Fuel 89 (2010) 3042–3050

Contents lists available at ScienceDirect

Fuel

journal homepage: www.elsevier .com/locate / fuel

Characterization and evaluation of fly-ash from co-combustion of ligniteand wood pellets for use as cement admixture

Andrea Johnson a, Lionel J.J. Catalan a,*, Stephen D. Kinrade b

a Department of Chemical Engineering, Lakehead University, Thunder Bay, ON, Canada P7B 5E1b Department of Chemistry, Lakehead University, Thunder Bay, ON, Canada P7B 5E1

a r t i c l e i n f o

Article history:Received 29 May 2009Received in revised form 12 May 2010Accepted 18 May 2010Available online 1 June 2010

Keywords:Fly-ashBiomass co-combustionCement mortarEntrained airCompressive strength

0016-2361/$ - see front matter � 2010 Elsevier Ltd. Adoi:10.1016/j.fuel.2010.05.027

* Corresponding author. Tel.: +1 807 343 8573; faxE-mail addresses: abulanda@lakeheadu.ca (A. Joh

(L.J.J. Catalan), skinrade@lakeheadu.ca (S.D. Kinrade).

a b s t r a c t

Conventional coal fly-ash (CFA) and two coal-biomass fly-ashes (CBFAs) were obtained at a thermoelec-tric power station (Atikokan, Ontario) from combustion of undiluted lignite coal and co-combustion oflignite coal with up to 66% wood pellets (on a thermal basis). Fly-ashes were characterized and analyzedfor use as cement admixtures. Co-combustion did not markedly change the fly-ash composition, owing toan extremely low ash content of wood pellets compared to lignite coal; toxic metals and minor elementswere within ranges reported for other coal fly-ashes. All fly-ashes had losses on ignition (LOI) <1 wt% andtherefore complied with ASTM LOI regulations for use in concrete. All fly-ashes contained major amor-phous phases, along with quartz and periclase. Partial substitution of cement with fly-ash (up to 40wt%) had a moderate effect on the entrained air content of mortars (up to 2.5%), but this difference van-ished upon addition of air entraining agent (0.6 mL/kg of cementitious material). Substituted mortarsexceeded 75% of the strength of ash-free mortar after 28 days of curing (therefore meeting ASTM require-ments for strength development), and by 90 days, met or surpassed 100% of the strength of ash-free mor-tar. Amending mortar with 20 wt% CFA or CBFA had no effect on its durability following repeated freeze–thaw cycles when air content was kept constant. Also, no micromineralogical differences were observedbetween hydrated CFA- and CBFA-amended mortars, with fly-ash particles reacting with Ca ions originat-ing from dissolution of cement clinker or calcium hydroxide.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

The government of Ontario has committed to phasing out coalpowered electricity by 2014 [1]. This is an ambitious goal, as in2008 over 23% of the electricity produced in Ontario was derivedfrom fossil fuels, with the majority being coal [2]. In order to inves-tigate whether current coal burning facilities can be used to burnbiomass, several full-scale co-combustion tests of wood pelletswith lignite were carried out at the Atikokan Generating Station(AGS) in Ontario. The wood pellets originated primarily from Wes-tern Canadian forests affected by pine beetle infestation.

Although co-combustion of renewable biomass/waste with coalcan reduce coal consumption as well as emissions of CO2, nitrousoxides, sulfurous oxides, and mercury [3–5], the beneficial use offly-ash from co-firing remains a challenge in North America. Thislimits the environmental benefits of co-combustion becauselong-term storage of fly-ash consumes valuable land and increasesrisk of groundwater contamination by leached heavy metals. By far

ll rights reserved.

: +1 807 343 8928.nson), lcatalan@lakeheadu.ca

the most important use of traditional coal fly-ash is as partial sub-stitute for cement in concrete. Of the approximately 4.7 milliontonnes of fly-ash produced annually by coal-fired power generat-ing stations in Canada, 31% is recycled, thus contributing an impor-tant source of revenue for power plants [5,6]. In Atlantic andWestern Canada, more than 80% of the produced concrete containsfly-ash [7]. Current industry standards [8e], however, prohibit co-combustion fly-ash from being used as a cement admixture inNorth America because of its ‘‘non-coal” origin – despite the dem-onstrated pozzolanic activity of siliceous ash derived from biomasswaste such as rice husks, wheat straw, bagasse and wood [9–12].By contrast, the European Union has approved the use of fly-ashderived from the co-combustion of coal with wood, straw, olivehusks, green wood, cultivated biomass, animal meal, municipal so-lid waste and paper sludge as a cement admixture, as long as thefly-ash contains less than 5 wt% carbon, 5 wt% total alkali, and0.1 wt% chloride [13].

The composition of fly-ash derived from co-combustion can beaffected by differences in composition between the biomass andcoal [14,15]. Grammelis et al. [16] reported slightly decreased car-bon content and much lower free CaO content in fly-ash resultingfrom the co-combustion of 19 wt% pine or oak wood compared to

A. Johnson et al. / Fuel 89 (2010) 3042–3050 3043

traditional coal fly-ash. The physical characteristics of fly-ash (e.g.,particle size distribution) may also differ, owing to divergent com-bustion parameters, such as temperature and loading rate [17,18].When loads approach 100% (i.e., maximum fuel input to the bur-ner), soot blowers are used to introduce steam to blow off layersof ash that collect on the boiler tubes [18]. The upward force ofthe steam can entrain larger fly-ash particles and therefore affectits particle size distribution. Higher loads also raise the tempera-ture of the boiler and exit gases, which slows down the coolingof fly-ash and results in larger fly-ash particles [18].

The benefits of fly-ash as a mineral admixture in cement derivefrom its pozzolanic action [6,19,20]. Fly-ash reacts with calciumhydroxide from cement hydration reactions to form calciumsilicate hydrate (CSH) gel, thus contributing to the strength of con-crete [21–26]. Fly-ash substitution for structural concrete applica-tions typically ranges from 15 to 35 wt%, and substitutions up to 70wt% have been used for mass concrete in dams, walls, and roller-compacted concrete pavements [27]. Although fly-ash typicallyenhances long term strength of concrete, it also tends to delayearly strength development [28–33]. ASTM C618 [8e] stipulatesthat the strength of mortars containing fly-ash must be at least75% that of ash-free mortar after either 7 or 28 days of curing.Fly-ash from the co-combustion of switch grass with coal [5,34]and from the combustion of pure peat [3] were found to meet theserequirements at substitutions up to 59 and 33 wt%, respectively.The particle size distribution of fly-ash can affect compressivestrength development, with ashes more abundant in smaller parti-cles contributing more to long term strength [5].

The increase in strength that fly-ash provides to well-cured con-crete also enhances its structural resilience to repeated cycles offreezing and thawing if the concrete contains a suitable amount(4–6 vol%) of entrained air [35]. Air entraining agents (AEA) areusually added, consisting of aqueous mixtures of ionic or non-ionicsurfactants [36]. Their non-polar end associates with small air bub-bles, preventing them from coalescing and escaping the mixture[25,36]. Fly-ash can increase the AEA requirement due to compet-itive adsorption onto active carbon particles [7,36–39]. Biomass-derived fly-ashes may thus increase the AEA requirements, if theycontain a large carbon fraction [5].

The objectives of this research were (1) to characterize the com-positions, mineralogy, particle size distributions and specific grav-ities of fly-ashes generated from pure coal (CFA) and from the co-combustion of coal with wood pellets (CBFA) and (2) to comparethe AEA requirements, compressive strength development, dura-bility to freeze–thaw, and micromineralogy of mortar samples pre-pared with both types of fly-ash. All fly-ashes were obtained fromfiring tests at the Atikokan thermoelectric power plant. Hence, fly-ash composition and particle size distribution were truly represen-tative of full-scale equipment configuration and combustionconditions.

Table 1Mix proportioning of mortar samples.

Sample OPC (g) Fly-ash (g) Sand (g) Water (mL)

0% Fly-ash 500 0 1375 24220% Fly-ash 400 100 1375 22840% Fly-ash 300 200 1375 200

2. Materials and methods

2.1. Fly-ash characterization

Two coal-biomass fly-ashes (CBFAs) were produced at theAtikokan thermoelectric power station from test burns of 15%and 66% wood pellets on a thermal basis (which corresponds toapproximate mass ratios of 13:87 and 62:38 wood pellets and lig-nite coal), respectively designated 15CBFA and 66CBFA. The ashcontent of wood pellets (<0.5 wt%) was substantially lower thanthat of lignite (�10%). Moreover, the two fuels differed substan-tially in their mineral matter composition; calcium was the mostabundant mineral element in wood pellets, whereas silicon wasmost abundant in lignite. Conventional coal fly-ash (CFA) was col-

lected from a prior burn of pure lignite. The CFA and 66CBFA werecollected with the station running at approximately 50% load,whereas 15CBFA was collected at full load. Burns with biomass re-sulted in noticeable reductions of sulfurous and nitrogenous emis-sions (up to 67% and 45%, respectively), while the levels ofparticulate matter remained constant.

The elemental composition of the fly-ashes was analyzed byXRF (major elements and chloride), LECO (sulfur), ICP-AES follow-ing acid digestion (trace metals), and cold vapour AAS (mercury).The loss on ignition (LOI) was measured gravimetrically. Mineral-ogical composition was determined by XRD, particle size distribu-tions by laser particle size analysis (Malvern Mastersizer 2000),and specific gravity (in triplicate) by volume displacement ofkerosene.

Thin sections of epoxy-mounted fly-ash were cut, polishedusing an oil based medium, carbon coated, and viewed with a scan-ning electron microscope (JEOL JSM-5900LV) in back-scatteredelectron (BSE) mode to enhance contrast between mineral phases.Quantitative EDS analysis was performed using calibration stan-dards: garnet for Al, Fe, Mg and Si; orthoclase for K and Na; wollas-tonite for Ca; and barium sulfate for Ba and S.

2.2. Mortar preparation and testing

2.2.1. Mix proportioning and fresh mortar analysisMortars were prepared according to ASTM C109 [8a] in a Hobart

mixer with Type 10 ordinary portland cement (OPC: Lafarge Can-ada, Montreal) and sand (Hoskin Scientific, Burlington) conformingto ASTM C788 [8g]. Fly-ash substituted for 0, 20 or 40 wt% of thecement, as shown in Table 1. The water requirement of the mortarscontaining fly-ash was determined by adjusting the amount ofwater to achieve a flow ±5% that of the control mortar, in accor-dance with ASTM C311 [8d]. Flow was determined by ASTMC1437 [8h]. Increasing cement substitution by fly-ash decreasedthe amount of water required to achieve constant flow, but thewater requirement was independent of the type of fly-ash. This ef-fect could be due to the spherical shape of the fly-ash particles [21],or more likely, their weak chemical interaction with water initiallycompared to that of OPC.

Air entraining agent (Airextra, Euclid Admixture Canada, Toron-to) was only used during air entrainment tests. Airextra consists ofan aqueous solution of sulfonated fatty acids [40] conforming toASTM C260 [8c], and was added to the mixing water at concentra-tions of 0, 0.2, 0.6 or 1.2 mL/kg of cementitious material (fly-ash + OPC). The air content of fresh mortar samples was deter-mined in quadruplicate by ASTM C185 [8b] using a 400 mL cylin-der (Hoskin Scientific, Burlington).

2.2.2. Compressive strength analysisMortar samples containing no air entraining agent were cast

into two-inch cubes (American Cube Molds, Twinsburg, Ohio)and cured for 24 h in 100% humidity at 23 ± 2 �C. Next, the cubeswere removed from the molds and submerged in saturated limewater until testing in accordance with ASTM C109 [8a]. Unconfinedcompressive strength was measured in quadruplicate using a Ti-nius Olsen analyzer after 1, 3, 7, 28 and 90 days of curing. Priorto loading, the cubes were capped with polyurethane pads and

3044 A. Johnson et al. / Fuel 89 (2010) 3042–3050

retainers (American Cube Molds, Twinsburg, Ohio [41]) to improvethe reproducibility of strength measurements.

2.2.3. Durability to rapid freeze–thaw cyclesMortar cubes containing 0 or 20 wt% CFA or 15CBFA were cured

in saturated lime water for 14 days prior to freeze–thaw cycling, asprescribed in ASTM C666 [8f]. Air entraining agent was not used,since the air contents of the mortars were all within the targetrange of 2–6 vol%. For each cycle, the cubes were placed in a freezeruntil their core temperature decreased from 4 to �18 �C. Next, thecubes were removed from the freezer and kept in air at room tem-perature until their temperature returned to 4 �C. The temperaturewas monitored using a reference mortar cube with a thermocoupleimbedded at its centre. Each cycle lasted between 2 and 5 h, withno less than 20% of the time taken for thawing. The freezing por-tion of the cycle typically lasted 3 h, while the thawing portiontook approximately 1 h. All mortar cubes were subjected to the cy-cles simultaneously. Changes in dimension and mass were mea-sured every 35 cycles, as prescribed by ASTM C666 [8f].Compressive strength was measured every 35 cycles by ASTMC109 [8a].

2.2.4. MicromineralogyThin sections were prepared from material collected near the

centre of 28-day old mortar samples. The sections were lappedand polished using oil-based media so as not to alter the water-sol-uble minerals. After carbon-coating, the sections were imaged bySEM, and quantitative EDS analysis was performed using calibra-tion standards: garnet for Al, Fe, Mg and Si; orthoclase for K andNa; wollastonite for Ca; and barium sulfate for Ba and S.

Table 3Trace element content (mg/kg) of fly-ash samples.

CFA 15CBFA 66CBFA

Ag <2 <2 <2As <30 <30 36

3. Results and discussion

3.1. Fly-ash characterization

Table 2 shows the bulk composition and LOI of the fly-ashes.Co-firing biomass with coal did not markedly change the fly-ashcomposition, owing to the low ash content of wood pellets(0.02–0.5 wt%) compared with lignite (10.0 wt%). All fly-ashescontained 43–45 wt% SiO2, 21–21.5% Al2O3, 13.6–14.5% CaO and3.9–4.2% Fe2O3, putting them in the class C designation [8e]. TheLOI ranged from 0.4 to 0.9 wt%, which is much lower than the 6wt% limit prescribed by ASTM C618 [8e]. The lack of correlationbetween LOI values and the fraction of wood pellets in the fuelindicates that these low LOI values are not an accurate measureof unburned carbon in fly ash. Brown and Dykstra [42] found that

Table 2Bulk composition (wt%) of fly-ash samples.

CFA 15CBFA 66CBFA

Total 96.3 97.1 97.1SiO2 43.0 45.2 43.6Al2O3 21.0 21.5 21.0Fe2O3 4.2 4.0 3.9MgO 2.60 2.50 2.91CaO 14.5 13.6 14.5Na2O 7.50 7.30 7.46K2O 0.60 0.70 1.17TiO2 0.90 1.00 0.89P2O5 0.60 0.60 0.67MnO 0.02 0.02 0.12Cr2O3 <0.01 <0.01 <0.01V2O5 0.02 0.03 0.02S 0.62 0.26 0.51SiO2 + Al2O3 + Fe2O3 68.2 70.7 68.5LOI 0.8 0.4 0.9

the accuracy of LOI measurements is significantly affected by smallamounts of calcite and portlandite in the fly ash, which lose weightdue to calcination and dehydration under the high-temperatureoxidation conditions of the LOI test. The presence of calcite in theAtikokan fly ashes is suggested by the results of XRD analysespresented below. Fly-ash compositions closely resembled that offresh coal fly-ash collected from the same boiler at AGS at anearlier date (prior to biomass testing) and reported by Yeheyiset al. [43].

Table 3 shows the concentrations of trace elements in fly-ashsamples and their relative changes due to co-firing biomass Theconcentrations of Ba, Be, Cu, Pb, Y and Zn consistently increasedwith the percent biomass in the fuel. The largest increase occurredfor Zn: its concentration in 66CBFA was 305% higher than in CFA.However, the concentrations of all trace elements in coal-biomassfly-ashes were within the ranges reported for conventional coalfly-ashes in various countries [44,45].

XRD spectra of all fly-ashes exhibited a strong hump in the 18–38� 2h region, revealing the predominance of amorphous phases.The principal crystalline component were quartz (SiO2), followedby some periclase (MgO), and possible traces of anhydrite (CaSO4),anorthite ((Ca,Na)(Al,Si)2Si2O8), belite (Ca2SiO4), calcite (CaCO3),feldspar (K0.5Na0.5AlSi3O8), gehlenite (Ca2All2SiO7), hematite(Fe2O3), lime (CaO), and mullite (Al4.54Si1.46O9.73).

The specific gravities measured for CFA, 15CBFA and 66CBFAwere 2.404 ± 0.048, 2.361 ± 0.036 and 2.587 ± 0.064 g cm�3,respectively. The higher value obtained for 66CBFA was attributableto a comparatively low content of porous particles (cenospheres).The cenosphere fraction of fly-ash is affected by combustion tem-perature and fuel properties such as mineral impurities [46–48].

Particle size distributions (Fig. 1) show that 15CBFA had a great-er proportion of large (>50 lm) particles than either CFA or 66CBFAwhich, in turn, had very similar particle size distributions. CFA hada lower median diameter (D50 = 14 lm) than the fresh coal fly-ashexamined by Yeheyis et al. [43] (D50 = 20 lm), which illustrates thevariability in particle size distribution of fly-ash samples taken atdifferent times. Since 15CBFA was collected with the boiler run-ning at full load, larger particles may have been entrained in thesteam being forced through the soot blowers. Conversely, bothCFA and 66CBFA were collected at 50% load and probably cooled

Ba 3800 3900 (2)* 6700 (76)Be 3.7 4.2 (14) 4.7 (27)Bi <20 <20 <20Cd <2 <2 <2Co 17 16 (�6) <25Cu 37 41 (11) 49 (32)Hg <0.3 <0.3 <0.3Li 20 16 (�20) 57 (185)Mo 13 10 (�23) <15Ni 28 27 (�4) 31 (11)Pb 39 48 (23) 55 (41)Sb <10 <10 <10Se <30 <30 <30Sn <20 <20 <20Sr 3500 3300 (�6) 3600 (3)Tl <30 <30 <30U <20 <20 <20Y 41 45 (10) 47 (15)Zn 37 53 (43) 150 (305)Cl 30 20 (�33) 63 (110)

* Numbers in parentheses represent the relative percent changes in concentrationcompared to CFA.

0

1

2

3

4

0.1 1 10 100 1000 Particle Size (µm)

Volu

me

(%)

CFA

15CBFA

66CBFA

Fig. 1. Particle size distribution of fly-ash samples. Size measurements wereaveraged over three replicate samples, each of which was analyzed in triplicate.

A. Johnson et al. / Fuel 89 (2010) 3042–3050 3045

more rapidly, favouring formation of smaller fly-ash particles. Thefact that both the CFA and 66CBFA shared the same particle sizedistribution would imply that co-firing biomass with coal did notdirectly result in a disparity in particle size distribution. These re-sults suggest that combustion conditions (loading, temperature,

Fig. 2. Back-scattered electron image of 15CBFA.

(a) AlSiCa

<5um

>5um

AlSi

Ca0

0

0

20

20

2040

40

4060

60

60

80

80

80

100

100

100

Fig. 3. Ternary diagrams for CFA comparing the relative molar compositi

soot blowers, etc.) affected the particle size distribution of fly-ashto a much larger extent than fuel composition.

Fig. 2 shows a typical back-scattered electron image of 15CBFA.Most fly-ash particles were spherical as a result of rapid cooling inthe post-combustion zone of the boiler, and a significant propor-tion is porous. Differences in grey levels reveal compositional vari-ations; light-shaded particles tend to be high in calcium, iron,magnesium or titanium, while darker particles tend to have lessof these elements and more sodium. Back-scattered electronmages of CFA and 66CBFA (not shown) revealed similarmicrostuctures.

Over 85 quantitative EDS measurements of each fly-ash wereused to create ternary compositional diagrams (Figs 3–5). Thesediagrams show the relative proportion of each of three elements(e.g., Al, Ca and Si) for fly-ash particles with diameter smaller orgreater than 5 lm. Overall, particle compositions did not markedlychange when co-firing up to 62 wt% wood pellets. Moreover, thecompositions of small and large particles were statistically indis-tinguishable in all fly-ashes. Aluminosilicates, represented bypoints close to the Si–Al binary line, were abundant in all fly-ashes.Calcium aluminosilicates correspond to points on tie lines connect-ing the aluminosilicate binary and the Ca-apex in Si–Al–Ca dia-grams. Most calcium aluminosilicate particles fall on tie lineshaving Si/Al molar ratios ranging from 1 to 2. The Si–Al–Fe ternarydiagrams also reveal the presence of iron-rich aluminosilicate par-ticles, with some of these particles having a relative Fe content ashigh as 90%.

3.2. Effect of ash substitution and AEA addition on the air content ofmortars

Figs. 6 and 7 show the air content of mortars containing 0, 20and 40 wt% fly-ash as a function of AEA addition. The entrainedair content in ash-free mortar samples rose from about 5 to 10vol% upon addition of 0.6 mL/kg AEA, but did not increase furtherwhen the AEA dosage was doubled. The air content of AEA-amended mortar was totally unaffected by fly-ash substitution,presumably due to its very low carbon content. However, in the ab-sence of AEA, use of 20% ash decreased the air content by roughly1% (Fig. 6), whereas 40% ash caused it to increase 2–2.5% (Fig. 7). Asimilar trend of air content versus mixed cement substitution byfly-ash was reported by Chindaprasirt et al. [39], although mostother studies show a uniform decrease in air content with increas-ing fly-ash [5,7,49–51]. The fly-ash used in Chindaprasirt et al. [39]

(b) AlSiFe

<5um

>5um

AlSi

Fe0

0

0

20

20

2040

40

4060

60

60

80

80

80

100

100

100

ons of individual smaller and larger particles: (a) AlSiCa and AlSiFe.

(a) AlSiCa

<5um

>5um

AlSi

Ca0

0

0

20

20

2040

40

4060

60

60

80

80

80

100

100

100

(b) AlSiFe

<5um

>5um

AlSi

Fe0

0

0

20

20

2040

40

4060

60

60

80

80

80

100

100

100

Fig. 4. Ternary diagrams for 15CBFA comparing the relative molar compositions of individual smaller and larger particles: (a) AlSiCa and AlSiFe.

(a) AlSiCa

<5um

>5um

AlSi

Ca0

0

0

20

20

2040

40

4060

60

60

80

80

80

100

100

100

(b) AlSiFe

<5um

>5um

AlSi

Fe0

0

0

20

20

2040

40

4060

60

60

80

80

80

100

100

100

Fig. 5. Ternary diagrams for 66CBFA comparing the relative molar compositions of individual smaller and larger particles: (a) AlSiCa and AlSiFe.

0

5

10

15

0 0.2 0.6 1.2

AEA Addition (mL/kg of cemtitous material)

Air

Con

tent

(%

)

0% Fly-ash

20% CFA

20% 15CBFA

20% 66CBFA

Fig. 6. Air content of mortars containing 0 and 20 wt% fly-ash. Air content wasaveraged from quadruplicate analyses. Error bars represent ±0.5 standard deviation.

0

5

10

15

0 0.2 0.6 1.2AEA Addition (mL/kg cementitous material)

Air

Con

tent

(%

)

0% Fly-ash

40% CFA

40% 15CBFA

40% 66CBFA

Fig. 7. Air content of mortars containing 0 and 40 wt% fly-ash. Air content wasaveraged from quadruplicate analyses. Error bars represent ±0.5 standard deviation.

3046 A. Johnson et al. / Fuel 89 (2010) 3042–3050

had a similar composition to those used in the present study andan LOI of 2.07 wt%.

3.3. Compressive strength development

Figs. 8 and 9 illustrate the development of compressive strengthof the mortar specimens over a 90 days curing period. The strengthof the ash-free mortar increased much faster than that of the other

specimens over the first 7 days, but thereafter increased veryslowly. These observations are consistent with previous reports[3,28–34] and are explained by the slow pozzolanic reaction offly-ashes.

Mortars containing the same percent cement substitution byfly-ash initially gained strength at the same rate. At day 28, how-ever, the mortars containing CFA or 66CBFA were 6–17% strongerthan mortars containing 15CBFA. Since the bulk composition ofall fly-ashes was similar, this phenomenon may have been caused

50.4

50.5

50.6

50.7

50.8

0 25 50 75 100 125 150

Cycles

Leng

th (

mm

)

0% Fly-ash

20% CFA

20% 15CBFA

Fig. 10. Average length of mortar cubes subjected to repeated freeze–thaw cycles.Length was averaged from two measurements on quadruplicate analyses. Error barsrepresent ±0.5 standard deviation.

0

5

10

15

20

25

30

35

40

45

50

0 20 40 60 80 100

Time (days)

Com

pres

sive

Str

engt

h (M

Pa)

0% Fly Ash 20% CFA

20% 15CBFA 20% 66CBFA

Fig. 8. Compressive strength development for mortars having 0 or 20 wt% fly-ashsubstitution. Compressive strength was averaged from quadruplicate analyses.Error bars represent ±0.5 standard deviation.

0

5

10

15

20

25

30

35

40

45

50

0 20 40 60 80 100

Time (days)

Com

pres

sive

Str

engt

h (M

Pa)

0% Fly Ash 40% CFA

40% 15CBFA 40% 66CBFA

Fig. 9. Compressive strength development for mortars having 0 or 40 wt% fly-ashsubstitution. Compressive strength was averaged from quadruplicate analyses.Error bars represent ±0.5 standard deviation.

A. Johnson et al. / Fuel 89 (2010) 3042–3050 3047

by differences in their particle size distributions. Indeed, the CFAand 66CBFA had a larger proportion of small (<30 lm) particlesand thus a larger reactive surface area which may have enhancedthe pozzolanic reaction rate compared to mortars containing15CFA. Nevertheless, all fly-ashes met ASTM requirement C 618[8e], that is, the 28-day compressive strength of mortars contain-ing fly-ash exceeded 75% that of mortar containing no fly-ash. At90 days of curing, the compressive strength of fly-ash amendedmortars met or exceeded the strength of ash-free mortar.

3.4. Resistance to freezing and thawing

The average length, mass and compressive strength of mortarcubes subjected to repeated freeze–thaw cycles are shown inFigs. 10–12, respectively. The first two parameters decreasedslightly for all mortars after 140 cycles, whereas compressivestrength decreased by 40–46%. No significant differences infreeze–thaw resistance were observed between ash-free mortarsand mortars containing 20 wt% CFA or 15CBFA, presumably be-cause they all had similar air content (Fig. 6), i.e., consistent withprevious studies [5,21,52] showing that specimens with controlledair content have equal tolerance to rapid freeze–thaw cycles.

265

270

275

280

285

0 25 50 75 100 125 150Cycles

Mas

s (g

)

0% Fly-ash

20% CFA

20% 15CBFA

Fig. 11. Average mass of mortar cubes subjected to repeated freeze–thaw cycles.Mass was averaged from quadruplicate analyses. Error bars represent ±0.5 standarddeviation.

15

20

25

30

35

40

45

0 25 50 75 100 125 150

Cycles

Com

pres

sive

Str

engt

h (M

Pa)

0% Fly-ash

20% CFA

20% 15CBFA

Fig. 12. Average compressive strength of mortar cubes subjected to repeatedfreeze–thaw cycles. Compressive strength was averaged from quadruplicateanalyses. Error bars represent ±0.5 standard deviation.

3048 A. Johnson et al. / Fuel 89 (2010) 3042–3050

3.5. Microstructure and micromineralogy of mortar specimens

Figs. 13 and 14 show typical back-scattered electron images of28 days cured mortar containing 20 wt% CFA or 15CBFA, respec-tively. Partly reacted fly-ash particles are recognizable by theirspherical shapes. They generally consisted of an un-reacted coresurrounded by a layer of smooth-textured hydration products richin calcium and silicon. In some cases, fly-ash particles had com-pletely reacted and only a spherical layer of hydration products re-mained. This material is designated ‘‘fly-ash inner-CSH” because itsmorphology is similar to that of the inner-CSH layers surroundingpartly unhydrated cement grains. Fly-ash inner-CSH consistsmainly of Ca and Si, with an average molar Ca/Si ratio of2.22 ± 0.73 (determined from 38 independent EDS measurements),along with, in descending order of abundance, Al > Fe > Mg � Na.This composition is similar to that reported for inner-CSH in OPC[53]. The fly-ash inner-CSH layers are either in direct contact withthe un-reacted core (e.g., see the fly-ash particle on the left handside of Fig. 13) or separated by a thin gap.

Much of the void space that was originally filled with water isoccupied by calcium hydroxide (CH) crystals, irregularly textured‘‘outer-CSH”, and other hydration products which are collectivelytermed ‘‘dense hydration product” (DHP). CH also forms an irregularcoating on sand grains (e.g., bottom of Fig. 14), as observed by Dia-mond [53] in concrete. Outer-CSH has an average Ca/Si molar ratio

Fig. 13. Back-scattered electron image of 20% CFA mortar cured for 28 days.

Fig. 14. Back-scattered electron image of 20% 15CBFA mortar cured for 28 days.

of 1.96 ± 0.35 (determined from 65 independent EDS measure-ments), which is at the very high end of the range (0.5–1.7) reportedby Wang et al. [23]. DHP consists of smooth-textured, irregularlyshaped particles (e.g., see DHP particle near the left top corner ofFig. 13), varying widely in size and composition (generally withCa > Si > Al > Fe � S � Na, and often approaching that of ettringite,Ca6Al2(SO4)3(OH)12�26H2O). There were no distinguishable differ-ences between the DHP found in CFA- and CBFA-amended mortars.

Using EDS analysis, the Ca/Si molar ratio was determined for anumber of points situated along lines that transect individualCFA and 15CBFA particles along with their shells of inner-CSH(Figs. 15–17). The comparatively high Ca/Si ratio of the fly-ash in-ner-CSH shell indicates that it is formed by reaction of fly-ash withCa ions in the interstitial fluid originating from cement or CHdissolution.

4. Conclusions

1. Co-combustion of lignite coal with up to 62 wt% wood pelletsdid not markedly change the bulk composition of fly-ash northe composition of individual fly-ash particles, owing to the

Fig. 15. Back-scattered electron image of reacted fly-ash particle in 20% CFA mortar,showing positions at which EDS analysis was carried out. Sites 4, 5 and 6 are in theun-reacted core of a CFA particle that is encased by a shell of fly-ash inner-CSH.

Fig. 16. Back-scattered electron image of reacted fly-ash particle in 20% 15CBFAmortar, showing positions at which EDS analysis was carried out. Sites 4 and 5 arein the un-reacted core of a 15CBFA particle that is encased by a shell of fly-ashinner-CSH.

0

0.5

1

1.5

2

2.5

3

0 1 2 3 4 5 6 7 8 9

Position

Ca/

Si (

mol

ar)

20% CFA

20% 15CBFA

Fig. 17. Ca/Si molar ratio profile for reacted CFA and 15CBFA particles in Figs. 15and 16.

A. Johnson et al. / Fuel 89 (2010) 3042–3050 3049

very low ash content of wood pellets. All fly-ashes had LOI val-ues less than 1 wt% and thus complied with ASTM regulationsregarding LOI. Concentrations of toxic metals and other minorelements in all fly-ashes were within the ranges reported forother coal fly-ashes.

2. Combustion conditions (e.g., loading) affected the particle sizedistribution of fly-ash, with higher loads increasing the propor-tion of larger particles.

3. Substitution of cement with up to 40% fly-ash from coal orwood-coal combustion did not affect the entrained air contentof mortar containing P0.6 mL/kg AEA because of the low car-bon content of the fly-ashes. In the absence of AEA, however,use of 20% ash decreased the air content by roughly 1%, whereas40% ash caused it to increase 2–2.5%.

4. The compressive strength of mortars in which up to 40% of cementis substituted by co-combustion fly-ash exceeded 75% that of ash-free mortars by 28 days, and approached or even surpassed thecompressive strength of ash-free mortars after 90 days of curing.Hence, all co-combustion fly-ashes met ASTM C 618 strengthrequirements. Fly-ash that contained a higher proportion of large(>50 lm) particles exhibited the lowest 28-day strength.

5. Amending mortar with 20 wt% CFA or 15CBFA had no effect onits durability following repeated freeze–thaw cycles when aircontent was kept constant.

6. No micromineralogical differences were observed betweenhydrated CFA- and CBFA-amended mortars.

Acknowledgements

This work was supported by the Ontario Centres of Excellence(ABRC Theme 1 project), the Ontario Ministry of Energy, and theNatural Sciences and Engineering Research Council of Canada. Dr.Nabajyoti Saikia is acknowledged for his contributions to XRDand particle size distribution analyses.

References

[1] Independent electricity system operator. The Ontario reliability outlook,Toronto, Ontario. Available from: <http://www.ieso.ca/imoweb/pubs/marketReports/ORO_Report-Dec2008.pdf>; December, 2008 [accessed May2009].

[2] Ontario power generation. Sustainable development report. Available from:http://www.opg.com/pdf/sustainable%20development%20reports/sustainable%20.development%20Report%202007.pdf; 2007 [accessed May 2009].

[3] Ipatti A. Peat fly-ash as a supplementary cementing material in concrete.Nordic Concrete Res 1998;7:152–66.

[4] Senneca O. Characterization of meat and bone mill for coal co-firing. Fuel2008;87:3262–70.

[5] Wang S, Miller A, Llamazos E, Fonseca F, Baxter L. Biomass fly-ash in concrete:mixture proportioning and mechanical properties. Fuel 2008;87(3):365–71.

[6] Canadian industries recycling coal ash. Fly-ash: its origin, applications and theenvironment. Available from: <http://circainfo.ca/pdf/CIRCA%20Fact%20Sheet%201%20-%20CTFinal.pdf>; 2002 [accessed May 2009].

[7] Bouzoubaa N, Foo S. Use of fly-ash and slag in concrete: a best practiceguide. Canmet, Ottawa, ON, Canada: Canmet Materials Testing Laboratory;2004.

[8] American society for testing and materials. Concrete and aggregates, vol. 04.02.West Conshohocken, PA, United States, 2003. a C 109/C 109M-02 standard testfor compressive strength of hydraulic cement mortars (using 2-in or [50 mm]cube specimens). b C 185-02 standard test method for air content of hydrauliccement mortar. c C260-06 standard specification for air-entraining admixturesfor concrete. d C 311-07 standard test methods for sampling and testing fly ashor natural pozzolans for use in Portland-cement concrete. e C 618-03 standardspecification for coal fly-ash and raw or calcined natural pozzolan for use inconcrete. f C 666/C 666M-03 standard test method for resistance of concrete torapid freezing and thawing. g C 778-03 standard specification for standardsand. h C 1437-01 standard test method for flow of hydraulic cement mortar.

[9] Fouad FH, Copham CA, Donovan JM. Evaluation of concrete containing fly ashwith high carbon content and/or small amounts of wood. AL, UnitedStates: Department of Civil Engineering, The University of Alabama atBirmingham; 1998.

[10] Yu Q, Sawayana K, Sugita S, Shova S, Shova M, Isojima Y. The reaction betweenrice husk ash and Ca(OH)2 solution: nature of its products. Cement ConcreteRes 1999;29:37–43.

[11] Naik TR, Kraus RN. A new source of pozzolanic material. Concrete Int2003;25:55–62.

[12] Martirena F, Middendorf RL, Day M, Gehrke M, Roque P, Martinez L, et al.Rudimentary, low tech incinerators as a means to produce reactive pozzolanout of sugarcane straw. Cement Concrete Res 2006;36:1056–61.

[13] vom Berg W, Feuerborn HJ. Present situation and perspective of CCPmanagement in Europe. European Coal Combustion Products Association;2003.

[14] Masia AAT, Buhre BJP, Gupta RP, Wall TF. Use of TMA to predict depositionbehaviour of biomass fuels. Fuel 2007;86:2446–56.

[15] Bridgeman TG, Darvell LI, Jones JM, Williams PT, Fahmi R, Bridgewater AV,et al. Influence of particle size on the analytical and chemical properties of twoenergy crops. Fuel 2006;86:60–72.

[16] Grammelis P, Skodras G, Kakaras E. Effects of co-firing with coal on ashproperties, part I: characterization and PSD. Fuel 2006;85:2310–5.

[17] Berry EE. Fly-ash for use in concrete, part I – a critical review of the chemical,physical and pozzolanic properties of fly-ash. Canmet report. Ottawa, ON,Canada: Minerals Research Program; 1976.

[18] Gaudry D. Personal communication 2008.[19] American coal ash association. Fly ash facts for highway engineers. US

Department of Transportation, FHWA-IF-03-019, 2003.[20] Meij R, van den Berg J. Coal fly ash management in Europe: trends, regulations

and health and safety aspects. In: International ash utilization symposium.Available from: <http://www.flyash.info/2001/keynote/108meij.pdf>; 2001(accessed September 2008).

[21] Berry EE, Malhotra VM. Fly-ash for use in concrete, part II – a critical review ofthe effects of fly-ash on the properties of concrete. Canmet report. Ottawa, ON,Canada: Minerals Research Program; 1978.

[22] Mindness S, Young JF, Darwin D. Concrete. 2nd ed. Upper Saddle River, NJ,United States: Pearson Education; 2003.

[23] Wang S, Miller A, Fonseca F, Baxter L. Biomass fly-ash in concrete: SEM, EDXand ESEM analysis. Fuel 2008;87:372–9.

[24] Taylor HFW. Cement chemistry. 2nd ed. London, United Kingdom: ThomasTelford, Heron Quary; 1997.

[25] Hewlett P. Lea’s chemistry of cement and concrete. 4th ed. Jordan Hill, Oxford,United Kingdom: Butterworth-Heinemann Linacre House; 1997.

[26] Ghose A, Pratt PL. Studies of the hydration reactions and microstructure ofcement-fly-ash paste. In: Diamond S, editor. The conference ‘‘effects of fly-ashincorporation in cement and concrete”. 1981. p. 82–91.

[27] Manz OE. Coal fly-ash: a retrospective and future look. Fuel 1999;78(2):133–6.[28] Washa GW, Withey NH. Strength and durability of concrete containing

Chicago fly-ash. Proc Am Concrete Inst 1953;49:701–12.[29] Lamond JF. Twenty-five years’ experience using fly-ash in concrete in: fly-ash.

Silica fume, slag and other mineral by-products in concrete, vol. I SP 79. 1983.p. 47–70.

[30] Samarin A, Munn RL, Ashby JB. The use of fly-ash in concrete- Australianexperience. Fly-ash, silica fume, slag and other mineral by-products inconcrete, vol. I SP 79. 1983. p. 143–72.

[31] Costa U, Massazza F. Some properties of pozzolanic cements containing fly-ashes. Fly-ash, silica fume, slag and other mineral by-products in concrete, vol.I SP 79. 1983. p. 235–54.

[32] Abdun-Nur EA. Fly-ash in concrete: an evaluation. Bulletin no.284. Washington, DC, United States: Highway Research Board; 1961.

[33] Pasko TJ, Larson TD. Some statistical analysis of the strength and durability offly-ash concrete. Proc ASTM 1962;62:1054–67.

[34] Wang S, Baxter L. Comprehensive study of biomass fly-ash in concrete:strength, microscopy, kinetics and durability. Fuel Process Technol2007;88(11–12):1165–70.

3050 A. Johnson et al. / Fuel 89 (2010) 3042–3050

[35] Wang S, Miller A, Llamazos E, Fonseca F, Baxter L. Durability of biomass fly-ashconcrete: freezing and thawing and rapid chloride permeability tests. Fuel2008;87:359–64.

[36] Pedersen KH, Jensen AD, Skjoth-Rasmussen MS, Dam-Johansen K. A review ofthe interference of carbon containing fly ash with air entrainment in concrete.Prog Energy Combust Sci 2008;34:135–54.

[37] American concrete institute committee 201. Durability of concrete in service,chapter I: freezing and thawing. Proc Am Concrete Inst 1962;59:1771–84.

[38] Bloem DL. Effect of fly-ash in concrete. National ready mixed concreteassociation bulletin 48, 1954.

[39] Chindaprasirt P, Buapa N, Cao HT. Mixed cement containing fly ash formasonry and plastering work. Construct Build Mater 2005;19(8):612–8.

[40] Euclid Chemical Company. Airextra technical data sheet. Available from:<http://www.euclidchemical.com/fileshare/ProductFiles/techdata/airextra.pdf> (accessed May 2009).

[41] American Cube Mold. Unpublished: comparison study. Twinsburgh, OH,United States; 1994.

[42] Brown RC, Dykstra J. Systematic errors in the use of loss-on-ignition tomeasure unburned carbon in fly ash. Fuel 1995;74:570–4.

[43] Yeheyis MB, Shang JQ, Yanful EK. Characterization and environmentalevaluation of Atikokan coal fly ash for environmental applications. J EnvironEng Sci 2008;7:481–98.

[44] Sushil S, Batra VS. Analysis of fly ash heavy metal content and disposal in threethermal power plants in India. Fuel 2006;85:2676–9.

[45] Goodarzi F. Characteristics and composition of fly ash from Canadian coal-firedpower plants. Fuel 2006;85:1418–27.

[46] Ghosal S, Ebert JL, Self SA. Chemical composition and size distribution for flyashes. Fuel Process Technol 1995;44:81–94.

[47] Bibby D. Composition and variation of pulverized fuel ash obtained from thecombustion of sub-bituminous coals in New Zealand. Fuel 1977;56:427–31.

[48] Ngu L, Wu H, Zhang D. Characterization of ash cenospheres in fly ash fromAustralian power stations. Energy Fuel 2007;21:3437–45.

[49] Wesche K. Fly ash in concrete: properties and performance, 1st ed. London,United Kingdom: E&FN Spoon; 1991. p. 3–24, 42–62, 117–43.

[50] Helmuth R. Fly ash in cement and concrete 1st ed. Portland CementAssociation; 1987. p.58-64, 80-2.

[51] Zhang D. Air entrainment in fresh concrete with pfa. Cement Concrete Compos1996;18(6):409–16.

[52] Brown PW, Clifton JR, Frohnsdorff G, Berger RL. Limitations to fly-ash use inblended cements. In: Proceedings of the 4th international ash utilizationsymposium, St. Louis, MO, United States, March 24–25, ERDA MERC/SP-76/4;1976. p. 518–29.

[53] Diamond S. The microstructure of cement paste and concrete – a visual primer.Cement Concrete Compos 2004;26:919–33.