high‐ quality aggregates (avoiding limestone filler or heavily contaminated reclaimed asphalt pavement if acetates or formates will be used as DIAICs); and (4) testing the compatibility of the materials using the boiling test. Colorado DOT Study The Western Transportation Institute at Montana State University conducted a study of the effects of acetate‐based DIAICs on airfield pavements for the Colorado Department of Transportation (CDOT) using sodium chloride and magnesium chloride as the baseline DIAICs (Pan et al. 2008). The study evaluated potassium acetate, potassium formate, and sodium acetate/formate–blend DIAICs as alternatives to the sodium chloride, salt‐sand mixtures, and magnesium chloride that were CDOT’s standard treatments for both portland cement concrete and HMA pavements. The findings related to HMA pavement are summarized below. A boiling water test of a single PG 67‐22 binder mixed with limestone and crushed gravel aggregate was performed at various concentrations of sodium acetate. Significant stripping was observed with both aggregate types, indicating that aggregate type is a secondary factor. Stripping was 4 percent at a 0 percent concentration of sodium acetate and increased to 42 percent at a 40 percent concentration of sodium acetate. Emulsification of binder samples in solutions of varying concentrations of sodium acetate at varying temperatures was examined using low‐speed magnetic stirring. No emulsification was observed at the 0 percent sodium acetate concentration or at temperatures below 104° F. Emulsification increased with both sodium acetate concentration and temperature. No emulsification was seen with the sodium chloride or sodium hydroxide solutions, indicating that the emulsification phenomenon was due to the acetate anion, not the sodium cation or the pH value. Solutions of calcium magnesium acetate behaved similarly to the sodium acetate solutions. Of the three binders tested, the PG 58‐22 exhibited the highest degree of emulsification, whereas the PG 64‐22 and PG 67‐22 binders behaved similarly. Fourier transform infrared (FTIR) spectroscopy of the asphalt‐acetate emulsions showed that the alkane component of the binder was present, and the test did not show the presence of any new chemicals, indicating that no chemical reaction had taken place. The authors proposed a mechanism by which asphalt binder at high temperatures swells and “sucks in” acetate solution from pore spaces, which greatly increases the contact area between binder and acetate ions. Acetate ions are attracted to nonpolar molecules in the binder by van der Waals’ forces and form hydrogen bonds with water. 9 Airfield Asphalt Pavement Technology Program (AAPTP) Study A recently completed Airfield Asphalt Pavement Technology Program study conducted by Advanced Asphalt Technologies investigated the performance of HMA airfield pavement subjected to DIAICs (Advanced Asphalt Technologies 2009). The study used five common DIAICs (potassium acetate, sodium acetate, urea, and ethylene and propylene glycol), five types of aggregates, four different binders, two air void content levels, and one mixture additive to investigate the effects of various factors on DIAIC‐ related pavement damage. The immersion tension (IT) test, a modified version of the AASHTO T283 test, was developed in the study and used to evaluate the mechanical properties of various combinations of these factors. The FTIR test was used to investigate the potential chemical reactions between DIAICs and asphalt binders and aggregates. The surface tension between the binders and DIAIC solutions was measured using the nuclear magnetic resonance (NMR) imaging method. The study’s key findings included:  For some tested mixtures, the HMA pavement damage attributed to DIAICs is most likely a form of moisture‐induced damage. Therefore, it should not be considered a unique form of distress, but rather an accelerated type of moisture damage, and should be referred to as “DIAIC‐related damage” rather than “DIAIC‐induced damage.”  DIAIC‐related damage on HMA pavement was mainly limited to mixtures containing highly siliceous aggregates; however, many siliceous aggregates may not exhibit significant damage. In this study, DIAIC‐related damage was associated with relatively high testing temperatures when acetate‐ and formate‐based DIAICs were used. In addition, the softer the binder, the more severe the damage.  To minimize the effect of DIAIC‐related damage on HMA pavement, researchers recommended adding hydrated lime to the mixture, decreasing the mixture’s air void content, and using a stiffer binder.  The FTIR test showed that significant amounts of carboxylate salts were generated during conditioning of the HMA specimen in DIAIC solutions. However, this compound was observed both in mixes exhibiting DIAIC‐related damage and in those not exhibiting DIAIC‐related damage. Thus, this technique does not produce a reliable indicator of DIAIC‐related damage.  There was no conclusive evidence that PAHs are generated in HMA pavements subjected to DIAICs as was suggested by the Helsinki University of Technology study (Alatyppö and Valtonen 2007). Laboratory Study Implications These four research projects paint a potentially contradictory picture regarding DIAIC‐induced damage in HMA pavements. In general, laboratory testing has not indicated that DIAIC solutions significantly accelerate moisture damage from a mechanical testing standpoint. However, these tests have been limited in scope, and it is possible that if a wider range of asphalt binders, aggregates, and environments were tested using the full range of commercial DIAIC solutions, some contribution to pavement damage would be observed. Also, the Helsinki University of Technology study (Alatyppö and Valtonen 2007) suggests that elevated temperatures may be necessary for DIAICs to contribute significantly to HMA 10 pavement damage. Although that study did observe significant moisture damage in pavements exposed to DIAICs, it is not certain that this is a cause‐and‐effect relationship—the damage might have occurred even without DIAIC use. In other words, in this and other cases of apparent DIAIC‐induced damage in HMA pavements, the DIAIC use may be merely coincidental in a pavement prone to moisture‐induced damage. Field Study The Oregon Department of Transportation (ODOT) conducted a research project to evaluate the effects of DIAICs on open‐graded pavements (Martinez and Poecker 2006). The motivation for this study was the increased number of accidents on ODOT highways constructed with these pavements after winter maintenance chemicals had been applied. The accidents were believed to be related to loss of surface friction caused by the application of DIAICs to the road surface. The study consisted of skid tests on four test sections under three conditions: (1) no DIAIC application; (2) after DIAICs were applied at a rate of 15 gallons per lane mile; and (3) after DIAICs were applied at 30 gallons per lane mile. Researchers found that the application of DIAICs to either type of pavement at either application rate appeared to have little if any effect on the friction number of the pavement surface. The friction numbers obtained in the skid tests were also compared with those recommended in a Federal Highway Administration study, and they were found to be well above the recommended values. POSSIBLE MECHANISMS FOR DIAIC‐INDUCED DAMAGE The literature reviewed above suggests three potential mechanisms by which DIAICs can damage HMA: 1. DIAICs may promote emulsification of the asphalt binder, causing the binder to become soft and lose adhesion to aggregate surfaces. The CDOT study provides direct evidence of binder emulsification in acetate solutions under laboratory conditions, and the Helsinki University of Technology study provides indirect evidence of binder emulsification in acetate‐ and formate‐ contaminated pavements under field conditions. Both studies indicate that high temperatures, such as those experienced in hot summer weather or during repaving operations, are required for this phenomenon. The Helsinki University of Technology study did not address chloride‐based DIAICs, while the CDOT study found that sodium chloride does not promote binder emulsification. 2. DIAICs may promote stripping by causing chemical deterioration of the aggregate surface; by binding to aggregate surfaces in preference to asphalt, thereby weakening the bond between aggregate and binder; or by accelerating normal moisture damage processes by decreasing surface tension, thereby facilitating the penetration of water into the pavement. The Canadian airfields study, the Helsinki University of Technology study, and the AAPTP study all found that mixes with siliceous aggregates were more susceptible to DIAIC‐induced damage than mixes with limestone aggregates. The Helsinki University of Technology study and the AAPTP study also found that acetate and formate greatly reduced the surface tension between the DIAIC solution and the asphalt, promoting dispersion of the solution through the binder. 3. DIAICs may directly attack the asphalt binder. The Helsinki University of Technology study detected the presence of PAHs, which are presumed to be a product of binder decomposition, in 11 asphalt mixes exposed to DIAICs in both the laboratory and the field. The CDOT and AAPTP studies did not detect PAHs or other chemicals indicating decomposition of the asphalt binder. The Canadian airfields study found that the penetration of extracted binder increased following freeze‐thaw cycling and decreased following wet‐dry cycling in DIAIC solutions, which may indicate accelerated aging of the binder. DIAIC‐induced damage may involve a combination of the three mechanisms described above. The Helsinki University of Technology study noted no change in binder samples boiled in DIAIC solutions, but researchers found significant changes, including apparent organic decomposition products, in samples of asphalt‐coated aggregate that had been boiled in DIAIC solutions. This may indicate a complex chemical interaction among the asphalt, aggregate, and DIAIC. It is also possible that DIAICs do not significantly damage asphalt pavements under real‐world conditions, and that the relatively few field observations of damage in the literature are coincidental. It should also be noted that the DIAICs commonly used on airfields are the most heavily represented in the literature. Relatively little research has been done on the effects of chloride‐based DIAICs commonly used on highways. FACTORS AFFECTING DEICER‐INDUCED DAMAGE IN HMA PAVEMENTS Research on DIAIC‐induced damage in HMA pavements has been limited. Anecdotal evidence from airfields in the northern United States, Canada, and northern Europe seems to suggest that DIAIC solutions can cause damage to HMA pavements, while laboratory evidence is somewhat ambiguous. It appears that the most likely mechanism by which DIAICs cause or accelerate damage is by increasing the potential for moisture damage, which is related to a decrease in the surface tension of water in DIAIC solutions. Based on this proposed mechanism and the research summarized above, a number of factors may affect the extent to which DIAICs cause damage to HMA pavements:  Aggregate type: The Canadian airfields study, the Helsinki University of Technology study, and the AAPTP study all indicate that siliceous aggregates are more susceptible to DIAIC‐related damage than calcareous aggregates. Most DIAICs promote alkaline pore water environments, which may cause a reaction with siliceous aggregates similar to alkali‐silica reactivity.  Binder type: The Helsinki University of Technology study and the AAPTP study indicate that softer binders are more susceptible to DIAIC‐related damage than harder binders. However, in the CDOT study, all three binders had the same low temperature grade, but the binder with the lowest high temperature grade exhibited the most emulsification. The Helsinki University of Technology study also found that rubber‐ and gilsonite‐modified asphalts were the least susceptible to DIAIC‐related damage. The mechanisms behind these effects are uncertain.  Air void content: The AAPTP study is the only one reviewed that investigated the effect of air void content. The study found that mixes with higher air void contents experienced more DIAIC‐ related damage. Air voids increase the permeability of asphalt, which enhances the dispersion of the DIAIC solution through the pavement layers. Air voids also provide spaces where DIAICs can accumulate within the pavement. 12  DIAIC type: The studies reviewed were primarily focused on DIAICs used on airfield pavements, and investigation of the chlorides commonly used on highway pavements has been limited. Still, most studies found that DIAIC‐related damage varied with DIAIC type.  DIAIC concentration: The Canadian airfields study found that aggregate degradation was highest at DIAIC concentrations of about 1 to 2 percent. The Helsinki University of Technology study found that damage to asphalt mixes was similar at concentrations of 5 percent and 50 percent, which is not inconsistent with the Canadian airfields study. The CDOT study found that emulsification of asphalt binder increased with increasing DIAIC concentration. It is possible that DIAIC concentration has different (and possibly offsetting) effects on different damage mechanisms.  Temperature: The CDOT study found that emulsification of asphalt binder increased with increasing temperature, and was negligible below 104° F. The Helsinki University of Technology study found that PAH compounds were detectable at 104° F, but were significantly more concentrated at 212° F. TEST METHODS FOR EVALUATING RESISTANCE TO DEICER‐INDUCED DAMAGE As discussed above, it appears that the most likely mechanism for DIAIC‐induced damage to HMA pavements is an acceleration of moisture damage in mixes prone to this type of distress. The test for moisture damage susceptibility that is most widely used in the United States is AASHTO T283, also known as the modified Lottman procedure. Unfortunately, this test is not highly effective since it does not consistently and accurately identify HMA systems’ level of resistance (or lack thereof) to moisture damage. Additionally, the test results sometimes vary across different laboratories. Furthermore, AASHTO T283 is an expensive and time‐consuming procedure. It involves vacuum saturation of three HMA specimens followed by freezing and thawing. The specimens are then tested using the indirect tension (IDT) strength test. The IDT strength of these conditioned specimens is then compared to the strength of unconditioned specimens. The ratio of conditioned strength to unconditioned strength is called the tensile strength ratio (TSR), often expressed as a percentage. Values below 0.70 to 0.80 are generally considered to indicate susceptibility to moisture damage. In the AAPTP study, the research team developed a modified and simplified version of the AASHTO T283 procedure called the immersion tension (IT) test. This procedure allowed a relatively large number of HMA systems to be quickly evaluated. The IT test does not apply a vacuum saturation process prior to conditioning the specimen. Instead of conditioning the specimen in a freezing cycle, the test conditions the specimens in distilled water or DIAIC solutions at 60° C (140° F) for 4 days, since previous studies showed the combination of DIAICs and high temperatures can lead to HMA damage. The IT test procedure also omits the step of testing the dry specimen, instead focusing on comparing the tensile strengths of specimens conditioned in water with the strengths of those conditioned in DIAIC solution. The IDT strength test is performed on three replicate specimens for each conditioning level using the same procedure as in AASHTO T283. Retained tensile strength/DIAIC treatment (TSR/D) is calculated as the IDT strength after DIAIC conditioning divided by the IDT strength after conditioning in distilled water, 13 expressed as a percentage. The authors recommended that TSR/D values below 80 percent be considered evidence of DIAIC‐related damage. The AAPTP study focused on dense‐graded HMA mixes. Testing open‐graded friction course (OGFC) mixes using the AASHTO T283 method can be problematic since the material could be damaged during specimen conditioning. Therefore, a modified specimen conditioning procedure proposed by Birgisson et al. (2006) could be adopted. The major modification in this test method is to skip the step of vacuum saturation for the specimens prior to the hot water bath. Since OGFC mixes have air void levels of at least 16 percent, it is possible that the specimens would experience creep or failure during moisture conditioning. In order to eliminate the likelihood of premature specimen failure or damage during conditioning, the specimens were wrapped in 1/8‐inch wire mesh. Two clamps held the mesh in place without exerting pressure onto the specimens. A much simpler and quicker procedure for identifying susceptibility to moisture damage is the boiling test (ASTM D3625). In this test, a small sample of asphalt‐coated aggregate is boiled in distilled water for a set period of time. The water is then decanted, and the aggregate is poured out onto filter paper and examined to determine how much of the coating has been stripped from the aggregate. An analysis of data published by Kennedy and Ping indicates that the results of this test relate reasonably well to the results of the modified Lottman test (also known as AASHTO T283) (Kennedy and Ping 1991). A boiling test was used in the Helsinki University of Technology research on DIAIC damage, and it appeared to be more effective than the modified Lottman test at identifying DIAIC‐induced damage in HMA samples. The test is well suited to the study of DIAIC damage for several reasons:  The test can be run in distilled water and any number of DIAIC solutions.  The test is simple and quick, so it can be run on a large number of specimens. This facilitates a thorough study addressing a large number of factors.  The test is run at an elevated temperature, which might be necessary to promote DIAIC damage in HMA mixes.  After the test, the solution can be used for chemical analyses to determine whether PAHs or similar chemicals are present. These presence of these chemicals indicates that deicing compounds have reacted with the asphalt binder to cause HMA damage. Another test procedure that may be used to evaluate the effect of DIAICs on moisture damage in HMA is nuclear magnetic resonance (NMR) spectroscopy. NMR spectroscopy has been a powerful tool in the structural characterization of chemical compounds for a number of years. It has only recently been used in the characterization of asphalts. Work at the Western Research Institute has demonstrated that solid‐ state NMR imaging (NMRI) can be used for noninvasive chemical, physical, and morphological characterization of local regions in the interior of intact asphalt samples (Miknis and Netzel 1996). In addition, NMRI can be used to characterize chemical and physical processes noninvasively over time. With respect to DIAICs’ effect on moisture damage in HMA, the results from NMRI analysis could be used to measure the effect that the addition of DIAICs has on the penetration of water in asphalt concrete. Moisture damage is one of the major sources of failure in HMA pavements. As reported in the Helsinki University of Technology study, it is expected that mixing DIAICs with water will lower the 14 surface tension of the solution. This should result in greater penetration of the asphalt concrete by the solution and could produce a marked increase in water‐induced damage. Besides contributing to moisture damage in HMA pavements, it is also possible that DIAICs might in some cases accelerate age hardening or interact with compounds produced in asphalt binders during long‐term oxidative aging. To evaluate this hypothesis, asphalt binder/sand briquettes could be conditioned using a modified version of the pressure‐aging vessel test (AASHTO PP1). The briquettes would be made in accordance with the ADOT job mix formulas for conventional and rubberized asphalt concrete. The briquettes would be completely saturated by immersion in various DIAIC solutions while being held in a vacuum desiccator. Next, the briquettes would be artificially aged in the pressure‐aging vessel at 60° C (140° F) in the presence of water for 20 hours. Once the artificial aging of the briquettes is completed, they would be chemically analyzed using Fourier transform infrared (FTIR) analysis to determine whether the conditioning produced any unusual chemical compounds or changed the chemistry of the asphalt in a characteristic way. FTIR analysis and related chemical procedures are discussed in more detail below. IDENTIFYING DEICER‐INDUCED DAMAGE IN HMA It would be extremely useful if one or more chemical “flags” could be identified that indicate DIAIC‐ induced damage has occurred in HMA pavements. For example, the results of the Helsinki University of Technology study suggest that PAHs are the result of reactions occurring between DIAICs and asphalt binder, and indicate DIAIC‐related damage in HMA pavements. Confirmation of this finding would mean that gas chromatography tests could be used on field samples of airfield pavements to confirm that damage was caused (or accelerated) by DIAIC use. Although the AAPTP study was not able to confirm the presence of PAH molecules, it is possible that the conditioning temperature used in the AAPTP study was not high enough to trigger the chemical interaction. It is possible that other, or perhaps more specific, flags exist. In the asphalt chemistry field, there is significant interest in further basic research geared toward identifying and verifying chemical flags that indicate DIAIC‐induced damage in HMA pavements. The primary method for identifying these chemical flags would be FTIR analyses. Asphalts are composed of a variety of different types of compounds. These include relatively nonpolar compounds like aliphatics and aromatics as well as more polar compounds like carboxylic acids, sulfoxides, ketones, and nitrogen‐containing compounds. FTIR analysis can be used to identify and quantify the latter types of compounds (Petersen 1986). To obtain quantitative information on the specific carbonyl compounds (such as carboxylic acids and ketones) in asphalt, a series of infrared spectra that are the result of selective chemical reactions and the evaluation of differential spectra must be recorded. From this series of spectra, sulfoxides, ketones, carboxylic acids, carboxylic acid salts, 2‐ quinolones, anhydrides, pyrroles, and phenols can be identified and quantified. In addition, the extent of hydrogen bonding can be quantitatively estimated. Note that asphalt is composed of a mixture of mono‐, di‐, and polyfunctional compounds (Petersen 1986). 15 SUMMARY Anecdotal field evidence in Europe, Canada, and the United States suggests that some DIAIC compounds might cause or contribute to the damage of HMA pavements. However, there is a lack of laboratory evidence that DIAIC compounds can cause such damage. This might be because of the limited scope of laboratory testing conducted to date, or it could be because factors critical to DIAIC‐related damage have not been included in previous studies. It is also possible that HMA pavements subject to DIAIC use are experiencing other types of damage, such as moisture damage or snowplow damage. This pavement damage may in fact be coincidental to the use of DIAIC compounds, not the result of a cause‐and‐effect relationship. In moving forward with this project, the technical advisory committee recommended a focus on DIAICs’ impacts on Arizona’s pavements, since DIAIC‐related damage is less well understood and harder to remedy than snowplow damage. There are several possible mechanisms by which DIAIC compounds might damage HMA pavements. DIAICs may:  Act to accelerate moisture damage in HMA.  Interact with the asphalt binder to cause emulsification of the binder.  Damage aggregates as a result of wet‐dry cycling.  Accelerate age‐hardening damage to the binder. Of these, the most likely cause of damage, and the mechanism focused on by most researchers, is moisture damage. Researchers have suggested that high temperatures—those in the range of 158° F to 212° F—might be necessary in order for DIAICs to damage HMA pavements (Alatyppö and Valtonen 2007, Pan et al. 2008). One study (Alatyppö and Valtonen 2007) suggested that resurfacing HMA pavements that have been exposed to DIAICs is especially damaging to the underlying pavement because of the extremely high temperatures that are produced during paving. Given the proposed mechanisms, factors that might affect DIAIC damage to HMA pavements include DIAIC type, DIAIC concentration, asphalt binder chemistry and grade, use of crumb rubber modified binder, aggregate type, HMA permeability, and temperature. ADOT most commonly uses potassium chloride and sodium chloride for deicing and anti‐icing. The majority of pavement surfaces in Arizona are rubberized open‐graded friction courses (OGFC), and the potential for DIAIC‐induced damage is higher for this type of mix compared to conventional dense mixes. Since OGFC mixes have a high air void content, it is easier for the DIAIC solution to penetrate into the pavement structure. Therefore, a larger portion of the pavement would be exposed to DIAIC solution as compared to a conventional dense mix. If the surface tension between the DIAIC solution and the asphalt is high, the DIAIC solution could penetrate the thin film of asphalt binder on the aggregate and gradually strip it away. In addition, if a pavement that has been exposed to DIAICs is subjected to overlay, it is possible that the elevated temperature could trigger chemical reactions or between the DIAIC solution and the asphalt binder or emulsification of the binder. However, studies have also indicated that stiffer binders or polymer‐modified binders may be better able to resist DIAIC‐related damage. As with any pavement, the performance of rubberized OGFC may also be impacted by the aggregate type 16 used. In general, studies have shown that mixes containing siliceous aggregates are more susceptible to moisture damage than those made with caustic aggregates. In the studies identified in this literature review, a variety of test procedures were used to evaluate DIAIC‐induced damage in HMA pavements. Because the primary mechanism of DIAIC‐induced damage appears to be moisture damage, many of the test methods used are those normally used to evaluate moisture resistance in asphalt concrete pavements. These include boiling tests and modifications of the AASHTO T283 procedure. Gas chromatography has been used to identify PAHs in HMA pavements that appear to have been damaged by DIAICs; FTIR could be used for the same purpose. It is recommended that ADOT consider using the boiling test as an initial screening tool to evaluate combinations of aggregates, binder types, and mixture structures for moisture damage susceptibility. If the boiling test results show a significant difference before and after the test, the modified AASHTO T283 procedure could be used as a supplementary test to verify the changes in mix mechanical properties. 17 18 CHAPTER 3. LABORATORY TESTING Based on the literature review, the researchers determined that DIAIC‐induced damage to HMA mixtures is influenced by aggregate and binder properties, as well as by interactions among these properties. To investigate the extent of the phenomenon in typical ADOT materials and under typical ADOT practices, a laboratory study was undertaken with a full‐factorial experimental design. EXPERIMENTAL DESIGN Asphalt mixes were prepared using four aggregate sources and two asphalt binder sources selected from those ADOT typically uses in areas that require winter storm maintenance. Throughout this report, the aggregate sources are referred to as CM‐0323, CM‐0343, CM‐2176, and CM‐2058, and the binder sources are labeled A and B. Susceptibility to moisture damage was calculated using the boiling test by comparing samples tested in distilled water with samples tested in three chemical solutions. Table 1 shows the experimental matrix. Three DIAICs, magnesium chloride (MgCl2), potassium chloride (KCl), and sodium chloride (NaCl), were used in the study to investigate the effect of DIAICs on the various mixture types. Distilled water was used as the control to simulate the effect of moisture damage alone, without the presence of DIAIC compounds. MATERIALS Aggregate Sources Aggregates from four sources were used to construct the mixes. The aggregates were from ADOT‐ approved sources that are commonly used in the production of rubberized mixes. Their properties are shown in Table 2. As shown, the general aggregate type in CM‐0343 is basalt. The remaining aggregates are alluvial, with chemical compositions that are difficult to predict. 19 Table 1. Experimental Matrix Aggregate Source Binder Source DIAIC Distilled water Magnesium chloride A Potassium chloride Sodium chloride CM‐0323 Distilled water Magnesium chloride B Potassium chloride Sodium chloride Distilled water Magnesium chloride A Potassium chloride Sodium chloride CM‐0343 Distilled water Magnesium chloride B Potassium chloride Sodium chloride Distilled water Magnesium chloride A Potassium chloride Sodium chloride CM‐2176 Distilled water Magnesium chloride B Potassium chloride Sodium chloride Distilled water Magnesium chloride A Potassium chloride Sodium chloride CM‐2058 Distilled water Magnesium chloride B Potassium chloride Sodium chloride 20 Table 2. Aggregate Properties Aggregate Source ADOT Specification CM‐0323 CM‐0343 CM‐2176 CM‐2058 Requirement Property Mass loss at 100 revolutions (%) 4 4 3 4 9 max. Mass loss at 500 revolutions (%) 20 15 17 20 40 max. Sand equivalent value (%) 77 77 82 87 45 min. 100 100 96 100 85 min. 17 15 12 16 25 max. Percentage of aggregates with at least two fractured faces (%) Flakiness index (%) Percentage of carbonates (%) 0.7 1.1 0.8 0.5 Combined specific gravity 2.586 2.768 2.637 2.550 2.35–2.85 Corrected combined specific gravity 2.582 2.761 2.632 2.555 ‐‐ Combined water absorption (%) 1.04 1.70 1.02 1.65 2.50 max. Alluvial Basalt Alluvial Aggregate classification Alluvial 30 max. ‐‐ Asphalt Binder and Crumb Rubber Two asphalt‐rubber blends were used in this experiment. The blends were prepared by heating a known quantity of asphalt binder PG 58‐22 to 400° F and slowly stirring the crumb rubber into the hot asphalt cement. Table 3 shows the percentages of crumb rubber in each blend by weight, and Table 4 shows the rubber gradations in each blend. Table 3. Asphalt‐Rubber Binder Properties PG 58‐22 binder source Percentage crumb rubber by weight (%) 21 Blend 1 Blend 2 A B 20.0 22.7 Table 4. Crumb Rubber Gradation Percentage Passing Sieve Size Blend 1 Blend 2 ADOT Specification Requirement No. 10 100 100 100% No. 16 83 77 65–100% No. 30 44 27 20–100% No. 50 5 6 0–45% No. 200 0 0.4 0–5% Several tests were conducted to evaluate asphalt‐rubber binder stability and retention properties. The tests were timed to identify binder properties after four reaction periods, covering a time range designed to represent changes that could occur following completion of field mixing (60 minutes), after a possible job delay (4 hours), and one day after mixing. The test results are presented in Tables 5 and 6. Table 5. Physical Properties of Asphalt‐Rubber Binder (Blend 1) Minutes of Reaction Test Performed 60 120 240 1700 2000 2300 2200 1500–4000 Penetration at 39.2° F, 200 g, 60 sec (0.1 mm) 29 29 25 30 15 min. Resilience at 77° F (rebound percentage) 50 49 51 48 20 min. 148.2 150.0 148.5 152.1 130 min. Rotational viscosity at 350° F (poise) Softening point (° F) 22 1440 ADOT Specification Requirement Table 6. Physical Properties of Asphalt‐Rubber Binder (Blend 2) Minutes of Reaction Test Performed 60 135 240 1600 2500 2800 3100 1500–4000 Penetration at 39.2° F, 200 g, 60 sec (0.1 mm) 19 19 ‐‐ 24 15 min. Resilience at 77° F (rebound percentage) 46 51 ‐‐ 52 20 min. 146.0 146.5 155.5 130 min. Rotational viscosity at 350° F (poise) Softening point (° F) 146.5 1440 ADOT Specification Requirement Mix Design The asphalt mixtures were designed according to Section 414 of the ADOT Standard Specifications, “Asphaltic Concrete Friction Course (Asphalt‐Rubber).” Table 7 shows the mix properties for each aggregate type. Table 7. Mix Properties Property Design binder content (%) Bulk density (pcf) Asphalt absorption (%) Aggregate Source CM‐0323 CM‐0343 CM‐2176 CM‐2058 9.8 9.9 9.5 9.5 115.6 130.6 116.1 113.2 0.26 0.61 0.28 0.69 DIAIC Solutions Three deicer chemical solutions were used to evaluate the moisture susceptibility of the rubberized asphalt mixes: magnesium chloride (MgCl2), potassium chloride (KCl), and sodium chloride (NaCl). All solutions were 20 percent salt solutions by weight and were made with technical‐grade salts. Distilled water was used as the control solution. Test Procedure Moisture susceptibility was evaluated using the boiling test (ASTM D3625). This test is a rapid procedure for visually observing the loss of adhesion in uncompacted bituminous‐coated aggregate mixtures due 23 to the action of boiling water. The procedure involves adding loose HMA to boiling water (or in this case, boiling DIAIC solutions) and bringing the liquid back to boiling after this addition. After 10 minutes, the mixture is allowed to cool while the stripped asphalt is skimmed away. The water is drained, and the wet mixture is placed on a paper towel and allowed to dry. The percentage of the total visible area of aggregate surface that retains its asphalt binder coating is measured. RESULTS The results of the boiling test are shown in the Appendix to this report, which includes photographs of the portions of the mixture that were stripped and those that were not stripped. An analysis of variance (shown in Table 8) indicated that at a 95 percent confidence level, the amount of stripping was significantly influenced by all factors and factor interactions (p‐values < 0.001). Table 8. Analysis of Variance of Boiling Test Results Factor F‐Statistic p‐Value 1567.31 < 0.001 Binder source 268.66 < 0.001 DIAIC solution 970.15 < 0.001 Aggregate x binder 224.59 < 0.001 Aggregate x solution 144.61 < 0.001 Binder x solution 24.33 < 0.001 Aggregate x binder x solution 79.67 < 0.001 Aggregate source The percentage of stripped material differed significantly across the different aggregate types (see Figure 2). Source CM‐0323 was significantly more susceptible to moisture damage than the other three sources, with an average of 65.6 percent of the aggregate stripped. Source CM‐2176 had the best results, with an average of 10.3 percent stripped. 24 Figure 2. Perce entage of Material Strippeed by Aggregaate Source wed that mixtu ures containin ng the binderr blend from SSource A had d significantly The analyysis also show higher mo oisture suscep ptibility (on average 10.1 percent p higheer than mixtu ures containin ng the binder blend from m Source B, as a shown in Figure 3). This was expecte d, because in n general the Source A blend had lowerr viscosity thaan the Source e B blend. The e lower viscossity can be atttributed to th he lower crum mb rubber co ontent; the PG G 58‐22 binde er made with the Source A blend contained 20 perceent crumb rubber, while the PG 58‐22 bin nder made with the Source e B blend conntained 22.7 p percent crumb rubber. As reported in the literatu ure, stiffer bin nders and polymer‐modifi ed binders arre more resisttant to DIAIC‐ related daamage. 25 Note: The So ource A binde er contained 20 2 percent crrumb rubber; the Source B binder contained 22 2.7 percent crumb rubber. M Strip ped by Bindeer Source Figure 3. Perrcentage of Material The potasssium chloride e and sodium m chloride solu utions causedd significantlyy more moistu ure damage tto the rubberized asphalt mixxtures than th he other two solutions (se e Figure 4). TThe least amo ount of stripping was obtained with the magnesium chloride soluttion, which (uunexpectedlyy) caused lesss damage than n the control so olution of disttilled water. 26 Figure 4. Perrcentage of Material M Strip pped by Type of DIAIC ustrate the interactions am mong the diffeerent factors. As indicated d by the analyysis of Figures 5 through 7 illu d earrlier, the interraction termss are highly siggnificant. This means thatt the differencces in variance discussed the levels of one factor depend on the t levels of the t other facttors. For examp ple, the perce entage of stripped materiaal varied for a particular biinder blend d depending on the aggregate e source used d (see Figure 5). 5 Only the CM‐2058 C aggrregate did nott show a significant differeence in results across binder blends. The CM‐0323 agggregate show wed a particullarly high percentage of m regardless of the binder blend used. Regardding the interraction of agggregate type aand stripped material DIAIC solu ution, the CM M‐2176 aggreggate was the least influencced by the typ pe of DIAIC so olution (see FFigure 6). On the e other hand, the percentaage of strippe ed material vaaried greatly w with the solution type when the CM‐03 323 aggregate was tested. The interaaction betwee en binder ble end and DIAIC C solution typpe was also an nalyzed (see FFigure 7). Botth binder ble ends showed a large variattion in the pe ercentage of sstripped mateerial when diffferent DIAIC solutions were used. However, H two o trends clearlly stand out. Regardless off the binder b blend, a lower percentagge of stripped d material was observed when w water annd magnesium m chloride so olutions were used than when potassium chloride and sodium chlorride solutionss were used. 27 Figu ure 5. Interacction Between Aggregate Source and B Binder Blend of DIAIC Solu Figure 6. Interaction Between Agggregate Sourcce and Type o ution 28 Figure e 7. Interactio on Between Binder B Blend and Type of DIAIC Solutio on To examin ne this effect more closelyy, Figures 8 an nd 9 present tthe data on p percentage off stripped material as three‐d dimensional diagrams. d Figure 8 shows the perrcentages of stripped s mate erial for four mix combina tions (aggreggates CM‐034 43 and CM‐0 0323 blended with Source A and B binders) and the ffour DIAIC solutions. The m mixes made w with aggregate e CM‐0323 de emonstrated the t highest moisture m dam age potential. Stripping leevels were especiallyy high when the mixes werre conditioned in the potasssium chlorid de and sodium m chloride solutions (an average of o 91.7 percent stripped material). m Strippping decreassed when CM M‐0323 mixes were 0323 conditioned in the maggnesium chloride solution or in water. In addition, the mixes containing CM‐0 aggregate e exhibited a high level of stripping s not only when thhey were conditioned with h potassium aand sodium ch hloride solutio ons, but also when they were w conditionned with watter. This reveaals that thesee mixes also o had high mo oisture susceptibility. As mentioned m ea rlier, mixes b blended with tthe Source B binder exhibited less sttripping. For CM‐0343 mixxes, stripping levels were vvery high regaardless of wh hich M‐0343 aggreggate is more h hydrophilic th han the otherr binder waas used, which might implyy that the CM aggregate es used in thiss study. On th he other hand d, mixes compposed of CM‐‐0343 aggregate showed lo ower percentagges of strippin ng than CM‐0 0323 mixes in general. Thee exceptions w were when th he Source A binder was mixed d with CM‐03 343 and conditioned in the e potassium cchloride and ssodium chloriide solutions (percentaages of strippe ed material were w 98.1 percent and 50.33 percent, resspectively). This also show ws that usingg a stiffer bind der (the Source B binder in n this experim ment) can help decrease sttripping levels. 29 urprisingly, strripping levelss were higher when the miixes were con nditioned in w water than wh hen Finally, su they were e conditioned d in the magnesium chlorid de solution. Fiigure 8. Perce entage of Maaterial Strippe ed by DIAIC TType (CM‐03443 and CM‐03 323 Mixes) Figure 9 shows the perrcentages of stripped s mate erial for the rremaining fou ur mix combin nations (aggregates CM‐2176 and a CM‐2058 8 blended with Source A annd B binders)) and the fourr DIAIC solutio ons. Stripping levels were much m lower fo or these mixe es than for thee four mixes sshown in Figu ure 8. For theese four mixes, the highestt stripping levvel was 39.3 percent; p this occurred wheen the mix made with the CM‐2058 aggregate an nd the Source e B binder blend was condiitioned in thee sodium chlo oride solution. The lowest strripping level was w 1.7 perce ent, yielded by the combinnation of CM‐‐2176 aggregaate and Sourcce A binder conditioned in the t magnesiu um chloride so olution. As inn Figure 8, two o groups werre still distinctt from each h other: Mixes conditioned d in the potasssium and soddium chloridee solutions haad higher strip pping levels (an average of 25 percent stripped material) than mixees conditioned d in the wateer and magnesium s (an average a of 6.6 percent stripped materi al). The beneefit of using a stiffer binderr was chloride solutions not signifiicant for thesse four mixes.. 30 entage of Maaterial Strippe ed by DIAIC TType (CM‐21776 and CM‐20 058 Mixes) Fiigure 9. Perce Stripping levels across all mix types and DIAIC so olutions are shhown in Tablee 9. For the p purpose of compariso on, the perce entage of strip pped material observed in the DIAIC so olutions was n normalized with respect to o the percentage of strippe ed material observed o in w water. A value greater than n 1 indicates tthat the combination of thaat mix type an nd DIAIC solution yielded aan increased level of moisture damage d with conditiioning that mix m in water. Conversely, C a value less thaan 1 indicatess that the compared combination of mix typ pe and DIAIC solution yield ded less moissture damagee than conditioning that mix in water did. Although th he overall anaalysis of varian nce showed tthat strippingg levels were h highly affecteed by all factorss (binder type e, aggregate type, and DIAIC type), the most severe stripping wass observed wh hen the sodium chloride an nd potassium chloride solu utions were ccombined with the CM‐034 43 and CM‐03 323 d the Source A binder. Usin ng magnesium m chloride ass a DIAIC soluttion appeared d to yield thee least mixes and moisture damage, exce ept when it was w used with h the Source B binder blended with the CM‐2176 aggregate n sodium and e. Note that even e in this caase, stripping levels were sstill much low wer than when d potassium m chloride solutions were used. u 31 Table 9. Level of Stripping Caused by DIAIC Solutions vs. Water Solution Mix Type (Binder + Aggregate Source) Water MgCl2 NaCl KCl Source A + CM‐0323 1 0.60 1.51 1.49 Source B + CM‐0323 1 0.29 0.81 1.20 Source A + CM‐0343 1 0.21 2.05 1.80 Source B + CM‐0343 1 0.41 6.19 5.13 Source A + CM‐2058 1 0.18 16.75 32.70 Source B + CM‐2058 1 0.75 4.01 2.45 Source A + CM‐2176 1 0.77 0.47 17.00 Source B + CM‐2176 1 1.81 5.02 3.78 32 CHAPTER 4. SUMMARY AND RECOMMENDATIONS SUMMARY The literature review and the results of the laboratory experiment conducted for this project reflect both the complexity and the challenges of evaluating the effects of various DIAICs on HMA pavements. While the literature is inconclusive regarding DIAICs’ impact on skid resistance, DIAICs have been shown to affect pavement structure and to cause a loss of strength in HMA pavements. Most studies on DIAICs’ effects on HMA pavement have focused on formate‐ and acetate‐based deicers used at airports, finding that these chemicals have significantly damaged the pavements by causing emulsification of the asphalt or accelerating moisture damage. The hypothesized damage mechanism of these deicers is a combination of chemical reactions, emulsifications, and distillations as well as the generation of additional stress inside the HMA pavements. However, little is known about the effects of chloride‐based deicers, which are widely used in ADOT’s winter maintenance efforts on HMA pavements. Based on the results of the laboratory experiments conducted for this project on the interactions between DIAIC solutions and HMA pavement, all tested factors (binder type, aggregate type, and DIAIC solution type) and the interactions of these factors significantly affect the extent of DIAIC‐related moisture damage. The most significant factor in these experiments was aggregate type, which presented differences of up to 55 percent in stripping levels as measured with the boiling test. However, the most severe stripping was observed when sodium chloride and potassium chloride solutions were used to condition mixes made with CM‐0323 and CM‐0343 aggregates and Source A binder. The Source A binder has lower viscosity than the Source B binder, and this result is consistent with the literature. The CM‐0323 and CM‐0343 aggregates are alluvial and basalt sources, respectively. For the CM‐0323 mix, because of its high moisture susceptibility, even when a stiffer binder with a higher percentage of crumb rubber was used, the tests still showed very high stripping levels. Based on these test results, the following mix types and DIAIC solutions are recommended for areas where deicing chemicals will be used on rubberized asphalt pavements:  DIAIC solution: Magnesium chloride is recommended as the best DIAIC solution for use on rubberized asphalt pavement. Potassium chloride and sodium chloride are not recommended, as both increase the potential for moisture damage.  Aggregate source: Source CM‐2176 aggregate is recommended for use in the rubberized asphalt mix. CM‐0323 aggregate is not recommended unless it is blended with high‐viscosity binder; in addition, magnesium chloride should be the only DIAIC used with CM‐0323.  Binder type: Stiffer binder is recommended, as it apparently helps mitigate moisture damage. Table 10 summarizes the best and the worst combinations of mix types and DIAICs. 33 Table 10. Best and Worst Combinations of Mix Types and DIAICs DIAIC Solution Mix Type (Binder + Aggregate Source) MgCl2 NaCl KCl Source A + CM‐0323    Source B + CM‐0323    Source A + CM‐0343    Source B + CM‐0343    Source A + CM‐2058    Source B + CM‐2058    Source A + CM‐2176    Source B + CM‐2176     Highly recommended (0–15% stripping)  Recommended (15–30% stripping)  Avoid (31–60% stripping)  Definitely avoid (61–100% stripping) RECOMMENDATIONS Because of the complexity of the interactions among DIAICs, asphalt binder, and aggregate and the lack of literature on the effect of chloride‐based deicers on rubberized HMA pavements, the research team suggests conducting the following experiments to obtain more detail on the interactions among the factors. Experiment 1: Evaluate the interactions between DIAIC solutions and ADOT‐approved crumb rubber modified binders using the boiling test (ASTM D3625). This experiment should test several DIAIC solution concentrations, ranging from 1 percent to 10 percent. The proposed experiment design is presented in Table 11. These tests would verify the existence of chemical interactions between DIAIC solutions and ADOT crumb rubber modified binders. The tests would also clarify the results of the experiments conducted during this project, which showed that all factors examined were significant. Experiment 2: From the results of Experiment 1, combinations of binders and DIAICs that create a high potential for moisture damage should be identified. For Experiment 2, selected aggregates should be incorporated as a third factor for testing with these identified combinations. The selected aggregates should be tested prior to mixing with the binders to determine their chemical compositions. The aggregate–binder mixes should be tested as loose mixes using the boiling test, or through modified immersion tension testing as proposed in the AAPTP study (Advanced Asphalt Technologies 2009), using the modified specimen conditioning procedure for OGFC mixes proposed by Birgisson et al. (2006). 34 Table 11. Proposed Experiment Design for Testing Binder–DIAIC Interaction Binder Type DIAIC Solution Concentration Water N/A Binder Type DIAIC Solution Concentration Water N/A 1% NaCl Source A KCl MgCl2 1% 2% NaCl 5% 2% 5% 10% 10% 1% 1% 2% Source B 5% KCl 2% 5% 10% 10% 1% 1% 2% MgCl2 5% 10% 2% 5% 10% The results of these follow‐up experiments could be used to make more detailed recommendations to ADOT as to the best combination of binder type, aggregate classification, and DIAIC solution for ADOT’s winter maintenance program. 35 36 REFERENCES Advanced Asphalt Technologies. 2009. Effect of Deicing Chemicals on HMA Airfield Pavements. Project 05‐03, Final Report. Auburn, Alabama: Airfield Asphalt Pavement Technology Program, Auburn University. Alatyppö, V., and J. Valtonen. 2007. “Experiences on the Effects of De‐Icing Chemicals on Bituminous Airfield Runways in Finland.” Presented at the Federal Aviation Administration Worldwide Airport Technology Transfer Conference, Atlantic City, NJ, April 2007. Birgisson, B., R. Roque, A. Varadhan, T. Thai, and L. Jaiswal. 2006. Evaluation of Thick Open Graded and Bonded Friction Courses for Florida. Gainesville: University of Florida. Curtis, C. W., R. L. Terrel, L. M. Perry, S. Al‐Swailm, and C. J. Braanan. 1991. “Importance of Asphalt‐ Aggregate Interactions in Adhesion.” Journal of the Association of Asphalt Paving Technologists 60: 476–532. Farha, M. H., Y. Hassan, A. O. Abd El Halim, A. G. Razaqpur, A. El‐Desouky, and A. Mostafa. 2002. “Effects of New Deicing Alternatives on Airfield Asphalt Concrete Pavements.” Presented at the Federal Aviation Administration Technology Transfer Conference, Atlantic City, NJ, 2002. Hassan, Y., A. O. Abd El Halim, A. G. Razaqpur, and M. Farha. 2001. “Laboratory Investigation of Effect of Deicing Chemicals on Airfield Asphalt Concrete Pavement Materials.” Proceedings of 2nd International Conference on Engineering Materials, Vol. 1: 299–308. Kennedy, T. W., and W. V. Ping. 1991. “An Evaluation of Effectiveness of Antistripping Additives in Protecting Asphalt Mixtures from Moisture Damage.” Journal of the Association of Asphalt Paving Technologists 60: 230–263. Kim, O., C. A. Bell, and R. G. Hicks. 1985. “The Effect of Moisture on the Performance of Asphalt Mixtures.” Special Technical Publication 899: Evaluation and Prevention of Water Damage to Asphalt Pavement Materials: 51–72. West Conshohocken, PA: American Society for Testing and Materials. Majidzadeh, K., and F. N. Brovold. 1968. State of the Art: Effect of Water on Bitumen‐Aggregate Mixtures. Highway Research Board Special Report 98. Washington, D.C.: Highway Research Board. Martinez, F. C., and R. A. Poecker. 2006. Evaluation of Deicer Applications on Open Graded Pavements. Salem: Oregon Department of Transportation. Miknis, F. P., and D. A. Netzel. 1996. “Use of Nuclear Magnetic Resonance Imaging to Study Asphalt.” Preprints of Papers, American Chemical Society, Division of Fuel Chemistry 41(4): 1327–1331. 37 Pan, T. S., X. He, and X. Shi. 2008. “Laboratory Investigation of Acetate‐Based Deicing/Anti‐Icing Agents Deteriorating Airfield Asphalt Concrete.” Journal of the Association of Asphalt Paving Technologists 77: 773–794. Thelen, E. 1958. “Surface Energy and Adhesion Properties in Asphalt‐Aggregate Systems.” Highway Research Board Bulletin 192: 63–74. Tunnicliff, D. G., and R. E. Root. 1982. “Antistripping Additives in Asphalt Concrete—State‐of‐the‐Art 1981.” Proceedings of the Association of Asphalt Paving Technologists Technical Sessions 53: 265–293. 38 APPENDIX: BOILING TEST RESULTS This appendix presents the results of the boiling test (ASTM 3625) laboratory experiments. Researchers conducted the test on 32 combinations of aggregates, binders, and DIAICs (see Table 1 for the experimental matrix). For each combination, the test was replicated twice with different samples. CM‐0323 Aggregate Stripped Source A Binder Not stripped Stripped Replicate 1—63.7% Stripped CM‐0323 Aggregate Stripped Distilled Water Not stripped Replicate 2—67.1% Stripped Source A Binder Not stripped Magnesium Chloride Stripped Replicate 1—39.4% Stripped Not stripped Replicate 2—39.5% Stripped 39 CM‐0323 Aggregate Stripped Source A Binder Not stripped Stripped Replicate 1—96.9% Stripped CM‐0323 Aggregate Stripped Source A Binder Not stripped CM‐0323 Aggregate Not stripped Replicate 2—97.7% Stripped Sodium Chloride Stripped Replicate 1—99.3% Stripped Stripped Potassium Chloride Not stripped Replicate 2—98.1% Stripped Source B Binder Not stripped Distilled Water Stripped Replicate 1—45.2% Stripped Not stripped Replicate 2—43.4% Stripped 40 CM‐0323 Aggregate Stripped Source B Binder Not stripped Stripped Replicate 1—4.6% Stripped CM‐0323 Aggregate Stripped Source B Binder Not stripped CM‐0323 Aggregate Not stripped Replicate 2—14.1% Stripped Potassium Chloride Stripped Replicate 1—80.2% Stripped Stripped Magnesium Chloride Not stripped Replicate 2—78.9% Stripped Source B Binder Not stripped Sodium Chloride Stripped Replicate 1—96.1% Stripped Not stripped Replicate 2—85.8% Stripped 41 CM‐0343 Aggregate Stripped Source A Binder Not stripped Stripped Replicate 1—3.6% Stripped CM‐0343 Aggregate Stripped Not stripped Magnesium Chloride Stripped Replicate 1—0.2% Stripped Stripped Not stripped Replicate 2—2.4% Stripped Source A Binder CM‐0343 Aggregate Distilled Water Not stripped Replicate 2—0.9% Stripped Source A Binder Not stripped Potassium Chloride Not stripped Replicate 1—98.7% Stripped Stripped Replicate 2—97.5% Stripped 42 CM‐0343 Aggregate Stripped Source A Binder Not stripped Stripped Replicate 1—49.1% Stripped CM‐0343 Aggregate Stripped Source B Binder Not stripped CM‐0343 Aggregate Not stripped Replicate 2—51.4% Stripped Distilled Water Stripped Replicate 1—2.1% Stripped Stripped Sodium Chloride Not stripped Replicate 2—0.9% Stripped Source B Binder Not stripped Magnesium Chloride Stripped Replicate 1—1.2% Stripped Not stripped Replicate 2—1.1% Stripped 43 CM‐0343 Aggregate Stripped Source B Binder Not stripped Stripped Replicate 1—25.5% Stripped CM‐0343 Aggregate Stripped Potassium Chloride Replicate 2—25.5% Stripped Source B Binder Not stripped Not stripped Sodium Chloride Not stripped Replicate 1—0.5% Stripped CM‐2176 Aggregate Not stripped Stripped Replicate 2—0.9% Stripped Source A Binder Stripped Distilled Water Not stripped Replicate 1—8.8% Stripped Stripped Replicate 2—2.4% Stripped 44 CM‐2176 Aggregate Stripped Source A Binder Not stripped Stripped Replicate 1—2.1% Stripped CM‐2176 Aggregate Stripped Source A Binder Not stripped CM‐2176 Aggregate Not stripped Replicate 2—1.2% Stripped Potassium Chloride Stripped Replicate 1—6.8% Stripped Stripped Magnesium Chloride Not stripped Replicate 2—6.6% Stripped Source A Binder Not stripped Sodium Chloride Stripped Replicate 1—6.3% Stripped Not stripped Replicate 2—2.8% Stripped 45 CM‐2176 Aggregate Stripped Source B Binder Not stripped Stripped Replicate 1—5.7% Stripped CM‐2176 Aggregate Stripped Not stripped Magnesium Chloride Stripped Replicate 1—2.4% Stripped Stripped Not stripped Replicate 2—4.4% Stripped Source B Binder CM‐2176 Aggregate Distilled Water Not stripped Replicate 2—1.7% Stripped Source B Binder Not stripped Potassium Chloride Stripped Replicate 1—28.7% Stripped Not stripped Replicate 2—23.1% Stripped 46 CM‐2176 Aggregate Stripped Source B Binder Not stripped Stripped Replicate 1—30.0% Stripped CM‐2058 Aggregate Not stripped Stripped Distilled Water Stripped Replicate 1—6.5% Stripped Stripped Not stripped Replicate 2—32.5% Stripped Source A Binder CM‐2058 Aggregate Sodium Chloride Not stripped Replicate 2—8.7% Stripped Source A Binder Not stripped Magnesium Chloride Stripped Replicate 1—13.1% Stripped Not stripped Replicate 2—14.4% Stripped 47 CM‐2058 Aggregate Stripped Source A Binder Not stripped Stripped Replicate 1—27.4% Stripped CM‐2058 Aggregate Stripped Source A Binder Not stripped CM‐2058 Aggregate Not stripped Replicate 2—30.1% Stripped Sodium Chloride Not stripped Replicate 1—36.6% Stripped Stripped Potassium Chloride Stripped Replicate 2—39.7% Stripped Source B Binder Not stripped Stripped Replicate 1—7.2% Stripped Distilled Water Not stripped Replicate 2—12.4% Stripped 48 CM‐2058 Aggregate Stripped Source B Binder Not stripped Magnesium Chloride Stripped Replicate 1—6.8% Stripped CM‐2058 Aggregate Stripped Replicate 2—7.9% Stripped Source B Binder Not stripped Stripped Potassium Chloride Stripped Replicate 1—22.0% Stripped CM‐2058 Aggregate Not stripped Not stripped Replicate 2—26.1% Stripped Source B Binder Not stripped Sodium Chloride Stripped Replicate 1—40.0% Stripped Not stripped Replicate 2—38.5% Stripped 49