FLY ASH AMENDED SOILS AS HIGHWAY BASE MATERIALS

FLY ASH AMENDED SOILS AS HIGHWAY BASE MATERIALS ABSTRACT Ahmet H. Aydilek., Member, ASCE 1, Sunil Arora, Member, ASCE 2 Class F fly ash cannot be used alone in soil stabilization applications as it is not selfcementing. An activator such as Portland cement or lime must be added to produce cementitious products often called pozzolan stabilized mixtures. The developed mixture must possess adequate strength and durability, should be easily compacted, and most importantly should be economical and environmentally friendly. Roadways have a high potential for large volume use of the fly ash stabilized soils. The main objective of this study is to investigate the beneficial reuse of Class F fly ash amended soil-cement or soil-lime as base layers in highways. A battery of tests was conducted on soil-fly ash mixtures prepared with cement and lime as activators. Unconfined compression, California bearing ratio, and resilient modulus tests were conducted. Results of the study show that the strength of a mixture is highly dependent on the curing period, the compactive energy, cement content, and water content at compaction. Lime treatment does not provide sufficient strength for designing the mixtures as highway bases. A power function in terms of bulk stress used for granular soils can model the resilient moduli. INTRODUCTION Approximately 90% of the coal used in United States is burned to produce electricity and the process results in the production of about 107 million metric tons of coal combustion products (CCPs). At present, 58% of the CCPs are fly ash, and only 32% of this ash is beneficially reused in construction. The overall percentage of CCP use in construction in the United States has increased from 12.3% in 1966 to more than 31% in 2001. Unless alternative energy sources are introduced, an increase in CCPs is inevitable under present circumstances. As the production of CCPs increases, it is essential to use the materials in applications that are technically sound, commercially competitive, and environmentally safe. Among different types of CCPs, Class F fly ash is the least commonly used ash, mainly due to its lack of self-cementitious properties. It consists of siliceous and aluminous materials (pozzolans) that lack cementitious value by themselves, but chemically react with calcium oxide in the presence of moisture to form cementitious compounds. Class F fly ash is usually activated by lime or cement to create pozzolanic stabilized mixtures (PSMs). Pozzolanic activity is initiated by the addition of water, and resulting in the formation of cementitious compounds while modifying the engineering properties of the soil. One way of handling waste fly ash is to use it as a base material in highways. This decreases the cost of landfilling and, therefore, is an environmentally friendly option. 1 Assistant. Professor, Dept. of Civ. and Env. Engrg., University of Maryland, 1163 Glenn Martin Hall, College Park, MD, 20742. E-mail: aydilek@eng.umd.edu. 2 Project Engineer, Hayward Baker Inc., 1780 Lemonwood Drive, Santa Paula, CA 93060. E-mail: sarora@haywardbaker.com. 1

Significant efforts have been made in the recent years to use fly ashes in soil stabilization and highway applications (Vishwanathan et al. 1997, Bergeson and Barnes 1998, Consoli et al. 2001, Parsons and Milburn 2003); however, a large percentage of reuse of fly ash in these mixes has not been studied. The objective of this study is to investigate the beneficial reuse of Class F fly ash amended mixtures for base layers in highways. To achieve this objective, a battery of tests was conducted on soil-fly ash mixtures prepared with cement and lime as activators. Unconfined compressive strength (q u ), California bearing ratio (CBR), and resilient modulus (M R ) tests were conducted to investigate the effect of fines content, curing time, molding water content, activator type, and soil cohesion on engineering parameters. MATERIALS AND TEST PROCEDURES Locally available sandy soil was used in the current study. The soil was classified as light brown silty sand (SM) according to the Unified Soil Classification System (USCS) and A-2-4 according to the American Association of State Highway and Transportation Officials (AASHTO) Classification System. The soil has approximately 18% particles passing the U.S. No. 200 sieve. Specific gravity (G s ) of the material is 2.68, and it does not exhibit any plasticity. The fly ash used in this study was low calcium Class F fly ash (0.74% CaO) obtained from Indian River Power Plant in Millsboro, Delaware. The fly ash had a dark grayish color, and a carbon content of 6-8%. Approximately 86% of the particles were finer than U.S. No. 200 sieve size. The ash had a ph of 7.9 and was insoluble in water. The specific gravity of the ash was 2.24. Type I Portland cement and high calcium (95%) quicklime from Pennsylvania Lime, Inc., were used as activators for the PSM. In order to investigate the effect of cohesion on engineering parameters, kaolinite was added to some mixtures. The kaolinite was obtained from Burgess Pigmet Inc. of Sandersville, Georgia, and had a cation exchange capacity (CEC) of 0.03-0.1 meq/g. Mixes possessing large fractions of ash was used in the testing program. Varying percentages of silty (cohesionless) fines and kaolinite were used and specimens were compacted at optimum moisture content, 4% wet and 4% dry of optimum to examine the effect of molding water content on the strength parameters. Table 1 provides a summary of the mixes used in the study along with the optimum water contents (OMC) and maximum dry unit weights ( dm ) of the mixtures based on compaction tests (ASTM D 698 and D 1557). The prepared mixes were subjected to unconfined compression and California Bearing Ratio (CBR) tests. ASTM D 1633 and D 5102 were used to determine the unconfined compressive strength of the cement and lime treated specimens, respectively, After compaction, the specimens were extruded with a hydraulic jack, sealed in plastic wrap, and cured for 1, 7, 28, and 56 days at 100% relative humidity and controlled temperature before testing. The CBR testing was conducted following the procedures listed in AASHTO designation T-193 and ASTM D 1883. The specimens were compacted at optimum moisture content (OMC) using the standard 2

Proctor effort and were cured for 7 days at 100% relative humidity and controlled temperature before testing. The procedure described in AASHTO T-294 was followed for the resilient modulus test. Specimens of 71.1 mm in diameter and 152.4 mm in height were compacted at their OMC in five layers, and were cured for 7 days at 100% relative humidity following the compaction. Details of the resilient modulus test apparatus can be found in Aydilek et al. (2003). RESULTS AND DISCUSSION Unconfined Compression and CBR Tests Non-plastic (silty) fines of the soil were varied from 6% to 30% by weight of soil particles with increments of 12%. That is, soil with 6%, 18%, and 30% fines was used. Only unconfined compression tests were used to evaluate the effect of fines content. As seen in Table 2, a consistent trend is not observed between q u and cohesionless fines content. All the specimens invariably show a similar trend of strength increase with increasing curing period. The postulated mechanism is that the release of calcium hydroxide (Ca(OH) 2 ) by Portland cement on crystallization reacts with the fly ash to form calcium aluminium silicates, which in turn hardens the specimen. The high temperature of the curing chamber and availability of 100% Table1. Legend and the composition for the mix designs. Silty Fly Specimen Soil fines Cement Lime Kaol. ash name (%) (% of (%) (%) (%) (%) soil) Compac. effort dm (kn/m 3 ) OMC (%) FA1-C7 60 6 40 7 - - Standard 15.88 16.8 FA3-C7 60 18 40 7 - - Standard 15.46 18.1 FA5-C7 60 30 40 7 - - Standard 15.34 17.7 FA3K-C7 50 18 40 7-10 Standard 15.88 16.7 FA3-C1 60 18 40 1 - - Standard 15.45 16.5 FA3-C2 60 18 40 2 - - Standard 15.5 17.5 FA3-C4 60 18 40 4 - - Standard 15.46 17.2 FA3-C5 60 18 40 5 - - Standard 15.39 17.2 FA3-L4 60 18 40-4 - Standard 15.47 17.7 FA3-L7 60 18 40-7 - Standard 15.36 17.7 FA3-L10 60 18 40-10 - Standard 15.03 18.2 FA3K-L7 50 18 40-7 10 Standard 15.45 16.8 FA1-C7-M 60 6 40 7 - - Modified 16.92 13.4 FA3-C7-M 60 18 40 7 - - Modified 17.15 13.2 FA5-C7-M 60 30 40 7 - - Modified 17.16 13.0 Note: dm = Maximum dry unit weight; A total percentage of 100% of soil and fly ash was considered as the base mix, and cement or lime was added at a certain percentage by weight of this base mix. All the specimens were compacted at their optimum moisture contents (OMC). The specimens FA1-C7, FA3-C7, FA5-C7, and FA3K-C7 were also compacted at OMC +4 and OMC -4 molding water contents, but not shown herein. 3

relative humidity enhances these cementitious reactions. Similarly, Vishwanathan et al. (1997) reported an increase of 50% in q u from 7 to 28 days for fly ash amended mixtures. The effect of water content on unconfined compressive strength of specimens with 6%, 18% and 30% fines is presented in Table 2. The effect is not prominent for the specimen cured for 28 days; however, for the same mix design, the variation in q u with compaction water content can be observed more clearly for the specimens cured for 7 days. For instance, a q u of 5.3 MPa was obtained for the specimen compacted at the drier side of optimum and 5.0 MPa for the specimen compacted at the wet side at 28 days. The values were 4.4 MPa and 0.9 MPa, respectively, for 7-day cured specimens. The effect of water content on strength can be explained by the characteristics of cementitious reactions. The water-to-cement (W/C) ratio is important in these reactions, even though it cannot always be optimized in solidification/stabilization work. At W/C > 0.48, cement is over-hydrated, leaving free water (pore water), and bleed water that appears as standing water on the surface of the solid mass (Conner 1990). The observed decrease in unconfined compressive strength with increasing molding water content was attributed to Table 2. Summary of unconfined compressive strengths (q u ) of all the specimens tested. Unconfined compressive strength (MPa) CBR Specimen Name Days of curing at 7 days of 1 7 28 56 curing (MPa) FA1-C7 1.2 3.8 5.2 7.5 FA1-C7+4 0.4 1.0 5.0 5.5 FA1-C7-4 1.8 4.4 5.4 6.5 FA3-C7 1.2 3.2 5.0 5.9 140 FA3-C7+4 0.5 1.2 3.1 4.8 FA3-C7-4 1.8 3.8 6.0 7.2 FA5-C7 1.0 1.6 6.9 7.5 FA5-C7+4 0.5 1.9 3.4 3.4 FA5-C7-4 1.5 4.4 6.2 6.9 FA3-C1-0.6 0.8-53 FA3-C2-1.4 1.8-80 FA3-C4-2.8 3.5-93 FA3-C5-3.2 4.5-133 FA3-L4-0.4 0.7 - FA3-L7-0.3 0.4 0.49 36 FA3-L10-0.1 0.2 - FA3K-L7-0.2 0.4 0.5 26 FA3K-C7 2.0 5.5 6.6 8.1 FA3K-C7+4 1.1 4.4 3.1 4.1 FA3K-C7-4 2.0 4.1 4.0 4.9 FA1-C7-M - 5.3 9.2 - FA3-C7-M - 5.1 11.2 - FA5-C7-M - 5.4 7.6-4

relatively high W/C ratios for the specimens used in the current study, since the ratio W/C was greater than 0.5 even for the mixes compacted at optimum water content using standard effort. Table 2 also shows the effect of compaction energies on q u for specimens cured 7 and 28 days. The strength of specimens cured for 7 days and compacted with modified Proctor effort are 29% to 70% higher than those compacted using standard Proctor effort. The same trend is observed for 28-day cured specimens, however, the difference is 10% to 55%. The fines present in the natural soil were non-plastic. However, it is well known that plastic fines play a major role in defining the strength of soil. Previous research indicated that the strength of fly ash-stabilized soil decreases with increasing plasticity index (PI) (Acosta 2002), and cement is generally considered as a good activator when used for stabilization of granular soils (Parsons and Milburn 2003). To determine the effect of cohesive fines on the unconfined compressive strength, mixes with 10% of kaoliniteby weight were prepared. The q u of cohesionless specimens tend to decrease with increasing water content. However, a similar conclusion can not be drawn for the specimens containing 10% kaolinite (Table 2). It has generally been accepted that cement treatment is effective for mixes containing cohesionless fines, and the presence of fines (clay and silt) is believed to reduce the ultimate strength by producing cracking and spalling, i.e., fine particulates act as inhibitors in the cementitious reactions (Conner 1990). Varying percentages of cement (i.e., 1, 2, 4, 5, and 7%) were added to a selected mix in order to investigate the effect of cement content on the unconfined compressive strength. As shown in Figure 1a, the strengths initially increase with increasing cement content. For instance, the 7-day strength of FA3-C2 (i.e., 2% cement) is about twice that of FA3-C1 (i.e., 1% cement); however, no further increase in strength is observed beyond 5% cement. Similarly, the 28-day strength increases with increasing cement content up to 5% cement and the rate decreases beyond this point. It should be noted that cement contents up to 7% were considered for the current study. Higher cement contents would probably lead to much higher strength values but also more expensive designs. For instance, a maximum strength of 5.0 MPa is observed for FA3-C7 after 28 days of curing, which is sufficient for a rigid base application in highway pavements. Fly ash and soil have limited silica or alumina available for reactions, and thus the addition of an activator such as cement is necessary. The typical Portland cement to fly ash ratio is 1:3 to 1:5 (TFHRC 2002). The addition of cement above this ratio does not contribute to strength, as the mix does not contain all the elements for the cementation. This is believed to be the main reason for not observing a strength increase at higher cement contents, and therefore a value of 5% can be termed as a cement demand for the soil-fly ash mixture. Table 2 presents the results of the CBR tests conducted on specimens with varying cement contents. The CBR consistently increases with increasing cement amount. The CBR of the 1% cement specimen (FA3-C1) is 53, which increases to 80 for FA3- C2. A CBR of 93 was determined for FA3-C4, which indicates that the mixture may be suitable for use as a base layer since a CBR of 80 is generally acceptable for bases 5

(Asphalt Institute 1991). The CBR for FA3-C5 and FA3-C7 are greater than 100, which indicates superior bearing capacities. Swell was not observed when the specimens were kept submerged for five days; hence, mixes appear not to have longterm swell potential (Aydilek et al. 2003). Figure 1b shows the results of unconfined compression tests performed on the lime treated specimens. The q u decreases with increasing lime content. For instance, the 7-day strength of specimens with 4% lime is approximately 448 kpa while the specimens with 7 and 10% lime exhibit a q u of 276 and 138 kpa, respectively. Similar results were determined after conducting unconfined compression tests on the Unconfined compressive strength (kpa) 6000 5000 4000 3000 2000 1000 0 FA3 (7 DAYS) FA3 (28 DAYS) 0 2 4 6 8 Cement content (%) Unconfined compressive strength (kpa) 800 700 600 500 400 300 200 100 0 FA3 (7 DAYS) FA3 (28 DAYS) 2 4 6 8 10 12 Lime content (%) Figure 1. Effect of cement and lime contents on unconfined compressive strengths specimens after 28 days of curing. The q u of the specimen with 4% lime is 724 kpa while that with 10% lime is only 241 kpa. Lime stabilization is usually used for high plastic clays to decrease plasticity, and to increase the strength of the mix. As a result, shear strength parameters of cohesive soils are generally improved by lime stabilization. The addition of lime increases shear strength but decreases plasticity index. For instance, Little (2000) reported an increase in q u from 160 kpato 2,275 kpa, and a decrease in PI from 38 to 10 with 6% hydrated lime treatment. Lime stabilization can only be beneficial for high plastic clays with a PI greater than 10 (Department of the Army, 1983). The fines of natural soil used in the current study were non-plastic, and this is believed to be the reason for the observed detrimental effect of lime on strength. 6

A comparison of the strengths of lime treated cohesive and cohesionless specimens (FA3K-L7 and FA3-L7, respectively) in Table 2 indicates that the presence of kaolinite was expected to increase the strength with lime treatment, but no significant gain can be observed. The q u of both mixes is comparable. Similar behavior is valid from the CBR test results. CBR of 26 and 36 is observed for the mixes containing cohesive (FA3K-L7) and cohesionless fines (FA3-L7), respectively. Here it should be noted that the effectiveness of lime-treatment is dependent on the PI of the mix and the significant presence of clay particles. The q u and CBR of the mixes did not increase in presence of cohesive fines since 10% kaolinite was not a sufficient amount to increase the plasticity of the mixture. It is believed that small amounts of clay did not increase the plasticity significantly in the current study. This is somewhat comparable with the findings of Baykal and Metehan (2002) that lime decreased plasticity and caused the formation of a granular mixture. Furthermore, the presence of cohesive fines might have inhibited cementitious reactions between lime and fly ash (Conner 1990). Resilient Modulus Tests The resilient modulus (M R ) describes the deformation behavior of a material subjected to repetitive loading and, hence, is similar to actual loading conditions due to vehicle traffic on pavements. The nonlinear behavior of the resilient modulus of a soil is dependent on the stress level. This was defined in this study using the common model: M R = K 1 2 (1) where, M R is resilient modulus, K 1 and K 2 are constants, ( = d + 3 c ) is bulk stress, c is the isotropic confining pressure, and d is the deviator stress. A bulk stress of 140 kpa was used in the current study to verify the power model. This stress level is typical in the middle of a granular layer (Chen et al. 1995) and is close to the minimum bulk stress applied during resilient modulus tests. High R 2 values were obtained from regression analyses performed on the transformed log model (Arora 2003). The good fit indicated that the mixtures have a response similar to that of granular materials. As shown in Figure 2, an increase in resilient modulus with increasing bulk stress is observed for most of the specimens, which is in agreement with the behavior generally observed for granular soils (AASHTO T 294). The optimum water content of specimens compacted using standard Proctor effort range from 16.7% to 18.2%, and the difference between the maximum dry densities of the specimens is insignificant. Therefore, the difference in resilient modulus is attributed solely to the variation of cement contents. As expected, resilient modulus increases with increasing cement content at a constant bulk stress. Increase in cement content results in production of more cementitious compounds and consequently increases the strength. The specimens with 1% and 2% cement (FA3-C1 and FA3- C2) show comparable M R values. Similarly, the magnitude of difference of resilient modulus between the specimens with 4%, 5%, and 7% cement is relatively small, but a large change in M R can be observed between the mixes containing 2% and 4% 7

Resilient Modulus, M R (kpa) 4 10 5 3.5 10 5 3 10 5 2.5 10 5 2 10 5 1.5 10 5 1 10 5 FA3-C7 FA3-C5 FA3-C3 FA3-C1 FA3-C2 Resilient Modulus, M R (kpa) 3 10 5 2.5 10 5 2 10 5 1.5 10 5 1 10 5 FA3-C7 FA3K-C7 FA3K-L7 5 10 4 (a) 5 10 4 (b) 0 0 0 50 100 150 200 250 300 350 400 Bulk stress, (kpa) 0 50 100 150 200 250 300 350 400 Bulk stress, (kpa) Figure 2. Resilient modulus of the specimens with varying bulk stresses. cement. It can be concluded that the increase in cement content leads to increase in M R, but the rate decreases beyond 4% cement. This supports observations made concerning 5% cement being satisfactory for cementitious reactions between fly ash and cement. Attempts were also made to determine the resilient modulus of the specimen with 7% lime (FA3-L7); however, strength gains were not observed even after curing for a period of 7 days. Figure 2b shows the resilient modulus of the selected specimens as a function of bulk stress. A comparison of FA3K-C7 and FA3- C7 gives a clear indication that cement treatment is more effective for cohesionless soils as the M R values for the FA3-C7 are much higher than FA3K-C7. Some loss in M R of the mix that contained kaolinite may be due to the presence of higher fines content in the mix (Conner 1990). CONCLUSIONS Due to the lack of self-cementitious characteristics, Class F fly ash needs an activator (e.g. cement), and currently only 32% of this ash has been beneficially reused. Roadways are the biggest application area and use of this ash could save millions of dollars annually as most of the ash is landfilled today. A study was conducted to promote the use of Class F fly ash and to investigate the effect of fines content, curing period, molding water content, compactive effort, cohesion, and cement or lime addition on geomechanical parameters of fly ash amended highway bases. The variation of q u with varying amounts of cohesionless fines was not consistent. On the other hand, addition of 10% kaolinite generally increased the strength of a mixture. It should be noted that for this study, cohesionless fines were varied from 8

6% to 30% of sand by weight and hence the observed results should be interpreted only for this range. An increase in strength can be obtained in the field by compacting the soil using higher compactive efforts (i.e., modified versus standard effort). The increase in strength with curing time was determined for all specimens irrespective of the molding water content. The highest strength was observed at 56 days; however, strength seemed to increase beyond this curing time. The test results showed that the water content at compaction could affect the q u of the mix design considerably. The performance of the fly ash, soil, and cement mix can be significantly increased by preventing the intrusion of excess water in the field. It is recommended to compact the base layer at dry of optimum for higher strength. Alternatively, compaction may be performed at optimum water content; however, engineers should be careful concerning rain or any other addition of unwanted water at the time of compaction. CBR, q u, and M R increased with increasing cement content; however, the rate decreased beyond 5% cement. That is, the strength of the mix did not increase proportionally beyond 5%. This was true for all three test methods. Lime treatment had a detrimental effect on the mix designs. An increase in lime content decreased the q u of the specimens for both 7-day and 28-day old specimens. On the hand, the increase in curing period had a positive effect and increased the q u of the lime treated mixes. The presence of cohesion lowered the q u as well as CBR values during lime treatment, while the M R of the specimen with kaolinite (FA3K-L7) was higher than that of its cohesionless companion (FA3-L7). As part of the study, the thicknesses of highway base layers with different mix designs were calculated using the q u, CBR and M R values but not shown herein. Lower thicknesses were required when higher amount of cement is used or higher compactive energies are employed. Presence of lime or cohesive fines generally required higher base thicknesses indicating that use of cohesionless fines, such as sandy soils, should be preferred. REFERENCES Acosta H. A. (2002). Stabilization of Soft Subgrade Soils Using Fly Ash, M.S. Thesis, University of Wisconsin, Madison, WI, 125 p. Arora, S. (2003). Suitability of Fly Ash Stabilized Soils as Highway Base Material, M.S. Thesis, University of Maryland, College Park, MD, 141 p. Asphalt Institute (1991). Thickness Design Asphalt Pavements for Highways and Streets, Manual Series No. 1, Asphalt Institute, Lexington, Kentucky. Baykal, G. and Metehan, T. (2002). The Effect of Lime Treatment on the Shear Strength Parameters of The Clay-Concrete Interface, Transportation Research Board, 81 st Annual Meeting, Washington, D.C. Bergeson, K.L., and Barnes, A.G. (1998). Iowa Thickness Design for Low Volume Roads Using Reclaimed Hydrated Class C Fly Ash Bases, ISUERI-Ames 98401, Iowa State University, Ames, Iowa. 9

Chen, D., Zaman, M., and Laguras, J. (1995). Characterization of Base/Subbase Materials under Repetitive Loading, Journal of Testing and Evaluation, ASTM, Vol. 23, No. 3, pp.180-188. Conner, J.R. (1990). Chemical Fixation and Solidification of Hazardous Wastes, Van Nostrand Reinfold, New York, 692p. Consoli, N. C., Prietto P. D. M., Carraro, J. A. H., Heineck, K., S. (2001). Behavior of Compacted Soil-Fly Ash-Carbide Lime Mixtures." Journal of Geotechnical and Geoenvironmental Engineering, ASCE, Vol. 127, No. 9, pp. 774-782. Department of the Army (1983). Soil Stabilization for Pavements, online document (http://www.army.mil/usapa/eng/) Little, D.N. (2000). Evaluation of Structural Properties of Lime Stabilized Soils and Aggregates, Mixture Design and Testing Protocol for Lime Stabilized Soils, Prepared for the National Lime Association, Vol. 3. 16 p. Parsons, R.L. and Milburn, J.P. (2003). Engineering Behavior of Stabilized Soils, Transportation Research Board, 82 nd Annual Meeting, Washington, D.C., CD- ROM, 29 p. Turner-Fairbanks Highway Research Center (TFHRC) (2002). Coal Fly Ash User Guideline Stabilized Base (http://www.tfhrc.gov/hnr20/recycle/waste/cfa55.htm) Vishwanathan, R., Saylak, D., and Estakhri, C. (1997). Stabilization of Subgrade Soils Using Fly Ash, Ash Utilization Symposium, CAER, Kentucky, pp. 204-211. 10