Wet storage effects on fly ash properties relevant to its use in air-entrained concrete

Recent changes in electricity generation have seen reductions in burning of coal, with demand being increasingly met by other means (BEIS, 2021). The situation is likely to develop further with international efforts to reduce greenhouse gas emissions (Rogelj et al., 2016) and plans to withdraw coal-fired power stations in many countries (Europe Beyond Coal, 2021). As a result, fly ash availability for concrete construction is being affected (Diaz-Loya et al., 2019). One option for sourcing material in future is the reserves held in wet storage areas near power stations (BEIS, 2017) – for example, stockpiles, available because of variations in supply and demand with time (more than 100 Mt of this fly ash are believed to be accessible in the UK (UKQAA, 2020)).

Investigations of wet-stored fly ash indicate that changes can occur when material is held under these conditions. For example, particle agglomeration/modified surface morphology (Eze et al., 2013) owing to physico-chemical processes (Donahoe, 2006; Sear, 2001) affecting the reactivity of the material and its performance as a cement component in concrete have been noted (McCarthy et al., 2000). Factors including fly ash properties and exposure conditions can influence the effects occurring. Hence, these can be variable (Robl et al., 2008; Zevenbergen et al., 1999) and evaluation of material in wet storage areas on an individual basis, to establish recovery and processing strategies, is likely to be necessary (McCarthy et al., 2017).

A key property of fly ash concrete that has received coverage in the literature is air entrainment (Folliard et al., 2009; Hill et al., 1997; Külaots et al., 2004; Pedersen et al., 2008). Air-entraining admixtures (AEAs, surfactants) used to achieve this contribute freeze–thaw resistance to concrete (Dyer, 2014) as well as other effects (e.g. increased cohesiveness, reduced bleeding (Hewlett et al., 2019)). The interest in air entrainment with fly ash relates to residual carbon present in the material, its adsorption of AEA and influences on the air–water/cement interface stability and admixture dose requirements (Concrete Society, 2011; Gao et al., 1997; Hower et al., 2017; Pedersen et al., 2008). The literature above suggests that several factors associated with the physical and chemical properties of fly ash control the process.

Related studies indicate that dye removal from water by fly ash carbon (analogous to AEA uptake) is influenced, among other factors, by adsorbate pH and its effects on surface charge (Janoš et al., 2003; Wang et al., 2008). Research on wet-stored fly ash has shown that triboelectric carbon separation requires reversal of the electrode polarity (following drying) compared with dry material (Baker et al., 2015). This appears to relate to the various chemical effects (Donahoe, 2006; Eze et al., 2013) occurring in fly ash during wet storage, which modify particle surface composition and charge (Baltrus et al., 2002; Cangialosi et al., 2009). Given these and the physical effects noted above, it seems possible that wet fly ash may affect air entrainment. This is also suggested in preliminary tests on processed stockpile fly ash concrete (McCarthy et al., 2022a). With the changing production situation and likely sourcing of wet-stored fly ash (including processing) for concrete construction (Cooke, 2018) in future, a research programme was established to investigate this.

The presence of carbon in dry fly ash can affect air entrainment and there is evidence to suggest that with wet storage the physico-chemical changes occurring may also affect this aspect of behaviour. A range of fly ash properties were considered, with recognised markers of air entrainment behaviour examined and air content measurements made in mortar. Experiments were carried out on fly ash during wet storage and following processing. The study was aimed at (i) developing an understanding of the effects occurring and (ii) providing practical information. The research has relevance to the future application of wet-stored fly ash in concrete construction.

In studying wet-stored fly ash, samples covering a range of properties either (i) moistened and kept in the laboratory or (ii) sourced from site stockpiles were examined. This enabled control of the storage conditions and means of validating the results. During the initial studies, fly ash was dried following laboratory storage/on receipt from stockpiles, before testing, with any de-agglomeration owing to handling being minor.

The fly ashes were physically and chemically characterised to establish effects during wet storage and assist with data interpretation. Properties of fly ash known to influence behaviour with AEAs or provide a measure of this were investigated for wet material. These include loss on ignition (LOI) and carbon content, median particle size (d₅₀, LASER diffraction), specific surface area (SSA, nitrogen adsorption), foam index and dye adsorption (Acid Blue 80 (AB80)) of fly ash. The last of these is one of several adsorption-based methods identified for characterising carbon in fly ash (Sutter and Bentz, 2017). Previous research examining fly ash/AEA behaviour has found good agreement between mortar and concrete (McCarthy et al., 2012a; Spörel et al., 2009), and air content studies on the former were also carried out.

Additional tests were made to identify fly ash wet storage effects on admixture accessibility to particles and their surface chemistry, known to influence air entrainment. Practical issues were studied by examining stockpile fly ash following processing during a pilot-scale trial (with techniques including particle separation (air-classifying), size reduction (micronising) and carbon removal (electrostatic)), which may be used in sourcing wet-stored fly ash, for concrete.

Low lime fly ashes, from bituminous (DFA1 and DFA3) and anthracite (DFA5) coals, were obtained from three UK power stations. A further three fly ash samples were also sourced from stockpiles at these sites (SFA1, 2 and 4, respectively). Their main physical and chemical characteristics were determined using Standard or in-house tests, as summarised in Table 1.

The characteristics, given in Table 2, indicate that the dry fly ashes had fineness ranging from 6 to 34% (retained on a 45 μm sieve) and LOI from 5.6 to 13.6%, with chemistry typical of that for the coal used. The stockpile fly ashes on receipt had moisture contents between 12.7 and 21.1% by mass of sample. Effects noted previously for wet fly ash, including increased coarsening and LOI (compared with dry material), reflecting probable agglomeration and product formation (McCarthy et al., 2017), are likely. Although production and storage histories were not known, the stockpile fly ashes had similar bulk oxide and mineral compositions to the dry materials. Minor levels of sulfate-based products were found in these fly ashes, reflecting reactions under the wet conditions, as noted previously (Donahoe, 2006).

A Portland cement (PC, CEM I) of strength class 52.5N to BS EN 197-1 (BSI, 2011) was used during foam index tests. This was also adopted with fly ash and CEN Standard sand to BS EN 196-1 (BSI, 2016) for the mortar tests (water requirement, activity index and air content).

A standard reagent, sodium dodecyl benzene sulfonate (AEA S), adopted previously in foam index tests (Spörel et al., 2009), at 0.01 mol/l concentration was used. A commercial AEA (AEA C, composite synthetic/tall oil fatty acid/anionic surfactant) was selected and included in the mortar mixes as received. This was a yellow liquid with a pH of 10 and conformed to BS EN 934-2 (BSI, 2009). The AB80 dye for fly ash adsorption tests was a standard reagent used at 100 mg/l concentration. Details of its characteristics can be found in McCarthy et al. (2012b).

The laboratory-stored fly ashes were moistened, as described previously, in a 25 l mixer with tap water (pH ≈ 7.0) at 10% by dry mass – that is, the drier end of practical moisture levels and in the range found to give maximum agglomeration with wet storage (McCarthy et al., 2017). The materials were sealed in plastic bags/air-tight containers and kept at 20°C for 730 days, allowing effects to develop, with samples taken frequently during this period. At these times, moisture contents were measured and drying carried out at 105°C. Following cooling, the materials were resealed (in fresh bags/containers) and kept at 20°C until testing. The three stockpile fly ashes were obtained in wet form and similarly dried and stored as described above.

The LOI of fly ash was tested on duplicate samples following BS EN 196-2 (BSI, 2013) and is summarised in Table 1. This involved measuring weight changes of 1.0 g samples after ignition at 950 ± 25°C for 1 h (BSI, 2012). The other tests described in this section were carried out on single samples, with repeats made for confirmatory purposes. The carbon content of dry and laboratory wet-stored fly ashes was measured from the carbon dioxide (CO₂) evolved during combustion, at an external laboratory.

The SSA of dry and wet fly ash was measured by nitrogen adsorption using a Quantachrome NOVA 3000e analyser. Fly ash samples of 3.0 to 5.0 g were introduced to the test cells, de-gassed and heated (77 Pa vacuum and 105°C) overnight. Following nitrogen adsorption and desorption cycles, SSA was determined in m²/g, using Brunauer–Emmett–Teller (BET) theory, directly from the equipment.

The foam index test examines fly ash/AEA behaviour, providing a measure of admixture dose requirements. A 2.0 g sample of fly ash, combined with 8.0 g of PC, was added to 25 ml of de-ionised water in a 40 mm diameter × 110 mm (round bottom) vessel and mixed using an IKA MS3 digital vortex shaker for 50 s. AEA S was then introduced in 20 μl increments by micropipette and (vortex) shaking applied for 25 s, after the admixture addition (similar to McCarthy et al. (2012a)). Following each cycle, the mixture was examined until a stable foam formed over the whole surface and was maintained for 45 s. The quantity of AEA required to achieve this was the foam index given in μl/g fly ash.

AB80 dye was used to measure fly ash adsorption, as described by McCarthy et al. (2012b). A 2.0 g fly ash sample was added to 100 ml of AB80 solution and mixed by magnetic stirrer (400 rpm) for 1 h. The mixture was passed through medium flow, cellulose, filter paper and the filtrate analysed by spectrophotometer (Jenway 7315). Calibration was made with distilled water and AB80 solution at 626 nm (peak AB80 wavelength (Valix et al., 2004)). The absorbance of filtered AB80 solution was measured, corrected for filter paper dye loss, with the results given as mg of adsorbed dye per g of fly ash.

Fly ash/AEA mortar tests were carried out to determine air content at fixed admixture dose. The mortars contained 20% fly ash combined with PC in cement (total = 450 g), w/c ratio of 0.50 and a cement to (Standard) sand ratio of 1 : 3. The mortar was mixed as described by McCarthy et al. (2012a) with a 1.0 g dose of AEA C (0.22% by mass cement) added with the second half of the mix water. This was established following trial mixing to give a reasonable air content range, with measurements made using the BS EN 1015-7 mortar test (pressure method; BSI, 1999).

The fineness by 45 μm sieve retention and median particle size (d₅₀) of dry and wet fly ash, shown against storage period, are given in Figure 1. These indicate increases for both properties, suggesting agglomerate development with time. Between initial and final tests, differences of approximately 40% and 28.0 μm and 25/15% and 6.0/6.0 μm for sieve retention and d₅₀ were obtained for DFA3 and DFA1/DFA5, respectively, with greater changes for finer, more reactive fly ash and less noticeable effects for the two coarser materials.

As shown in Figure 2, for dry material, DFA1 surprisingly gave the lowest water requirement, with those of DFA3 and DFA5 following fineness and LOI, as might be expected (Thomas, 2013), but with none meeting the 95% limit in BS EN 450-1 (BSI, 2012) for category S fly ash. The property increased with wet storage, giving a change in ranking, reflecting coarsening/agglomeration and particle roughening, and showing general agreement with behaviour noted previously (McCarthy et al., 2017).

The activity index of the dry fly ashes, also given in Figure 2, met the 75/85% BS EN 450-1 (BSI, 2012) limits at 28/90 days and generally followed fineness, as noted previously (Sear, 2001), with DFA3 highest, and similar results for DFA1 and DFA5. Activity index reduced with moistening compared with dry fly ash, with greatest effects for DFA1 and DFA3 and less for DFA5, generally increasing with storage period. This appears to be influenced by fine particle levels in the fly ashes – most affected by wet storage and influencing reactivity (McCarthy et al., 2022b). For many tests, wet-stored fly ash was outwith the BS EN 450-1 (BSI, 2012) limits.

An example of bulk oxide and mineral compositions for DFA5 is given in Table 3. With one or two exceptions, appearing to reflect material variations, the bulk oxide compositions for dry and wet fly ashes (stored for up to 730 days) were similar. Given the storage conditions, changes in form rather than gains and losses of components are likely, and are reflected in the minor differences noted. A comparison of mineral compositions indicates no consistent effect between storage periods. Minor levels of calcium sulfate were observed for some laboratory wet-stored materials – corresponding to products forming (Donahoe, 2006; Sear, 2001); however, there were variations in these with time.

The LOI of dry and wet-stored fly ashes are shown in Figure 3. As noted above, the dry values provided a range from medium to relatively high. With wetting and storage, the property either gave little change compared with dry material (i.e. was within about 0.4%; DFA3), or else increased by up to around 1.0% (DFA1 and DFA5). Where differences were noted, they mainly occurred shortly after moistening, with small changes thereafter. These types of variation have been found previously for material similarly moistened and stored (McCarthy et al., 2017) and are likely to reflect decomposition of carbonates, sulfates and release of combined/adsorbed water present (Donahoe, 2006; Eze et al., 2013; McCarthy et al., 1999) during ignition.

The carbon contents of dry and 180-day wet-stored fly ash are also shown in Figure 3. For dry fly ash, these were 0.3, 0.7 and 2.1% lower for DFA3, DFA1 and DFA5 than the LOI values, with greatest differences as the latter increased, probably owing to certain minor constituents and combined water in residual clay minerals, which contribute to LOI (Alonso and Wesche, 1991). Carbon contents were up to 0.3% higher in wet-stored fly ash than dry, which may reflect carbonation during storage, albeit this was not detected by X-ray diffraction.

Studies of dry fly ash (Külaots et al., 2004) indicate that the mineral component has a relatively narrow SSA (nitrogen adsorption) range of approximately 0.7 to 0.8 m²/g between materials, with differences from this corresponding to carbon present and its structure (internal porosity). SSA has also been found to give correlations with different measures of fly ash/AEA behaviour (Ley et al., 2008; Spörel et al., 2009). This property for the dry fly ashes at 5.4, 3.0 and 3.3 m²/g for DFA1, DFA3 and DFA5, respectively, suggests that in spite of a higher LOI, DFA5 has less porous carbon than DFA1 and may give better air entrainment performance.

Wet storage effects on the SSA of fly ash are shown as changes with respect to dry fly ash in Figure 4. The results relate to the different fly ash characteristics and correspond to product and agglomerate structure formation. Between initial and 730 day values for wet-stored fly ash, changes ranged from 1.0 to 6.1 m²/g, with greater increases for material giving more noticeable coarsening with wet storage, and lower SSA (nitrogen adsorption) in dry form. Changes in carbon, following weathering, with variations in morphology and fewer surface pores, have been noted (Wirth et al., 2019). It is also possible with fly ash/carbon agglomerates forming that nitrogen uptake could be less (owing to restricted access to carbon), affecting differences compared with dry fly ash.

Wet storage effects are also shown in the scanning electron microscopy (JEOL SM-7400F) and computed tomography (Nikon XT H 225 ST) images in Figure 5. These demonstrate particle surface activity (Eze et al., 2013) and pores in agglomerates formed during wet storage, which may affect fly ash surface area. The agglomeration is similar to the physical effects noted previously in the combustion of brown coal (Sarbak et al., 2004), where fly ash clusters developed, giving many small channels and increased surface area. The results suggest, from a physical point of view, that the processes referred to above could influence AEA behaviour in cementious systems.

Foam index and AB80 adsorption results are shown with wet storage period for the three fly ashes in Figure 6, along with SSA. For dry fly ash, DFA3 had the lowest values for all three parameters, agreeing with the LOI ranking, indicating it is likely to adsorb least AEA and perform best with respect to air entrainment. As noted above, despite DFA1 having a lower LOI than DFA5, the results suggest more porous carbon for the former. This is reflected in the foam index and AB80 adsorption, where similar results between the two fly ashes, or higher values with the low LOI fly ash (DFA1), were obtained.

With moistening of fly ash, the results show that foam index and AB80 adsorption generally reduced with increasing storage period, with all materials following similar trends. These ranged from 30 to 20, and 120 to 50 μl/g for foam index between DFA3 and DFA5 (lowest and highest LOI) during storage, with reductions from 1.9 to 0.6 and 2.4 to 1.8 mg/g for AB80 adsorption, suggesting that wet storage may reduce air entrainment requirements. The opposite was generally noted with SSA, with increases from 3.0 to 9.1 and 3.3 to 5.5 m²/g for DFA3 and DFA5, respectively. This differs from dry fly ash, where agreement between parameters has been found (Ley et al., 2008; McCarthy et al., 2012b), suggesting internal pores of agglomerated structures and product formation/surface roughening (reflected by nitrogen adsorption) do not significantly affect AEA behaviour.

To further examine this, the foam index results are compared against d₅₀ during storage in Figure 7. The results suggest that as fly ash coarsens with wet storage (increasing d₅₀), the foam index reduces, which is similar to AEA requirements noted with coarser PC (Du and Folliard, 2005). The rate of change of foam index, however, was material dependent (e.g. greatest d₅₀ change was in DFA3, which gave least change in foam index). Similarly, the most noticeable differences in foam index tended to occur in fly ash with highest SSA when dry. It seems that admixture accessibility to carbon particles may be inhibited following moistening/agglomeration with greatest effects in high adsorption materials. At the same time, it is possible that chemical changes noted in fly ash following moistening, referred to above, may also influence behaviour.

The various properties measured for dry, laboratory- and site-stored fly ash are shown in Figure 8. This indicates similar behaviour for wet fly ash, whether stored in the laboratory or stockpile (SFA1, 2 and 4), compared with dry material, with LOI and SSA giving higher values for the former, while foam index and AB80 adsorption were generally lower. The exception to this was the latter two tests for SFA4, which gave values similar to and higher than DFA5 (foam index, 120 μl/g; AB80 adsorption 2.4 mg/g). These are likely to relate to sample variability and the high LOI and changes in/condition of the fly ash following wet storage.

The air content results of mortars containing both dry and wet-stored fly ash with fixed AEA dose are given in Figure 9. As shown, the air content of dry fly ash mortar was highest for DFA3, with similar values for DFA1 and DFA5. This corresponds to the data noted earlier for these fly ashes in foam index, AB80 adsorption and SSA tests and highlights the role of carbon content and its characteristics on the process.

Following 180 days wet storage at 10% moisture, the air contents of all fly ash mortars were higher, by between about 2 and 6%, than those with dry material. The air contents of stockpile-stored material (SFA1, 2 and 4) were between the dry and laboratory wet-stored data except for SFA4, which was slightly lower than the corresponding dry fly ash. The results, therefore, give general agreement with those obtained earlier (foam index and AB80 adsorption; see Figure 8).

Studies on air entrainment with fly ash suggest factors including (i) carbon content/characteristics, (ii) SSA, (iii) AEA accessibility to particle surfaces and (iv) their surface chemistry (particle charge) influence the process. Therefore, the effects noted in the data above are likely to correspond to the influences that wet storage has on fly ash and, consequently, on (i) to (iv).

Initially, following water addition to fly ash, cohesion occurs between particles (Clarke, 1992), with chemical processes taking place thereafter (Sear, 2001). Products form and agglomerates develop, depending on factors including fly ash composition, particle size, pore solution chemistry (Donahoe, 2006; McCarthy et al., 1999) and exposure conditions – for example, moisture levels, temperature and holding period (McCarthy et al., 2017, 2019). Minerals identified in the material following wet storage include calcium sulfate hydrates and carbonates, aluminosilicate hydrates and others (Anthony et al., 2002; Eze et al., 2013; Fruchter et al., 1990; Georgakopoulos et al., 2002; Janssen-Jurkovicova et al., 1994; Wirth et al., 2019; Zevenbergen et al., 1999).

Given the increasing SSA during wet storage (by nitrogen adsorption), but coarsening (increasing d₅₀) and reduced AEA requirements with foam index, AB80 adsorption and mortar air content tests, the results suggest products forming and internal agglomerate surfaces do not significantly affect behaviour. Of the other factors above, the carbon content is unlikely to change (with the LOI/carbon increases, reflecting other factors), although, as noted above, admixture accessibility to these particles could be important.

To investigate wet storage effects on admixture accessibility, DFA1 and DFA3 were moistened and stored (as above) for 120 days in the laboratory, before grinding for different periods in a ball mill, following a similar approach to that referred to by Pedersen et al. (2008). With increasing grinding time, anticipated effects for wet-stored fly ash include gradual agglomerate breakdown with fracturing of the particles thereafter. Given the friable nature of carbon (Hurt et al., 1995), this may also be affected from an early stage. In other unpublished work, the authors have noted that grinding of wet fly ash can give small increases in LOI (up to 0.5%), which may reflect carbon release from agglomerates. This has also been observed in dry material following processing by other techniques (Baltrus et al., 2001), with similar effects being found for adsorption by fly ash (Yu et al., 2000).

Foam index and d₅₀ results for wet-stored DFA1 and DFA3 are given in Figure 10. These show appreciable increases in foam index and reductions in d₅₀ with initial grinding and for DFA1 – that is, higher LOI/SSA. Beyond this, the materials gave continued changes for both properties, which were most noticeable between 40 and 120 min. For DFA3, similar behaviours, but at lower levels with grinding, were noted. The greatest reductions in particle size occurred with early grinding, increasing admixture accessibility, with longer-term effects owing to increased surface area and particle breakdown. The data, therefore, suggest that the behaviour occurring is material dependent.

Research indicates that adsorbancy of fly ash is influenced by the pH of the solution the material is in contact with, affecting surface charge (at low pH, fly ash is positive, with the reverse at high pH (Wang et al., 2008)). In addition, when fly ash is moistened it undergoes a charge reversal, affecting the efficiency of electrostatic carbon separation (Xing et al., 2019). It has been noted that lagooning and stockpiling can considerably reduce soluble salt levels (Haynes, 2009), with greater effects likely for the former in an excess of water, and that the surface charge of fly ash depends on the quantities and type of ions present on particles (Baltrus et al., 2002).

As mentioned above, the surface composition of wet fly ash is influenced by the material properties and the exposure conditions. While drying wet fly ash can give charge distributions similar to recently produced material, other wet storage effects – for example, enhanced surface concentrations of soluble ions – are not reversed (Cangialosi et al., 2009). Experiments by Baltrus and LaCount (2001) involving fly ash washing prior to air entrainment showed removal of calcium (Ca²⁺) and magnesium (Mg²⁺) ions, that may interfere with the process, gave reductions in AEA requirements, irrespective of LOI. However, the addition of sodium (Na⁺) and potassium (K⁺) (as nitrates) were found to have little effect on foaming.

Tests to investigate these influences were carried out on DFA1 and DFA3 by measuring the effect of (i) the wet storage solution pH and (ii) the composition of fly ash particle surfaces on foam index. The experiments for (i) involved exposing fly ash to different pH solutions (reflecting possible field variations) (nitric acid (HNO₃) and potassium hydroxide (KOH) diluted with de-ionised water to give the target pH: 4.0 and 11.5, respectively) at 10% moisture. The results, following 120 days storage, are given in Table 4 and show differences of 10 μl/g with solution pH prior to grinding. Following this, the foam index increased slightly with reducing pH for DFA1 (higher LOI/SSA), with little variation noted for DFA3. The results suggest minor effects of moistening solution pH, with the data generally following the changes noted as admixture accessibility increased.

Given the changing chemistry with moistening and possible effects on fly ash/AEA behaviour, attempts were made during (ii) to examine the surface composition of DFA1 and DFA3 following storage at different pH for 120 days. Results from energy-dispersive X-ray spectroscopy (EDX) are shown in Table 5, normalised to dry fly ash carbon content. As indicated, different effects were noted between fly ashes and storage conditions and there did not appear to be consistent behaviour in the data. In general, as noted previously (Cangialosi et al., 2009), there were increases in soluble components on wet-stored fly ash particles compared with dry.

To examine fly ash particle surface chemistry, the sum of components, previously referred to as influencing air entrainment (Ca²⁺, K⁺, Na⁺ and Mg²⁺; Külaots et al., 2003) were examined against foam index, as shown in Figure 11. This indicates that as concentrations of these components increased, there was a reduction in foam index of around 20 to 30 μl/g. Consideration of various combinations of these components was found to give only minor changes in behaviour. The different relationships between the two fly ashes may be related to the other factors influencing the process – for example, carbon content, characteristics and accessibility effects.

In practice, wet-stored fly ash is likely to be processed to achieve suitable properties for use in concrete. In addition to drying, this may involve particle size separation or reduction and carbon removal. The studies of accessibility indicate breakdown of agglomerates may increase admixture requirements. To examine this further, foam index, AB80 adsorption and tests for other properties were made on stockpile fly ash, processed during a pilot-scale trial. Following recovery, the stockpile fly ash (SFA8) was flash dried/de-agglomerated in a pin mill (PFA2), and then either air classified (PFA3, material split with finer fraction considered) or micronised (PFA4, material reduced in size). Some of these materials also underwent carbon removal by triboelectric separation (PFA2 LC (low carbon) and PFA4 LC) at small-scale (bench-top unit) (Bittner et al., 2014).

The fly ash properties before and after processing, including foam index and AB80 adsorption, are given in Table 6. The initial de-agglomeration gave little change in d₅₀, SSA or LOI and hence foam index or AB80 adsorption. While d₅₀ reduced with air classifying, SSA, LOI, foam index and AB80 adsorption were generally unchanged, suggesting little influence of the process on air entrainment. Following micronising, reductions in d₅₀, SSA (increased particle accessibility) and minor changes in LOI and carbon were noted, with foam index and AB80 adsorption increasing.

These data show similarities to the accessibility tests with wet-stored DFA3, which was of comparable LOI (to that in the pilot-scale trial, see Figure 10). In practical terms, the results suggest processing may influence admixture accessibility for wet-stored fly ash, depending on the technique used. The fly ashes with carbon removed gave little change in foam index with lower LOI/SSA. AB80 adsorption was higher in both cases. This represents an area currently receiving further attention. Research is also ongoing to investigate the behaviour of wet-stored fly ash in air-entrained concrete.

Wet storage of fly ash in the laboratory was found to give agreement with effects noted previously for material held under similar conditions. These included agglomeration of particles, increased water requirement, reduced reactivity and the presence of sulfate-based components in fly ash. For the materials tested, it was noted that effects occurring tended to be greater in finer fly ash (i.e. material with increased levels of fine/reactive particles).

Tests on factors influencing air entrainment in concrete, including LOI/carbon content and SSA (by nitrogen adsorption), indicate that these increased with wet storage. For the former, processes involving sulfate and carbonate product formation and their decomposition during testing appear to be responsible. For the latter, effects appear to be owing to product formation on particle surfaces and agglomerate structures development. Similar types of results were obtained between materials under laboratory or stockpile storage.

Foam index and AB80 adsorption tests indicate that less AEA was required or dye adsorbed when fly ash had been stored wet and then dried prior to use. These effects generally became greater with increasing wet storage period and were different to that noted for SSA (by nitrogen adsorption). For a given material, foam index could be related to d₅₀ following wet storage. There was general agreement between foam index, AB80 adsorption and mortar air content for laboratory- and stockpile-stored fly ashes compared with those kept dry.

Additional studies suggest that increasing accessibility by grinding had greatest effects on foam index for the coarse/high LOI/SSA fly ash. The pH of the moistening solution was generally found to have a small effect, with similar behaviour noted following different periods of grinding. Increased concentrations of Ca²⁺, K⁺, Na⁺ and Mg²⁺ were found in many cases on fly ash particle surfaces by EDX following wet storage. Some reductions in foam index were noted as the sum of these components increased.

Pilot-scale processing of wet-stored fly ash by drying/de-agglomeration and air classifying (reductions in d₅₀ but little or no change in LOI or SSA) had no real effect on foam index and AB80 adsorption. With micronising (reductions in d₅₀ and SSA) there were increases in foam index and AB80 adsorption, highlighting processing influences on fly ash/AEA behaviour. Carbon removal from wet-stored fly ash by electrostatic separation had little influence on foam index but gave increased AB80 adsorption.

Thanks are given to the Engineering and Physical Science Research Council (Doctoral Training Award) and the UK Quality Ash Association for funding the research. The helpful advice of Drs Robert Carroll and Nigel Cooke during the project is greatly appreciated. Acknowledgement is also given to (i) Attritor Ltd (UK) and STET (USA) for processing fly ash, (ii) Hanson Cement (UK) for supplying the CEM I used and (iii) STEAG's Building Materials Laboratory (Germany) for carrying out the carbon content tests.