Methods for Evaluating Fly Ash for Use in Highway Concrete (2013)

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

23 CFA Characterization Study Results obtained from characterization tests of the 30 sources of CFA are summarized in this section. Additional results are provided in Attachment C. Chemical Tests The bulk chemical properties of the CFA samples are sum- marized in Tables 4.1 through 4.3. The bulk chemical (XRF) test results for CFA are reported on an oven-dry basis in an oxide format. A comparison of reported and measured values is provided in Attachment C. Table 4.1 lists the major elements, moisture content, and LOI test results. The minor oxides and available alkali test results (i.e., lime-soluble alkali content at 28 days per ASTM C311) are given in Tables 4.2 and 4.3, respectively. In Table 4.3, aNa2O and aK2O denote the available sodium and potassium values determined by flame photometry, respec- tively. The available alkali content is expressed as equivalent sodium oxide (%Na2Oe = %Na2O + 0.658 × %K2O). The available alkali test results are reported on an as-received basis (no drying prior to the test). The bulk chemistry values for minor components and available alkali results for the 30 samples of CFA showed a good range of composition (see Tables 4.2 and 4.3). Seven Class C ashes and three Class F ashes failed the AASHTO M 295 requirement for available alkali not to exceed 1.50% wt. However, two of the Class C ashes failed the requirement by a very small margin (1.58% wt for sample FA-ZD and 1.52% wt for sample FA-W). The available alkali test exhib- its rather poor precision, and it is therefore unlikely these two CFAs were significantly above the specification limit. However, the total alkali data, primarily sodium oxide con- tent, indicated an increase in alkali content for the CFAs that failed the available alkali test as seen in Figure 4.1. Class F fly ashes containing potassium as the major alkali (below 1% on the graph) tend to exhibit little correlation between total and available alkali. In contrast, when sodium becomes the primary alkali, as is the case for most Class C ashes, there is a good correlation between total alkali and available alkali. This could be used to simplify the process of determining if a particular source of CFA meets the alkali requirements. The available alkali test takes about 35 days to complete. Total alkali could easily be determined on a daily basis and could be used to provide a quick determination of the alkali content of CFA. Another investigation was conducted to evaluate the amount of sodium and potassium that could be extracted from CFA when it was mixed with water (i.e., water-soluble alkali). This investigation included the eight CFA samples that had been selected for extensive ASR testing; the test results are sum- marized in Table 4.4. The water-soluble alkali test utilized a sample of 1.75 g of the as-received CFA mixed with 200 mL of water and stirred for 1 h. The suspension was filtered using a medium-texture filter paper; solids were then washed using room tempera- ture water. The effect of the number of wash cycles on the measured soluble alkali content was small for three, six, and nine wash cycles (±0.02% soluble alkali expressed as %Na2Oe). The test results obtained using three washes are given in Table 4.4. The results presented in Tables 4.3 and 4.4 indicate sol- uble alkali of about an order of magnitude lower than the available alkali, when expressed as %Na2Oe, except for two Class C fly ashes (FA-X and FA-ZA). Coal fly ash FA-ZA was obtained from a power plant that reportedly adds trona to the raw coal feed to enhance the performance of its electro- static precipitators and that appears to influence the amount of soluble sodium in the CFA. C H A P T E R 4 Findings

24 ID Class Sum of Oxides (% wt) Moisture (% wt) LOI (% wt) SiO2 (% wt) Al2O3 (% wt) Fe2O3 (% wt) CaO (% wt) FA-A F 92.52 0.04 0.94 61.6 27.9 3.02 0.82 FA-B C 67.62 0.14 1.19 39.2 20.3 8.12 14.3 FA-C F 88.99 0.10 1.22 59.8 22.2 6.99 2.24 FA-E F 87.76 0.01 1.84 55.5 24.6 7.66 2.78 FA-F F 90.62 0.04 2.26 56.9 27.3 6.42 1.09 FA-G F 89.89 0.02 2.32 53.9 27.7 8.29 1.45 FA-H F 91.26 0.02 0.25 60.9 25.7 4.66 3.46 FA-I F 89.45 0.09 2.19 62.4 20.1 6.95 1.81 FA-J F 91.90 0.09 1.59 46.0 23.6 22.3 1.28 FA-K F 83.50 0.08 1.61 46.9 23.2 13.4 6.85 FA-L F 84.80 0.05 0.94 47.2 19.3 18.3 6.79 FA-M F 81.85 0.07 0.27 60.3 16.6 4.95 7.17 FA-N F 86.90 0.03 0.80 46.6 19.9 20.4 5.33 FA-O F 79.81 0.03 1.43 58.9 16.2 4.71 10.2 FA-P F 73.34 0.05 0.13 50.2 16.9 6.24 14.0 FA-Q F 74.34 0.07 0.38 50.3 19.2 4.84 16.6 FA-R F 73.27 0.02 0.07 50.7 15.3 7.27 15.3 FA-S F 70.55 0.02 1.01 43.2 20.4 6.95 17.1 FA-T F 77.41 0.11 0.45 44.8 23.1 9.51 13.6 FA-U C 65.80 0.02 0.54 39.4 19.4 7.00 21.9 FA-V C 63.00 0.06 0.50 38.0 19.3 5.70 24.8 FA-W C 62.83 0.03 0.33 35.7 20.0 7.13 24.2 FA-X C 61.63 0.06 0.42 36.7 19.5 5.43 19.3 FA-Y C 62.77 0.03 0.20 37.1 19.5 6.17 24.4 FA-Z C 61.21 0.04 0.17 34.4 20.0 6.81 26.5 FA-ZA C 55.32 0.02 0.27 32.8 16.8 5.72 27.3 FA-ZB C 61.66 0.05 0.16 37.2 19.3 5.16 25.7 FA-ZC C 53.09 0.02 0.16 31.4 15.9 5.79 30.2 FA-ZD C 54.27 0.02 0.20 30.8 17.6 5.87 29.2 FA-ZL C 61.52 0.05 0.32 36.3 19.4 5.82 18.0 Table 4.1. Summary of CFA chemical properties. Table 4.2. Summary of CFA minor elements. ID MgO (% wt) SO3 (% wt) Na2O (% wt) K2O (% wt) Total Alkali P2O5 (% wt) TiO2 (% wt) Na2Oe (% wt) FA-A 0.84 0.19 0.30 2.80 2.14 0.18 1.45 FA-B 3.53 3.22 5.84 1.10 6.56 0.56 1.24 FA-C 1.79 0.65 0.92 2.27 2.41 0.19 1.00 FA-E 1.23 0.70 1.02 2.09 2.40 0.23 1.20 FA-F 0.83 0.31 0.24 2.43 1.84 0.37 1.58 FA-G 1.15 0.24 0.38 2.88 2.28 0.37 1.36 FA-H 1.12 0.18 1.46 1.23 2.27 0.07 1.09 FA-I 1.33 0.33 0.89 1.93 2.16 0.17 0.95 FA-J 0.99 0.97 0.43 2.73 2.23 0.18 1.08 FA-K 1.79 1.27 0.83 1.68 1.94 0.66 1.16 FA-L 0.80 2.27 0.58 2.15 1.99 0.18 0.99 FA-M 2.58 1.07 3.34 1.36 4.23 0.30 0.87 FA-N 1.12 1.15 0.62 3.03 2.61 0.43 0.95 FA-O 3.13 0.86 1.19 1.29 2.04 0.46 0.85 FA-P 4.38 0.90 3.32 1.71 4.45 0.35 0.99 FA-Q 3.46 0.88 1.05 0.88 1.63 0.25 1.28 FA-R 5.24 0.83 1.51 2.23 2.98 0.13 0.59 FA-S 3.41 2.01 2.53 0.76 3.03 0.74 1.32 FA-T 2.97 0.96 1.00 1.44 1.95 0.74 1.50 FA-U 4.84 1.07 1.54 0.92 2.15 1.05 1.43 FA-V 4.64 1.41 1.62 0.50 1.95 1.03 1.55 FA-W 4.79 2.30 1.80 0.53 2.15 1.16 1.54 FA-X 5.21 2.56 6.33 0.79 6.85 0.56 1.32 FA-Y 5.45 1.24 1.65 0.52 1.99 1.22 1.48 FA-Z 4.78 1.93 1.70 0.39 1.96 1.33 1.64 FA-ZA 6.41 3.56 3.70 0.41 3.97 0.94 1.28 FA-ZB 4.84 1.71 1.13 0.44 1.42 1.70 1.45 FA-ZC 7.93 2.64 2.32 0.27 2.50 0.93 1.32 FA-ZD 7.64 3.25 2.19 0.30 2.39 0.86 1.33 FA-ZL 4.44 3.43 7.45 0.86 8.02 0.60 1.29

25 ID Mn2O3 (% wt) SrO (%wt) BaO (%wt) aNa2O (%wt) aK2O (%wt) AA (%wt) FA-A 0.01 0.07 0.10 0.12 0.74 0.61 FA-B 0.04 0.54 1.15 4.21 0.65 4.64 FA-C 0.06 0.05 0.10 0.31 0.70 0.77 FA-E 0.04 0.10 0.16 0.45 0.73 0.93 FA-F 0.02 0.11 0.12 0.10 0.66 0.54 FA-G 0.03 0.12 0.14 0.13 0.77 0.64 FA-H 0.02 0.10 0.18 0.48 0.32 0.69 FA-I 0.06 0.08 0.19 0.35 0.70 0.81 FA-J 0.03 0.04 0.07 0.18 0.89 0.77 FA-K 0.03 0.15 0.23 0.28 0.44 0.56 FA-L 0.04 0.05 0.05 0.25 0.77 0.76 FA-M 0.05 0.20 0.56 1.61 0.58 1.99 FA-N 0.05 0.05 0.19 0.21 0.85 0.76 FA-O 0.02 0.12 0.16 0.46 0.42 0.73 FA-P 0.08 0.31 0.63 1.59 0.68 2.04 FA-Q 0.13 0.31 0.38 0.57 0.36 0.81 FA-R 0.07 0.29 0.55 0.42 0.54 0.77 FA-S 0.03 0.35 0.70 1.53 0.37 1.77 FA-T 0.03 0.29 0.47 0.50 0.59 0.89 FA-U 0.04 0.34 0.70 1.02 0.52 1.37 FA-V 0.06 0.39 0.77 1.20 0.35 1.43 FA-W 0.03 0.37 0.78 1.28 0.36 1.52 FA-X 0.04 0.69 1.57 4.57 0.42 4.88 FA-Y 0.02 0.37 0.81 1.14 0.31 1.34 FA-Z 0.04 0.44 0.84 1.23 0.27 1.41 FA-ZA 0.03 0.47 0.83 2.16 0.22 2.31 FA-ZB 0.02 0.44 0.97 0.73 0.26 0.90 FA-ZC 0.02 0.54 0.90 1.62 0.18 1.74 FA-ZD 0.03 0.55 0.95 1.46 0.18 1.58 FA-ZL 0.05 0.67 1.51 5.67 0.54 6.03 AA = available alkali Table 4.3. Summary of CFA minor elements and available alkali results. 0 1 2 3 4 5 6 7 0 2 4 6 8 10 A va ila bl e Al ka li (as % Na 2O e) Total Alkali (as %Na2Oe) Class C Class F Figure 4.1. Total alkali versus available alkali. ID Na2O (% wt) K2O (% wt) %Na2Oe (% wt) FA-H 0.07 < 0.01 0.07 FA-M 0.19 0.02 0.20 FA-O 0.06 0.01 0.06 FA-Q 0.06 0.01 0.07 FA-U 0.06 0.01 0.07 FA-X 1.05 0.08 1.10 FA-ZA 0.72 0.01 0.73 FA-ZC 0.18 0.01 0.19 Table 4.4. Water-soluble alkali for selected samples of coal fly ash after three wash cycles. Physical Tests The results of the physical tests conducted in accordance with ASTM C311 are presented in this section. The values for physical properties are the average of two tests conducted on different days. The values for SAI, PAI, and ASR tests were typically obtained from a single test. However, tests on six or seven of the mortar mixtures were repeated to pro- vide an estimate of the precision of the test results. Infor- mation on the precision of the various tests is provided in Attachment C. The soundness (i.e., autoclave expansion) values for the various CFA sources ranged from -0.03% to 0.12%, which is substantially lower than the specification limit given in AASHTO M 295 of 0.8% expansion. A good correlation was noted between expansion and bulk %MgO as shown in Figure 4.2. The mass retained on a #325 mesh sieve as determined from the fineness tests ranged from 10% to 27%. Class C fly ashes tended to have lower fineness values than Class F fly ashes as shown in Figure 4.3. All of the samples easily met the 34% maximum fineness limit. Particle size distributions for the thirty sources of CFA, PC-1, INF-1, INF-2, two additional CFA sources (FA-ZM and FA-ZN), and a commercially available ground blast furnace slag used in strength modeling experiments were determined by scanning electron microscopy (SEM) and quantitative image analysis. The results are summarized in Attachment C together with a description of the methodology used. -0.04 -0.02 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0 2 4 6 8 10 A ut oc la ve E xp an si on (% ) MgO (% wt) Class C Class F Figure 4.2. Percentage of magnesium oxide versus autoclave expansion.

26 with LOI or fineness as shown in Figure 4.4. Similar to den- sity, AEA dosage is also a uniformity measure for CFA sources and, therefore, only a variability metric is specified. This test method needs to be updated to reflect the current practice of substituting CFA for an equal mass of cement rather than as a replacement of fine aggregate. This updated test method would effectively change the aggregate to cement ratio from 3.0 to 4.0. Additional testing would indicate if the change would impact the AEA dosage for the various CFAs. Density of the 30 CFA sources ranged from about 2.1 to 2.8 g/cm3. Class C fly ashes tended to have higher densities than Class F fly ashes. Because density is a uniformity mea- sure of CFA sources, only a variability metric is specified (i.e., ±5% relative to a moving average). Since only single samples were obtained from the various sources, the uniformity of each product stream could not be evaluated. This test method utilized a helium pycnometer for the determinations, which provided good repeatability [within lab (single-operator) standard deviation at about 0.005 g/cm3, suggesting a maxi- mum difference between duplicate determinations (d2s) of about 0.015 g/cm3]. Therefore, use of a helium pycnometer appears appropriate for evaluating product uniformity. Mortar Air Content Tests The results of the mortar air content tests conducted in accordance with ASTM C311 using PC-2 and 20-30 sand are also summarized in Table 4.5. At least four mortar mixtures were made for each CFA sample. The first mortar consisted of CFA, cement, sand, and water (i.e., no AEA was added). Sub- sequent mortar mixtures were made with increasing amounts of AEA. The target air content for the second and third mor- tar mixtures were 15% to 18% and 18% to 21%, respectively. The test results were then used to estimate by interpolation the amount of AEA needed to produce a mortar containing 18% air; additional mortar mixture was made on a different day to verify this value. All of the final mortar mixtures had air contents of 18 ± 1%. The dosage of AEA required to produce 18% mortar air content ranged from about 1.1 to 2.7 oz/cwt of cementitious material, which is within the “normal” dosage recommended by the manufacturer. Class C ashes, except FA-V, tended to require less AEA than Class F ashes to produce 18% mortar air. In addition, AEA dosage did not appear to correlate well 0 5 10 15 20 25 30 35 50 60 70 80 90 100 Fi ne ne ss (% R eta ine d) Sum of Oxides (SiO2 + Al2O3+Fe2O3) (% wt) C Ash F Ash Figure 4.3. Sum of the oxides versus fineness. ID Fineness (% retained) Density (g/cm3) Soundness (% expansion) AEA dosage (oz/cwt) FA-A 20.6 2.19 0.02 1.88 FA-B 13.2 2.56 0.04 1.59 FA-C 20.7 2.35 0.01 1.72 FA-E 19.5 2.36 0.01 1.62 FA-F 20.7 2.25 0.02 1.99 FA-G 19.1 2.31 0.03 1.98 FA-H 26.6 2.11 0.02 1.70 FA-I 15.8 2.43 0.01 2.10 FA-J 17.8 2.50 0.01 2.01 FA-K 18.6 2.53 0.01 2.01 FA-L 13.1 2.55 0.02 1.62 FA-M 22.8 2.41 0.05 1.39 FA-N 26.0 2.49 0.01 1.34 FA-O 25.7 2.41 0.06 2.63 FA-P 19.8 2.58 0.04 1.21 FA-Q 16.6 2.47 0.01 2.18 FA-R 25.9 2.45 0.02 1.24 FA-S 18.6 2.53 0.03 2.16 FA-T 23.6 2.48 0.00 1.24 FA-U 15.0 2.61 0.02 1.59 FA-V 17.0 2.68 0.03 2.15 FA-W 12.6 2.71 0.01 1.60 FA-X 12.3 2.66 0.07 1.24 FA-Y 16.9 2.62 0.04 1.20 FA-Z 17.6 2.70 0.03 1.12 FA-ZA 14.2 2.73 0.08 1.22 FA-ZB 12.6 2.61 0.03 1.20 FA-ZC 13.0 2.77 0.12 1.20 FA-ZD 10.9 2.71 0.10 1.34 FA-ZL 9.8 2.57 0.10 1.37 Table 4.5. Physical properties of the CFA samples. R2 = 0.429 1.0 1.4 1.8 2.2 2.6 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 A EA D os ag e (o z A EA / c wt ce me nt ) LOI (% wt) Class C Class F Figure 4.4. LOI versus AEA dosage.

27 ID Compressive Strength Ratio (% control) Water Required (% control) 7-day SAI 91-day SAI FA-A 85 107 100.6 FA-B* 95 108 96.4 FA-C 86 110 99.4 FA-E 82 104 99.2 FA-F 83 98 98.9 FA-G 83 103 98.9 FA-H* 79 107 98.2 FA-I 81 107 98.9 FA-J 81 98 98.3 FA-K 79 99 96.4 FA-L 77 96 98.3 FA-M 78 105 98.3 FA-N 79 96 99.2 FA-O 79 100 98.3 FA-P 85 107 96.9 FA-Q 87 111 96.9 FA-R* 84 102 96.1 FA-S 88 99 96.9 FA-T 82 95 96.9 FA-U 88 102 96.9 FA-V 93 107 96.9 FA-W 95 108 96.9 FA-X* 99 95 96.1 FA-Y 91 105 96.4 FA-Z 90 108 96.9 FA-ZA 92 100 96.9 FA-ZB* 96 112 96.2 FA-ZC* 97 103 95.8 FA-ZD* 95 103 96.1 FA-ZL 104 28-day SAI 91 99 90 89 84 90 83 91 86 89 87 91 83 85 90 96 87 93 86 106 100 110 97 100 103 93 101 100 98 99 104 95.3 * Denotes the average of two batches Table 4.6. SAI test results. SAI Tests The results of the standard SAI test are given in Table 4.6. The test specimens consist of mortar cubes that are broken in unconfined compression after curing for specific periods of time. The test results are expressed in terms of a compres- sive strength ratio (expressed as a percentage) to the portland cement-only control mixture, which was prepared at a fixed water-cement ratio (0.485). The SAI tests are formulated on a mass basis. Coal fly ash is used to replace 20% of the cement in the mixture and water (expressed as a percentage of the control mixture water) is added as required to produce a flow within ±5% of the con- trol mortar. The test results showed that all CFA mixtures met both the 7- and 28-day specification limits (i.e., 75% minimum). Test results also showed reasonable trends with respect to CFA classification (i.e., the sum of silicon, alumi- num, and iron oxides) that indicated Class F ashes had a lower rate of strength gain at early ages (see Figure 4.5). However, after extended moist curing (e.g., 90 days), the trend is no longer evident. None of the mixtures approached the maxi- mum water requirement limit of 105%. SAI tests were also conducted on mixtures made with ground quartz to determine if finely ground materials without poz- zolanic properties meet the specification requirements. Tests conducted using ground quartz materials INF-1 and INF-2 produced 7- and 28-day test results of 81% and 74%, respec- tively, for INF-1, and 74% and 73%, respectively, for INF-2. These values are very close to the specification limit. Further discussion of this issue is provided later in this chapter. PAI Tests Like the SAI tests, the PAI tests use mortar specimens, but the mixture formulation and curing regime differ between the two tests. The PAI tests were formulated on a volume basis. Coal fly ash replaced 35% of the volume of cement in the mixture and water is added to produce a flow within 100% to 115%. Cubes were cured at 38°C ± 2°C in sealed containers (i.e., wide-mouth mason jars). The PAI test is slightly acceler- ated in comparison to the SAI test; results are given in Table 4.7. The PAI test results showed reasonable trends with respect to CFA classification (i.e., the sum of silica, alumina and iron oxide) that indicated Class F ashes had a lower rate of strength gain at early ages (see Figure 4.6). However, after 28 days of moist curing at 38°C, the trend was no longer evident. All CFA mixtures met the 28-day specification limit of at least 75% of the control strength. However, some Class C ashes showed little or no strength gain from 7 to 28 days as shown in Figure 4.7, which is somewhat different than the SAI test results. PAI tests were also conducted using INF-1 and INF-2. The 7- and 28-day test results using INF-1 were 57% and 58%, respec- tively, and 51% and 56%, respectively, for INF-2. Therefore, this test method was able to discriminate between ground quartz and CFA samples. None of the mixtures approached the maximum water requirement limit of 105%. ASR Mortar Bar Tests The results of the ASR mortar bar tests conducted in accor- dance with ASTM C311 are summarized in Table 4.8. 70 80 90 100 110 120 50 60 70 80 90 100 St re ng th A ct iv ity In de x (% of co nt ro l) Sum of Oxides (SiO2 + Al2O3+Fe2O3) (% wt) 7d 28d 90d Figure 4.5. Strength activity index versus sum of the oxides.

28 The values listed in Table 4.9 were calculated using the raw data in Table 4.8. Because these were calculated, the test results in Table 4.8 were reported to an extra significant figure (i.e., three decimal places rather than the normal two commonly used for this type of test result). The values in Table 4.9 are reported as integers as that is roughly the precision of the base measurements. The values in Table 4.9 are the measured expansions relative to the low-alkali control cement PC-1, low-alkali lab cement PC-2, and high-alkali control cement PC-3. This particular table will be used to illustrate an anom- ID Expansion (%) Difference 14 days 28 days 56 days (56 – 14 days) PC-1 only 0.202 0.262 0.285 0.083 PC-2 only 0.306 0.331 – – PC-3 only 0.488 0.579 0.610 0.122 FA-A 0.147 0.152 0.152 0.005 FA-B 0.193 0.234 0.262 0.069 FA-C* 0.174 0.185 0.188 0.015 FA-E 0.184 0.190 0.191 0.007 FA-F 0.192 0.200 0.200 0.008 FA-G 0.183 0.193 0.196 0.013 FA-H 0.179 0.187 0.188 0.009 FA-I* 0.158 0.167 0.170 0.014 FA-J 0.197 0.208 0.209 0.012 FA-K 0.179 0.195 0.199 0.020 FA-L 0.224 0.246 0.253 0.029 FA-M 0.192 0.221 0.238 0.046 FA-N 0.190 0.203 0.205 0.015 FA-O 0.199 0.216 0.220 0.021 FA-P* 0.227 0.254 0.270 0.054 FA-Q 0.159 0.166 0.170 0.011 FA-R 0.193 0.209 0.216 0.023 FA-S 0.214 0.239 0.252 0.038 FA-T 0.189 0.201 0.202 0.013 FA-U* 0.284 0.312 0.325 0.045 FA-V 0.285 0.305 0.314 0.029 FA-W* 0.253 0.278 0.287 0.026 FA-X 0.220 0.259 0.282 0.062 FA-Y 0.283 0.305 0.312 0.029 FA-Z 0.286 0.306 0.316 0.030 FA-ZA 0.275 0.302 0.319 0.044 FA-ZB 0.276 0.304 0.316 0.040 FA-ZC* 0.350 0.380 0.400 0.040 FA-ZD 0.372 0.394 0.409 0.037 FA-ZL 0.205 0.273 0.313 0.108 * Denotes the average of two batches Table 4.8. Expansion measured in ASR tests. ID Compressive Strength Ratio (% control) Water Required (% control) 7-day PAI 28-day PAI FA-A 64 93 99.8 FA-B 94 95 96.5 FA-C* 72 102 99.0 FA-E 70 92 99.0 FA-F 64 91 100.6 FA-G 67 95 99.0 FA-H 66 91 102.3 FA-I 80 99 97.7 FA-J 67 88 97.3 FA-K 68 90 95.7 FA-L 77 92 96.9 FA-M 76 98 99.0 FA-N* 62 91 99.0 FA-O 69 88 99.8 FA-P 80 103 94.4 FA-Q 83 109 94.4 FA-R 78 94 92.4 FA-S 82 98 95.7 FA-T 72 92 95.7 FA-U* 86 105 95.7 FA-V 96 103 93.6 FA-W* 112 112 93.6 FA-X* 95 87 93.6 FA-Y 88 99 93.6 FA-Z 91 89 93.2 FA-ZA 90 85 93.2 FA-ZB 94 100 94.0 FA-ZC* 86 80 91.5 FA-ZD 84 86 92.8 FA-ZL 87 84 92.8 * Denotes the average of two batches Table 4.7. PAI test results. 60 70 80 90 100 110 120 50 60 70 80 90 100 Po zz ol an ic A ct iv ity In de x (% of co nt ro l) Sum of Oxides (SiO2+Al2O3+Fe2O3) (% wt) 7-d PAI 28-d PAI Figure 4.6. Sum of the oxides versus pozzolanic activity index. -10 -5 0 5 10 15 20 25 30 35 40 50 60 70 80 90 100 Ch an ge in P AI (% 28 d - % 7d ) Sum of Oxides (SiO2+Al2O3+Fe2O3) (% wt) Class C Class F Figure 4.7. Sum of the oxides versus change in pozzolanic activity index from 7 to 28 days of curing.

29 ID Relative Expansion (% control mixture) Reduction in Expansion (% control mixture) PC-1 control cement Na2Oeq = 0.51% PC-2 lab cement Na2Oeq = 0.53% PC-3 control cement Na2Oeq = 1.04% 14 days 28 days 56 days 14 days 28 days 14 days 28 days 56 days FA-A 73 58 53 48 46 70 74 75 FA-B 96 89 92 63 71 60 60 57 FA-C 86 70 66 57 56 64 68 69 FA-E 91 73 67 60 57 62 67 69 FA-F 95 76 70 63 60 61 65 67 FA-G 91 74 69 60 58 63 67 68 FA-H 89 71 66 58 56 63 68 69 FA-I 78 64 59 51 50 68 71 72 FA-J 98 79 73 64 63 60 64 66 FA-K 89 74 70 58 59 63 66 67 FA-L 111 94 89 73 74 54 58 59 FA-M 95 84 84 63 67 61 62 61 FA-N 94 77 72 62 61 61 65 66 FA-O 99 82 77 65 65 59 63 64 FA-P 112 97 95 74 77 54 56 56 FA-Q 79 63 60 52 50 67 71 72 FA-R 96 80 76 63 63 60 64 65 FA-S 106 91 88 70 72 56 59 59 FA-T 94 77 71 62 61 61 65 67 FA-U 141 119 114 93 94 42 46 47 FA-V 141 116 110 93 92 42 47 49 FA-W 125 106 101 83 84 48 52 53 FA-X 109 99 99 72 78 55 55 54 FA-Y 140 116 109 92 92 42 47 49 FA-Z 142 117 111 93 92 41 47 48 FA-ZA 136 115 112 90 91 44 48 48 FA-ZB 137 116 111 90 92 43 47 48 FA-ZC 173 145 140 114 115 28 34 35 FA-ZD 184 150 144 122 119 24 32 33 FA-ZL 101 104 110 67 82 58 53 49 Table 4.9. ASR test results expressed as relative expansion. aly in the current testing scheme, although the test method has positive aspects. Results from the ASR mortar bar tests, Table 4.9, are illus- trated in Figures 4.8 and 4.9 for Class C and Class F fly ashes. The 14-day expansion relative to the expansion of the low-alkali control cement is plotted on the x-axis and the 14-day reduc- tion in expansion relative to the high-alkali cement is plotted on the y-axis. There is no specification limit for the percentage of reduction in expansion for CFA for the Pyrex glass-mortar bar test method but earlier versions of AASHTO M 295 used an absolute maximum expansion limit of 0.020%, except for 20 30 40 50 60 70 80 60 80 100 120 140 160 180 200 R ed uc tio n Re la tiv e to H ig h- Al ka li PC -3 (% ) Expansion Relative to Low-Alkali PC-1 (%) Class C Class F FailPass Figure 4.8. Expansion relative to PC-1. 98.5, 59.2 97.5, 59.6 55 60 65 85 90 95 100 105 110 115Re du ct io n Re la tiv e to H ig h- Al ka li PC -3 (% ) Expansion Relative to Low-Alkali PC-1 (%) Class C Class F FailPass Figure 4.9. Expansion relative to PC-1 near failure/ acceptance limit.

30 20 30 40 50 60 70 80 40 50 60 70 80 90 100 110 120 130 R ed uc tio n Re la tiv e to P C- 2 (% ) Expansion Relative to PC-2 (%) Class C Class F FailPass Figure 4.10. Expansion relative to PC-2. 0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 Ex pa ns io n at 1 4 da ys (% ) Total Alkali as Na2Oe (%) Roy, 2002 This Study Figure 4.11. Expansion in the Pyrex mortar bar test. natural pozzolans for which a minimum reduction in expan- sion of 75% was the acceptance criterion. The current limit of 100% expansion relative to the low-alkali cement and the proposed 60% reduction in expansion limit relative to the high-alkali cement are in general agreement as they tend to rank the 30 CFA sources used in this project in a similar order. Class C ashes are generally on the lower right-hand side of Figure 4.8 and Class F ashes are generally on the upper left- hand side. Only two Class F CFAs failed the 60% reduction criterion while passing the current specification limit of 100% maximum, low-alkali control (see Figure 4.9). Therefore, the two failure limits are in reasonable agreement, especially if the precision of the test method is considered. To further analyze the results with consideration of the alkali content of the cement, test results using PC-2 as the low-alkali control cement are also presented in Table 4.9 and Figure 4.10. These results show that only two CFAs (both Class C) failed the specification limit given in AASHTO M 295, which is substan- tially different from the results obtained when cement PC-1 was used in the experiments. Considering the various performance criteria for expan- sion at 14 days, the results are summarized as follows: • Fourteen of thirty CFAs (three Class F and eleven Class C) failed the current specification limit of 100% (relative to low-alkali cement PC-1). • Two Class C fly ashes (FA-ZC and FA-ZD) failed the cur- rent specification limit of 100% (relative to low-alkali cement PC-2). • None of the CFAs (either Class F or Class C) passed the old specification limit of 0.020% (absolute expansion) at a replacement level of 25%. • Fifteen of the CFAs (four Class F and eleven Class C fly ashes) would fail the assumed criteria of 60% reduction in expan- sion criterion (relative to high-alkali cement PC-3). Clearly, the apparent performance of a CFA changes depend- ing on the cement used in the test because of the inherent variability in cement chemistry as illustrated in Figure 4.11 (Roy, 2002). The figure illustrates that a 0.02% limit cannot be obtained with cement alkali contents greater than 0.4% expressed as Na2Oe. In addition, low-alkali cements (Na2O < 0.6%) tend to exhibit expansions ranging from 0% to 0.2% at 14 days. Therefore, the current specification limit differs depend- ing on cement selection. All 30 CFA samples reduced the expan- sion to below that obtained using ground quartz and, therefore, they appear to have potential for reducing ASR expansion. X-ray Diffraction, Thermal Analysis, and CFA Mineralogy Material characterization was performed using qualita- tive and quantitative x-ray diffraction, and thermal analysis methods to establish the mineralogy of the CFA sources. The results are presented in Attachment C. Summary of Precision Estimates for the Methods The precision of each test method was evaluated using a procedure described by Youden et al. (1951) and Taylor (1990). This method utilizes the difference between duplicate tests to estimate the pooled standard deviation of the analytical method; the results are presented in Attachment C. Discussion of Characterization Test Results The bulk chemistry values presented in this chapter indi- cated the CFA samples exhibited a wide range of compo- sition and physical properties. Eighteen of the CFAs were designated as Class F and twelve were designated as Class C

31 ash. In addition, the CFAs exhibited a good range of physi- cal properties. All of the CFA samples met the mandatory chemical requirements given in AASHTO M 295 (see Table 4.10); the majority of the CFAs met the mandatory physi- cal requirements; and some of the CFAs failed the optional requirements (see Table 4.11). The majority of the failures pertained to alkali content or the effectiveness of control- ling ASR. Some of the test requirements (e.g., the multiple factor) are not supported by the test results in Table 4.12). For example, the maximum LOI and fineness values observed during this study did not approach the specification limits. The requirements for a maximum fineness of 34% and a maximum LOI limit of 5% constrain the multiple factor to a maximum value of 170, which is substantially lower than the specification limit of 255 (the multiple factor is no longer part of AASHTO M 295). Current class limits for CFA are based on the sum of oxides (i.e., silicon, aluminum, and iron oxides). However, some researchers suggest this approach has shortcomings in two gen- eral areas related to performance: the classification does not include calcium oxide and is based on bulk chemistry, not crys- talline and glass phase content (Diamond, 1981; Manz, 1986; Mehta, 1986; Bumrongjaroen et al., 2011). The data in Table 4.12 for the 30 ash sources used in this study are sorted by the sum of oxides. Also, the results for CaO and CaO + MgO content are presumed as examples of other compositional parameters that could be used to classify CFA. However, these variables are related to each other, as shown in Figures 4.12 and 4.13. As shown in Figure 4.12, there is a clear linear relation- ship between the sum of oxides and CaO content and using either produces the same results as a criterion for classifying the CFAs. The addition of MgO as shown in Figure 4.13 incrementally improves the linear regression (based on the R2 value) as a result of adding an additional analyte to the regression analysis but does not change the classification of the ashes. The three Class C ashes falling slightly below the regression line (FA-B, FA-X, and FA-ZL) show the same trend as the other CFAs but are biased low because elements present in significant concentrations are not included in the regres- sion. For the 30 ashes analyzed, adding BaO and Na2O into the sum of oxides, and plotting against CaO, results in an R2 value of 0.99 with no change in CFA classification. The ultimate goal of a classification method is to group CFAs that have similar physical and chemical properties with- out excessive testing, and then to measure and report other properties that are known to affect performance within a spe- cific class. The existing AASHTO M 295 classification system is adequate at grouping similar materials but needs refinement with respect to reporting other properties. For example, particle characterization based on crystalline composition (Bumrongjaroen et al., 2011) provides a rigorous analysis of the fly ash microstructure, but the inherent variability within a given source makes characterization at the particle level impractical. Classification based on bulk properties (e.g., composition, fineness) has worked for many years, and correlations between bulk properties and performance have been developed by highway agencies. It is recommended to establish distinct limits for Class F and Class C ashes based on the sum of the oxides and to report the CaO content and the total alkali content. Fly Ashes Not Meeting Requirements Test Spec. Limit Class F Class C SiO2 + Al2O3 + Fe2O3 Only defines class SO3 5.0% max. None None Moisture Content 3.0% max. None None Loss on Ignition 5.0% max. None None Fineness 34% max. None None Strength Index 75% min. None None Water Requirement 105% max. None None Soundness 0.8% max. None None Table 4.10. Test results for mandatory AASHTO M 295 requirements. Fly Ashes Not Meeting Requirements Test Spec. Limit Class F Class C Available Alkali 1.5% max. M, P, S B, W, X, ZA,ZC, ZD, ZL Multiple Factor 255% max. None Not applicable Drying Shrinkage 0.03% max. Not tested Not tested Uniformity of Air- Entraining Dosage 20% max. Not tested Not tested Effectiveness in Controlling ASR 100% max. L, P, S All failed except B Sulfate Resistance 0.05% for high resistance Not tested Not tested Table 4.11. Test results for optional AASHTO M 295 requirements.

32 Moisture content and LOI tests are currently limited to a sample mass of 1 g. The small sample mass can affect the pre- cision of the determinations particularly for test results of low magnitudes, as is the case when analyzing CFA; this defi- ciency could be eliminated by simply increasing the sample mass. LOI test results generally are not significantly influ- enced by the mass of the CFA sample as seen in Figure 4.14 for a fixed ignition time of 45 min at 720°C, with no reheat cycle. These results are likely because the determinations were made using porcelain crucibles, each having a mass of approximately 10 to 12 g. The LOI test can be further improved by removing the reheat cycle required by ASTM Sum of Oxides CaO CaO + MgO Multiple ID (SiO2 + Al2O3 + Fe2O3) Class (% wt) (% wt) Factor FA-ZC 53.2 C 30.2 38.1 2.1 FA-ZD 54.2 C 29.2 36.9 2.2 FA-ZA 55.3 C 27.3 33.7 3.8 FA-Z 61.2 C 26.5 31.2 3.0 FA-ZL 61.5 C 18.0 22.5 3.1 FA-X 61.6 C 19.3 24.5 5.2 FA-ZB 61.7 C 25.7 30.5 2.0 FA-Y 62.8 C 24.4 29.9 3.4 FA-W 62.8 C 24.2 29.0 4.2 FA-V 62.9 C 24.8 29.4 8.5 FA-U 65.7 C 21.9 26.8 8.1 FA-B 67.7 C 14.3 17.8 15.7 FA-S 70.6 F 17.1 20.5 18.8 FA-R 73.2 F 15.3 20.6 1.8 FA-P 73.3 F 14.0 18.3 2.6 FA-Q 74.3 F 16.6 20.1 6.3 FA-T 77.4 F 13.6 16.6 10.6 FA-O 79.8 F 10.2 13.4 36.8 FA-M 81.9 F 7.17 9.8 6.2 FA-K 83.5 F 6.85 8.6 31.4 FA-L 84.8 F 6.79 7.6 12.3 FA-N 87.0 F 5.33 6.5 20.8 FA-E 87.8 F 2.78 4.0 35.9 FA-C 89.0 F 2.24 4.0 17.8 FA-I 89.5 F 1.81 3.1 34.6 FA-G 89.9 F 1.45 2.6 44.3 FA-F 90.7 F 1.09 1.9 46.8 FA-H 91.2 F 3.46 4.6 6.7 FA-J 91.9 F 1.28 2.3 28.3 FA-A 92.5 F 0.82 1.7 20.3 Table 4.12. Summary of CFA properties (calculated from test results). R2 = 0.96 0 5 10 15 20 25 30 35 50 60 70 80 90 100 Ca O (% ) Sum of Oxides (SiO2+Al2O3+Fe2O3) (% wt) Class C Class F Figure 4.12. Correlation between CaO and the sum of the oxides. R2 = 0.97 0 5 10 15 20 25 30 35 40 45 50 60 70 80 90 100 Ca O + M gO (% ) Sum of Oxides (SiO2+Al2O3+Fe2O3) (% wt) Class C Class F Figure 4.13. Correlation between CaO + MgO and the sum of the oxides.

33 C114, which is very time consuming when working with porcelain crucibles that cool much slower than platinum crucibles typically used with cement. Coal fly ash may con- tain a significant amount of carbon that can destroy plati- num crucibles. The thermal analysis experiments clearly indicated most CFA samples lost mass but some samples gained small amounts of mass unrelated to the carbon content of the sample. All the thermal gravimetric analysis curves indicated that the CFA samples started to rapidly lose mass at 500°C to 600°C and then remained relatively stable up to approximately 900°C. Additional discussion of the thermal analysis is presented in Appendix C. Sulfur (SO3) determination tests indicated that values obtained from the fused disk technique tended to be in very poor agreement with SO3 values determined using pressed pellets, particularly for Class F fly ash that had LOI greater than 1% and total SO3 values less than 1%. However, test results were within the specification limit but with large dif- ferences between labs. Because it is common practice to use a portion of the ignited sample (i.e., constant mass) obtained from the LOI test to manufacture fused disks for XRF analy- sis, analysts need to be aware of this potential error. Further tests indicated the majority of the sulfur was lost during the LOI test, as shown in Figure 4.15. CFA Pozzolanic Reactivity SAI Tests Conventional SAI and PAI tests were performed on all 30 CFA sources and also using an inert filler. These tests indicated an inert material could meet SAI test requirements. To explore improvements to the SAI and PAI tests, a series of modified tests were conducted. These included an SAI test performed with 20% and 35% by weight replacement of cement with the quartz filler (INF-1). In addition to the higher substitu- tion rates, all specimens were prepared using a constant w/cm rather than a constant flow. Because fly ash particles generally increase the flow of a mortar due to their spherical shape, a reduction in water content was necessary to maintain a constant flow, which potentially enables a non-pozzolanic fly ash to meet the SAI requirements in AASHTO M 295. Strengths and SAI values using the inert filler are shown in Table 4.13. Table 4.13 shows that mortars containing 20% replace- ment of cement with non-pozzolanic ground quartz filler met or exceeded the AASHTO M 295 SAI requirement (i.e., 75%) at both 7 and 28 days, even at a constant w/cm ratio. This result is likely due to the effect of a filler on nucleation and acceleration of cement hydration. This ground quartz would meet the SAI requirements for a Class F fly ash despite not having any pozzolanic value. However, at a replacement level of 35%, this non-pozzolanic filler did not meet the 75% limit at either 7 or 28 days. For evaluating the SAI and KHI tests, the effect of different quartz fillers was investigated. The compressive strength of ASTM C109 mortar cubes made with quartz filler INF-1 were compared to that obtained using INF-2 quartz filler; results are shown in Figure 4.16. At a 20% replacement of cement, there was a negligible difference in strength between the mor- tars prepared with either cement type or inert filler, with the exception of a lower 28-day strength for the PC-2 and INF-1 combination. At a 35% replacement of cement, the effect of filler type was more pronounced when PC-2 was compared with PC-3, with INF-1 mortars having higher strengths than comparable INF-2 mortars. The modified SAIs for eight CFAs at 20% replacement are presented in Table 4.14. The SAIs for four Class C CFAs at 35% replacement are presented in Table 4.15. 0 2 4 6 8 10 0 1 2 3 4 5 LO I (% ) Sample Mass (g) CCRL 29 CCRL 40 CCRL 37 CCRL 43 Figure 4.14. Influence of sample mass on %LOI. R2 = 0.998 (before LOI data only) 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 0 10 20 30 40 50 60 SO 3 (% ) Net Sulfur X-ray Intensity Measured (kcps) Before LOI After LOI Figure 4.15. Sulfur loss during the LOI test.

34 ment. However, the KHI values range from approximately 0 (for inert materials) to over 100% (equal performance to the cement). Figures 4.21 and 4.22 show the SAI and KHI values versus compressive strength for combinations of eight CFA sources and three cements. These figures show that both methods provide a similar trend for strength development. Also both methods indicate a clear effect of cement type with both PC-2 and PC-3 yielding higher values of SAI and KHI than PC-1 for the same CFA. Effect of Carbon on Air Entrainment To evaluate the application of the developed tests for assess- ing the effects of CFA on air entrainment, tests were performed and correlated with known LOI values for the various CFAs, The tests conducted to evaluate whether the SAI and PAI indicate the minimum strength level, CFA replacement level, or both need to be modified so that inert materials will not meet the specification requirements. KHI Tests Results of the KHI tests are shown in Table 4.16 for the three cements and 20% replacement by the eight CFAs, and 35% replacement by the four Class C CFAs, at 7, 28, and 56 days. As indicated by the negative values, fly ashes FA-H and FA-O had 7-day strengths lower than obtained with the INF-1 filler. These results indicate the pozzolanic reaction of FA-H and FA-O did not lead to strength gain at 7 days but were likely acting as a filler and providing nucleation sites for hydration of cement. This demonstrates the KHI test’s ability to separate filler effects from pozzolanic contributions. The SAI versus KHI values at 7 days of age are presented in Figures 4.17 and 4.18 for 20% and 35% CFA replacement of cements, respectively; the values at 56 days of age are presented in Figures 4.19 and 4.20, respectively. As can be seen, the two tests appear to provide similar prediction of strength develop- 1 2 3 4 5 6 0 28 56 Co m pr es si ve S tre ng th (k ps i) Age (days) PC-2, Control PC-3, Control PC-2, 20% - INF-1 PC-3, 20% - INF-1 PC-2, 35% - INF-1 PC-3, 35% - INF-1 PC-2, 20% - INF-2 PC-3, 35% - INF-2 Figure 4.16. Cube strengths for different fillers, cements, and replacement levels. 7-Day SAI 28-Day SAI 56-Day SAI Cement Cement Cement ID PC-1 PC-2 PC-3 PC-1 PC-2 PC-3 PC-1 PC-2 PC-3 FA-H 79 80 81 95 98 95 92 108 98 FA-M 85 85 90 103 89 90 93 106 90 FA-O 85 78 83 85 94 91 91 98 86 FA-Q 92 88 90 106 100 101 104 111 104 FA-U 103 88 98 113 94 111 107 108 95 FA-X 102 100 97 99 84 96 106 103 99 FA-ZA 103 89 101 113 100 108 110 104 88 FA-ZC 117 96 92 106 104 105 108 112 101 Table 4.14. Strength activity indices of coal fly ashes at 20% replacement. 7-Day SAI 28-Day SAI 56-Day SAI Cement Cement Cement ID PC-1 PC-2 PC-3 PC-1 PC-2 PC-3 PC-1 PC-2 PC-3 FA-U 87 70 81 101 81 97 107 101 109 FA-X 96 90 103 106 89 98 92 105 94 FA-ZA 93 75 89 107 82 105 106 101 100 FA-ZC 113 79 82 95 92 94 101 100 99 Table 4.15. Strength activity indices of Class C coal fly ashes at 35% replacement. Cement Type Age (Days) 100% Cement 20% Replacement 35% Replacement Strength (psi) Strength (psi) SAI Strength (psi) SAI PC-1 7 4,554 3,829 84 3,075 68 PC-2 7 4,293 3,408 79 2,640 62 PC-3 7 4,090 3,539 87 2,886 71 PC-1 28 5,715 4,815 84 3,945 69 PC-2 28 5,526 4,235 77 3,655 66 PC-3 28 5,134 4,351 85 3,307 64 Table 4.13. Strengths and SAI for the INF-1 filler at 20% and 35% cement replacement.

35 7-day KHI (%) 28-day KHI (%) 56-day KHI (%) ID PC-1 PC-2 PC-3 PC-1 PC-2 PC-3 PC-1 PC-2 PC-3 20% Replacement FA-H 31 4 43 71 91 66 60 162 88 FA-M 7 28 26 119 55 34 66 143 50 FA-O 10 6 24 7 73 39 57 84 26 FA-Q 53 44 26 135 102 109 120 185 121 FA-U 121 40 84 184 75 171 133 158 73 FA-X 115 101 80 96 30 72 127 126 96 FA-ZA 122 46 110 184 99 153 150 132 38 FA-ZC 203 83 41 138 119 130 140 193 106 35% Replacement FA-U 60 21 35 102 44 93 121 102 126 FA-X 89 74 110 118 68 94 78 114 82 FA-ZA 80 35 63 124 46 114 116 102 101 FA-ZC 140 45 39 83 75 82 102 99 96 Table 4.16. Keil hydraulic indices for 20% and 35% CFA replacement of cements and using filler INF-1. 60% 80% 100% 120% -50% 0% 50% 100% 150% 200% St re ng th A ct iv ity In de x Keil Hydraulic Index PC-1 PC-2 PC-3 Figure 4.17. SAI versus KHI values at 7 days for a 20% replacement level. 60% 80% 100% 120% -50% 0% 50% 100% 150% 200% St re ng th A ct iv ity In de x Keil Hydraulic Index PC-1 PC-2 PC-3 Figure 4.18. SAI versus KHI values at 7 days for a 35% replacement level. 60% 80% 100% 120% -50% 0% 50% 100% 150% 200% St re ng th A ct iv ity In de x Keil Hydraulic Index PC-1 PC-2 PC-3 Figure 4.19. SAI versus KHI values at 56 days for a 20% replacement level. 60% 80% 100% 120% -50% 0% 50% 100% 150% 200% St re ng th A ct iv ity In de x Keil Hydraulic Index PC-1 PC-2 PC-3 Figure 4.20. SAI versus KHI values at 56 days for a 35% replacement level.

36 and also correlated with each other. Also, a series of mortar and concrete mixtures were prepared to demonstrate the application of the various tests. Key results are summarized here; additional results are presented in Attachment C. Correlation among the Tests Foam Drainage Test Test results indicated that in most cases, the foam drainage test does not adequately characterize the interaction of AEAs with CFAs having significantly different levels of LOI. The high ratio of AEA to water used in the test may explain the similarity in results obtained with different CFAs. Published results for the foam drainage test use cementitious combi- nations (i.e., CFA plus cement). The test was ineffective in evaluating the impact of fly ash only. Foam Index Test An example of the correlation between the measured foam index number and LOI is shown in Figure 4.23. Relationships for the six principal AEAs tested and the numeric results are summarized in Attachment C. The relationship between LOI and the foam index test is reasonable but the variability in the results is high compared to the results from either the CFA iodine number or the direct adsorption isotherm tests. One source of this variability is the subjective nature of the test. Also, for some fly ash sources there is anomalous behavior with respect to correlation between the foam index test and the LOI results. Another source of vari- ation is that LOI measures the mass of carbon and does not account for the adsorption potential of the carbon. The foam index test is fundamentally different from the iodine and direct isotherm tests. For the latter two tests, a CFA is equilibrated with a high concentration of adsorbate (relative to the foam index test). However, in the foam index test, the sys- tem starts with no adsorbate and then AEA (i.e., adsorbate) is introduced in an incremental manner. The AEA concentration increases from zero until a stable foam is formed but equilib- rium conditions may never be achieved. The foam index test is dynamic and if the system was left to equilibrate, the foam might likely disappear because adsorption continues to take place. In comparison, isotherms are based on equilibrium conditions after which no significant change in the concentration occurs. CFA Iodine Number Test CFA iodine number tests were performed on 14 CFA sources and the results are summarized in Table 4.17. The CFA iodine number versus LOI is shown in Figure 4.24, which presents a linear relationship. Figure 4.25 presents the Figure 4.21. SAI versus compressive strength values for a 20% replacement level. Figure 4.22. KHI versus compressive strength values for a 20% replacement level. R2 = 0.760 0.00 0.01 0.02 0.03 0.04 0.05 0 2 4 6 8 10 12 A bs ol ut e Vo lu m e of A EA -1 (m L) LOI (% wt) Figure 4.23. Absolute volume of AEA added (mL) versus measured LOI.

37 iodine number versus LOI on a semi-log plot for the same CFA sources. This figure highlights the ability of the test to detect differences in adsorption for relatively low LOI ashes. For exam- ple, FA-T (LOI = 0.45%) has a higher LOI than FA-H (LOI = 0.25%), yet FA-H exhibited a higher CFA iodine number (i.e., adsorption); a similar observation can be made for FA-J, FA-O, and FA-G. This observation can be attributed to the nature of the LOI test because many factors, such as decomposition of carbonate minerals (e.g., CaCO3) and portlandite [Ca(OH)2] and combustion of carbon affect the mass loss of fly ash as a result of burning. Also, a gain in mass may occur due to the oxi- dation of sulfur and iron. CFA sources with the same LOI may have different adsorption properties depending on the form of available carbon. For high LOIs, the majority of the mass loss is due to carbon volatilization and therefore errors with the LOI test tend to be high at lower LOI values. *FA-ZF/FA-ZE blends ID LOI (%) Iodine No. (mg/g CFA) FA-H 0.25 0.004 FA-T 0.45 0.001 FA-A 0.94 0.013 FA-J 1.59 0.545 FA-O 1.43 0.535 FA-G 2.32 0.354 FA-ZN 3.41 2.619 FA-ZF 6.06 3.761 25-75 Blend* 10.37 10.989 FA-ZM 10.69 7.266 50-50 Blend* 14.68 16.857 75-25 Blend* 18.99 25.205 FA-ZJ 21.34 35.603 FA-ZE 23.30 30.583 Table 4.17. Results of CFA iodine number test. R2 = 0.947 0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 Io di ne N um be r ( mg io din e/g C FA ) LOI (% wt) Figure 4.24. CFA iodine number versus LOI. Direct Adsorption Isotherm Test The final evaluation of the direct adsorption isotherms was conducted on six AEAs (AEA-1 through AEA-6 in Table 3.2) and eight CFAs listed in Table 3.1. Adsorption isotherms for eight CFAs and AEA-1 are pre- sented in Figure 4.26 and in Attachment C for the other five AEAs. These isotherms quantify the amount of AEA adsorbed by the CFA as a function of the AEA concentration. The AEA solution concentration in the cement-only mixture is determined by dividing the volume of AEA in the mixture by the volume of water used in the concrete mixture. The CFA adsorption capacity at this concentration can then be determined from the isotherm graph. For example, the adsorption capacity of FA-T (LOI = 0.45%) for AEA-1 at an initial concentration (C0) of 0.3% volume of AEA is 0.0023 mL of AEA per gram of 0.0001 0.0010 0.0100 0.1000 1.0000 10.0000 100.0000 0 5 10 15 20 25 lo g I od in e Nu m be r (m g i od ine /g CF A ) FA-H FA-T LOI (% wt) Figure 4.25. CFA iodine number versus LOI (on a semi-log scale).

38 CFA (Figure 4.26). This capacity multiplied by the total mass of CFA in the mixture determines the volume of AEA-1 adsorbed by FA-T for the specified initial concentration. The capacity is the first estimate of the correction to the initial AEA dosage (i.e., cement-only mixture) to compensate for the adsorption of AEA-1 by FA-T. As shown in Figure 4.26, the capacity can be determined graphically or by fitting a power line to the isotherm points and determining the Freundlich equation coefficients. LOI Correlation to Adsorption Tests Table 4.18 and Figure 4.27 illustrate the relationship between LOI and the CFA iodine number, foam index, and CFA capacity measured by direct adsorption isotherms. The foam index and adsorption isotherm tests in Figure 4.27 are performed with AEA-1 at a concentration of 0.4% vol. Similar relationships for AEA-1 through AEA-6 are provided in Attachment C. Figure 4.27 shows the correlation of all adsorption-based tests to LOI for AEA-1. The foam index test shows the most deviation and scatter but generally correlates with the other results. However, the LOI does not provide a good correlation to adsorption capacity for many of the CFAs with a moderate level of LOI (FA-O LOI = 1.43% and FA-G LOI = 2.32%). All three adsorption-based tests show FA-G has less adsorption capacity than FA-O suggesting the LOI value does not reflect the true adsorption potential of the CFA. The same is true for FA-H (LOI = 0.25%) and FA-T (LOI = 0.45%) where FA-H has a higher capacity than FA-T as shown by the three adsorption- based tests. These observations agree with an earlier finding that, depending on the composition of the CFA, using LOI as a measure of carbon content could be as accurate as 99% or have as much as a 75% error between carbon content and LOI (Brown and Dykstra, 1995). Based on these results, it appears that LOI is not a reliable test for predicting AEA adsorption, particularly for lower-carbon-content CFAs because the non- carbon-related mass change can be a significant fraction of the reported LOI while not affecting the adsorption behavior of the CFA. However, LOI provides a good indication of carbon con- tent, and therefore adsorption capacity, for high-carbon CFAs, because the loss in mass due to burning carbon is substantially higher than the mass loss/gain by other mechanisms. The iodine number test assesses the adsorption properties of a CFA using iodine as an adsorbate; it does not measure AEA adsorption but classifies the CFA in terms of its performance as an adsorbent. The difference in adsorption capacity for a given CFA is dependent upon the AEA. The direct adsorption iso- therm test measures the adsorption of AEA for a specific AEA and CFA combination, which can be expressed as a function of the CFA iodine number. The foam index test measures a physi- cal property of the AEA-water solution that is affected by the adsorption of AEA from that solution, which can be expressed in terms of either the CFA iodine number or the direct adsorp- tion isotherm. Any one of these tests can predict the effect of CFA addition on air entrainment in concrete by measuring 0.001 0.01 0.1 0.01 0.1 AEA Concentration in Water (% vol.) 1 Fl y As h Ca pa ci ty (m L A EA /g C FA ) 0.25 0.45 0.94 1.43 1.59 2.32 3.41 10.69 LOI 0.3 0.0023 Capacity FA-T = 0.0068 C00.9081 Figure 4.26. Adsorption isotherms of AEA-1 with eight coal fly ashes. ID LOI (%) Foam Index* (mL) Capacity* (mL/g CFA) CFA Iodine No. (mg I/g CFA) FA-H 0.25 0.01 0.004125 0.00397 FA-T 0.45 0.0076 0.002959 0.00076 FA-A 0.94 0.0154 0.004737 0.01279 FA-O 1.43 0.0292 0.009677 0.53464 FA-J 1.59 0.011 0.006458 0.54499 FA-G 2.32 0.0123 0.004244 0.35415 FA-ZN 3.41 0.0285 0.016040 2.61915 FA-ZM 10.69 0.0468 0.054647 7.26624 * Determined for AEA-1 Table 4.18. LOI, foam index, CFA capacity, and CFA iodine number. Capacity Foam Index Iodine Number 0 2 4 6 8 0 0.01 0.02 0.03 0.04 0.05 0.06 0 2 4 6 8 10 12 Io di ne N um be r ( mg io din e/g C FA ) Ca pa ci ty (m L/g ) a nd Fo am In de x ( mL ) LOI (% wt) FA-G FA-O Figure 4.27. LOI correlations to CFA iodine number, foam index, and CFA capacity for AEA-1.

39 adsorption properties—not the physical property of LOI that may or may not correlate to adsorption capacity. Foam Index Test Correlation to Adsorption Tests The foam index test, in its improved form where the total time of the test is consistent, provides a better representation of the adsorption capacity of CFA than LOI. Figure 4.28 pres- ents a comparison between the foam index test results, direct adsorption measurements for AEA-1, and the CFA iodine number for eight CFAs. The results, although more variable, show the foam index test provides a clear trend following the CFA iodine number and correlating to adsorption capac- ity. The CFA iodine number test results correlate better with adsorption capacity measurements than the foam index test results because of the low resolution, non-equilibrium condi- tions, and the subjective nature of the foam index test. CFA Iodine Number Correlation to Direct Adsorption Isotherms The CFA iodine number is a measure of the capacity of CFA for iodine. Iodine is a single solute and has a much smaller molecular size than a typical AEA and is easily adsorbed onto carbon. Figure 4.29 illustrates the relationship between direct adsorption isotherms and CFA iodine number for AEA-5 and eight CFA sources. Figure 4.30 shows the CFA iodine number and AEA adsorption capacity for several AEAs. The figure shows that the relative adsorption affinity of each AEA can be established with the lowest capacity at a given iodine number being the AEA with the least tendency to adsorb. The figure also shows, for iodine numbers below 0.1 mg/g, the AEA adsorption capacity is low and does not vary signifi- cantly over this CFA iodine number range. For CFA iodine number values between 0.1 and 1, the AEA capacity begins to rapidly increase in magnitude. CFA iodine numbers greater than 1 indicate a very significant level of adsorption and the adsorption capacity increases with small changes in the CFA iodine number. This relationship can be used in modifications of CFA specifications. The relationships shown in Figures 4.29 and 4.30 illustrate a simple means for determining the capacity of a CFA based on the iodine number. By selecting a suite of CFAs with vary- ing adsorption potential and determining the CFA capacity based on the direct adsorption isotherm test for a given AEA, as well as determining the iodine number for those CFAs, a plot similar to Figure 4.29 would be generated. The user would then determine adsorption capacity for the AEA corresponding to the iodine number for the CFA. For example, if AEA-5 (as 0.000 0.010 0.020 0.030 0.040 0.050 0 2 4 6 8 0 0.01 0.02 0.03 0.04 0.05 0.06 Fo am In de x (m L A EA -1) Io di ne N um be r ( mg io din e/g C FA ) Capability of AEA-1 (mL/g CFA) Capacity vs Iodine Number Capacity vs Foam Index Figure 4.28. Relationship between CFA iodine number and foam index test versus capacity of AEA-1. y = 2.4083ln(x) + 14.971 0.0001 0.0010 0.0100 0.1000 1.0000 10.0000 0 0.01 0.02 0.03 0.04 0.05 Io di ne N um be r ( mg io din e/g C FA ) Capacity (mL AEA/g CFA) Figure 4.29. Relationship between CFA iodine number and direct adsorption isotherm capacity for AEA-5. 0.0001 0.0010 0.0100 0.1000 1.0000 10.0000 0 0.02 0.04 0.06 0.08 0.1 Io di ne N um be r ( mg io din e/g C FA ) AEA-1 AEA-2 AEA-3 AEA-4 AEA-5 AEA-6 Capacity (mL AEA/g CFA) Figure 4.30. Relationship between CFA iodine number and direct adsorption isotherm capacity for six AEAs and eight CFA sources.

40 shown in Figure 4.29) is used with a CFA that has an iodine number of 3, the capacity of that CFA for AEA-5 would be approximately 0.01 mL/g of CFA. Correlation with Mortar and Concrete Mortar Experiments Application of the direct adsorption isotherm test for pre- dicting the adjustment to AEA dosage was investigated using cement-only control mixtures and separate batches of cement and fly ash mixtures. The AEA dosage was determined by (1) trial and error and (2) using the direct adsorption isotherm; the results are shown in Figures 4.31 through 4.36. The air contents of cement-only control mixtures are shown as “Base- line Air Content.” The error bars indicate the range in air con- tent obtained with 16 repetitions (i.e., 17 total batches) of each mixture combination. The variability in air content obtained in the control mixtures was typically ±1.5% to 2% total air content. The AEA dosages determined using the trial-and-error approach are listed in Table 4.19 and the dosages determined using the direct adsorption isotherm test are presented in Table 4.20. Table 4.21 shows the change in the AEA dosages from the baseline dosages. Test results show that if the mixtures made using AEA-5 are disregarded, in 73% of the cases, the air content obtained by the two approaches agree within ±1.5% to 2% total air content of the baseline air content. FA-ZM, an extremely high LOI material (LOI = 10.69% wt), was included to inves- tigate the use of the direct adsorption isotherm test for these materials. In spite of the high adsorption capacity of this ash, predictions for two of the five AEAs (excluding AEA-5) were within the margin of error and one estimate was slightly out- side this range. 0 1 2 3 4 5 6 7 8 FA-A FA-G FA-H FA-J FA-O FA-T FA-ZM FA-ZN A ir Co nt en t (% v o l.) Ash Source Trial & Error Isotherm Prediction Baseline Air Content 5.2% % LOI 0.94 2.32 0.25 1.59 1.43 0.45 10.69 3.41 Figure 4.32. Air contents for mortar mixtures with AEA-2. 0 1 2 3 4 5 6 7 8 9 FA-A FA-G FA-H FA-J FA-O FA-T FA-ZM FA-ZN A ir Co nt en t (% v o l.) Ash Source Trial & Error Isotherm Prediction Baseline Air Content 6.1% % LOI 0.94 2.32 0.25 1.59 1.43 0.45 10.69 3.41 Figure 4.33. Air contents for mortar mixtures with AEA-3. 0 2 4 6 8 10 12 FA-A FA-G FA-H FA-J FA-O FA-T FA-ZM FA-ZN A ir Co nt en t (% v o l.) Ash Source Trial & Error Isotherm Prediction Baseline Air Content 7.6% % LOI 0.94 2.32 0.25 1.59 1.43 0.45 10.69 3.41 Figure 4.34. Air contents for mortar mixtures with AEA-4. 0 2 4 6 8 10 12 FA-A FA-G FA-H FA-J FA-O FA-T FA-ZM FA-ZN A ir Co nt en t (% v o l.) Ash Source Trial & Error Isotherm Prediction Baseline Air Content 8.7% % LOI 0.94 2.32 0.25 1.59 1.43 0.45 10.69 3.41 Figure 4.31. Air contents for mortar mixtures with AEA-1.

41 0 1 2 3 4 5 6 7 8 9 FA-A FA-G FA-H FA-J FA-O FA-T FA-ZM FA-ZN A ir Co nt en t (% v o l.) Ash Source Trial & Error Isotherm Prediction Baseline Air Content 6.6% % LOI 0.94 2.32 0.25 1.59 1.43 0.45 10.69 3.41 Figure 4.35. Air contents for mortar mixtures with AEA-5. 0 2 4 6 8 10 12 FA-A FA-G FA-H FA-J FA-O FA-T FA-ZM FA-ZN A ir Co nt en t (% v o l.) Ash Source Trial & Error Isotherm Prediction Baseline Air Content 7.4% % LOI 0.94 2.32 0.25 1.59 1.43 0.45 10.69 3.41 Figure 4.36. Air contents for mortar mixtures with AEA-6. Figure 4.35 shows the direct adsorption isotherm test did not accurately predict the adsorption of AEA-5 with any of the CFAs; the required dosage to achieve the target air content was twice the manufacturer’s recommended maximum for cement-only control mixtures (Table 3.5). Direct adsorption isotherms were performed for AEA-9 (a benzene sulfonate). A series of control mortars were prepared using AEA-9 and PC-1 to determine a baseline dosage (i.e., 1.15 oz/cwt and 7.4% vol. air). The trial-and-error approach and the direct adsorption isotherm test were used to determine AEA-9 dos- ages for CFA mortar mixtures with FA-J and FA-O; the results are summarized in Table 4.22. The values shown for the air content obtained by trial and error were measured while the values shown for the air content obtained using the direct adsorption isotherm were predicted from the AEA dosage versus air content relationship established by trial and error. The variability of these control mixtures was not estab- lished but is assumed to be similar to the other control mor- tar mixtures (i.e., ±1.5% to 2% vol. air). With this variability, the predicted air content for the FA-J mixture was near the limits of variation and the predicted air content for the FA-O mixture was almost identical to that determined by trial and error for mortar mixtures. AEA Dosage (oz AEA/cwt cementitious) Baseline Dosage 2.12 0.81 0.61 1.77 1.84 1.54 ID LOI AEA-1 AEA-2 AEA-3 AEA-4 AEA-5 AEA-6 FA-A 0.94 4.24 2.24 1.84 5.30 5.07 1.54 FA-G 2.32 4.24 2.04 1.07 2.65 3.96 3.07 FA-H 0.25 3.18 1.63 1.23 6.18 4.61 3.07 FA-J 1.59 3.44 2.24 2.00 3.53 5.07 1.92 FA-O 1.43 4.13 2.64 3.68 5.47 5.99 3.07 FA-T 0.45 1.80 0.94 0.61 1.32 2.30 0.77 FA-ZM 10.69 20.13 15.87 62.02 107.69 37.30 13.05 FA-ZN 2.41 7.94 6.92 5.83 9.71 11.97 5.37 Table 4.19. AEA dosages based on the trial-and-error approach. The data presented in Table 4.21 show that changes in dos- age for FA-T (a borderline Class C/Class F ash with low LOI) mixtures between the baseline dosages and those determined empirically were quite small or negative, which is common with Class C ashes. However, neither adsorption-based tests nor existing LOI testing predicted a decrease in AEA dosage. CFA sources FA-A, FA-G, and FA-H represent benefici- ated ashes: FA-A is a triboelectrostatic processed ash, FA-G is a carbon burn-out ash, and FA-H an air classified ash. The air content for mortars prepared using the change in dosage predicted by the direct adsorption isotherm fell within the margin of error in 93% (i.e., 14/15) of the cases (excluding AEA-5). These ashes represent the type of fly ash that may become more common in the future. Although FA-A and FA-G are processed to remove car- bon, there was still a significant increase in AEA demand for these CFA sources, likely because removal of carbon from CFA may be based on lowering the measured LOI and not the bulk adsorption potential. The processes used likely will remove larger carbon particles first, thereby lowering the LOI of the ash. However, the smaller carbon particles con- tribute most to AEA adsorption due to their high specific surface and these carbon particles can impact AEA demand

42 significantly. As processing is adapted to remove carbon from CFA, it will be increasingly important to measure the efficacy of that process with respect to its impact on AEA adsorption, not LOI. The CFA iodine number and direct adsorption isotherm tests are new approaches to quantify AEA adsorption by CFA. Concrete Experiments The purpose of the concrete experiment was to provide additional evaluation of the direct adsorption isotherm test. Control mixtures were prepared with an air content of 6.5% ± 1.5% (AEA dosage was determined by trial and error); the results are listed in Table 4.23. For the test mixtures with 25% fly ash, the AEA dosage used was based on the results for mortars that used the same combi- nation of AEA and CFA. For example, if a particular CFA-AEA combination required a 50% larger AEA dosage compared to that required for a mortar mixture, then a 50% increase was applied to the AEA dosage for the concrete mixtures of the same combination. The estimated dosages are listed in Table 4.23 for the control mixtures and those with 25% CFA. As seen in Table 4.23, the trial-and-error results for mor- tar were successfully used to predict AEA dosages for concrete, except when AEA-1 is used with mixtures containing FA-ZM (LOI = 10.69%) and FA-O (LOI = 1.43%). The differences with FA-ZM were not unexpected because of the high LOI content and AEA-1 was shown to be one of the most adsorbable AEAs. Selected concrete mixtures were prepared using the direct adsorption isotherm test results to correct the baseline AEA dosage determined for concrete. As was done with the mor- tar mixtures, the adsorption capacity determined by the direct AEA Dosage (oz AEA/cwt cementitious) Baseline Dosage 2.12 0.81 0.61 1.77 1.84 1.54 ID LOI AEA-1 AEA-2 AEA-3 AEA-4 AEA-5 AEA-6 FA-A 0.94 3.66 1.55 1.18 2.98 2.39 2.26 FA-G 2.32 3.49 1.69 0.88 2.80 2.26 2.19 FA-H 0.25 3.51 1.38 1.08 3.11 2.13 2.12 FA-J 1.59 4.30 1.69 1.09 3.32 2.55 2.28 FA-O 1.43 5.53 2.54 1.82 4.96 2.69 2.70 FA-T 0.45 2.97 1.08 1.03 2.42 2.24 1.86 FA-ZM 10.69 21.76 11.12 21.26 16.98 8.06 6.62 FA-ZN 2.41 7.92 4.49 4.45 7.52 4.47 3.55 Table 4.20. AEA dosages based on the direct adsorption isotherm test. Percentage Change in AEA Dosage ID LOI AEA-1 AEA-2 AEA-3 AEA-4 AEA-5 AEA-6 FA-A 0.94 100 176 200 200 175 0 FA-G 2.32 100 151 75 50 115 100 FA-H 0.25 50 100 100 250 150 100 FA-J 1.59 63 176 225 100 175 25 FA-O 1.43 95 225 500 210 225 100 FA-T 0.45 15 15 0 25 25 50 FA-ZM 10.69 850 1,851 10,000 6,000 1,925 750 FA-ZN 2.41 275 751 850 450 550 250 Table 4.21. The relative change in AEA dosage from the baseline dosages. CFA ID AEA Dosage (oz/cwt) Air Content (% vol.) Baseline Trial & Error Direct Adsorption Isotherm Baseline Trial & Error Direct Adsorption Isotherm (Predicted) FA-J 1.15 1.74 2.59 7.40 7.44 8.96 FA-O 1.15 2.93 3.05 7.40 7.45 7.49 Table 4.22. AEA dosages and air contents for mortar mixtures with AEA-9.

43 adsorption isotherm test and the known weight of CFA in the concrete mixtures was used to determine an adjusted AEA dos- age. The results, shown in Table 4.24, indicate that the direct adsorption isotherm estimation produced an air content within the desired range in nine of twelve cases (i.e., 75%). There are many factors that affect air entrainment in con- crete (e.g., mixing time, temperature, aggregate type), and there is significant variance in the ability to measure the air content in fresh concrete. The direct adsorption isotherm test only addresses the issue of AEA adsorption; in some cases, other factors will dominate. Nevertheless, the results of the mortar and concrete experiments indicate the test could be used to predict AEA adsorption and provide an estimate of the associated adjustment in AEA dosage. Influence on Air-Void System Parameters Mortars containing CFA required higher dosages of AEA to yield total air contents approximating those of the control mor- tars. An investigation was performed to identify any alteration of the air-void system that may be associated with an increased AEA dosage. Hardened mortars made with FA-G, FA-H, FA-J, and their respective control mortars were analyzed using an auto- mated air-void analysis system. Results of this analysis are pre- sented in Attachment C. Example data are shown in Table 4.25. Figure 4.37 presents the cumulative chord length distributions for control mortars and their respective test mortars. These results indicate slightly different air-void systems, but it is not possible to establish if this variability is statistically relevant because of the small sample size used for each condition. Prediction of ASR Mitigation Pyrex Mortar Bar Test (ASTM C441) Fly ash characterization was performed using the Pyrex mortar bar expansion test (ASTM C441) currently specified in AASHTO M 295. Expansions were measured at 14 and 28 days for 25% by mass fly ash replacement of both low-alkali cement PC-1 and high-alkali cement PC-3. In Table 4.26, these expansions are expressed as (1) a percentage of expansion of low-alkali portland cement, PC-1 mortar bars (max. 100% is allowed as per AASHTO M 295), and (2) as a percentage reduc- tion in expansion relative to high-alkali cement PC-3 mortar bars (typically reductions should be greater than 60%). Based on these results, only fly ashes FA-H, FA-M, FA-O, and FA-Q would be considered effective in mitigating ASR expansions when used at 25% replacement. In terms of effec- tiveness in reducing ASR expansions, based on these data alone, AEA Cementitious AEA Dosage (oz/cwt) Total Air Content (% vol.) AEA-1 PC-1 Control 1.9 6.9 PC-1 + 25% FA-A 3.8 6.9 PC-1 + 25% FA-G 3.8 8.1 PC-1 + 25% FA-O 3.7 3.9 PC-1 + 25% FA-ZM 18.2 3.4 PC-1 + 25% FA-ZN 7.2 5.4 AEA-2 PC-1 Control 1.7 5.8 PC-1 + 25% FA-A 4.2 5.6 PC-1 + 25% FA-G 4.6 6.5 PC-1 + 25% FA-O 5.5 6.3 PC-1 + 25% FA-ZM 32.9 5.4 PC-1 + 25% FA-ZN 14.3 5.9 AEA-6 PC-1 Control 2.1 7.1 PC-1 + 25% FA-A 4.3 5.5 PC-1 + 25% FA-G 2.1 7.6 PC-1 + 25% FA-O 3.1 6.8 PC-1 + 25% FA-ZM 18.3 5.5 PC-1 + 25% FA-ZN 7.5 6.5 Table 4.23. AEA dosages and air contents for concrete test mixtures. AEA Cementitious AEA Dosage (oz/cwt) Total Air Content (% vol.) AEA-1 PC-1 Control 1.9 6.9 PC-1 + 25% FA-A 3.5 6.4 PC-1 + 25% FA-O 5.3 8.5 PC-1 + 25% FA-ZN 7.7 7.1 AEA-2 PC-1 Control 1.7 5.8 PC-1 + 25% FA-A 3.0 5.3 PC-1 + 25% FA-O 4.2 4.9 PC-1 + 25% FA-ZN 6.4 4.0 AEA-6 PC-1 Control 2.1 7.1 PC-1 + 25% FA-A 3.2 6.9 PC-1 + 25% FA-O 3.5 6.6 PC-1 + 25% FA-ZN 4.5 5.4 Table 4.24. AEA dosages and air contents for baseline concrete mixtures. Material Average Chord Length (mm) Specific Surface (mm 1) Void Frequency (voids/mm) Air Content (%) Total Intercepts Total Chord Length (mm) Control 0.17 23.95 0.69 11.5 3,356 561.3 FA-G 0.16 25.48 0.79 12.4 3,778 591.3 Table 4.25. Air-void parameters for control and mixtures with 25% replacement with FA-G and AEA-1.

44 FA-O would be the most effective, followed by FA-M, FA-Q, FA-H, FA-X, FA-ZA, FA-U, and FA-ZC. Concrete Prism Tests With the exception of long-term outdoor exposure sites or known performance history with the same materials combi- nation, the ASTM C1293 concrete prism test is considered to be the most reliable indicator of the effectiveness of a given level of CFA for mitigating deleterious alkali-silica reactiv- ity. The test recommends that expansions be monitored for 2 years. The concrete prism test serves as a good reference for evaluating the effectiveness of both the current and proposed accelerated test methods. Concrete prism expansions after 12, 18, and 24 months at 38°C are given in Tables 4.27 through 4.29. In Figure 4.38, the 24-month expansions are plotted for concretes made with alkali-reactive Sudbury aggregate using PC-3 and each of the eight CFAs at a 20% and 30% replacement of cement, and for the four Class C CFAs at a 40% replacement of cement. The 12-month results indicated the four Class F CFAs (i.e., FA-H, FA-M, FA-O, and FA-Q) appeared to be effective when used at 20% replacement (i.e., the expansions were less than 0.04% at 12 months), and FA-U appeared effective when used at 30% replacement. For FA-X, a 40% replacement was required to meet the 12-month requirement, but the expansion exceeded 0.04% after 2 years. For the two Class C CFAs with the lowest sum of the oxides (i.e., FA-ZA, and FA-ZC), a 40% replacement of cement was inadequate. After 24 months, only FA-H and FA-Q maintained expan- sions below 0.04% at a 20% replacement of cement. At a 30% replacement of cement, CFAs FA-O and FA-U were adequate. 0% 20% 40% 60% 80% 100% 1 10 100 1000 10000 Cu m ul at iv e Pe rc en t F in er (a re a % ) Chord length (µm) 25% replacement with FA-G No replacement Figure 4.37. Cumulative chord length distribution for control and 25% replacement with FA-G and AEA-1. ID (Class) Age (Days) Expansion (%) Low-Alkali Cement Expansion (%) Reduction from High-Alkali Cement Expansion (%) FA-H (F) 14 0.179 89 63 28 0.187 71 68 FA-M (F) 14 0.192 95 61 28 0.221 84 62 FA-O (F) 14 0.199 99 59 28 0.216 82 63 FA-Q (F) 14 0.159 79 67 28 0.166 63 71 FA-U (C) 14 0.284 141 42 28 0.312 119 46 FA-X (C) 14 0.220 109 55 28 0.259 99 55 FA-ZA (C) 14 0.275 136 44 28 0.302 115 48 FA-ZC (C) 14 0.350 173 28 28 0.380 145 34 Table 4.26. ASTM C441 Pyrex mortar bar expansion data for 25% CFA replacement. Material Expansions at 12 Months (%) 0% Replacement 20% Replacement 30% Replacement 40% Replacement PC-3 Control 0.100 – – – PC-3 Control with Non- reactive Aggregate 0.030 – – – FA-H – 0.007 0.012 – FA-M – 0.032 0.022 – FA-O – 0.033 0.026 – FA-Q – 0.019 0.012 – FA-U – 0.037 0.023 0.018 FA-X – 0.074 0.048 0.032 FA-ZA – 0.073 0.060 0.049 FA-ZC – 0.076 0.069 0.055 Table 4.27. Twelve-month expansions (ASTM C1293 concrete prisms).

45 With a replacement level of 30%, FA-M was right at the 0.04% expansion limit suggesting a 35% replacement would be adequate. At 40% replacement of cement, the low sum-of- the-oxides CFAs (FA-X, FA-ZA, and FA-ZC) exceeded 0.04% after 12, 18, and 24 months. Concrete prism expansions at 3, 6, 9, 12, 15, 18, and 24 months are provided in Attachment C. The 24-month concrete expansions versus the sum of the oxides (i.e., SiO2 + Al2O3 + Fe2O3) for each CFA, presented in Figure 4.39, show the differences between the Class F and Class C CFAs, in terms of effectiveness in mitigating deleterious ASR expansions. Accelerated Mortar Bar Tests The 14-day and 28-day ASTM C1567 mortar bar expansion data are shown in Tables 4.30 and 4.31 and in Figures 4.40 and 4.41; expansion data at intermediate test ages are provided in Attachment C. Although the limit used in AASHTO PP 65 is 0.10% at 14 days, the tests were continued to 28 days. Four of Material Expansions at 18 Months (%) 0% Replacement 20% Replacement 30% Replacement 40% Replacement PC-3 Control 0.116 – – – PC-3 Control with Non- reactive Aggregate 0.029 – – – FA-H – 0.020 0.019 – FA-M – 0.049 0.038 – FA-O – 0.047 0.039 – FA-Q – 0.039 0.021 – FA-U – 0.049 0.031 0.021 FA-X – 0.098 0.068 0.048 FA-ZA – 0.082 0.069 0.057 FA-ZC – 0.085 0.081 0.064 Table 4.28. Eighteen-month expansions (ASTM C1293 concrete prisms). Material Expansions at 24 Months (%) 0% Replacement 20% Replacement 30% Replacement 40% Replacement PC-3 Control 0.141 – – – PC-3 Control with Non- reactive Aggregate 0.028 – – – FA-H – 0.019 0.018 – FA-M – 0.055 0.040 – FA-O – 0.053 0.033 – FA-Q – 0.034 0.020 – FA-U – 0.059 0.036 0.020 FA-X – 0.110 0.079 0.057 FA-ZA – 0.095 0.082 0.069 FA-ZC – 0.100 0.097 0.076 Table 4.29. Twenty-four-month expansions (ASTM C1293 concrete prisms). 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 20 30 40 Pr is m e xp an si on a t 2 4 m on th s (% ) Replacement of Cement with CFA (% wt) FA-H FA-M FA-O FA-Q FA-U FA-X FA-ZA FA-ZC 0.141% Expansion for Control @ 24 Months Figure 4.38. Effect of CFA replacement of cement PC-3 on 24-month concrete prism expansions.

46 0.00 0.02 0.04 0.06 0.08 0.10 0.12 50 60 70 80 90 100 Pr is m e xp an si on a t 2 4 m on th s (% ) Sum of Oxides (SiO2+Al2O3+Fe2O3) (% wt) 20% Cement Replacement 30% Cement Replacement 40% Cement Replacement Class C Class F Figure 4.39. Concrete prism expansions at 24 months versus the sum of the oxides. Material Expansions at 14 Days (%) 0% Replace. 10% Replace. 20% Replace. 30% Replace. 40% Replace. 50% Replace. PC-3 Control 0.23 – – – – – PC-3 Control with Non-reactive Sand 0.05 – – – – – FA-H – 0.15 0.04 0.03 – – FA-M – – 0.15 0.07 0.11 – FA-O – – 0.28 0.19 – – FA-Q – – 0.07 0.05 0.04 – FA-U – – 0.18 0.11 0.07 – FA-X – – 0.21 0.14 0.13 0.11 FA-ZA – – 0.26 0.23 0.19 0.11 FA-ZC – – 0.30 0.28 0.25 – Table 4.30. Accelerated mortar bar expansions at 14 days (ASTM C1567). Material Expansions at 28 Days (%) 0% Replace. 10% Replace. 20% Replace. 30% Replace. 40% Replace. 50% Replace. PC-3 Control 0.39 – – – – – PC-3 Control with Non-reactive Sand 0.08 – – – – – FA-H – 0.30 0.10 0.05 – – FA-M – – 0.30 0.18 0.23 – FA-O – – 0.45 0.34 – – FA-Q – – 0.20 0.10 0.06 – FA-U – – 0.27 0.18 0.12 – FA-X – – 0.41 0.26 0.23 0.19 FA-ZA – – 0.44 0.38 0.34 0.22 FA-ZC – – 0.46 0.42 0.34 – Table 4.31. Accelerated mortar bar expansions at 28 days (ASTM C1567). the mixtures (FA-O, FA-ZA, FA-ZC, and FA-X) with a 20% CFA replacement level and two mixtures (FA-ZA and FA-ZC) with a 30% replacement level showed higher expansions than the PC-3 control mixture. The 12-month expansions for all mixtures decreased as the replacement level increased above 20%, and all exhibited less expansion than the PC-3 cement control mixture at the highest replacement levels. However, the data presented in Tables 4.27 through 4.29 show that all prisms with CFA exhibited less expansion than the control with increased reduction in expansion as the replacement level increased. For Class F CFAs, the replacement levels needed to miti- gate deleterious ASR expansion based on 14-day ASTM C1567 expansions were equal or higher than those indi- cated by the 12-month ASTM C1293 tests and higher than indicated by the current AASHTO M 295 Pyrex mortar bar

47 tests. For Class C CFAs, the levels indicated by the 14-day ASTM C1567 expansions were equal or higher than the levels indicated by the 12-month ASTM C1293 tests. If a 0.10% limit at 28 days was required by ASTM C1567, for three Class F ashes, 40% replacement of cement is not ade- quate. This is not in agreement with the 2-year expansions measured for the concrete prism tests (Table 4.29) at 30% replacement for Class F CFAs, suggesting an expansion limit of 0.10% at 28 days is not appropriate. Alkali Leaching Test The alkali leaching test was developed by Shehata and Thomas (2006) to evaluate the effectiveness of combina- tions of cement and CFA in binding alkalis to mitigate alkali-silica reaction. The results indicated no clear alkali release threshold for predicting concrete expansions that correlate with the ASTM C1293 expansion limit of 0.04%. Therefore, the test was not considered for further evaluation. 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 10 20 30 40 50 Ex pa ns io n at 1 4 da ys (% ) Replacement of Cement with CFA (% wt) FA-H FA-M FA-O FA-Q FA-U FA-X FA-ZA FA-ZC 0.23% Control @ 14 Days Figure 4.40. Expansions at 14 days versus cement replacement with CFA (ASTM C1567). 0.00 0.10 0.20 0.30 0.40 0.50 10 20 30 40 50 Ex pa ns io n at 2 8 da ys (% ) Replacement of Cement with CFA (% wt) FA-H FA-M FA-O FA-Q FA-U FA-X FA-ZA FA-ZC 0.39% Control @ 28 Days Figure 4.41. Expansions at 28 days versus cement replacement with CFA (ASTM C1567).

48 The methodology and results of these test experiments are provided in Attachment C. Comparison of ASR Test Results Table 4.32 lists the CFA replacement levels required for meeting ASR expansion limits for various tests. When test results did not provide a definitive CFA replacement level, “greater than” (>) symbols are used. For FA-H, a 20% replace- ment of high-alkali cement was adequate to mitigate deleteri- ous ASR expansion, except when a 0.10% expansion limit is used for the ASTM C1567 test at 28 days. In Figure 4.42, 14-day mortar bar expansions (ASTM C1567) are plotted against 24-month concrete prism expansions (ASTM C1293). In AASHTO PP 65, the maximum mortar bar expansion limit for ASTM C1567 is 0.10% and for ASTM C1293 the maximum concrete prism expansion limit is 0.04% at 24 months. 0.0 0.1 0.2 0.3 0.4 0.5 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 M or ta r B ar E xp an si on a t 1 4 da ys (% ) Concrete Prism Expansion at 24 months (%) 20% Cement Replacement Control 30% Cement Replacement 40% Cement Replacement Figure 4.42. 14-day mortar bars expansions (ASTM C1567) versus 24-month concrete prism expansions (ASTM C1293). Component FA-H FA-M FA-O FA-Q FA-U FA-X FA-ZA FA-ZC (% wt) Sum of Oxides 91.26 81.85 79.81 74.34 65.8 61.63 55.32 53.09 CaO 3.46 7.17 10.2 16.6 21.9 19.3 27.3 30.2 Na2Oe 2.27 4.23 2.04 1.63 2.15 6.85 3.97 2.5 Test CFA Replacement Level (%) ASTM C441 25 25 25 25 >25 >25 >25 >25 ASTM C1293 (24 months) 20 >30 30 20 40 >40 >40 >40 ASTM C1567 (14 days) 20 30 >30 20 40 >50 >50 >40 (28 days) 30 >40 >40 40 >40 >50 >50 >40 AASHTO PP 65 min.% CFA* 25 30 25 25 X** X** X** X** * Based on Prevention Level Y ** Must be determined by test Table 4.32. Levels of CFA required to mitigate ASR from different tests.