SUMMARY

Use of Marginal and Unconventional-Source Coal Ashes in Concrete

The objectives of this research were (1) to propose revisions to the AASHTO M 295 Standard Specification for Coal Fly Ash and Raw or Calcined Natural Pozzolan for Use in Concrete that will ensure the desired properties of the fresh and hardened concrete and (2) to provide guidelines for potential use of coal ash that does not meet the recommended revised specification. These objectives are accomplished through a testing plan designed to fundamentally understand both the science behind coal ash reactivity and adsorption and the contributions of these measurements to fresh and hardened concrete properties. A total of 22 coal ashes—16 Class F ashes and six Class C ashes—were collected and tested throughout the study to represent a broad distribution of ash sources and properties. The performance of these coal ashes was also compared to three inert filler materials: limestone, basalt, and quartz. The coal ashes were separated into two groups: eight standard ashes and 14 unconventional ashes. The standard ashes are coal ashes that meet the AASHTO M 295-21 specification; the unconventional coal ashes are either harvested, off-specification, have been beneficiated, have marginal properties, or do not fall under the specification scope.

Physicochemical Characterization

The chemical composition, mineralogical composition, moisture content, loss on ignition, fineness, morphology, and specific gravity of the ashes were determined using a variety of experimental techniques. Some ashes displayed different properties compared to standard in-specification coal ashes, including high loss on ignition (LOI), high sulfate content (SO₃), coarse particles, and different morphologies. However, the beneficiated ashes and many of the unconventional harvested ashes did not have properties substantially different from standard ashes. On average, the Class C ashes measured had 63% amorphous content and the Class F ashes had 81% amorphous content. Differences between ashes seemed to be driven more by ash class, i.e., Class C vs. Class F, than by whether they were standard or not. The three measures of fineness or particle size—median particle size (d₅₀), specific surface area (SSA), and Blaine fineness—correlated well with each other, suggesting any one measurement could be used to estimate the others. LOI values measured using furnace and thermogravimetric analysis methods generally correlated, but both temperature and furnace atmosphere affected the measured values due to oxidation/reduction of sulfur and iron phases. Electron microscopy provided important information regarding particle sizes, shapes, and potential contaminants for the unconventional coal ashes. From the microscopy, “spongy” particles, very large and very fine particles, and irregular-shaped particles were all observed in unconventional coal ashes. The similarities between unconventional and standard coal ashes, in addition to their performance, highlight their potential to be specified for use in concrete.

Coal Ash Reactivity

The reactivity of numerous unconventional coal ashes—harvested, beneficiated, off-specification, and marginal—and other supplementary cementitious materials (SCMs) and fillers was studied using the R³ test (ASTM C1897) and a modified R³ test conducted at a higher temperature of 50°C without added sulfates and carbonates. Heat release, calcium hydroxide consumption, and bound water were measured for these materials. For siliceous materials with relatively low CaO + Al₂O₃ content, curing temperature had a dominant effect on the heat release, whereas for materials with a higher CaO + Al₂O₃ content, the effects of sulfates and carbonates dominated the effect of temperature due to secondary reactions. The slow but sustained heat release of Class F coal ashes and other siliceous materials highlighted the importance of kinetic corrections or extrapolations to the measures of reactivity measured in the standard R³ test. However, when performing tests at 50°C, the heat release curves for siliceous materials plateaued at the end of 10 days, and kinetic corrections were not required. All tested coal ashes, irrespective of the test, were reactive and passed reactivity thresholds, showing the importance of allowing their use in concrete. When using the R³ test, the Class C coal ashes showed a much greater 7-day heat release than the Class F coal ashes. This was not the case with the modified R³ test, where the classes showed similar levels of heat release. The heat release was not strongly influenced by whether a coal ash was unconventional or not. Some of the highest-heat-release ashes were unconventional, but some of the ashes with lower heat release were also unconventional. Fineness had a clear effect on heat release, with the coal ashes off-specification for fineness by having > 34% retained on the 45 µm sieve showing the lowest values of heat release. Although the chemical composition and fineness have important effects on reactivity, the link to standard or unconventional ash nature is not obvious.

Cement Paste Properties

The impacts of coal ashes on early- and later-age properties of cementitious paste, including heat-release characteristics, bulk resistivity (BR), compressive strength, calcium hydroxide content, and bound water content were studied using a variety of experimental techniques. These properties were assessed and compared to the impacts of inert fillers. Differences between unconventional and standard coal ashes were in most cases minimal, but substantial differences between Class C and Class F ashes were observed. This finding is in line with the findings from the physicochemical characterization and reactivity testing. The coal ash CaO content was positively correlated with several early-age paste properties, especially heat release and compressive strength. Class F coal ashes showed significantly higher BR at 91 days, while Class C coal ashes showed higher bound water contents. By 91 days, no significant difference in strength was observed between classes of ash. Strength and resistivity in ashes were significantly higher than in mixtures with inert fillers. Given the stark differences in BR, the test appears promising for distinguishing inert and reactive materials. Interestingly, even at 91 days, most coal ashes showed lower strength than the control mixture without any coal ash. Critically, it was found that all tested unconventional coal ashes were reactive, contributed to strength development, and did not negatively affect cement hydration. Unconventional ashes did in a few cases show outlier properties for time to peak heat flow, heat release, bulk resistivity, and other properties. These differences were generally due to high LOI, low specific surface area/coarse ashes, high sulfate content, or due to different phases and morphology. Correlations between reactivity and early- and later-age paste properties were shown—specifically, greater reactivity as measured using heat release leads to greater strength; as measured using calcium hydroxide consumption, it leads to greater BR.

Modified Strength Activity Index and Bulk Resistivity Index

The current ASTM C311 strength activity index (SAI) test is unable in some cases to successfully distinguish inert materials from reactive materials. The early age of testing, low coal ash replacement level, variable water-to-cementitious materials ratio, and low specification limits all contribute to this limitation. Test modifications using higher ash replacement levels (50% by mass) and higher temperatures (50°C) were evaluated in this study for three coal ashes and two inert materials. In addition to strength, the BR of the mortars was also measured. To explain the findings, heat release and calcium hydroxide were also measured. At 7 days, at room temperature, most tests failed to distinguish Class F coal ashes from inert materials, presumably due to the slow reaction of the Class F ashes. Using higher replacement levels and especially performing the test at higher temperatures provides improved differentiation of inert and reactive materials when using SAI, but the magnitude of differences was generally small, especially considering the inherent variability of strength measurements. On the other hand, higher ash replacements and especially higher temperatures resulted in substantial (orders of magnitude) increases in BR for the coal ashes, but not for inert materials. Therefore, a bulk resistivity index (BRI) test carried out on mortar specimens cured at high temperatures appears to provide a clear differentiation between inert and reactive materials. For this reason, standardization/specification for this test needs to be considered and subject to additional robustness and round-robin testing. Carrying out the test at 38°C could be especially attractive, given a number of labs have storage at such temperatures. The calcium hydroxide content and heat release measured in the corresponding cement pastes clearly showed that elevated temperature increased coal ash reaction but did not substantially affect inert materials. The reason the BR increases so significantly with temperature could be due to pore depercolation.

Mortar Testing

The impacts of unconventional coal ashes on early- and later-age properties of mortar—water requirement, SAI, modified SAI (MSAI, using a constant water-to-cementitious ratio), Keil hydraulic index (KHI), bulk uniaxial resistivity, and total efficiency (TE)—were investigated, assessed, and compared to the impacts of the control and inert fillers. For most standard and unconventional ashes, the measured water demand was within ±5% of the control mixture, except for a few off-specification ashes. An SAI limit of 75% at 7, 28, or 56 days using constant flow was chosen to remain in the standard after finding that a proposed SAI limit of 80% at those ages or adjustment to use a constant water-to-cementitious materials (w/cm) ratio restricted viable, reactive coal ashes. Later-age BR measurements in the SAI tests showed greater differentiation and significantly higher values compared to the control than their mortar strength counterparts. BR may be useful for inclusion in a future specification after further testing. The KHI may also be a potential future test to include in specifications as it was able to detect low reactivity for certain ashes including coarse or off-specification fineness ashes. The TE test results were inconsistent for the tested ashes and were therefore less meaningful than the other reactivity tests. Overall, SAI, MSAI, KHI, and TE testing on the ashes showed varied success in differentiating ash reactivity and influence on water demand regardless of ash source—standard or unconventional.

Concrete Testing

The impacts of unconventional coal ashes on early- and later-age mechanical and durability properties of concrete (slump, water demand, compressive strength, bulk uniaxial resistivity, chloride penetration resistance, and sulfate attack resistance) were evaluated and

compared to the impacts of the control. Generally, both standard and unconventional ashes showed increased workability or slump compared to the control. High water demand with some coal ashes may be mitigated by adjusting admixture dosage during mixing. All other fresh properties showed comparable values among the ashes by classes and sources and with respect to the control concrete mixture. In terms of compressive strength, most standard and unconventional ashes generated 28- and 180-day concrete strengths comparable to those of the control concrete. Certain coal ashes, both standard and unconventional, exhibited lower later-age strength compared to the control due to varying properties. By 180 days on average, both standard and unconventional ashes generated comparable BR values, all of which were significantly greater than the control. Rapid chloride permeability test (RCPT) results showed that all investigated ashes reduced chloride ingress significantly more than the control concrete. No significant differences were identified in the RCPT results between standard and unconventional ash sources at 28 and 91 days. Also, most coal ashes better resisted sulfate-attack expansion than the control mortar and filler materials with notable exceptions of Class C coal ashes from both standard and unconventional sources. These Class C ashes all contain higher CaO, CaSO₄, and C₃A, resulting in significantly higher expansions and earlier deteriorations due to sulfate attack (SA). Overall, the fresh and hardened properties and the durability performance measurements showed that most unconventional ashes are a viable option for concrete use as a partial replacement for cement.

Adsorption Testing

Adsorption of air-entraining admixtures (AEAs) with unconventional coal ash sources was assessed using a variety of methods: LOI, foam index testing (FIT), fluorescence method, iodine number, mortar air, and concrete air testing, with the goal of determining both the likelihood of adsorption variances with unconventional sources and understanding the applicability—precision, accuracy, ease of use—of the existing adsorption methods. Additional work was performed to understand how differences in chemistry of unconventional source coal ashes could affect precipitation and foaming of AEAs if calcium, alkalis, or sulfate from those ashes dissolved into concrete mixing water.

Most of the unconventional ashes were insignificantly different from standard ashes with regard to adsorption tracked with LOI and FIT. The exceptions were the cyclone collector ash and the ash with very high carbon and LOI content, which showed much higher adsorption levels in all test methods. Mortar air content and FIT testing indicated high AEA demand in the cyclone collector ash, despite its low LOI levels, suggesting that both these tests may provide better identification of potentially problematic materials. However, no issue was found in adequately entraining air and obtaining spacing factors to meet the recommendations of ACI Committee 201 (2016) in any of the coal ash concrete mixtures tested, including mixtures using the cyclone collector ash and high LOI ash, apart from higher AEA dosage requirements. Testing of air content stability in cement-AEA slurries and mortars showed no increased likelihood of air loss due to use of unconventional ashes. Indeed, foam drainage testing indicated that many of the unconventional coal ash samples were more stable than the control (portland limestone cement, PLC) mixture. FIT was found to provide consistent indication of increased adsorption and was able to account for AEA-calcium precipitation effects, chemisorption, and physisorption in the mixtures. These results suggest that changes or increased adsorption can be tracked using FIT. The research team proposes using a FIT limit equal to the FIT dosage required for a control (ordinary portland cement [OPC] or PLC) mixture plus two standard deviations. This translates to a mixture with FIT greater than 160% of the control FIT. This “limit” is proposed to help identify coal ashes with adsorption greater than that of the control mixture. It is not intended to exclude use of higher FIT ashes, but instead to identify ashes potentially requiring higher

AEA dosages in concrete. Ashes exceeding this limit should be investigated further with mortar air or concrete testing. Of the other adsorption test methods, LOI, fluorescence, and the mortar air test were found to correlate with each other, suggesting that any of these methods can be employed to identify changes in coal ash sample adsorption. Iodine number testing had several artifacts, especially when used with Class C coal ashes, and was difficult to use. As a result, the team does not recommend its further use with coal ashes.

A variety of ionic solutions were tested with the AEAs and calcium was found to have the most significant impact on AEAs, inducing precipitation of a significant fraction of the AEA from solution. Adsorption test methods that do not standardize or account for the interaction with calcium will underestimate the dosage of AEA required to obtain adequate air generation. These test methods could then provide indications of differences in adsorption of coal ash sources that will not manifest in cementitious mixtures. Addition of potassium, sodium, and sodium sulfate to AEA solutions resulted in increases in the required AEA dosages by approximately 10%. Similar increases can be expected in AEA requirements from coal ashes with higher-than-typical water-soluble alkali and sulfate contents.

Use of PLC instead of OPC required on average a one-third lower AEA dosage to reach the FIT endpoint. In addition, while AEA dosage varied consistently among the coal ash types, different AEAs required different dosages to reach the FIT endpoint. If agencies wish to use the FIT to track changes in adsorption, they should require use of a standard AEA, or use increases in AEA relative to (%) a cement control slurry. Additionally, due to differences in FIT when using different AEAs and cements, a standardization procedure should be investigated. One approach may be to classify adsorption on the basis of percentage change from the dosage required to reach the FIT endpoint in a control (cement + AEA) mixture. This project’s limits, given in the discussion of FIT and differences in adsorption across standard and unconventional ashes at the start of this section, suggest that FIT dosages in coal ash + cement mixtures > 160% of the cement-only FIT mixture indicate significant increases in adsorption.

Uniformity Testing

Coal ash suppliers provided uniformity datasets for two standard and two unconventional coal ashes. The required composite sampling frequency (monthly or 3,200 tons) specified in ASTM C311 for ash properties measured in AASHTO M 295 was assessed using the power analysis, t-test, uniformity assessment, and the Levene test for all four coal ashes. Furthermore, regular sampling frequency (daily or per 400 tons) was assessed on fineness and foam index measurements for one unconventional coal ash. This was the only regular sampling data provided. Conclusions drawn here are only applicable to the tested ashes. The results of the power analysis and t-test on composite sampling (per 3,200 tons) conducted on pertinent properties (moisture content, density, LOI, fineness, AEA dosage, foam index) of two standard and two unconventional coal ashes indicated that a reduction of testing frequency from every 3,200 tons to every 6,400 tons did not negatively affect the ability of testing to catch outlier data nor fail to assess properties with respect to given limits. However, more data are needed to assess how regular sampling for moisture content, LOI, and fineness is affected by unconventional source ashes.

When comparing the uniformity of the moving mean for the composite samples and subsequently using the Levene test to compare variances, the unconventional ashes had comparable or lower variance in fineness and density than the standard ashes, but much higher variance in moisture content, except in one instance. When comparing foam index uniformity, the untreated unconventional coal ash Y had much higher variance compared

to the same coal ash after treatment and another standard ash. The beneficiation significantly improved uniformity variance for foam index measurements. Additional uniformity testing on regular sampling data (daily or 400 tons) for an unconventional ash showed it had much higher variability in fineness and foam index measurements compared to composite sampling (monthly or 3,200 tons). This suggests that load-to-load variability may not be captured by current uniformity testing, which is performed on composite sampling. More regular sample data are needed to further investigate load-to-load variability.

Proposed Draft Language for AASHTO M 295 Standard Specification

The specification can be harmonized with ASTM C618-23e1 by making the following changes:

1. Change scope to allow harvested, bottom, and processed ashes.
2. Add no. 100 sieve fineness limit to a maximum of 10% for harvested and bottom ash samples.
3. Remove the soundness limit as measured by autoclave.
4. Change terminology from “coal fly ash” to “coal ash.”
5. Increase maximum allowable LOI from 5% to 6%.
6. Allow use of Class F coal ashes with up to 12% LOI with accompanying successful field performance records or laboratory testing of air entrainment in mortar or concrete.

The team recognizes that the AASHTO M 295 specification has undergone some of these changes before the publication of this document. Beyond harmonization, the specification could:

1. Change water requirement to “report only.”
2. Add an R³ reactivity bound water limit of 3.5 g/100 g dry paste when tested by ASTM C1897-Procedure B.
3. Add the FIT as “report only” under optional physical requirements in accordance with ASTM C1827. In the future, after further research, the specification could include the BRI test, the KHI test, and assessment of uniformity using the FIT.