ABSTRACT Title of Document: STABILIZATION OF RECYCLED BASE MATERIALS WITH HIGH CARBON FLY ASH Bora Cetin, Master of Science, 2009 Directed By: Associate Professor, Ahmet H. Aydilek, Civil and Environmental Engineering Department A study was conducted to stabilize low stiffness road surface material with high carbon fly ash. The non-cementitious Maryland fly ash was activated with another recycled material, lime kiln dust (LKD). California bearing ratio (CBR) and resilient modulus tests were conducted to determine the strength and stiffness, respectively, of the stabilized materials. Addition of LKD and curing of specimens generally increased CBR and summary resilient modulus (SMR) and lowered plastic strains, whereas fly ash addition alone decreased the strength and stiffness due to the non-cementitious nature of the ash. CBR increased with increasing CaO content as well as with CaO/SiO2 and CaO/(SiO2 + Al2O3) ratio of the mixtures; however, these parameters could not be correlated with the SMR. The unpaved road materials stabilized with LKD and fly ash is expected to lose 31 to 67% of their initial moduli after twelve cycles of freezing and thawing. Finally, required base thicknesses were calculated using the laboratory-based strength parameters. STABILIZATION OF RECYLED BASE MATERIALS WITH HIGH CARBON FLY ASH By Bora Cetin Thesis submitted to the Faculty of the Graduate School of the University of Maryland, College Park, in partial fulfillment of the requirements for the degree of Master of Science 2009 Advisory Committee: Associate Professor Dr. Ahmet H. Aydilek, Advisor Professor Dr. Sherif M. Aggour, Committee Member Professor Dr. Deborah J. Goodings, Committee Member ? Copyright by Bora Cetin 2009 ii ACKNOWLEDGEMENT First and foremost, I would like to thank my advisor, Professor Ahmet H. Aydilek, for his enthusiasm and guidance throughout my graduate studies at the University of Maryland ? College Park. Thanks also to Professors Deborah J. Goodings and M. Sherif Aggour for being part of my committee. I also would like to thank Professor Yucel Guney for his effort in microscopy analysis of cured specimens. Thanks to Maryland State Highway Administration to help us to run resilient modulus test and provide us necessary information?s about highway construction that are used in State of Maryland. iii TABLE OF CONTENTS ACKNOWLEDGEMENT ..................................................................................................ii TABLE OF CONTENTS...................................................................................................iii LIST OF TABLES .............................................................................................................iv LIST OF FIGURES.............................................................................................................v 1. INTRODUCTION.........................................................................................................1 2. MATERIALS................................................................................................................6 3. METHODS....................................................................................................................8 3.1 CALIFORNIA BEARING RATIO TEST..............................................................8 3.2 RESILIENT MODULUS TEST .............................................................................9 3.3 MICROSCOPY ANALYSIS????????????????????.11 4. RESULTS AND ANALYSIS.....................................................................................12 4.1 RESULTS OF CALIFORNIA BEARING RATIO ..............................................12 4.2 RESULTS OF RESILIENT MODULUS TEST...................................................15 5.2.1 Effects of Curing Time, LKD addition on Summary Resilient Moduli and Plastic Strain..............................................................................................................15 5.2.2 Effect of Freeze ? Thaw on Summary Resilient Moduli ................................19 5. PRACTICAL IMPLICATIONS .................................................................................21 5.1 HIGHWAY BASE DESIGN ................................................................................21 5.2 COST CALCULATIONS.....................................................................................23 6. CONCLUSIONS...........................................................................................................25 REFERENCES..................................................................................................................27 TABLES............................................................................................................................33 FIGURES ..........................................................................................................................42 APPENDIX A ...................................................................................................................59 APPENDIX B ...................................................................................................................62 APPENDIX C ...................................................................................................................70 APPENDIX D ...................................................................................................................73 APPENDIX E?????????????????????????????76 iv LIST OF TABLES TABLE 1 a. Index properties of the materials used in the current study. TABLE 1 b. Chemical compostion of the fly ashes. Concentrations of major minerals determined by X ? ray fluorences spectroscopy analysis. TABLE 2. Legend and compositions of the mixtures. TABLE 3. Power model fitting parameters and measured plastic strains for resilient modulus tests. TABLE 4. CBR and summary resilient modulus values. TABLE 5. Chemical compositons of mixtures prepared with three different fly ashes. TABLE 6. Effect of freezing and thawing cycles on summary resilient moduli.. TABLE 7. Required base thickness for different mixture designs for two traffic conditions under excellent drainage condition. TABLE 8. Cost analysis for different mixture designs for two different traffic conditions under excellent drainage condition. TABLE A.1. Properties of Maryland fly ashes with ASTM C 618 chemical and physical criteria for class C and class F fly ash. TABLE B.1. AASHTO 307 ? 99 resilient modulus testing sequence for base and subbase materials. TABLE C.1. Predicted SMR values based on CBR values. TABLE D.1. Base thickness values based on SMR values for traffic case I TABLE D.2. Base thickness values based on SMR values for traffic case II TABLE E.1. Required base thickness for different mixture designs for two different traffic conditions under good drainage condition. TABLE E.2. Required base thickness for different mixture designs for two different traffic conditions under fair drainage condition. TABLE E.3. Required base thickness for different mixture designs for two different traffic conditions under poor drainage condition. TABLE E.4.Cost analysis for different mixture designs for two different traffic conditions under good drainage condition. TABLE E.5.Cost analysis for different mixture designs for two different traffic conditions under fair drainage condition. TABLE E.6.Cost analysis for different mixture designs for two different traffic conditions under poor drainage condition. v LIST OF FIGURES FIGURE 1. Comparison of unpaved road material and the two materials used in base construction in Maryland to (a) VDOT base materials specification, and (b) AASHTO specification. FIGURE 2. Effect of curing time on CBR of mixtures prepared with (a) Brandon Shores fly ash, (b) Paul Smith fly ash, (c) Dickerson Precipitator fly ash. FIGURE 3. SEM photograph of (a) Brandon Shores fly ash, (b) unpaved road material amended with 10% Brandon Shores fly ash and 2.5% LKD by weight, (c) Dickerson precipitator fly ash, and (d) unpaved road material amended with 20% Dickerson Precipitator fly ash and 5% LKD by weight. FIGURE 4. EDX plot of the SEM photograph of (a) Dickerson Precipitator fly ash, and (b) 7-day cured unpaved road material amended with 20% Dickerson Precipitator fly ash and 5% LKD by weight. FIGURE 5. Effect of LKD contents on (a) CBR and (b) SMR of 28-day cured specimens. FIGURE 6. Effect of CaO content, (b) CaO/SiO2, (c) CaO/(SiO2 + Al2O3), and fineness on CBR. FIGURE 7. Effect of (a) Silica ratio, (b) alumina ratio, (c) lime saturation factor, and (fly ash percentage on CBR. FIGURE 8. CBR versus measured and predicted SMR for (a) 1 day cured specimens, and (b) 7 days cured specimens. FIGURE 9. Effect of curing time on SMR of mixtures prepared with (a) Brandon Shores fly ash, (b) Paul Smith fly ash, and (c) Dickerson Precipitator fly ash. FIGURE 10. Resilient Modulus of the 28-daycured specimens with varying bulk stresses: Mixtures prepared with (a) Paul Smith fly ash, and (b) Dickerson Precipitator fly ash. FIGURE 11. Effect of CaO content, (b) CaO/SiO2, (c) CaO/(SiO2 + Al2O3), and fineness on SMR. FIGURE 12. Effect of (a) Silica ratio, (b) alumina ratio, (c) lime saturation factor, and (d) fly ash percentage on SMR. FIGURE 13. Effect of freeze and thaw cycles on SMR values. FIGURE 14. Effect of freeze and thaw cycles on water contents. FIGURE 15. Summary resilient modulus as a function of base layer thickness. FIGURE B.1.Compaction curves for (a) conventional base materials, (b) mixtures prepared with Brandon Shores fly ash, (c) mixtures prepared with Paul Smith fly ash, and (d) mixtures prepared with Dickerson Precipitator fly ash. FIGURE B.2. Photo of Resilient Modulus Testing Equipment. FIGURE B.3. GAB (a), URM (b), and BRG (c) . vi FIGURE C.1. CBR versus measured SMR vs. CBR graph for (a) 1 day cured specimens, and (b) 7 days cured specimens. FIGURE D.1. Required thickness vs. number of freeze and thaw cycles (a) traffic case I, and traffic case II. FIGURE E.1. Effect of LKD addition (a), (b), (c) and traffic conditions (d), (e), (f) on total construction fee of the highway base layer. 1 1. INTRODUCTION American Society of Civil Engineers estimates that $2.2 trillion is needed over a five- year period to bring the nation?s infrastructure to a good condition. Establishing a long- term development and maintenance plan is a national priority. Large volumes of earthen materials are used in construction each year in the United States. In many cases, these materials can be replaced with reclaimed highway paving materials, secondary materials, suitable waste materials and construction debris that are normally disposed in landfills, and can generate millions of dollars savings to taxpayers. Reuse in construction has several benefits, including reduction in solid waste disposal costs incurred by industry, reduction in landfill requirements, minimization of damage to natural resources caused by excavating earthen materials for construction, obtaining added value from waste materials, conservation of production energy, and ultimately providing sustainable construction and economic growth. Legislations have been promulgated in many states that remove barriers to large- scale beneficial reuse of recycled materials, to reduce construction costs and increase sustainability. As a result, there is a policy shift, both nationally and at the state level, aimed at substantially increasing the use of such materials in geotechnical construction. For instance, one material that has been increasingly dealt with is road surface material from an unpaved road or a road undergoing rehabilitation and uses it as the base layer for newly paved roads (Hatipoglu et al. 2008). Due to its low strength and stiffness, the material often has to be stabilized by adding good quality granular material, or by blending with hydrated lime and fly ash. 2 Over 60% of the electricity generated in the United States is produced by coal combustion, with resulting abundant quantities of fly ash as residue, which presents another environmental challenge. Fly ash has been used as bulk fill material in geotechnical fill, such as in construction of embankments, dikes, and road subgrade (DiGioia and Nuzzo 1972, Gray and Lin 1972). The advantages of using fly ash as a bulk fill material include low cost, low unit weight, and good strength. In Eastern parts of the United States, anthracite and bituminous coals are burned by the power plants and, as a result, non-cementitious ashes (Class F or off-spec fly ashes) are produced. These fly ashes contain high amounts of SiO2 and Al2O3, which can react with an activator rich in CaO (e.g., lime, cement, lime or cement kiln dust) in the presence of moisture to form cementitious compounds for stabilization applications where additional strength gain is needed (e.g., base stabilization). Fly ash is generally reused in concrete production. However, the fly ashes produced by several power plants in United States occasionally contains significant amounts of unburned carbon (i.e., high loss on ignition) due to the increasingly common use of low nitrogen oxide (NOx) and sulphur oxide (SOx) burners in recent years. This ash has a carbon content of 12-25%, cannot be efficiently re-burnt by using current technology, and has no value as a concrete additive as the unburned carbon tends to adsorb the air entrainment admixtures that are added to the cement to prevent crack formation and propagation. These ashes are typically classified as off-spec fly ashes meaning that they do not meet the physical and chemical requirements criteria outlined in ASTM C 618. Recent data indicate that approximately 68% of this high-carbon fly ash (HCFA) is placed in landfills, thereby consuming valuable land space and creating the 3 potential to impact terrestrial and aquatic resources in Maryland. Roadways have high potential for large volume use of HCFA. HCFA can be activated with lime kiln dust (a disposed residue of lime production plants) and used as the base layer for newly paved roads. Significant efforts have been made to use fly ashes in stabilization of highways base structures, unpaved roads and soil stabilization. Arora and Aydilek (2005) evaluated the engineering properties of Class F fly ash amended soils as highway base materials. Cement-activated fly ash increased the California bearing ratio (CBR), unconfined compression strength, and resilient modulus (Mr) of sandy soils with plastic fines contents ranging from 18 to 30%. Similar observations were made Vishwanathan et al. (1997) when silty and sandy soils were stabilized with lime-activated-Class F fly ash for their possible use in highway bases. Hatipoglu et al. (2008) showed through unconfined compression, CBR and resilient modulus tests that self-cementitious Class C fly ash can be a viable binder for stabilization of recycled asphalt pavement material (RPM) for base applications. Li et al. (2007) conducted laboratory tests to evaluate the use of RPM blended with fly ash as base course. CBR of RPM increased from 3-17 to 70-94 with the addition of fly ash. Similarly, addition of fly ash caused more than two-fold increase in Mr of laboratory RPM specimens. Camargo (2008) showed that addition of 10-15% by weight of Class C fly ash increases the CBR and resilient modulus of recycled pavement material (RPM) and road surface gravel by 3 to 6 and 9 to 22 times, respectively. Camargo (2008) has also observed a 6 to 11 and 34 to 57 times increase in CBR and resilient modulus of road surface gravel when stabilized with 10 and 15% Class C fly ash, respectively. In a study conducted by Wen et al. (2007, 2008) high carbon self- 4 cementitious fly ash was shown to increase the strength and stiffness of RPM. CBR and Mr of fly ash-stabilized RPM were higher than CBR and Mr for RPM without fly ash; both engineering properties were comparable to the CBR of conventional crushed aggregate. The plastic deformations for RPM were generally decreased by addition of fly ash. Previous research has shown that self-cementing fly ash can be an effective binder for stabilizing soils for highway bases (Consoli et al. 2001, Zaman et al. 2003, Arora and Aydilek 2005, Edil et al. 2006, Kumar et al. 2007, Buhler and Cerato 2007, Hatipoglu et al. 2008, Saylak et al.2008, Shao et al. 2008, Wen et al. 2008, Camargo et al. 2008). However, limited information exists on the reuse of high carbon off-spec fly ash in construction of highway pavements. This is particularly important when high carbon fly ash is non-cementitious (e.g., Maryland fly ashes) and calcium-rich activators are required to generate pozzolanic reactions. Thus, there is a need to evaluate the strength and stiffness of base layers stabilized with high carbon fly ash. To respond to this need, a battery of tests was conducted on unpaved road surface material-fly ash mixtures amended with lime kiln dust for its possible use in highway base construction. California bearing ratio (CBR) and resilient modulus (MR) as well as scanning electron microscopy (SEM) analyses were conducted to investigate the engineering properties of granular soil- fly ash mixtures with and without lime kiln dust (LKD), and to study the effect of curing time on soil-fly ash-LKD mixtures. The effect of winter conditions were also evaluated by performing resilient modulus tests on the specimens after a series of freeze-thaw cycles. 5 Another issue that impedes soil stabilization with fly ash is the potential for groundwater and other environmental impacts caused by metals in the fly ash. Fly ash contains a small amount of trace metals that can have environmental consequences when fly ash is used in geotechnical applications. Even though an environmental impact analysis was necessary, it was left out of the scope of the current project. 6 2. MATERIALS An unpaved road material (URM) and two conventional base materials were used in this study. The URM was collected from a highway construction site in Caroline County, Maryland. Any debris and foreign materials in the soil were removed by hand and, by sieving through the 19 mm sieve. The soil is classified as poorly graded sand with gravel (SP) according to Unified Soil Classification System (USCS), and A-1-b (0) according to the American Association of State Highway and Transportation Officials (AASHTO) Classification System. The material did not exhibit any plasticity per ASTM D 4318. Two base materials, Bank Run Gravel (BRG) and Graded Aggregate Base (GAB), used in highway construction in Maryland were tested as control soils. GAB meets the Maryland State Highway Administration (MDSHA) and AASHTO M-147 specifications and is termed as a high quality base material in Maryland. BRG is less commonly used in highway construction but was selected due its comparable particle size distribution with URM. GAB was excavated from an underground limestone mine located in Frederick, Maryland. The material was crushed upon mining, passed through a series of sieves to meet the gradations given in AASHTO M-147, and stockpiled in pits. BRG is originally mined from a sandstone mine located in Middletown, Maryland and was stockpiled in pits. Both materials were collected directly from the pits and delivered to the laboratory. The soils did not contain any organic matter or exhibit plasticity in Atterberg limit tests (ASTM D 4318). The fines content of BRG was 12%, and it was classified as SP-SM and A-1-b (0) according to the USCS and AASHTO, respectively. GAB included 4% fines by weight and was classified as SP and A-1-a (0) according to 7 the USCS and AASHTO, respectively. The two base materials and URM were stored in airtight buckets upon transfer to the laboratory in order to preserve their natural water content. Particle size distribution curves for the unpaved road material and conventional base materials are shown in Figure 1 along with the AASHTO M-147 and Virginia Department of Transportation (VDOT) specifications used for base construction. MDSHA specifications are the same of AASHTO M-147 specifications, thus not included in the figure. The obtained particle size distributions indicated that GAB satisfied the AASHTO M-147 and VDOT particle size distribution limits for highway bases whereas BRG and URM are tend to be outside of the limits. Physical properties of the two soils are summarized in Table 1. The fly ashes used in this study were obtained from three power plants in Maryland: Brandon Shores, Paul Smith and Dickerson Precipitator. All three fly ashes consisted primarily of silt-size particles and contained 79 to 91% fines (passing the 75- mm sieve). Specific gravity (Gs) of fly ashes ranged between 2.17 and 2.37 per ASTM D 854. The fly ashes investigated in this study were classified as off-specification fly ashes (neither C or F type according to ASTM C 618) due their high loss on ignition (LOI>6). The chemical compositions of all three fly ashes are provided in Table 1. Since the three fly ashes do not have high cementing potential (i.e., low CaO), lime kiln dust (LKD) was used to initiate pozzolanic reactions for stabilization of the soil. LKD was obtained from Carmeuse Lime and Stone Company, Pittsburgh, Pennsylvania, and contained approximately 60% CaO by weight. The specific gravity of LKD is 2.97. 8 3. METHODS 3.1 California Bearing Ratio Test The California bearing ratio (CBR) test is a penetration test for evaluation of the mechanical strength of road subrgrades and base courses. Soil-fly ash mixtures used in the CBR tests were prepared by mixing air-dried soil with a specified percent fly ash by weight. Fly ash percentages were selected as 10 and 20% to cover the typical range used in soil stabilization (ACAA 1999, Edil et al. 2002, Bin-Shafique et al. 2004). Initially, high percentages by weight of LKD (10 to 15%) were used as the activator for large volume use of this recycled material. However, due to extremely high strength values (CBR>150), more modest percentages of 2.5 and 5% by weight were selected. All specimens for the CBR tests were compacted at their optimum moisture contents (OMC) using the standard Proctor effort (ASTM D 698 Method B). Table 2 provides the OMC and maximum dry unit weights (?dm) of the mixtures based on compaction tests. After compaction, the specimens were extruded with a hydraulic jack, sealed in plastic wrap, and cured for 1, 7, and 28 days at 100% relative humidity and controlled temperature (21? 2 OC) before testing. CBR tests on specimens without fly ash/LKD were tested immediately after compaction (i.e., no curing). All CBR tests were conducted by following the methods outlined in AASHTO T-193 and ASTM D 1883. The specimens were unsoaked and the tests were performed with 1.27 mm/min strain rate using the Geotest Instrument S5840 Multi-Loader loading frame. The equipment had a maximum 9 loading capacity of 44.8 kN. Duplicate specimens were tested for CBR tests as quality control, and the averages of these two tests were reported as results. 3.2 Resilient Modulus Test Resilient modulus test provides the stiffness of a soil under a confining stress and a repeated axial load. The procedures outlined in AASHTO T 307-99, a protocol for testing of base and highway base and subbase materials, were followed for resilient modulus tests. Unpaved road material and the two conventional base materials were mixed with fly ash and LKD at 10-20% and 2.5-5% by weight, respectively, and specimens of 101.6 mm in diameter and 230.2 mm in height were compacted in split molds at their OMC in eight layers using the standard Proctor energy. After compaction, the specimens were removed from the molds, sealed in plastic wrap, and were cured for 1, 7, and 28 days at 100% relative humidity and controlled temperature (21? 2 OC) before testing. The testing procedures were the same for BRG and GAB, except the specimens were compacted in split molds 152 mm in diameter and 305 mm in height. A Geocomp LoadTrac-II loading frame and associated hydraulic power unit system was used to load the specimens. Conditioning stress was 103 kPa. Confining stress was kept between 20.7 and 138 kPa during loading stages, and the deviator stress was increased from 20.7 kPa to 276 kPa and applied 100 repetitions at each step. The loading sequence, confining pressure, and data acquisition were controlled by a personal computer equipped with RM 5.0 software. Deformation data were measured with 10 external linear variable displacement transducers (LVDTs) that had a measurement range of 0 to 50.8 mm. Resilient moduli from the last five cycles of each test sequence were averaged to obtain resilient modulus for each load sequence. The resilient modulus of soil is usually nonlinear and is dependent on the stress level. This nonlinear behavior was defined in this study using the common model developed by Moosazadh and Witczak (1981): 21RM KK ?= (1) where MR is resilient modulus, K1 and K2 are constants, ? (= d? + 3? c) is bulk stress, ?c is the isotropic confining pressure, and ?d is the deviator stress. A summary resilient modulus (SMR) was computed at a bulk stress of 208 kPa, following the guidelines provided in NCHRP 1-28A. The approach is also consistent with the suggestions in the recent mechanistic-empirical design guide on new and rehabilitated pavement structures to provide a constant resilient modulus for chemically stabilized materials (ARA 2004). The same bulk stress level was used in this study to verify the model. The power-model based values are summarized in Table 3. With few exceptions high R2 values (R2 >0.8) were obtained from regression analyses performed on the model, indicating that the mixtures have a response similar to that of granular materials. To observe the effect of winter conditions on resilient moduli, some of the mixtures were subjected to strength and hydraulic conductivity tests after a series of freeze-thaw cycles. Specimens with varying fly ash and LKD contents and fly ash types 11 were compacted at their optimum moisture contents and 100% of maximum standard Proctor dry unit weight and following the procedures outlined in ASTM D 698. After 7 days of curing, the specimens were frozen in a temperature chamber at -23? 1 OC for 24 hours and then thawed in a humidity chamber at 100% relative humidity and controlled temperature (21? 2 OC) for 23 hours per ASTM D 560. Specimens were frozen and thawed at zero overburden stress. The water content and resilient modulus were measured at the end of 4, 8, and 12 freeze-thaw cycles. Resilient modulus tests were conducted as described previously. Duplicate specimens were tested for most of the resilient modulus tests as quality control, and the averages of these two tests are reported as results. 3.3 Microscopy Analysis Scanning electron microscopy (SEM) analyses were conducted on 7-day cured specimens. The specimens were initially treated with acetone, and a critical point drying apparatus was utilized to replace the acetone with CO2. The specimens were held on an aluminum sample holder with adhesive tape. Later, they were coated with gold to minimize any charge build-up. Microstructure and chemical composition of the samples were examined under LEO 440 Model SEM using the energy dispersive X-ray (EDX) technique. 12 4. RESULT AND ANALYSIS 4.1 RESULTS OF CBR TESTS The CBR test results are provided in Table 4 and Figure 2. Both BRG and GAB have higher CBR than unpaved road material (URM). The URM has significantly lower CBR than 50, a generally accepted limit for base applications (Asphalt Institute 2003), and thus required stabilization for use in highway construction. In all cases, CBR of stabilized mixtures is higher than that of URM and is comparable with or higher than the CBR of the two conventional base materials even after 1 day of curing. The data also indicate that mixing with only fly ash is not sufficient enough to increase the strength due to its non-cementitious nature (i.e., low free lime, CaO) of the ashes, and addition of a lime source, such as LKD, is necessary to start pozzolanic reactions. The SEM photographs in Figure 3 show the coating of fly ash particles as a result of cement or lime kiln dust (LKD) addition, which may be an indicator of an increase in CBR (Conner 1990). Relatively higher amounts of calcium are evident as a result of the addition of LKD, as shown in the EDX plots of Figure 4. CBR of LKD amended soil-fly ash mixtures also increase with increasing curing time (Figure 2). As LKD is mixed with moist soil, the hydration of calcium oxide (CaO) causes the formation of (Ca(OH)2), and disassociation of (Ca(OH)2) favors dissolution of silica and alumina in fly ash. This phenomenon gives rise to formation of calcium silicate hydrate (CSH) and calcium aluminate silicate hydrate gels (CASHs) around soil particles. It is speculated that the delayed release of CaO in LKD caused these increases and the temperature of the curing chamber and availability of 100% relative humidity 13 also contributed to the cementitious reactions. The increase in CBR after 1 day of curing is relatively modest; however, CBR of the 7 and 28-day cured specimens increased up to 6 and 7 times, respectively. Similar increases in strength with increasing curing period were reported by previous researchers (Vishwanathan et al. 1997, Arora and Aydilek 2005, Guney et al. 2006). Two different amounts of lime kiln dust (LKD) were added to the soil specimens. As seen in Figure 5, an increase in LKD amount increases the CBR values significantly due to cementation of the particles by the LKD. The rate of increase is higher initially, and increasing the LKD amount from 0% to 2.5% by weight had a greater effect than increasing the amount from 2.5% to 5% by weight. The CBR of unpaved road material increases at least three and five times due to addition of 2.5% and 5% LKD by weight, respectively. Similar trends were observed by Consoli et al. (2001) during strength testing of soils stabilized with fly ash and carbide lime. The CBR of all 7-day and 28-day cured specimens tested in the current study exceeds 50. In order to evaluate the effect of mixture chemical composition on observed strength, a paired t-test was conducted at significance level of 0.05, corresponding to tcritical =tcr = 2.06 for CBR test results and tcr=2.09 for resilient modulus test results. CBR values are plotted against CaO content, and CaO/SiO2 and CaO/ (SiO2 + Al2O3) ratios and fineness of fly ash in Figure 6. The two ratios stay below 1.0, and are well below 3 and 2.5, the ratios documented for Portland cement (Table 4). As expected, the data in Figure 6 suggest that CBR increases with increasing CaO content and CaO/SiO2 and CaO/(SiO2 + Al2O3) ratios and the best fit curves to the data produced modest R2 values (0.79, 0.66 and 0.73, respectively). Janz and Johansson (2002) indicated that the 14 CaO/SiO2 ratio can be a good indicator of pozzolanic reactions, and larger CaO/SiO2 generally yields higher strength values. Tastan et al. (2009) showed that the cementing potential of materials can be strongly related to CaO content of the binder as well to these ratios. Tastan et al. (2009) also reported that the CaO/SiO2 and CaO/ (SiO2 + Al2O3) ratios typically range from 0.5 to 1.0 and from 0.4 to 0.7, respectively, for fly ash- stabilized subgrade soils. These observations are, in general, consistent with the findings obtained in the current study. Fineness of fly ash refers to particles retaining on the 45 ?m sieve (U.S. No. 325 standard sieve size) and defines the surface area of fly ash particles present in per unit weight. If the fly ash is self-cementitious (i.e., Class C), higher fineness percentages typically enhance the reaction rate which result in faster gain of strength at earlier stages. As seen in Figure 6d, the correlation between the fineness and CBR is poor (t < 1.96), mainly due to non-cementitious nature of the Maryland fly ashes. Attempts were also made to relate CBR to three commonly used ratios in cement production: silica ratio (SR), alumina ratio (AR) and lime saturation factor (LSF). The silica ratio (SR=SiO2/ (Al2O3 + Fe2O3)) represents the required energy to combine raw materials in a stabilization application. When SR increases, it becomes harder to combine the raw materials whereas a decrease in SR suggests an increase in the ability of solid materials to become liquid. The alumina ratio (Al2O3/Fe2O3) is important as it alumina- to-iron ratio in cement is known to be an indicator of sulfate resistance, heat generation, and admixture compatibility issues. The lime saturation factor (LSF) is dependent on the C3S-to-C2S ratio in the finished cement, where the early and delayed age strength development is governed by C3S and C2S, respectively (PCA 2009). LSF typically 15 remains between 0.95 and 0.98 for Portland cement and higher LSF indicates the presence of excess free lime which is likely to remain unreacted in the mixture (Taylor 1997). No clear trends or correlations can be visualized in Figure 7. This is not surprising as these ratios are generally not used as indicators of pozzolanic reactions. As mentioned before, increase in fly ash content only does not change stresses and no correlation can be observed between fly ash percentage and CBR (Figure 7d). 4.2 RESULTS OF RESILIENT MODULUS TESTS 4.2.1 Effects of Curing Time, LKD Addition on Summary Resilient Moduli and Plastic Strain Summary resilient moduli of the tested specimens are given in Table 4. GAB has the highest CBR of all three unstabilized materials due to its high gravel content. The measured SMR of GAB falls into the range of suggested SMR for SP or SP-SM soils reported in the Mechanistic-Empirical Design Guide (165 to 228 MPa), whereas the same is not true for BRG and URM (ARA 2004). Similar to CBR data, URM has low summary resilient modulus, justifying the need for stabilization with a calcium-rich binder. The order of SMR is compatible with that of CBR; however, a strong correlation was not obtained when CBR was plotted against SMR (Figure C1 in Appendix). Camargo (2008) also attributed the difference to the application of different magnitudes of deformations to the specimens during testing and measurement of two separate 16 geomechanical properties (i.e., small deformations and measurement of stiffness during resilient modulus test versus large deformations and bearing capacity determination during CBR test). Attempts were also made in the past to correlate CBR to resilient modulus data by using two well-known empirical equations by Powell et al. (1984) and Haukelom and Foster (1960), respectively: SMR = 7.6CBR0.64 (2) SMR = 10CBR (3) where SMR is summary resilient modulus in MPa. As seen in Figure 8, both equations overpredict the measured resilient modulus data. Similar observations were also made by Sawangsuriya and Edil (2005) and Acosta et al. (2006). Due to low correlations observed between CBR and SMr (R2=0.33 and 0.32 for 1 and 7-day cured specimens, respectively), no further attempt was made to develop an empirical equation to predict resilient modulus from the CBR data. Average plastic strains were calculated for all base materials during resilient modulus testing using data from the LVDTs. Plastic strain for a resilient modulus test was calculated as the sum of the plastic strains for each loading sequence, excluding the plastic strains in the conditioning phase. The plastic strains (?plastic) for two Maryland base materials and URM, along with stabilized soils are summarized in Table 3. BRG and GAB showed average plastic strains of 1% and 0.94%, respectively, whereas URM 17 showed an average plastic strain of 0.97%. The calculated strains for the high quality Maryland base materials are comparable with those of conventional base aggregates reported in the literature (Camargo 2008). The high plastic strain of BRG and URM is attributed to their relatively higher sand content (68% and 67%, respectively) as compared to GAB (51%), consistent with the observations of Camargo (2008). A variation of SMR with curing time is shown in Figure 9. SMR increases with increasing curing time. The average change in SMR caused during the 1-day curing was about 35%, and increasing the curing time from 0 to 1 day had a slightly greater effect than increasing the curing time from 1 to 7 or 7 to 28 days (assuming that the SMR of mixtures were no different than that of URM at 0 days). Curing of specimens longer periods also resulted in lower plastic strains (Table 3). Similar to CBR test results, the SMR increased with increasing LKD amount due to cementitious reactions formed between fly ash and LKD, and the summary resilient modulus of unpaved road material increased 1.4 to 4 times as a result of stabilization (Figure 5). The SMR ranged between 157 and 510 MPa, and the maximum SMR was recorded for BS20+5 LKD (unpaved road material mixed with 20% Brandon Shores fly ash and 5% LKD by weight) upon 28 days of curing. It can be concluded that an LKD content of 5% by weight leads to reasonably high SMR values and addition of LKD beyond that amount may not be necessary. Moreover, Cross and Young (1997) reported higher initial cracking of recycled pavement materials with increasing cementitious fly ash contents. On the other hand, the maximum LKD content was set at 5% in the current study solely considering the cost-effectiveness; however, additional testing is required to obtain the LKD contents beyond which SMR no longer increases. 18 Table 3 summarizes the two constants, K1 and K2, for the resilient modulus power function model along with the best fit correlation coefficients. The resilient modulus is plotted versus bulk stresses for the mixtures prepared with two of the fly ashes in Figure 10. An increase in resilient modulus with increasing bulk stress is observed for all specimens, which is in agreement with the behavior generally observed for granular soils (AASHTO T-307-99). The optimum moisture content of the mixtures compacted using standard Proctor effort range from 9% to 13%, and the difference between the maximum dry densities of the specimens is insignificant (Table 2). Therefore, the difference in resilient modulus is attributed solely to the variation in LKD content. As expected, resilient modulus increases with increasing LKD content at a given bulk stress due to production of more cementitious compounds with LKD addition. Figure 10 further suggests that fly ash generally acts as a bulking agent and does not contribute to resilient modulus, as the resilient modulus at a given LKD content and bulk stress either stays the same or decreases with increasing fly ash content, with the exception of the specimen prepared with 10% Paul Smith fly ash. LKD addition caused an increase in SMR but also decreased the plastic strain of URM. In general, addition of non-cementitious fly ash did not cause significant changes in plastic strains (Table 3). Figure 11 suggests that SMR tends to increase with increasing mixture CaO content, and CaO/SiO2 and CaO/(SiO2 + Al2O3) ratios, even though CaO/SiO2 ratio is the only one that exhibited correlation with resilient modulus (t>tcritical=2.09). Furthermore, SMR was not correlated with silica ratio, alumina ratio, lime saturation factor, or fly ash percentage (Figure 12). 19 4.2.2 Effect of Freeze-Thaw on Summary Resilient Moduli Stabilized highway construction material should be able to resist against climatic stresses, especially freeze-thaw cycles (TFHRC 2002). Subjecting the specimens to strength after freeze-thaw (F-T) cycles and recording the change in weight have been reported as indicators of durability. However, the evaluation of durability by weight loss as a result of freeze-thaw cycles (ASTM D 560) has been dropped by some state agencies as the procedure is overly severe, and does not totally simulate field conditions. Previous research indicates that 8 to 12 cycles of freezing and thawing could be considered adequate in investigating the effect of F-T cycles on different engineering parameters including strength (Zaman and Naji 2003). In this study, six different mixtures were cured for 7 days as normally practiced in pavement construction. The specimens were subjected to resilient modulus tests following a series of freezing and thawing cycles. The test results are summarized in Table 6 and Figure 13. The summary resilient modulus ratio (SMR Ratio = SMRn/ SMRi) is the ratio of summary resilient modulus after n freeze-thaw cycles (SMRn) to the initial summary resilient modulus (SMRi). The specimens either gain strength or lose only 3 to 12% of their initial resilient modulus after four cycles, and then SMR starts to decrease indicating the detrimental effects of freeze and thaw cycles. The highest decreasing rate of SMR can be observed between the fourth and eighth cycle, and the specimens lose 31 to 67% of their initial moduli after twelve cycles of freezing and thawing. Similar trends were observed for 20 unbound materials by Simonsen et al. (2002). Rosa (2006) also reported a 20 to 66% reduction in SMR of various coarse and fine-grained soils. The effect of freeze-thaw on resilient modulus can be explained in terms of retardation or acceleration of the cementitious reactions. Freezing action retards the cementitious reactions, which causes a reduction in stiffness: conversely, thawing action contributes to an increase in SMR via accelerating the cementitious reactions. The freezing and thawing compensated each other in first four cycles of soil specimens. It is believed that between cycles of 4 and 12, freezing caused the breaking of the cement bonds in the mixture and resulted in a significant decrease in SMR. As compared to previous studies conducted on sandy soils and soils with some plasticity (Arora and Aydilek 2005, Rosa 2006, Camargo 2008), a larger change can be observed in the moduli of specimens tested in the current study. This may be attributed to the high gravel content and nonplastic nature of the current mixtures, which have relatively high porosity and susceptibility to frost action. The effect of freeze-thaw is consistent with the volume changes of specimens. The volume changes remain nearly constant within the first four cycles, after which increases significantly evidenced by the changes in water contents shown in Figure 14. It is believed that freezing process caused breakage of the chemical bonds and allowed water to freely penetrate into the pores, thereby causing large increases in water contents, i.e., up to 89% increase in water content after 12 cycles of freezing and thawing. 21 5. PRACTICAL IMPLICATIONS 5.1 Highway Base Design CBR and resilient modulus test results were used to estimate the thickness of the base layer in a pavement by following the procedures defined in the AASHTO Guide (1993). Low traffic (Case I) and high traffic (Case II) conditions were simulated by using 5 million and 50 million equivalent single-axle loads (ESALs or W18), respectively. The overall standard deviation (So) and reliability (ZR) were assumed to be 0.35 and 95%, respectively. Structural numbers (SN) for two traffic conditions were back-calculated using Equation 2. (4) where ?PSI is design serviceability loss and MR is the roadbed material effective resilient modulus. The values were selected as 1.9 and 34.5 MPa, respectively, based on Huang (1993). An asphalt layer thickness of 102 mm for Case I and 152 mm for Case II was selected. The resilient modulus of asphalt was assumed to be 2965 MPa, which corresponded to a structural coefficient of a1 = 0.44 according to the AASHTO Guide (1993). A resilient modulus of 103 MPa (corresponding to a structural coefficient of a3 = ( ) ( ) ( ) ( )( ) ( ) 07.8Mlog32.21SN10944.0 5.12.4PSIlog20.01SNlog36.9SZWlog R1019.510100R18 ??+++ ??+?+?+?= 22 0.11) and a thickness of 406 mm (D3) were assumed for the subbase layer for both cases. The structural coefficient of the base layer (a2) was calculated for its corresponding CBR or SMR values using the procedure given in the AASHTO Guide (1993). The CBR and SMR of 28-day cured specimens were used due to their common use in highway construction. Finally, the base thicknesses were calculated using the following formula: m2a mD3a1D1aSN 2D 2 33 ??= (5) where m2 and m3 are the drainage modification factors for the base and subbase layer, respectively, and were chosen as 1.2, 1.0, 0.8, 0.6 for excellent, good, fair, poor drainage conditions, respectively, within the pavement system (Huang 1993). Table 7 shows the required base thicknesses for Cases I and II under excellent drainage conditions. Furthermore, the required base thickness decreased with increasing resilient modulus, as seen in Figure 15. The change in drainage conditions require higher base thickneeses as reflected in Tables E.1-E.3 in Appendix E. The base layer thicknesses of all stabilized mixtures based on SMR are consistently lower than that of unpaved road material and generally comparable or lower than that of the two typical Maryland base materials. Increase in fly ash content at fixed LKD content results in higher base thicknesses due to non-cementitious nature of the ashes used. In addition, a decrease in amount of lime kiln dust (LKD) and increase in traffic load required thicker base layers in highway construction. An analysis of curing time effect on required base thicknesses were not conducted; however, it is well-known 23 that increase in curing time is likely to yield lower base thickness since the specimens gain strength at later stages. As discussed before, winter conditions generally lead to a decrease in strength of lime-treated mixes. This would generally require larger base thicknesses, in particular after 4 cycles of freezing and thawing as presented in Tables D1 and D2 and Figure D1 of Appendix. However, it should be noted that the climatic stresses may have unexpected effects on the soil mixtures, particularly in short term (i.e., during construction), and therefore precautions should be taken to protect specimens from in-situ freezing conditions. 5.2 Cost Calculations A simple cost analysis, considering the material, hauling and transportation costs only, was performed to on all soils and mixtures. In the current study, three different fly ashes and one type of lime kiln dust were used at varying percentages by weight. The road construction site was assumed to be the Route 1 expansion project site located in College Park, Maryland. The fly ashes were available at no cost. The Brandon Shores, Paul Smith and Dickerson power plants were located about 43 km, 101 km, 52 km, respectively, from the construction site. The URM, a material commonly used in unpaved road construction in Caroline County, Maryland, was available at $4/t. The distance between the construction site and URM plant was 64 km. Lime kiln dust material was available from the manufacturer for about $16/t. However the closest lime kiln dust supplier was in Pittsburgh, PA which was 390 km away from the construction site. The GAB plant was 77 km away from the construction site, and the cost of the GAB material was $10/t. The 24 BRG plant was 83.2 km away from the construction site and the cost of the BRG materials was $10/t. A fuel charge of $0.5/t and hauling costs of $0.25/t were assumed for all materials. Lane widths in the United States can range from 3 m (low volume roads) to 5 m (highway ramps) in width, and a typical design width of 4 m was selected for the Route 1 expansion project. To represent a typical roadway, a four-lane roadway was considered with two 2-m shoulders. The detailed cost analysis summarized in Table 8 indicates that using stabilized URM in a roadway base application has a clear advantage over using other conventional earthen base materials. The factors that mainly affect the cost are traffic volume and drainage conditions. The required base layer thicknesses increase due to a decrease in drainage quality of highways (Table E4-E6). Moreover, increasing traffic volume increases the total construction cost increases significantly (Figure E1). Addition of the activator (i.e., LKD) decreases the cost, even under low drainage conditions, due to its positive effect on required base thickneesses. 25 6. CONCLUSIONS Roadways are one of the largest construction fields, and reuse of suitable waste materials in their construction can provide significant cost savings while meeting the objectives of the United States Federal Highway Administration Green Highways Partnerships initiative. A laboratory study was conducted to investigate the feasibility of reusing chemically stabilized road surface material in construction of highway bases. Non- cementitious off-spec high carbon fly ash was activated lime kiln dust and used to stabilize an unpaved road material (URM) collected from Maryland. The effects of lime kiln dust (LKD) and fly ash addition, and curing time on strength and stiffness of highway bases were studied. The effects of winter conditions on stiffness were examined by performing resilient modulus tests on the specimens after a series of freeze-thaw cycles. The base thicknesses were calculated for all mixture designs by using their CBR and summary resilient moduli (SMR) values. The observations are summarized as follows: 1) Addition of lime kiln dust (LKD) and curing of specimens increase CBR and SMR significantly, whereas increase in fly ash content generally decreases the strength and stiffness due to the non-cementitious nature of the high carbon Maryland fly ash and that the fact that ash acts as a bulking agent. Almost all specimens have CBR values higher than 50, a limit typically considered for construction of base layers. Measured SMRs were comparable with the ones reported for highway bases in previous studies. 26 2) Both BRG and GAB exhibited higher CBR and SMR than unpaved road material (URM). The order of SMR was also compatible with that of CBR in spite of the differences in nature of the two test methods. The two Maryland base materials, BRG and GAB, had comparable plastic strains with those documented for conventional base aggregates in the literature. In general, lower plastic strains were obtained by addition of LKD as well as with increasing curing time whereas fly ash addition did not affect the strains significantly. 3) CBR increased with increasing CaO content as well as with CaO/SiO2 and CaO/(SiO2 + Al2O3) ratios; however, these parameters could not be correlated with the SMR. Moreover, silica and alumina ratios and lime saturation factor, three common parameters for definition of cementitious activity, are not likely to affect the CBR or SMR. 4) Early stages of freezing and thawing did not cause detrimental effects on resilient modulus; however, the unpaved road materials stabilized with LKD and fly ash lost 31 to 67% of their initial moduli after twelve cycles of freezing and thawing. Such high changes in SMR are attributed to the frost susceptibility of mixtures due to their high gravel content and nonplastic nature 5) Lower base thickness would be required if higher amount of lime kiln dust materials is used during construction of the base layer. It should also be noted that an increase in LKD content and decrease in traffic load are likely to decrease the required thicknesses. 27 6) Significant cost reduction be expected by stabilizing unsuitable roadway materials with high carbon fly ash and lime kiln dust while the total cost is mainly affected by the traffic volume and pavement drainage conditions 28 REFERENCES AASHTO Guide (1993). ?Guide for Design of Pavement Structures?, American Association of State Highway and Transportation Officials, Washington, D.C. ACAA (2003). ?2001 Coal Combustion Product Production and Use?, American Coal Ash Association, Denver, Colorado ARA (2004). ?Guide for Mechanistic-Empirical Design on New and Rehabilitated Pavement Structures,? NCHRP Project 1-37A. Prepared for National Cooperative Highway Research Program, Washington, D.C. Arora, S. and Aydilek, A.H. (2005). ?Class F Fly Ash Amended Soils as Highway Base Materials?, Journal of Materials in Civil Engineering, Vol. 17, No. 6, pp. 640- 649. Asphalt Institute (2003). ?Thickness Design ?Highways and Streets?, Manual Series, No. 1, Asphalt Institute, Lexington, Kentucky, 110 p. Bin-Shafique, S., Edil, T., Benson, C., and Senol, A. (2004). ?Incorporating a fly ash stabilized layer into pavement design-Case study.? Geotechnical engineering, Institution of Civil Engineers, London, Vol. 157, No. GE4, 239?249. Buhler, S.R. and Cerato, A.B. (2007). ?Stabilization of Oklahoma Expansive Soils Using Lime and Class C Fly Ash?, ASCE Geotechnical Special Publication, 162, pp. 1- 10. Camargo, F.F. (2008). ?Strength and Stiffness of Recycled Base Materials Blended with Fly Ash?, M.S. Thesis, University of Wisconsin-Madison, 123 p. 29 Conner, J.R. (1990). ?Chemical Fixation and Solidification of Hazardous Wastes?, Van Nostrand Reinfold, New York, 692 p. Consoli, N. C., Prietto P. D. M., Carraro, J. A. H., and Heineck, K., S. (2001). ?Behavior of Compacted Soil-Fly Ash-Carbide Lime Mixtures." Journal of Geotechnical and Geoenvironmental Engineering, Vol. 127, No. 9, pp. 774-782. Cross, S. A. and Young, D. A. (1997). Evaluation of Type C Fly Ash in Cold-in-Place Recycling, Journal of the Transportation Research Board, 1583, pp. 82?90. DiGioia, A. M. and Nuzzo, W. L. (1972). ?Fly Ash as Structural Fill,? J. Power Div., ASCE, New York, Vol. 98 (1), pp. 77-92. Edil, T.B., Acosta , A.A., and Benson, C.H. (2006). ?Stabilizing Soft Fine-Grained Soils with Fly Ash?, Journal of Materials in Civil Engineering, Vol. 18, No. 2, pp. 283- 294. Edil, T. B., Benson, C., Bin-Shafique, M., Tanyu, B., Kim, W. and Senol, A. (2002). ?Field evaluation of construction alternatives for roadways over soft subgrade?, Journal of the Transportation Research Board, 1786, pp. 36-48. Gray, D. H., and Lin, Y. K. (1972). ?Engineering Properties of Compacted Fly Ash,?J. of Soil Mech. and Found. Engrg., ASCE, New York, Vol. 98 (4), pp. 361-380. Guney, Y., Aydilek, A.H., and Demirkan, M.M. (2006). ?Geoenvironmental Behavior of Foundry Sand Amended Mixtures for Highway Subbases?, Waste Management, Vol. 26, pp. 932-945. Hatipoglu, B., Edil, T., and Benson, C. (2008). ?Evaluation of base prepared from road surface gravel stabilized with fly ash?, ASCE Geotechnical Special Publication, 177, pp. 288-295. 30 Heukelom, W., and Foster, C. (1960). ?Dynamic Testing of Pavements.? Journal of Soil Mechanics and Foundation Division, Vol. 86, No. 1, pp.1?28. Huang, Y. H. (1993). ?Pavement Analysis and Design?, Prentice-Hall, Inc., New Jersey, 805 p. Janz, M. and Johansson, S.-E. (2002). "The Function of Different Binding Agents in Deep Stabilization," Report 9. Swedish Deep Stabilization Research Centre, Linkoping, Sweden. Kumar, A., Walia, B.S., and Bajaj, A.(2007). ?Influence of Fly Ash, Lime, and Polyster Fibers on Compaction and Strength Properties of Expansive Soils?, Journal of Materials in Civil Engineering, Vol. 19, No. 3, pp. 242-248. Li, L., Benson, C. H., Edil, T. B., Hatipoglu, B., and Tastan, E. (2007). ?Evaluation of recycled asphalt pavement material stabilized with fly ash?, ASCE Geotechnical Special Publication (CD-ROM), 169. Moosazedh, J., and Witczak, M. (1981). ?Prediction of Subgrade Moduli for Soil that Exhibits Nonlinear Behavior?, Journal of the Transportation Research Board, 810, pp. 10-17. PCA (2009). Portland Cement Association, www.cement.org Powell, W., Potter, J., Mayhew, H., and Nunn, M. (1984). The Structural Design of Bituminous Roads, TRRL Laboratory Report 1132, Transportation and Road Research Laboratory, Crowthorne, Berkshire, U.K, 62 p. Rosa, M. (2006). ?Effect of Freeze and Thaw Cycling on Soils Stabilized using Fly Ash,? MS Thesis, University of Wisconsin-Madison, Madison, WI. 31 Saylak, D., Mishra, S.K., Mejeoumov, G.G., and Shon, C.C. (2008). ?Fly Ash-Calcium Chloride Stabilization in Road Construction?, Proceedings of 87th Annual Meeting (CD ROM), Transportation Research Board, Washington, D.C. Sawangsuriya, A. and Edil, T. B. (2005), ?Evaluating stiffness and strength of pavement materials, geotechnical engineering.? Geotechnical Engineering, Institution of Civil Engineers, London, Vol. 158, No. GE4, 217?230. Shao, L., Liu, S., Du, Y., Jing, F. and Fang, L. (2008). ?Experimental Study on the Stabilization of Organic Clay with Fly Ash and Cement Mixed Method?, ASCE Geotechnical Special Publication, 179, pp.20-27. Simonsen, E., Janoo, V., and Isacsson, U. (2002). ?Resilient properties of unbound road material during seasonal frost conditions?, Journal of Cold Region Eng., 16(1), pp. 28-50. Tastan, O., Benson, C.H., Edil, T.B., and Aydilek, A.H. (2009). ?Stabilization of Organic Soils with Fly Ashes?, Journal of Geotechnical and Geoenvironmental Engineering, submitted. Taylor, H.F.W. (1997). ?Cement Chemistry?, Second edition, Thomas Telford, Inc. Vishwanathan, R., Saylak, D., and Estakhri, C. (1997). ?Stabilization of Subgrade Soils Using Fly Ash?, Proceedings of the Ash Utilization Symposium, CAER, Kentucky, pp. 204-211. Wen, H., Baugh, J., and Edil, T. (2007). ?Use of Cementitious High Harbon Fly ash to Stabilize Recycled Pavement Materials as Pavement Base Material?, Proceedings of 86th Annual Meeting (CD ROM), Transportation Research Board, Washington, D.C. 32 Wen, H., Warner, J., and Edil, T. (2008). ?Laboratory comparison of crushed aggregate and recycled pavement material with and without high-carbon fly ash?, Proceedings of 87th Annual Meeting (DVD), Transportation Research Board, Washington, D.C. Zaman, M., and Naji, K. (2003). ?Effect of freeze-thaw cycles on class C fly ash stabilized aggregate base?, Proceedings of 82nd Annual Meeting (CD ROM), Transportation Research Board, Washington, D.C. 33 TABLES 34 TABLE 1a. Index properties of the materials used in current study Classification Sample Cu Gs wopt (%) ?d (kN/m3) LL (%) PI (%) Gravel Content (%) Fines Content (<75 ?m) (%) Fineness (>45 ?m) (%) USCS AASHTO BRG 25 2.52 8.0 20.4 NP NP 20 12 0 SP-SM A - 1 ? b (0) GAB 39 2.81 4.1 23.8 NP NP 45 4 0 SP A - 1 ? a (0) URM 6.7 2.64 13.4 18.8 NP NP 30 3 0 SP A - 1 ? b (0) BS 0.43 2.17 ? ? NP NP ? 91 25 ML A - 2 ? 4 (0) PS 11 2.2 ? ? NP NP ? 80 51 ML A - 2 ? 4 (0) DP 3.6 2.37 ? ? NP NP ? 79 34 ML A - 2 ? 4 (0) . TABLE 1b. Chemical composition of the fly ashes. Concentrations of major minerals were determined by X-ray fluorescence spectroscopy analysis. All concentrations are in percentage by weight. Chemical Composition Fly ash LOI (%) SiO2 (%) Al2O3 (%) Fe2O3 (%) CaO (%) K2O (%) TiO2 (%) MgO (%) Na2O (%) Cr2O3 (%) P2O5 (%) SrO (%) BaO (%) BS 13.4 45.1 23.1 3.16 7.8 1.7 1.4 0.8 0.3 0.02 0.09 0.06 0.06 PS 10.7 50.8 26.9 5.5 0.7 2.2 1.5 0.6 0.2 0.02 0.2 0.03 0.05 DP 20.5 34.9 24.4 12.6 3.2 1.1 1.3 0.5 0.3 0.03 1.0 0.2 0.11 BRG: Bank run gravel, GAB: Graded aggregate base, URM: Unpaved road material, PS: Paul Smith fly ash, DP: Dickerson Precipitator fly ash, BS: Brandon Shores fly ash, LOI: Loss on ignition. Gs: Specific gravity, Cu: coefficient of uniformity, Cc: coefficient of curvature, woptm: optimum water content, ?dmax: maximum dry unit weight, LL: liquid Limit, PL: plastic limit, NP: Nonplastic. 35 TABLE 2. Legend and compositions of the mixtures. Legend of Mixtures Fly Ash Content (%) LKD Content (%) Optimum Water Content (%) Maximum Dry Unit Weight (kN/m3) BRG 0 0 8.0 20.4 GAB 0 0 4.1 23.8 URM 0 0 13.4 18.8 10 BS + 2.5 LKD 10 2.5 10 19.2 10 BS + 5 LKD 10 5 9.5 19.2 20 BS + 2.5 LKD 20 2.5 10 18.5 20 BS + 5 LKD 20 5 13 17.4 10 PS + 2.5 LKD 10 2.5 9.0 18.8 10 PS + 5 LKD 10 5 10 18.8 20 PS + 2.5 LKD 20 2.5 12 17.3 20 PS + 5 LKD 20 5 13 17.0 10 DP + 2.5 LKD 10 2.5 9.0 19.1 10 DP + 5 LKD 10 5 10 19.4 20 DP +2.5 LKD 20 2.5 10 18.1 20 DP + 5 LKD 20 5 12 18.0 BS: Brandon Shores fly ash, PS: Paul Smith fly ash, DP: Dickerson Precipitator fly ash, LKD: Lime kiln dust, BRG: Bank run gravel, GAB: Graded aggregate base, URM: Unpaved road gravel. The numbers that follow the fly ashes and LKD indicate the percentages by weight of admixtures added to the soil. 36 TABLE 3. Power model fitting parameters and measured plastic strains for resilient modulus tests. Fly Ash Type Fly Ash Content (%) LKD Content (%) K1 K2 ?plastic (%) K1 K2 ?plastic (%) K1 K2 ?plastic (%) 1-Day Cured or no curing for BRG, GAB and URM 7-Day Cured 28-Day Cured 10 2.5 3260 0.57 0.83 6593 0.53 0.62 1412 0.93 0.49 10 5 6470 0.49 0.67 3595 0.69 0.5 1636 0.81 0.49 20 2.5 3260 0.57 0.74 1873 0.75 0.73 1300 0.84 0.63 Brandon Shores 20 5 85000 0.46 0.67 4225 0.59 0.59 36473 0.2 0.47 10 2.5 5896 0.49 0.63 8193 0.43 0.61 10670 0.33 0.54 10 5 6341 0.43 0.73 12410 0.41 0.4 16508 0.38 0.34 20 2.5 6022 0.47 1.07 4904 0.52 0.66 3623 0.57 0.66 Paul Smith 20 5 8158 0.42 0.83 12876 0.33 0.59 15629 0.32 0.49 10 2.5 3220 0.61 0.76 3973 0.55 0.72 2047 0.87 0.44 10 5 4768 0.56 0.59 4800 0.55 0.42 10310 0.53 0.29 20 2.5 3169 0.62 0.7 8830 0.41 0.57 2689 0.72 0.54 Dickerson Precipitator 20 5 6500 0.49 0.56 11329 0.41 0.43 7367 0.54 0.4 BRG 0 0 2450 0.60 1.0 - - - - - - GAB 0 0 7700 0.40 0.94 - - - - - - URM 0 0 1623 0.71 0.97 - - - - - - LKD: Lime kiln dust 37 TABLE 4. CBR and summary resilient modulus values Soil or Fly Ash Type Fly Ash Content (%) LKD Content (%) CBR (%) SMR (MPa) 1 Day Cured 7 Days Cured 28 Days Cured 1 Day Cured 7 Days Cured 28 Days Cured 10 2.5 70 108 > 131 157 228 260 10 5 73 142 > 156 179 245 274 20 2.5 45 112 > 121 147 157 169 Brandon Shores 20 5 76 133 > 152 215 284 510 10 2.5 83 115 > 120 189 230 243 10 5 95 135 > 164 216 350 418 20 2.5 34 71 100 172 198 207 Paul Smith 20 5 60 87 >105 238 278 323 10 2.5 44 69 > 120 174 180 277 10 5 93 129 > 143 219 322 433 20 2.5 65 84 111 183 217 243 Dickerson Precipitator 20 5 82 110 > 134 241 312 326 BRG 0 0 27 130 GAB 0 0 42 206 URM 0 0 24 127 BRG: Bank run gravel, GAB: Graded aggregate base, URM: Unpaved road gravel, LKD: Lime kiln dust 38 TABLE 5. Chemical compositions of mixtures prepared with three different fly ashes Mixtures Prepared with BS Mixtures Prepared with PS Mixtures Prepared with DP Chemical Constituents BS (%) PS (%) DP (%) Type I Portland Cement (%) Lime kiln dust (%) 10BS+ 2.5L 10BS+ 5L 20BS+ 2.5L 20BS+ 5L 10PS+ 2.5L 10PS+ 5L 20PS+ 2.5L 20PS+ 5L 10DP+ 2.5L 10DP+ 5L 20DP+ 2.5L 20DP+ 5L SiO2 45.1 50.8 34.9 22 7.5 37.6 32.6 40.9 37.6 42.1 36.4 46 42.1 29.4 25.8 31.9 29.4 Al2O3 23.1 26.9 24.4 5 2.7 19 16.3 20.8 19 22 18.8 24.2 22 20.1 17.2 22 20.1 Fe2O3 3.2 5.5 12.6 3 1.1 2.7 2.5 2.9 2.7 4.6 4 5 4.6 10.3 8.8 11.3 10.3 CaO 7.8 0.7 3.2 67 61.6 18.6 25.7 13.8 18.6 12.9 21 7.5 12.9 14.9 22.7 9.7 14.9 MgO 0.8 0.6 0.5 1 2.5 1.2 1.4 1 1.2 1 1.2 0.8 1 0.9 1.2 0.7 0.9 SiO2+ Al2O3 + Fe2O3 71.4 83.2 71.9 30 11.3 59.3 51.3 64.7 59.3 68.8 59.2 75.2 68.8 59.8 51.7 65.2 59.8 Silica Ratio (SR) 1.7 1.6 0.9 2.8 2 1.7 1.7 1.7 1.7 1.6 1.6 1.6 1.6 1 1 1 1 LSF 0.05 0.006 0.03 1 2.6 0.1 0.2 0.1 0.1 0.1 0.2 0.05 0.1 0.1 0.2 0.1 0.1 Alumina Ratio (AR) 7.3 4.9 1.9 1.7 2.5 6.9 6.6 7.1 6.9 4.8 4.7 4.8 4.8 2 2 1.9 2 CaO/SiO2 0.2 0.01 0.1 3 8.2 0.5 0.8 0.3 0.5 0.3 0.6 0.2 0.3 0.5 0.9 0.3 0.5 CaO/ (SiO2+Al2O3) 0.1 0.01 0.05 2.5 6 0.3 0.5 0.2 0.3 0.2 0.4 0.1 0.2 0.3 0.5 0.2 0.3 LOI 13.4 10.7 20.5 0 26.7 16.1 17.8 14.9 16.1 13.9 16 12.5 13.9 21.7 22.6 21.2 21.7 BS: Brandon Shores fly ash, PS: Paul Smith fly ash, DP: Dickerson Precipitator fly ash, L: Lime kiln dust (LKD), LOI: Loss of ignition, Silica ratio=SiO2/(Al2O3 + Fe2O3), LSF (Lime saturation factor) =CaO/(2.8 SiO2+1.2Al2O3+0.65Fe2O3) , Alumina ratio=Al2O3/Fe2O3 39 TABLE 6. Effect of freezing and thawing cycles on summary resilient moduli. Fly Ash Type Fly Ash Content (%) LKD Content (%) SMR (Mpa) Water Content (%) Number of Freeze and Thaw Cycles 0 4 8 12 0 4 8 12 10 2.5 228 307 155 103 9 9 11 17 Brandon Shores 20 2.5 157 247 169 95 10 10 14 15 10 5 350 340 222 165 9 11 14 15 Paul Smith 20 5 278 245 105 93 13 13 16 22 10 2.5 180 166 144 117 11 11 14 12 Dickerson Precipitator 10 5 217 218 158 152 11 10 18 15 LKD: Lime kiln dust 40 TABLE 7. Required base thickness for different mixture designs for two traffic and excellent drainage conditions (all thickness values are in mm). Based on CBR Based on MR Soil or Fly Ash Type Fly Ash Content (%) LKD Content (%) Case I Case II Case I Case II 10 2.5 212 474 205 459 10 5 176 395 192 431 20 2.5 219 491 289 647 Brandon Shores 20 5 198 445 176 395 10 2.5 219 491 205 459 10 5 176 395 159 356 20 2.5 227 508 235 527 Paul Smith 20 5 227 508 176 395 10 2.5 219 491 265 594 10 5 205 459 167 374 20 2.5 223 501 227 508 Dickerson Precipitator 20 5 212 474 167 374 BRG 0 0 382 846 335 750 GAB 0 0 301 678 212 475 URM 0 0 397 890 353 791 BS: Brandon Shores fly ash, PS: Paul Smith fly ash, DP: Dickerson Precipitator fly ash, LKD: Lime kiln dust, BRG: Bank run gravel, GAB: Graded aggregate base, URM: Unpaved road gravel. The numbers that follow the fly ashes and LKD indicate the percentages by weight of admixtures added to the soil. Minimum thickness requirement (AASHTO Guide 1993) for ESALs greater than 5,000,000 is 152.4 mm. a-) 41 TABLE 8. Cost analysis for all mixture designs under excellent drainage conditions FlyAsh Type Fly Ash Content (%) Lime Kiln Dust Content (%) Thickness of the Base Layer (mm) Cost of Base Materials (x1000$) Cost of LKD (x1000$) Cost of Fly Ash (x1000$) Total Cost of Construction (x1000$) Case I Case II Case I Case II Case I Case II Case I Case II Case I Case II 10 2.5 187 419 102.2 228.9 15.5 34.7 7.2 16.2 124.5 388.8 10 5 182 407 101.9 228.1 30.0 67.3 5.6 12.6 137.5 433.0 20 2.5 265 593 123.6 276.7 21.1 47.4 17.8 39.9 162.5 486.6 Brandon Shores 20 5 159 356 69.3 155.2 24.6 55.0 11.5 25.7 105.3 315.6 10 2.5 193 432 102.2 228.8 15.5 34.7 16.3 36.5 134.0 397.5 10 5 167 375 92.1 206.2 27.2 60.8 11.4 25.6 130.1 397.8 20 2.5 227 509 100.8 225.8 17.3 38.7 32.7 73.2 150.8 415.3 Paul Smith 20 5 177 396 75.2 168.5 26.6 59.7 28.0 62.8 130.0 358.1 10 2.5 179 400 96.2 215.5 14.6 32.7 8.2 18.2 119.0 367.1 10 5 167 375 95 212.8 28.0 62.8 6.2 14.0 129.3 404.9 20 2.5 193 432 87.9 196.9 15.1 33.7 15.1 33.8 118.1 348.7 Dickerson Precipitator 20 5 177 396 78.9 176.9 28.0 62.6 15.6 34.9 122.5 362.0 BRG 0 0 353 791 230.6 516.3 0 0 0 0 230.5 746.7 GAB 0 0 318 712 233.3 522.5 0 0 0 0 233.3 755.8 URM 0 0 398 890 252.4 565.3 0 0 0 0 252.2 565.4 Note: Material cost includes only the material purchase, transportation, fuel and hauling costs. 42 FIGURES 43 0 20 40 60 80 100 0.010.1110100 URM VDOT Limits BRG GAB Pe rce nt Fin er (% ) Particle Size (mm) (a) 0 20 40 60 80 100 0.010.1110100 URM AASHTO Limits BRG GAB Pe rce nt Fin er (% ) Particle Size (mm) (b) FIGURE 1. Comparison of unpaved road material and the two materials used in base construction in Maryland to (a) VDOT base materials specification, and (b) AASHTO specification. Note: URM: Unpaved road material, BRG : Bank Run Gravel, GAB: Graded aggregate base. 44 0 5 0 1 0 0 1 5 0 1 7 2 8 1 7 2 8 2 . 5 % L K D 5 % L K D Ca lifo rni a B ea rin g R ati o, CB R 2 0 % B S b y w e ig h t 1 0 % B S b y w e ig h t (a ) 0 5 0 1 0 0 1 5 0 1 7 2 8 1 7 2 8 Ca lifo rni a B ea rin g R ati o, CB R 2 0 % P S b y w e ig h t 1 0 % P S b y w e ig h t (b ) 0 5 0 1 0 0 1 5 0 1 7 2 8 1 7 2 8 Ca lifo rni a B ea rin g R ati o, CB R C u r i n g T i m e ( d a y s ) 2 0 % D P b y w e ig h t 1 0 % D P b y w e ig h t (c ) FIGURE 2. Effect of curing time on CBR of mixtures prepared with (a) Brandon Shores fly ash, (b) Paul Smith fly ash, and (c) Dickerson Precipitator fly ash. 45 FIGURE 3. SEM photograph of (a) Brandon Shores fly ash, (b) unpaved road material amended with 10% Brandon Shores fly ash and 2.5% LKD by weight (10BS+2.5LKD), (c) Dickerson Precipitator fly ash, and (d) unpaved road material amended with 20% Dickerson Precipitator fly ash and 5% LKD by weight (20DP+5LKD). All specimens were cured for 7 days. LKD: Lime kiln dust. a-) (a) (b) (d) (c) 20?m 20?m 1?m 1?m 46 (a) (b) FIGURE 4. EDX plot of the SEM photograph of (a) Dickerson Precipitator fly ash, and (b) 7-day cured unpaved road material amended with 20% Dickerson Precipitator fly ash and 5% LKD by weight (20DP+5LKD). 47 20 40 60 80 100 120 140 160 180 0 1 2 3 4 5 10 BS 20 BS 10 PS 20 PS 10 DP 20 DP CB R LKD (%) (a) r=0.75t=3.54 100 200 300 400 500 600 0 1 2 3 4 5 10 BS 20 BS 10 PS 20 PS 10 DP 20 DP SM R (M Pa ) LKD (%) (b) r=0.76 t=3.75 FIGURE 5. Effect of LKD contents on (a) CBR and (b) SMR of 28-day cured specimens (Note: 10 BS, 20 BS, 10 PS, 20 PS, 10 DP, and 20 DP designate the specimens with 10% and 20% Brandon Shores, Paul Smith, and Dickerson Precipitator fly ash respectively. LKD: Lime Kiln Dust). 48 90 100 110 120 130 140 150 160 170 5 10 15 20 25 30 C B R CaO Ratio (%) R2 = 0.79 (a) r=0.89 t=6.13 90 100 110 120 130 140 150 160 170 0 0.2 0.4 0.6 0.8 1 C B R CaO/SiO2 R2 = 0.66 (b) r=0.78 t=3.98 90 100 110 120 130 140 150 160 170 0 0.1 0.2 0.3 0.4 0.5 0.6 C B R CaO/(SiO2 + Al2O3) R2 = 0.74 (c) r=0.84 t=4.81 90 100 110 120 130 140 150 160 170 45 50 55 60 65 70 75 80 C B R Fineness (%) (d) r=0.34 t=1.16 FIGURE 6. Effect of (a) CaO content, (b) CaO/SiO2, (c) CaO/(SiO2 + Al2O3), and (d) fineness on CBR. 49 90 100 110 120 130 140 150 160 170 0 0.05 0.1 0.15 0.2 0.25 C B R Silica Ratio (SR) (a) r=0.79 t=4.09 90 100 110 120 130 140 150 160 170 1 2 3 4 5 6 7 8 C B R Alumina Ratio, (AR) (b) r=0.22 t=0.72 90 100 110 120 130 140 150 160 170 0 0.1 0.2 0.3 0.4 0.5 0.6 C B R Lime Saturation Factor, (LSF) (c) r=0.17 t=0.55 90 100 110 120 130 140 150 160 170 0 5 10 15 20 25 C B R Fly ash (%) (d) r=-0.47 t=-1.68 FIGURE 7. Effect of (a) Silica ratio, (b) alumina ratio, (c) lime saturation factor, and (d) fly ash percentage on CBR. 50 0 200 400 600 800 1000 30 40 50 60 70 80 90 100 Experimental SM R (M Pa ) CBR SMR = 10.34 x CBR SMR = 17.64 x CBR0.64 (a) 0 500 1000 1500 60 80 100 120 140 160 Experimental SM R (M Pa ) CBR SMR = 10.34 x CBR SMR = 17.64 x CBR0.64 (b) FIGURE 8. CBR versus measured and predicted SMR for (a) 1 day cured specimens, and (b) 7 days cured specimens 51 0 1 0 0 2 0 0 3 0 0 4 0 0 5 0 0 1 7 2 8 1 7 2 8 2 . 5 % L im e K i ln D u s t 5 % L im e K iln D u s t SM R (M Pa ) 2 0 % B S b y w e ig h t1 0 % B S b y w e ig h t a ) 0 1 0 0 2 0 0 3 0 0 4 0 0 5 0 0 1 7 2 8 1 7 2 8 SM R (M Pa ) 20% P S b y w eig h t 10% P S b y w eig h t b ) 0 1 0 0 2 0 0 3 0 0 4 0 0 5 0 0 1 7 2 8 1 7 2 8 SM R( MP a) C u r in g T im e ( d a y s ) 2 0 % D P b y w e ig h t1 0 % D P b y w e ig h t c ) FIGURE 9. Effect of curing time on SMR of mixtures prepared with (a) Brandon Shores fly ash, (b) Paul Smith fly ash, and (c) Dickerson Precipitator fly ash. 52 0 100 200 300 400 500 600 700 0 200 400 600 800 1000 10 PS + 2.5 LKD 10 PS + 5 LKD 20 PS + 2.5 LKD 20 PS + 5 LKD Field Material GAB BRG Re sil ien t M od ulu s M R, (M Pa ) Bulk Stress (kPa) (a) 0 200 400 600 800 1000 0 200 400 600 800 1000 10 DP + 2.5 LKD 10 DP + 5 LKD 20 DP + 2.5 LKD 20 DP + 5 LKD Field Material GAB BRG Re sil ien t M od ulu s, M R (M Pa ) Bulk Stress (kPa) (b) FIGURE 10. Resilient modulus of the 28-day cured specimens with varying bulk stresses: Mixtures prepared with (a) Paul Smith fly ash, and (b) Dickerson Precipitator fly ash 53 150 200 250 300 350 400 450 500 550 5 10 15 20 25 30 S M R ( M P a ) CaO Ratio (%) (a) r=0.54 t=2.03 150 200 250 300 350 400 450 500 550 0 0.2 0.4 0.6 0.8 1 S M R ( M P a ) CaO/SiO2 (b) r=0.56 t=2.11 150 200 250 300 350 400 450 500 550 0 0.1 0.2 0.3 0.4 0.5 0.6 S M R ( M P a ) CaO/(SiO2 + Al2O3) (c) r=0.54 t=2.05 150 200 250 300 350 400 450 500 550 45 50 55 60 65 70 75 80 S M R ( M P a ) Fineness (%) (d) r=0.04 t=0.12 FIGURE 11. Effect of (a) CaO content, (b) CaO/SiO2, (c) CaO/(SiO2 + Al2O3), and (d) fineness on SMR. 54 150 200 250 300 350 400 450 500 550 0.8 1 1.2 1.4 1.6 1.8 S M R ( M P a ) Silica Ratio, (SR) (a) r=-0.09 t=-0.29 150 200 250 300 350 400 450 500 550 0 1 2 3 4 5 6 7 8 ( S M R ) ( M P a ) Alumina Ratio, (AR) (b) r=-0.09 t=-0.27 150 200 250 300 350 400 450 500 550 0 0.05 0.1 0.15 0.2 0.25 S M R ( M P a ) Lime Saturation Factor, (LSF) (c) r=0.48 t=1.71 150 200 250 300 350 400 450 500 550 0 5 10 15 20 25 S M R ( M P a ) Fly ash (%) (d) r=-0.11 t=-0.35 FIGURE 12. Effect of (a) Silica ratio, (b) alumina ratio, (c) LSF , and (d) fly ash percentage on SMR. 55 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 0 2 4 6 8 10 12 10 BS + 2.5 LKD 20 BS + 2.5 LKD 10 PS + 5 LKD 20 PS + 5 LKD SM R Ra tio = SM Rn / S M R i Freeze - thaw cycles (a) 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1 1.05 0 2 4 6 8 10 12 10 DP + 2.5 LKD 10 DP+ 5 LKD SM R Ra tio = SM Rn / S M R i Freeze - thaw cycles (b) FIGURE 13. Effect of freeze and thaw cycles on SMR values (Note: 10 BS, 20 BS, 10 PS, 20 PS, 10 DP, and 20 DP designate the specimens with 10% and 20% Brandon Shores, Paul Smith, and Dickerson Precipitator fly ash respectively. LKD: Lime Kiln Dust). 56 0 5 1 0 1 5 2 0 2 5 10 B S + 2 .5 LK D 20 B S + 2 .5 LK D 10 P S + 5 LK D 20 P S+ 5 LK D 10 D P + 2 .5 LK D 10 D P + 5 LK D 0 c y c le s 4 c y c le s 8 c y c le s 1 2 c y c le s W ate r C on ten t (% ) FIGURE 14. Effects of freeze and thaw cycles on water contents (Note: 10 BS, 20 BS, 10 PS, 20 PS, 10 DP, and 20 DP designate the specimens with 10% and 20% Brandon Shores, Paul Smith, and Dickerson Precipitator fly ash respectively. LKD: Lime Kiln Dust). 57 100 150 200 250 300 350 400 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Case I Case II Su mm ary re sil ien t m od ulu s, SM R (M Pa ) Thickness, t (m) SMR = 91.45t-1.23 SMR = 33.84t-1.23 FIGURE 15. Summary resilient modulus as a function of base layer thickness. 58 APPENDIX 59 APPENDIX A CHEMICAL AND PHYSICAL PROPERTIES OF FLY ASHES USED IN THE CURRENT STUDY 60 Table A.1 Properties of Maryland fly ashes with ASTM C 618 chemical and physical criteria for Class C and Class F fly ash. ASTM Requirements Chemical Requirements Class F Class C BS DP PS SiO2 + Al2O3 + Fe2O3, min (%) 70 50 71.4 71.9 83.2 SO3, max (%) 5 5 - - - Moisture Content (as- received), max (%) 3 3 1.8 7.7 3.5 Loss on Ignition, max (%) 6 6 13.4 20.5 10.7 ASTM Requirements Physical Requirements Class F Class C BS DP PS Fineness, max (%) 34 34 25 34 51 Strength Activity @ 7 Days, min (%) 75 75 - - - Strength Activity @ 28 Days, min (%) 75 75 - - - Water Requirement, max (%) 115 115 - - - Autoclave Expansion, max (%) 0.8 0.8 - - - Density Variation, max (%) 5 5 - - - Variation of % Retained on 45-?m filter, max (%) 5 5 - - - 61 A 1. STATISTICAL ANALYSIS The relationship between CBR or resilient modulus and each of the fly ash chemical characteristics was tested for statistical significance by determining whether the Pearson correlation coefficient between CBR or resilient modulus and each of the fly ash variables is statistically different from zero. For this statistical analysis, the t-statistic (t) is computed from the correlation coefficient (r) as: 2 1 2 ? ? ?= n r rt ? where ? is the population correlation coefficient (assumed to be zero) and n is the number of degrees of freedom (24 for CBR test data and 20 for resilient modulus test data in this analysis). A comparison then is made between t and the critical t (tcr) corresponding to a significance level ?. If t > tcr, then the Pearson correlation coefficient is significantly different from zero and a significant relationship exists between CBR or resilient modulus and the fly ash property. In this analysis, ? was set at 0.05 (the commonly accepted significance level), which corresponds to tcr = 2.06 for CBR test results and tcr=2.09 for resilient modulus test results. 62 APPENDIX B BASE COURSE TESTING PROTOCOL 63 B.1 COMPACTION CURVES FOR ALL SPECIMENS 17 18 19 20 21 22 23 24 25 0 5 10 15 20 URM ZAV GAB BRG Dr y U nit W eig ht (kN / m 3 ) Moisture Content (%) (a) Gs = 2.67 16 17 18 19 20 21 22 5 10 15 20 URM ZAV 10 BS + 2.5 LKD 10 BS + 5 LKD 20 BS + 2.5 LKD 20 BS + 5 LKD Dr y U nit W eig ht (kN / m 3 ) Moisture Content (%) (b) Gs = 2.67 64 15 16 17 18 19 20 21 22 4 6 8 10 12 14 16 18 20 URM ZAV 10 PS + 2.5 LKD 10 PS + 5 LKD 20 PS + 2.5 LKD 20 PS + 5 LKD Dr y U nit W eig ht (kN / m 3 ) Moisture Content (%) (c) Gs = 2.70 16 17 18 19 20 21 22 4 6 8 10 12 14 16 18 URM ZAV 10 DP + 2.5 LKD 10 DP + 5 LKD 20 DP + 2.5 LKD 20 DP + 5 LKD Dr y U nit We igh t (k N / m 3 ) Moisture Content (%) (d) Gs = 2.69 Figure B. 1 Compaction curves for (a) conventional base materials, (b) mixtures prepared with Brandon Shores fly ash, (c) mixtures prepared with Paul Smith fly ash, and (d). mixtures prepared with Dickerson Precipitator fly ash (Note: 10 BS, 20 BS, 10 PS, 20 PS, 10 DP, and 20 DP designate the specimens with 10% and 20% Brandon Shores, Paul Smith, and Dickerson Precipitator fly ash respectively. LKD: Lime Kiln Dust, GAB: Graded Aggregate Base, RGB: Bank Run Gravel). 65 B.2 RESILIENT MODULUS TEST PROTOCOL AND LOADING SEQUENCE SUMMARY TABLE The resilient modulus test procedure was based on the AASHTO T 307-99 a protocol for testing base and highway base and subbase materials. The specimens of 101.6 mm in diameter and 230.2 mm in height were compacted in split molds at their OMC in eight layers using the standard Proctor energy. The deformation was measured externally with two spring-loaded linear variable differential transducers (LVDT). A Geocomp LoadTrac-II loading frame and associated hydraulic power unit system was used to load the specimens. Conditioning stress was 103 kPa. Confining stress was kept between 20.7 and 138 kPa during loading stages, and the deviator stress was increased from 20.7 kPa to 276 kPa and applied 100 repetitions at each step. The loading sequence, confining pressure, and data acquisition were controlled by a personal computer equipped with RM 5.0 software. The base and subbase testing sequence is shown in Table B.1 The resilient modulus of soil was computed by using the common model defined by Moosazadh and Witczak (1981). A summary resilient modulus (SMR) was computed at a bulk stress of 208 kPa, following the suggestions provided in NCHRP 1-28A. 66 Table B.1 AASHTO T 307-99 resilient modulus testing sequence for base and subbase materials. Sequence No. Confining Pressure, S3 (kPa) Maximum Axial Stress, Smax (kPa) Cyclic Stress, Scyclic (kPa) Constant Stress, 0.1 Smax (kPa) Cycles 0 103.4 103.4 93.1 10.3 500 1 20.7 20.7 18.6 2.1 100 2 20.7 41.4 37.3 4.1 100 3 20.7 62.1 55.9 6.2 100 4 34.5 34.5 31 3.5 100 5 34.5 68.9 62 6.9 100 6 34.5 103.4 93.1 10.3 100 7 68.9 68.9 62 6.9 100 8 68.9 137.9 124.1 13.8 100 9 68.9 206.8 186.1 20.7 100 10 103.4 68.9 62 6.9 100 11 103.4 103.4 93.1 10.3 100 12 103.4 206.8 186.1 20.7 100 13 137.9 103.4 93.1 10.3 100 14 137.9 137.9 124.1 13.8 100 15 137.9 275.8 248.2 27.6 100 67 B.3 STEP-BY-STEP RESILIENT MODULUS TEST PROCEDURE 1) Turn on Geocomp Load Trac II 2) Turn the air pressure pump on 3) Measure the specimen height and diameter 4) Place the porous stone on bottom plate 5) Place the filter paper on bottom porous stone 6) Place the specimen on bottom plate 7) Place the filter paper on top of the specimen 8) Place the porous stone on filter paper 9) Place the top plate on top of the specimen 10) Place rubber membrane over specimen using a mold 11) Place two O- rings on both bottom and top of the plates to hold the membrane in place 12) Plug the drainage tubes on top plate. 13) Place the cell on bottom cap 14) Place cover plate, it should not be tight 15) Place LVDT on top of chamber 16) Screw cover plate with three rods carefully 17) Plug air supply hose into cell 18) Log into PC and open the Resilient modulus RM version 5.0 software 19) Input Specimen height, diameter, and weight. 20) Input the loading and pressure data which is designed for base and subbase test protocol 21) Click on the load calibration menu and check the applied load with the load data that you entered 22) Click run test and save the file. 68 B.4 RESILIENT MODULUS TEST APPARATUS Figure B.2 Photo of Resilient Modulus Testing Equipment 69 B.5 PHOTOS OF SOIL SAMPLES Figure B.3. GAB (a), URM (b), and BRG (c) (Note: GAB: Graded aggregate Base, URM: Unpaved road material, BRG: Bank run gravel.) (a) (b) (c) 70 APPENDIX C CBR AND RESILIENT MODULUS TEST RESULTS 71 C1. CBR VERSUS MEASURED SMR 140 160 180 200 220 240 260 30 40 50 60 70 80 90 100 SM R (M Pa ) CBR (a) R2 = 0.33 150 200 250 300 350 400 60 80 100 120 140 160 SM R (M Pa ) CBR (b) R2 = 0.32 Figure C.1. CBR versus measured SMR vs. CBR graph for (a)1 day cured specimens, and (b) 7 days cured specimens. 72 C2. CBR VERSUS PREDICTED SMR Table C.1. Predicted SMR values based on CBR values Fly Ash Type Fly Ash Content (%) Lime Kiln Dust Content (%) CBR SMR (Mpa) (SM R = 10.34 x CBR) SMR (Mpa) (SMR = (17.64 x CBR0.64) 1 Day Cured 7 Days Cured 1 Day Cured 7 Days Cured 1 Day Cured 7 Days Cured 10 3 70 108 724 1117 267 353 10 5 73 142 755 1468 275 420 20 3 45 112 465 1158 201 361 Brandon Shores 20 5 76 133 786 1375 282 403 10 3 83 115 858 1189 298 367 10 5 95 135 982 1396 325 407 20 3 34 71 352 734 168 270 Paul Smith 20 5 60 87 620 900 242 307 10 3 44 69 455 713 199 265 10 5 93 129 962 1334 321 395 20 3 65 84 672 869 255 300 Dickerson Precipitator 20 5 82 110 848 1137 296 357 GAB 0 0 42 434 193 BRG 0 0 27 279 145 URM 0 0 24 248 135 73 APPENDIX D BASE THICKNESS CALCULATIONS FOR SPECIMENS SUBJECTED TO FREEZE-THAW 74 D.1 BASE THICKNESS CALCULATIONS FOR SPECIMENS SUBJECTED TO FREEZE-THAW Table D.1. Base thickness values based on SMR values for traffic case I (all thickness values are in mm) Base thickness values based on SMR values Specimen Name Cycle Number 0 4 8 12 10 BS + 2.5 LKD 205 318 530 530 20 BS + 2.5 LKD 289 265 530 530 10 PS + 5 LKD 159 223 289 289 20 PS + 5 LKD 176 454 454 578 10 DP + 2.5 LKD 265 318 318 454 10 DP + 5 LKD 167 289 289 302 Table D.2. Base thickness values based on SMR values for traffic case II (all thickness values are in mm) Base Thickness values based on SMR values Specimen Name Cycle Number 0 4 8 12 10 BS + 2.5 LKD 459 712 1186 1186 20 BS + 2.5 LKD 647 593 1186 1186 10 PS + 5 LKD 356 501 647 647 20 PS + 5 LKD 395 1017 1017 1294 10 DP + 2.5 LKD 593 712 712 1017 10 DP + 5 LKD 374 647 647 678 BS: Brandon Shores fly ash, PS: Paul Smith fly ash, DP: Dickerson Precipitator fly ash, LKD: Lime kiln dust. The numbers that follow the fly ashes and LKD indicate the percentages by weight of admixtures added to the soil. Minimum thickness requirement (AASHTO Guide 1993) for ESALs greater than 5,000,000 is 152.4 mm. 75 100 200 300 400 500 600 0 2 4 6 8 10 12 10 BS + 2.5 LKD 20 BS + 2.5 LKD 10 PS + 5 LKD 20 PS + 5 LKD 10 DP + 2.5 LKD 10 DP + 5 LKD Ba se Th ick ne ss (m m) Number of Freeze and Thaw Cycles (a) 200 400 600 800 1000 1200 1400 0 2 4 6 8 10 12 14 10 BS + 2.5 LKD 20 BS + 2.5 LKD 10 PS + 5 LKD 20 PS + 5 LKD 10 DP + 2.5 LKD 10 DP + 5 LKD Ba se Th ick ne ss (m m) Number of Freeze and Thaw Cycles (b) Figure D.1 Required thickness vs. number of freeze and thaw cycles (a) for traffic case I, and (b) for traffic case II 76 APPENDIX E BASE THICKNESS AND COST CALCULATIONS FOR SPECIMENS UNDER DIFFERENT DRAINAGE CONDITIONS 77 E.1 BASE THICKNESS CALCULATIONS FOR SPECIMENS UNDER DIFFERENT DRAINAGE CONDITIONS Table E1. Required base thickness for different mixture designs for two traffic conditions under good drainage conditions (all thickness values are in mm). Fly Ash Type Fly Ash Content (%) Lime Kiln Dust Content (%) Thickness of the Base Layer Based on CBR Thickness of the Base Layer Based on SMR Case I Case II Case I Case II 10 2.5 277 555 277 555 10 5 262 524 269 539 20 2.5 314 629 392 787 Brandon Shores 20 5 262 524 235 472 10 2.5 314 629 285 572 10 5 262 524 248 497 20 2.5 336 674 336 674 Paul Smith 20 5 332 665 262 524 10 2.5 314 629 265 530 10 5 269 539 248 497 20 2.5 327 656 285 572 Dickerson Precipitator 20 5 277 555 262 524 BRG 0 0 523 1049 589 1180 GAB 0 0 428 858 471 944 URM 0 0 673 1349 589 1180 78 Table E2. Required base thickness for different mixture designs for two traffic conditions under fair drainage conditions (all thickness values are in mm). Fly Ash Type Fly Ash Content (%) Lime Kiln Dust Content (%) Thickness of the Base Layer Based on CBR Thickness of the Base Layer Based on SMR Case I Case II Case I Case II 10 2.5 412 760 412 760 10 5 389 718 400 738 20 2.5 467 861 584 1076 Brandon Shores 20 5 389 718 350 646 10 2.5 467 861 424 783 10 5 389 718 369 680 20 2.5 500 923 500 923 Paul Smith 20 5 493 910 389 718 10 2.5 467 861 393 726 10 5 400 738 369 680 20 2.5 486 897 424 783 Dickerson Precipitator 20 5 412 760 389 718 BRG 0 0 778 1435 875 1614 GAB 0 0 637 1174 700 1292 URM 0 0 1000 1845 875 1614 79 Table E3. Required base thickness for different mixture designs for two traffic conditions under poor drainage conditions (all thickness values are in mm). Fly Ash Type Fly Ash Content (%) Lime Kiln Dust Content (%) Thickness of the Base Layer Based on CBR Thickness of the Base Layer Based on SMR Case I Case II Case I Case II 10 2.5 637 1101 637 1101 10 5 601 1039 619 1069 20 2.5 722 1247 902 1559 Brandon Shores 20 5 601 1039 541 935 10 2.5 722 1247 656 1134 10 5 601 1039 570 985 20 2.5 773 1336 773 1336 Paul Smith 20 5 762 1318 601 1039 10 2.5 722 1247 608 1051 10 5 619 1069 570 985 20 2.5 752 1299 656 1134 Dickerson Precipitator 20 5 637 1101 601 1039 BRG 0 0 1203 2079 1353 2339 GAB 0 0 984 1701 1083 1871 URM 0 0 1547 2673 1353 2339 80 Table E4. Cost Analysis for different mixture designs for two traffic conditions under good drainage conditions (all thickness values are in mm). Fly Ash Type Fly Ash Content (%) Lime Kiln Dust Content (%) Thickness of the Base Layer (mm) Cost of Base Materials (x1000$) Cost of LKD (x1000$) Cost of Fly Ash (x1000$) Total Cost of Construction (x1000$) Case I Case II Case I Case II Case I Case II Case I Case II Case I Case II 10 2.5 277 555 151.3 303.3 229.6 46.0 10.7 21.5 185.0 534.4 10 5 269 539 150.8 302.4 445.0 89.2 8.3 16.7 203.7 595.2 20 2.5 392 787 183.0 366.7 313.4 62.8 26.4 52.9 240.7 670.3 Brandon Shores 20 5 235 472 102.7 205.8 363.4 72.8 17.0 34.1 156.0 434.6 10 2.5 285 572 151.3 303.2 229.5 46.0 24.2 48.4 198.4 547.6 10 5 248 497 136.4 273.4 402.3 80.6 16.9 33.9 193.5 547.5 20 2.5 336 674 149.3 299.3 255.8 51.3 48.4 97.1 223.3 573.9 Paul Smith 20 5 262 524 111.4 223.4 394.5 79.1 41.5 83.1 192.4 494.8 10 2.5 265 530 142.5 285.6 216.1 43.3 12.0 24.1 176.1 505.0 10 5 248 497 140.7 282.1 415.1 83.2 9.30 18.5 191.5 556.8 20 2.5 285 572 130.2 261.0 223.0 44.7 22.3 44.8 174.8 480.5 Dickerson Precipitator 20 5 262 524 117.0 234.4 414.0 83.0 23.1 46.2 181.4 498.8 BRG 0 0 523 1049 341.4 684.4 0 0 0 0 341.4 102.6 GAB 0 0 471 944 345.5 692.5 0 0 0 0 345.5 103.8 URM 0 0 589 1180 373.8 749.4 0 0 0 0 373.8 749.4 81 Table E5. Cost Analysis for different mixture designs for two traffic conditions under fair drainage conditions (all thickness values are in mm). Fly Ash Type Fly Ash Content (%) Lime Kiln Dust Content (%) Thickness of the Base Layer (mm) Cost of Base Materials (x1000$) Cost of LKD (x1000$) Cost of Fly Ash (x1000$) Total Cost of Construction (x1000$) Case I Case II Case I Case II Case I Case II Case I Case II Case I Case II 10 2.5 412 760 225.0 303.3 34.2 63.0 16.0 29.4 275.2 641.5 10 5 389 738 218.1 302.4 64.3 122.1 12.0 22.8 294.4 718.8 20 2.5 467 1076 217.7 366.8 37.3 86.0 31.4 72.3 286.3 739.0 Brandon Shores 20 5 389 646 169.6 205.8 60.0 99.7 28.1 46.6 257.7 563.1 10 2.5 467 783 247.5 303.2 37.6 63.0 39.5 66.2 324.5 690.7 10 5 389 680 214.1 273.4 63.2 110.3 26.6 46.4 303.8 687.5 20 2.5 500 923 222.1 299.3 38.0 70.2 72.0 132.8 332.1 701.6 Paul Smith 20 5 493 718 210.1 223.4 74.4 108.2 78.2 113.8 362.6 694.2 10 2.5 467 726 251.4 285.6 38.2 59.3 21.2 33.0 310.8 655.7 10 5 400 680 227.2 282.1 67.0 113.9 14.9 25.4 309.1 705.1 20 2.5 486 783 221.8 261.0 38.0 61.2 38.1 61.3 297.9 620.0 Dickerson Precipitator 20 5 412 718 184.1 234.4 65.2 113.6 36.3 63.2 285.6 633.5 BRG 0 0 778 1435 507.7 684.4 0 0 0 0 507.7 1192.1 GAB 0 0 637 1292 467.1 692.5 0 0 0 0 467.1 1159.6 URM 0 0 1000 1614 635.4 749.4 0 0 0 0 635.3 749.4 82 Table E6. Cost Analysis for different mixture designs for two traffic conditions under poor drainage conditions (all thickness values are in mm). Fly Ash Type Fly Ash Content (%) Lime Kiln Dust Content (%) Thickness of the Base Layer (mm) Cost of Base Materials ($) Cost of LKD ($) Cost of Fly Ash ($) Total Cost of Construction ($) Case I Case II Case I Case II Case I Case II Case I Case II Case I Case II 10 2.5 637 1101 347.9 601.3 52.8 91.2 24.7 42.7 425.360 1117.8 10 5 619 1069 346.8 599.3 102.3 176.8 19.1 33.1 468.184 1244.3 20 2.5 902 1559 420.6 726.9 72.1 124.5 60.6 104.8 553.300 1404.7 Brandon Shores 20 5 541 935 236.0 407.8 83.5 144.4 39.1 67.5 358.566 910.8 10 2.5 656 1134 347.8 601.0 52.8 91.2 55.5 95.9 456.056 1148.3 10 5 570 985 313.5 541.8 92.5 159.8 38.9 67.3 444.906 1146.6 20 2.5 773 1336 343.3 593.3 58.8 101.6 111.3 192.4 513.416 1208.3 Paul Smith 20 5 601 1039 256.2 442.7 90.7 156.7 95.4 164.9 442.247 1041.7 10 2.5 608 1051 327.5 566.0 49.7 85.9 27.7 47.8 404.890 1056.8 10 5 570 985 323.5 559.1 95.4 164.9 21.3 36.7 440.203 1164.2 20 2.5 656 1134 299.3 517.2 51.3 88.6 51.4 88.8 401.939 1007.8 Dickerson Precipitator 20 5 601 1039 268.8 464.6 95.2 164.5 53.0 91.6 416.995 1046.1 BRG 0 0 1203 2079 784.9 1356.5 0 0 0 0 784.887 2141.4 GAB 0 0 1083 1871 794.3 1372.7 0 0 0 0 794.246 2166.9 URM 0 0 1353 2339 859.4 1485.2 0 0 0 0 859.415 1485.3 83 10 BS + 2.5 LKD 10 BS + 5 LKD GAB ........ 0 5 10 15 20 25 0.0833 1 7 10 Excellent Good Fair Poor Time to drain water from the base layer (days) Quality of Drainage a) T o t a l C o s t x 1 0 0 , 0 0 0 $ Case II Traffic Conditions 0 5 10 15 20 25 0.0833 1 7 10 Excellent Good Fair Poor Time to drain water from the base layer (days) Quality of Drainage T o t a l C o s t X 1 0 0 , 0 0 0 $ 20 PS + 2.5 LKD 20 PS + 5 LKD GAB ........ (b) Case II Traffic Conditions 10 DP + 2.5 LKD 10 DP + 5 LKD GAB .......... 0 5 10 15 20 25 0.0833 1 7 10 Excellent Good Fair Poor Time to drain water from the base layer (days) Quality of Drainage T o t a l C o s t X 1 0 0 ,0 0 0 $ (c) Case II Traffic Conditions 0 5 10 15 20 25 0.0833 1 7 10 Excellent Good Fair Poor Time to drain water from the base layer Quality of Drainage 10 DP + 2.5 LKD GAB .......... Case II Case I T o t a l C o s t X 1 0 0 , 0 0 0 $ (d) 10 BS + 2.5 LKD GAB .......... 0 5 10 15 20 25 0.0833 1 7 10 Excellent Good Fair Poor Time to drain water from the base layer (days) Quality of Drainage T o t a l C o s t x 1 0 0 , 0 0 0 $ Case II Case I(e) 0 5 10 15 20 25 0.0833 1 7 10 Excellent Good Fair Poor T o t a l C o s t x 1 0 0 , 0 0 0 $ Time to drain water from the base layer (days) Quality of Drainage 20 DP + 2.5 LKD GAB . . . . . . Case II Case I (f) FigureE.1. Effect of LKD addition (a),(b),(c) and traffic conditions(d), (e), (f) on total construction fee of the highway base layer. BS: Brandon Shores fly ash, PS: Paul Smith fly ash, DP: Dickerson Precipitator fly ash, LKD: Lime kiln dust. The numbers that follow the fly ashes and LKD indicate the percentages by weight of admixtures added to the soil. 84 REFERENCES AASHTO Guide (1993). ?Guide for Design of Pavement Structures?, American Association of State Highway and Transportation Officials, Washington, D.C. ACAA (2003). ?2001 Coal Combustion Product Production and Use?, American Coal Ash Association, Denver, Colorado ARA (2004). ?Guide for Mechanistic-Empirical Design on New and Rehabilitated Pavement Structures,? NCHRP Project 1-37A. Prepared for National Cooperative Highway Research Program, Washington, D.C. Arora, S. and Aydilek, A.H. (2005). ?Class F Fly Ash Amended Soils as Highway Base Materials?, Journal of Materials in Civil Engineering, Vol. 17, No. 6, pp. 640- 649. Asphalt Institute (2003). ?Thickness Design ?Highways and Streets?, Manual Series, No. 1, Asphalt Institute, Lexington, Kentucky, 110 p. Bin-Shafique, S., Edil, T., Benson, C., and Senol, A. (2004). ?Incorporating a fly ash stabilized layer into pavement design-Case study.? Geotechnical engineering, Institution of Civil Engineers, London, Vol. 157, No. GE4, 239?249. Buhler, S.R. and Cerato, A.B. (2007). ?Stabilization of Oklahoma Expansive Soils Using Lime and Class C Fly Ash?, ASCE Geotechnical Special Publication, 162, pp. 1- 10. Camargo, F.F. (2008). ?Strength and Stiffness of Recycled Base Materials Blended with Fly Ash?, M.S. Thesis, University of Wisconsin-Madison, 123 p. 85 Conner, J.R. (1990). ?Chemical Fixation and Solidification of Hazardous Wastes?, Van Nostrand Reinfold, New York, 692 p. Consoli, N. C., Prietto P. D. M., Carraro, J. A. H., and Heineck, K., S. (2001). ?Behavior of Compacted Soil-Fly Ash-Carbide Lime Mixtures." Journal of Geotechnical and Geoenvironmental Engineering, Vol. 127, No. 9, pp. 774-782. Cross, S. A. and Young, D. A. (1997). Evaluation of Type C Fly Ash in Cold-in-Place Recycling, Journal of the Transportation Research Board, 1583, pp. 82?90. DiGioia, A. M. and Nuzzo, W. L. (1972). ?Fly Ash as Structural Fill,? J. Power Div., ASCE, New York, Vol. 98 (1), pp. 77-92. Edil, T.B., Acosta , A.A., and Benson, C.H. (2006). ?Stabilizing Soft Fine-Grained Soils with Fly Ash?, Journal of Materials in Civil Engineering, Vol. 18, No. 2, pp. 283- 294. Edil, T. B., Benson, C., Bin-Shafique, M., Tanyu, B., Kim, W. and Senol, A. (2002). ?Field evaluation of construction alternatives for roadways over soft subgrade?, Journal of the Transportation Research Board, 1786, pp. 36-48. Gray, D. H., and Lin, Y. K. (1972). ?Engineering Properties of Compacted Fly Ash,?J. of Soil Mech. and Found. Engrg., ASCE, New York, Vol. 98 (4), pp. 361-380. Guney, Y., Aydilek, A.H., and Demirkan, M.M. (2006). ?Geoenvironmental Behavior of Foundry Sand Amended Mixtures for Highway Subbases?, Waste Management, Vol. 26, pp. 932-945. Hatipoglu, B., Edil, T., and Benson, C. (2008). ?Evaluation of base prepared from road surface gravel stabilized with fly ash?, ASCE Geotechnical Special Publication, 177, pp. 288-295. 86 Heukelom, W., and Foster, C. (1960). ?Dynamic Testing of Pavements.? Journal of Soil Mechanics and Foundation Division, Vol. 86, No. 1, pp.1?28. Huang, Y. H. (1993). ?Pavement Analysis and Design?, Prentice-Hall, Inc., New Jersey, 805 p. Janz, M. and Johansson, S.-E. (2002). "The Function of Different Binding Agents in Deep Stabilization," Report 9. Swedish Deep Stabilization Research Centre, Linkoping, Sweden. Kumar, A., Walia, B.S., and Bajaj, A.(2007). ?Influence of Fly Ash, Lime, and Polyster Fibers on Compaction and Strength Properties of Expansive Soils?, Journal of Materials in Civil Engineering, Vol. 19, No. 3, pp. 242-248. Li, L., Benson, C. H., Edil, T. B., Hatipoglu, B., and Tastan, E. (2007). ?Evaluation of recycled asphalt pavement material stabilized with fly ash?, ASCE Geotechnical Special Publication (CD-ROM), 169. Moosazedh, J., and Witczak, M. (1981). ?Prediction of Subgrade Moduli for Soil that Exhibits Nonlinear Behavior?, Journal of the Transportation Research Board, 810, pp. 10-17. PCA (2009). Portland Cement Association, www.cement.org Powell, W., Potter, J., Mayhew, H., and Nunn, M. (1984). The Structural Design of Bituminous Roads, TRRL Laboratory Report 1132, Transportation and Road Research Laboratory, Crowthorne, Berkshire, U.K, 62 p. Rosa, M. (2006). ?Effect of Freeze and Thaw Cycling on Soils Stabilized using Fly Ash,? MS Thesis, University of Wisconsin-Madison, Madison, WI. 87 Saylak, D., Mishra, S.K., Mejeoumov, G.G., and Shon, C.C. (2008). ?Fly Ash-Calcium Chloride Stabilization in Road Construction?, Proceedings of 87th Annual Meeting (CD ROM), Transportation Research Board, Washington, D.C. Sawangsuriya, A. and Edil, T. B. (2005), ?Evaluating stiffness and strength of pavement materials, geotechnical engineering.? Geotechnical Engineering, Institution of Civil Engineers, London, Vol. 158, No. GE4, 217?230. Shao, L., Liu, S., Du, Y., Jing, F. and Fang, L. (2008). ?Experimental Study on the Stabilization of Organic Clay with Fly Ash and Cement Mixed Method?, ASCE Geotechnical Special Publication, 179, pp.20-27. Simonsen, E., Janoo, V., and Isacsson, U. (2002). ?Resilient properties of unbound road material during seasonal frost conditions?, Journal of Cold Region Eng., 16(1), pp. 28-50. Tastan, O., Benson, C.H., Edil, T.B., and Aydilek, A.H. (2009). ?Stabilization of Organic Soils with Fly Ashes?, Journal of Geotechnical and Geoenvironmental Engineering, submitted. Taylor, H.F.W. (1997). ?Cement Chemistry?, Second edition, Thomas Telford, Inc. Vishwanathan, R., Saylak, D., and Estakhri, C. (1997). ?Stabilization of Subgrade Soils Using Fly Ash?, Proceedings of the Ash Utilization Symposium, CAER, Kentucky, pp. 204-211. Wen, H., Baugh, J., and Edil, T. (2007). ?Use of Cementitious High Harbon Fly ash to Stabilize Recycled Pavement Materials as Pavement Base Material?, Proceedings of 86th Annual Meeting (CD ROM), Transportation Research Board, Washington, D.C. 88 Wen, H., Warner, J., and Edil, T. (2008). ?Laboratory comparison of crushed aggregate and recycled pavement material with and without high-carbon fly ash?, Proceedings of 87th Annual Meeting (DVD), Transportation Research Board, Washington, D.C. Zaman, M., and Naji, K. (2003). ?Effect of freeze-thaw cycles on class C fly ash stabilized aggregate base?, Proceedings of 82nd Annual Meeting (CD ROM), Transportation Research Board, Washington, D.C. 89