Waste Management & Research (1996) 14, 15–28

PHYSICO-CHEMICAL CHARACTERIZATION AND LEACHINGOF DESULPHURIZATION COAL FLY ASH

I. Lecuyer1, S. Bicocchi1, P. Ausset2 and R. Lefevre2

1Electricite de France, Direction des Etudes et Recherches, Departement Environnement, 6 QuaiWatier, 78400 Chatou, France and 2Laboratoire Interuniversitaire des Systemes Atmospheriques,

Universite Paris XII, Val de Marne, URA CNRS 1404, 61 Av. du General deGaulle, 94010 Creteil Cedex, France

(Received 26 October 1994, accepted in revised form 24 January 1995)

Fly ash produced by coal combustion using two types of desulphurization processwere studied: a conventional pulverized coal boiler equipped with lime injection(PCL ash), and a circulating fluidized bed combustion boiler with limestone injection(CFBC ash). The ashes were characterized completely: granulometry, morphology,mineralogy, chemical composition and behaviour to water contact.

Both PCL ash and CFBC ash present similar features: fine granulometry, presenceof anhydrite phase and sulphate content. However, PCL ash also shows lots ofspherical particles, unlike CFBC ash, and a much higher lime content, due to thelower desulphurization rate in PC boilers. Unlike CFBC ash, most of the traceelements in PCL ash show an inverse concentration–particle size dependence.

Leachates obtained from both samples are rich in soluble salts [CaSO4 andCa(OH)2] and arsenic and selenium are prevented from solubilizing by high limecontent.

In wetted PCL ash, the formation of ettringite crystals stabilizes calcium andsulphate ions. Simultaneously, arsenate, selenate and chromate anions are trappedin the crystal. CFBC ash does not really harden because the lime content is too low.However, the leached selenium concentration is cut down in wetted CFBC ashsamples. 1996 ISWA

Key Words—Fly ash, desulphurization, trace element, leaching, furnace sorbentinjection, circulating fluidized bed combustion, arsenic, selenium,chromate.

1. Introduction

Following the recent regulatory decisions concerning sulphur dioxide exhausts (SO2),coal-fired power plants have to be equipped with a desulphurization system. Thereduction of the SO2 emissions from Electricite de France (EDF) coal-fired powerplants has been achieved following two strategies:

(1) The equipment of existing units with:(i) Wet flue gas desulphurization for 600 MW unit; and(ii) Furnace sorbent injection for 250 MW units.

(2) The development of circulating fluidized bed combustion (CFBC) units.

The last two processes are based on injection of an alkaline reagent (lime or limestone)into the combustion furnace (direct furnace sorbent injection). These produce solid by-products (fly ash and bottom ash) which are significantly different from conventional

0734–242X/96/010015+14 $12.00/0 1996 ISWA

I. Lecuyer et al.16

pulverized coal (PC) combustion fly ash. These new types of ashes result from themixture of the coal combustion ash with the reacted lime or limestone. These by-products consist of 10–40% bottom ash and 60–90% fly ash which are collected by theflue gas particulate filters.

The enhancement of disposal regulations in France and Europe, together with theincreasing costs of disposal, make it necessary to know the physico-chemical propertiesand the environmental impact of these fly ashes during storage, even temporary.

A complete physico-chemical characterization of two types of desulphurization flyash, and one of conventional PC fly ash for comparison, was performed using differentanalytical methods. The environmental impact during storage was mainly approachedby means of the French standard leaching test.

2. Materials and methods

2.1 Sampling

The samples studied here come from two units:

(1) The 125 MW circulating fluidized bed combustion (CFBC) unit of the Emile Huchetpower plant (Charbonnages de France) situated in Carling (Lorraine). This unitburns coal residues from the mine close to the power plant (schlamms) and a slurryof coal and water. The sorbent used is limestone added with a Ca/S molar ratio of2.3, where S is the coal sulphur content. The ash samples coming from this unitare referred to as CFBC ash in the text.

(2) The 250 MW electric power plant of Loire-sur-Rhone (France) which is a pulverizedcoal (PC) unit with furnace sorbent injection. The fuel used is U.S. coal (13% ashand 0.97% sulphur) and the sorbent is hydrated lime [Ca(OH)2]. The fly ash obtainedby the desulphurization process (Ca/S>3.5–4.2), PCL ash in this text, was comparedto the PC ash collected when no sorbent injection was performed.

All samples were collected during stabilized runs of the units.

2.2 Particle size distribution

Particle size distribution of fly ash was determined by Fraunhofer diffraction of a laserbeam (Coulter LS 130), which allows analysis of particle sizes between 0.1 and 900 lm.

2.3 Analytical scanning electron microscopy: morphology, granulometry andsemi-quantitative individual analysis of particles

Morphology and granulometry of the three kinds of fly ash were observed by scanningelectron microscopy (SEM) (Jeol JSM 840A). Individual semi-quantitative elementaryanalysis of particles was performed by a device based on X-ray energy dispersivespectrometry (EDX) (Tracor TN 5400). This EDX system provides information aboutthe elements of atomic number equal to or higher than that of Na. EDX data isreported as a normalized weight percent value of oxides based on only the analysedelements. For this reason, all EDX analyses have been normalized to 100% of theelements detected and not to the total weight of the particle, as in bulk chemicalanalysis.

Desulphurization coal fly ash 17

TABLE 1Ash mineralization method test with certified sample NIST 1633B

Variation in %, withCertified values given Analysed values respect to certified

Elements by NIST in lg l−1 obtained in lg l−1 value

As 136.2±2.6 142.8±8.1 +4.9Cd 0.784±0.006 0.716±0.126 −8.7Cr 198.2±4.7 204±5 +3Cu 112.8±2.6 116.5±14 +3.3Ni 120.6±1.8 113.3±3.9 −6.1Pb 68.2±1.1 66.2±3.7 −3Zn 210∗ 207.2±12.1 −1.4

∗Non-certified value.

2.4 Mineralogy

The crystalline phases of fly ash were analysed by powder X-ray diffraction (XRD).XRD analyses of fly ash samples were performed with a Rigaku Miniflex 2005diffractometer.

2.5 Bulk chemical analysis

The complete chemical analysis of the major and minor components was achieved byspecific analyses which were controlled by several round-robin tests (Soreau & Jacob1993). Particular attention was paid to trace element analysis.

The trace element analysis in fly ash required a preliminary preparation of aciddigestion (mineralization) performed with a mixture of three acids (3 ml HCl 30%, 4 mlHF 40%, 2 ml HNO3 65%) for an ash sample weight of 100 mg. These acids of analyticalquality (Prolabo Normaton) were placed with the ash sample in a round flask with ateflon condenser. The mixture was heated in a Prolabo microwave digester A301. Thetrace element analysis was then performed by atomic absorption spectroscopy (AAS),using a Varian Spectra AA 800 with either flame or graphite furnace mode (VarianGTA 100) according to the concentrations.

Previous work has helped to optimize the trace element analysis by AAS in a matrixcontaining large quantities of salts (sulphates, chlorides, fluorides), lime and acids. Themethods used in this study are shown to be suitable for analysing these fly ashes.

This method was controlled by analysis of a sample of standard homogenized electricutility fly ash from the National Institute of Standards and Testing (NIST 1633b). Theresults presented in Table 1 indicate that this experimental method is suitable for theanalysis of As, Cd, Cr, Cu, Ni, Pb and Zn.

2.6 Leaching test

The environmental impact of fly ash storage was assessed by means of the Frenchstandardized leaching test, in order to meet regulatory requirements concerning landfillsand disposals.

I. Lecuyer et al.18

Vol

um

e (%

)

1000

4.0

0

Particle diameter (µm)400

1.0

10 20 100 20060401 2 640.2 0.4

1.5

2.0

2.5

3.0

3.5

0.5

Fig. 1. Particle size distribution of fly ash samples using laser diffraction technique. Solid line, pulverizedcoal combustion fly ash; dotted line, pulverized coal combustion with lime injection fly ash; dashed line,

circulating fluidized bed combustion with limestone injection fly ash.

The French standard leaching test (AFNOR X31-210) is similar to the EuropeanDIN test 38414-S4. This test uses demineralized water with a liquid:solid mass ratio of10:1. The mixture is shaken in a closed vessel for 24 h, the leachate is filtered andsubsequently analysed.

The trace elements, As, Cd, Cu, Cr, Hg, Ni, Pb, Se and Zn, were analysed by AASand CrVI by the colorimetric method according to the French standard test (AFNORNFT 90-043).

3. Results and discussion

3.1 Granulometry, morphology and micro-analysis

Particle size distribution of fly ash is of particular importance when the material is usedin the cement or concrete industries, which are the main users of fly ash.

Laser granulometry measurements were taken for the three ash samples. The resultsare shown in Fig. 1. One specific parameter of the distribution is d50, which is themedian diameter: 50% of particles (by volume) are smaller than d50.

The particle size distribution of PC ash had a single peak (at 67 lm), with d50>28 lm.The corresponding PCL ash had a granulometry shifted towards small sizes (d50>13 lmand maximum peak also at 13 lm). This was due to the small size (about 7 lm) of theinjected sorbent particles. The granulometric distribution of PCL ash seems to be dueto the combination of PC ash distribution and reacted sorbent distribution.

CFBC ashes are as fine as PCL ashes (d50>15 lm) and showed a bimodal distribution(peaks at 4 and 39 lm). This feature has been observed previously on sample ash ofthis unit (Carles-Gibergues & Delsol 1994). This was likely due to a difference ofgranulometry between the specific fuel (d50>80 lm) and the injected limestone(d50>40 lm).

Desulphurization coal fly ash 19

SEM observations showed that both PC ash and PCL ash are mainly composed ofmicrospheres of various diameters [Fig. 2(a, b) ]. These microspheres are characteristic ofcoal combustion ash when the temperature is high (T=1200–1400°C). Small irregularly-shaped particles were more abundant in the PCL ash sample than in the PC ash sample[Fig. 2(a) ].

The CFBC ash particle morphology was quite different from the previous samples:almost no microspheres could be identified [Fig. 2(c) ]. Most of the particles wereirregularly-shaped, with variable granulometry. The very low abundance of microspheresand the irregular shape of most particles is usually attributed (Behr-Andres et al. 1993)to the lower combustion temperature (T=850°C), which does not allow ash melting.More detailed description and analytical results will be given in a further paper.

It was not possible to single out a crystal of pure anhydrite (CaSO4). A wetted sampleof CFBC ash, in fact, the solid residue from leachate filtration, showed the formationof acicular crystals [Fig. 2(d)]. These needles are composed of Al, S and Ca, with tracesof Si likely due to the background matrix, which was close to ettringite composition[Ca6Al2(SO4)3(OH)12, 25H2O]. However, no ettringite could be detected by XRD as theconcentration was probably less than the XRD detection limit.

Ettringite crystals could not be observed on leached PCL ash samples neither bySEM nor by XRD. The general aspect of the sample remained unchanged throughhydration. No crystal of pure Ca(OH)2 could be detected either, even though portlanditewas detected by XRD.

3.2 Crystalline phases

The main crystalline phases observed by XRD are described in Table 2 as a functionof their relative abundances. In addition to the phases usually found in fly ash (quartz,mullite) issued from the quartz and kaolinite phases existing in the fuel, calciumcompounds were detected: anhydrite, calcite in PCL and CFBC ashes, and lime in PCLash. This lime was hydrated into portlandite in the leached PCL ash sample andanhydrite was totally solubilized. The CFBC ash contained illite rather than mullite:this should be due to the specific fuel used in this unit (coal residues). When wetted, itstill showed anhydrite but ettringite could not be detected using XRD because the timeof hydration was too short to form such a high concentration of this crystal. However,Delsol has shown by XRD that ettringite formation evolved in wetted PCL ash(unpublished).

The amorphous phase was also abundant in PC and PCL ash and less predominantin CFBC ash due to the lower combustion temperature. All these observations are ingood agreement with observations previously reported by other authors (Behr-Andreset al. 1993).

3.3 Bulk chemical analysis

The chemical compositions of the three samples are summarized in Table 3.The total Ca content was very high (25–30%) in PCL ash and was predominantly in

the form of free lime (about 20%) because of the high Ca/S molar ratio (3.5), and ofthe poor desulphurization yield (53%) in the furnace. Desulphurization also inducesthe presence of sulphur in sulphate form (3–4% in weight).

The Ca and CaO contents of CFBC ash were lower because (i) the Ca/S molar

I. Lecuyer et al.20

Fig

.2.

Scan

ning

elec

tron

mic

rosc

opic

pict

ures

.(a

,b)

pulv

eriz

edco

alco

mbu

stio

nw

ith

lime

inje

ctio

nas

h;(c

)ci

rcul

atin

gflu

idiz

edbe

dco

mbu

stio

nw

ith

limes

tone

inje

ctio

nas

h;(d

)ci

rcul

atin

gflu

idiz

edbe

dco

mbu

stio

nw

ith

limes

tone

inje

ctio

nas

h.

Desulphurization coal fly ash 21

TABLE 2Crystalline phases observed using X-ray diffraction

Crystalline RelativeSample phase Formula abundance

PC ash Mullite Si2Al6O13 V.AQuartz SiO2 A

PCL ash Lime CaO V.AMullite Si2Al6O13 AAnhydrite CaSO4 SQuartz SiO2 SCalcite CaCO3 e

Leached PCL ash Portlandite Ca(OH)2 V.AMullite Si2Al6O13 AQuartz SiO2 SCalcite CaCO3 e

CFBC ash Illite Si3AlO10(OH)2KAl2 V.AAnhydrite CaSO4 AQuartz SiO2 ACalcite CaCO3 e

Leached CFBC ash Illite Si3AlO10(OH)2KAl2 V.AQuartz SiO2 AAnhydrite CaSO4 SCalcite CaCO3 e

V.A, very abundant; A, abundant; S, scarce; e, traces; PC ash, pulverized coal combustion fly ash; PCLash, pulverized coal combustion with lime injection fly ash; CFBC ash, circulating fluidized bedcombustion with limestone injection fly ash.

TABLE 3Bulk chemical composition of fly ash

Oxides(% by weight) PC ash PCL ash CFBC ash

Si as SiO2 58 40.1 46.9Al as Al2O3 28.7 20.6 22.8Fe as Fe2O3 5.6 3.9 9.4Ti as TiO2 1.7 1.2 0.9Mg as MgO 0.9 0.9 4.1K as K2O 2.8 1.9 4.7Na as Na2O 0.3 0.2 0.2P as P2O5 0.3 0.2 0.2S as SO3 0.3 3.3 4Ca as CaO 1.4 27.7 7.2Free lime — 19.5 0.7CaO

PC ash, pulverized coal combustion fly ash; PCL ash, pulverized coalcombustion with lime injection fly ash; CFBC ash, circulating fluidizedbed combustion with limestone injection fly ash.

ratio was lower (2.3) and involved lesser quantities of sorbent, and (ii) the CFBCdesulphurization rate was much higher (90%) than for pulverized coal furnaces.

Table 4 gives the trace element contents of the three kinds of fly ash.

I. Lecuyer et al.22

TABLE 4Trace element concentrations in fly ash

Concentration in mg kg−1

Element PC ash PCL ash CFBC ash

As 41±3 (3) 38±5 (9) 34±3 (8)Cd 1.0±0.6 (4) 1.0±0.9 (11) 0.5±0.7 (9)Cr 220±5 (4) 115±2 (9) 215±10 (10)Cu 145±10 (5) 108±7 (11) 126±19 (9)Ni 112±3 (4) 76±10 (7) 99±15 (10)Pb 86±12 (4) 58±3 (7) 90±6 (10)Zn 118±4 (4) 93±6 (9) 258±24 (10)

The number of analysed samples is indicated between parentheses.PC ash, pulverized coal combustion fly ash; PCL ash, pulverizedcoal combustion with lime injection fly ash; CFBC ash, circulatingfluidized bed combustion with limestone injection fly ash.

The trace element concentrations were slightly lower in PCL ash than in PC ash, assorbent injection into the furnace induces a mass dilution effect on trace elementconcentrations (the lime contribution to the total quantities of trace element is quitenegligible). The mass dilution was estimated to be approximately 25% in this unit,which roughly corresponds with the variations between PC and PCL ash concentrations.As and Cd concentrations remained the same in both samples whereas Cr was ratherdepleted in PCL ash. This seems to indicate that lime may have an effect on thedistribution of some trace elements during combustion.

No comparison could be established between PCL and CFBC ash because both coalsand combustion processes are quite different. For example, the high Zn concentrationin CFBC ashes was due to the fuel, particularly rich in Zn (55 mg kg−1 in schlammsand 120 mg kg−1 in the slurry).

The trace element content was also studied as a function of particle size. The sampleswere passed through a series of nylon sieves and divided into seven granulometricranges for PCL ash and five granulometric ranges for CFBC ash.

Trace element contents of each class were determined. The results are illustrated inFig. 3. CFBC ash does not show any variation of trace element content vs. particlesize. On the other hand in PCL ash, Cu, Ni, Pb, Zn and Cr show increasing concentrationswith decreasing particle size.

Gutierrez et al. (1993) also observed that trace elements (Cu, Ni and Pb) tend to bemore concentrated in the smallest particles. Davison et al. (1974) observed the samebehaviour with other PC ashes. These results are in good agreement with the assumptionthat trace elements are volatilized during combustion and then re-condensed on particlesurfaces (Davison et al. 1974; Smith 1980; Gay 1989), the smaller the particle is, thelarger the specific area is and therefore, the more abundant this re-condensationphenomenon is. This explains the small particle enrichment by trace elements. Thecontribution of this volatilization–condensation mechanism is much smaller in a CFBCthan in a PC boiler. Other mechanisms can also take place (Gay 1989).

Desulphurization coal fly ash 23

250

0<5

Cr

150

50

200

100

<15 <25 <37 <50<15

0<25

0To

tal

300

0<5

Zn

250

200

150

100

<15 <25 <37 <50<15

0<25

0To

tal

200

0<5

Cu

150

100

50

<15 <25 <37 <50<15

0<25

0To

tal

140

0<5

Ni120

100

80

60

<15 <25 <37 <50<15

0<25

0To

tal

50

0<5

As

40

30

20

10

<15 <25 <37 <50<15

0<25

0To

tal

100

0<5

Pb

80

60

40

20

<15 <25 <37 <50<15

0<25

0To

tal

40

20

50

Con

cen

trat

ion

(m

g kg

–1)

Fig. 3. Trace element content of fly ash samples as a function of particle size (in lm). Solid bar, pulverizedcoal combustion with lime injection fly ash; open bar, circulating fluidized bed combustion with limestone

injection fly ash.

3.4 Fly ash leaching

The chemical composition of leachates obtained through the French standard test ispresented in Table 5. PCL ash leachates show high alkalinity (pH>12.5) and containlarge quantities of soluble salts, mainly CaSO4 and Ca(OH)2. Both leachates of PCLash and CFBC ash are generally saturated with sulphates (1–1.8 g l−1).

In these standardized leaching conditions, only As, Cr and Se were solubilized. Theother trace element concentrations remained under the detection limit at such high pHvalues. The high Cr concentration values are due to the easier solubilization of anionicforms [Cr(OH)4

2− or CrO42−] (Theis & Richter 1980). Colorimetric analysis established

that Cr is mainly present in hexavalent chemical form, such as CrO42−.

As a conclusion, only SO42− and hexavalent Cr concentrations seem to be worrying

compared to the European regulation concerning waste storage (EEC Draft Directive93/C 212/02).

The As and Se solubilization in the PC ash leachates seems to be hindered almostcompletely in the PCL ash leachates. Three further experiments have been performedto check that the high lime content is responsible for the absence of As (Fig. 4):

(1) 100 g of PC ash have been mixed with 24.3 g of CaO (equivalent to 19.6% CaO bymass), then submitted to the standard leaching test. The leachate As concentration

I. Lecuyer et al.24

TABLE 5Composition of fly ash leachates obtained with French standard test

Analysis PC ash PCL ash CFBC ash

pH 11.5 12.6 11.4Conductivity (mS cm−1) 0.9 8.8 3.4Sulphates (mg l−1) 197 1434–1824 658–1440∗Chlorides (mg l−1) 4.5 44–55 34–86∗Fluorides (mg l−1) 1 1.9–2.1 0.4–2∗Ca2+ (mgl−1) 164 1560–1610 425–820∗As (lg l−1) 60–80 <4.5 12–27Cd (lg l−1) <0.05 <0.05 <0.05Cr (lg l−1) 190 226–267 23–410CrVI (lg l−1) 81–151 23–236 18–276Cu (lg l−1) <1 <1 <1Ni (lg l−1) <2.5 <2.5 <2.5Pb (lg l−1) <1 1.5–2.4 <1Se (lg l−1) 421–511 21–34 20–136Zn (lg l−1) <5 <5 <5

∗Obtained after only 16 h mixing, PC ash, pulverized coal combustion fly ash; PCLash, pulverized coal combustion with lime injection fly ash; CFBC ash, circulatingfluidized bed combustion with limestone injection fly ash.

Leaching

Leaching

Leaching

Leaching

Leaching

Leaching

+ CaO1.5 g l–1 Ca2+

+ NaOH1.0 g l–1 Na+

+ CaSO41.5 g l–1 Ca2+

As = 64 µg l–1

Se = 508 µg l–1

pH = 11.3

As < DLSe = 41 µg l–1

pH = 12.5

As = 68 µg l–1

Se = 188 µg l–1

pH = 11.5

As < DLpH = 12.3(precipitation)

As = 21 µg l–1

pH = 12.4

As = 24 µg l–1

pH = 11.5

PC ash

PC ash+

19.6% CaO

PC ash+

2.9% CaSO4

PC ash

PC ash

PC ash

Fig. 4. Influence of CaO and CaSO4 on As and Se concentrations in fly ash leachates. DL, detection limit.

is lower than the detection limit, whereas when leaching the silico-aluminous PCash alone, it releases 64 lg l−1 of As.

(2) The same experiment with CaSO4 (2.9%) replacing CaO (19.6%) shows no influenceon As concentration (68 lg l−1).

Desulphurization coal fly ash 25

(3) 100 g of silico-aluminous PC ash have been submitted to the standard leaching test,then before acidification of the leachate with HNO3, 0.11 g of CaO was dissolvedin the leachate in order to get 1.5 g l−1 of Ca2+. A precipitate appears in the solutionand the As concentration decreases under the detection limit.

The influence of pH is not sufficient to explain the As concentration variation (Theis& Wirth 1977) between pH=11.3 and pH=12.5. The addition of NaOH instead ofCaO in the PC ash leachate induces the same pH variation (11.3–12.4) but a muchlower influence on As concentration (decrease from 64 to 21 lg l−1).

Adding 0.65 g CaSO4 in the same leachate in order to get the same concentration of1.5 g l−1 Ca2+ also has little effect. The As concentration decreases from 64 to 24 lg l−1.

So, the inhibition of arsenic solubilization is obviously linked with the large limecontent in the ash. The most likely mechanism is given by Sato (1986), which showsthat when bubbling CO2 in Ca2+/As solution, As concentration decreases along withCaCO3 precipitation because AsO4

3− can adsorb on the surface of CaCO3 precipitate.It is also possible to assume that As V forms insoluble arsenates with Ca2+ (Moretti

et al. 1988) or other cations, but experiments with CaSO4 addition showed that Asconcentration hardly decreases in such conditions. The co-precipitation of AsO4

3− withCaCO3 seems a better explanation.

The case of Se is still more complex because its concentration in leachates decreaseswhen adding CaO or CaSO4 (Fig. 4). The former arguments established for As remaintrue for Se but in addition, one should bear in mind that selenate ions can easilysubstitute for sulphate ions in crystalline structure (EPRI 1994).

3.5 Trace element behaviour in wet ashes

Before being disposed of, fly ashes are wetted in order to avoid dust emissions. However,calcic ash tends to solidify when mixed with water, just as a hydraulic binder. In such‘‘blocks’’ of matter, the leaching behaviour of trace elements should not be the sameas in dry fly ash. Moreover, the permeability of this kind of block has been shown todecrease dramatically with time (Jozewicz et al. 1991).

Series of samples made of fly ash and water were prepared in closed vessels, withPCL ash and CFBC ash, according to three ratios:

(1) 100 g dry ash+30 g water(2) 100 g dry ash+65 g water(3) 100 g dry ash+100 g water

With PCL ash, Sample 1 formed friable aggregates when ageing. Sample 2 became adense, hard ball and Sample 3 also hardened, although this was less compact thanSample 2.

CFBC ash hardly reacted with water because of the small CaO content. It slowlybecame more compact over a few months, but did not really harden, even thoughettringite formation could be observed (Carles-Gibergues & Delsol 1994).

These samples were crushed and submitted to the leaching test at successive ages: afew hours, 4 days, 15 days, 36 days and 50 days. The evolution of the leachates wasstudied for four compounds: Ca2+, SO4

2−, Cr and Se. The results are illustrated inFig. 5.

The pH value of the leachates remained alkaline and did not show any decrease (pH>12 for PCL ash and 10.6 for CFBC ash).

I. Lecuyer et al.26

Det

ecti

on li

mit

s: 0

.8 µ

g l–1

for

Cr

5

g l–1

for

Se

Cr–CFBC ash

As–CFBC ash

60

2000

Con

cen

trat

ion

(m

g l–1

)

50

1000

10 20 30 400

60

200

Time (days)

Con

cen

trat

ion

(µg

l–1)

50

100

10 20 30 400

Se–PCL ash

Se–CFBC ash

Cr–PCL ash

Ca–PCL ashSO4–PCL ashSO4–CFBC ashCa–CFBC ash

Fig. 5. Evolution of the chemical leachability of elements in wetted ash samples as a function of time (water:ash=0.65). Similar results are obtained for ratios 0.33 and 1. PCL ash, pulverized coal combustion with

lime injection fly ash; CFBC ash, circulating fluidized bed combustion with limestone injection.

With PCL ash, most salt concentrations decreased when the sample hardened in thecourse of time: in 50 days, Ca2+ decreased by a factor of 2 and Cl− by a factor of 4,whereas SO4

2−, CrVI and Se fell below the detection limit.In a 50-day-old block, only a few lg l−1 of Cr and Pb were leached, together with

the excess CaO. Moreover, the higher the water:ash ratio, the earlier the elements wereprevented from leaching.

Desulphurization coal fly ash 27

Carles-Gibergues (1978) observed that lime, anhydrite and alumino-silicates containedin ash can react with water to form an ettringite phase [Ca6Al2(SO4)3(OH)12, 25H2O].This crystalline phase stabilizes the SO4

2− and Ca2+ ions and traps most of the traceelements so they can no longer be leached. Kumarathasan et al. (1990) also showedthat oxyanions such as SeO4

2−, AsO42− and CrO4

2− can partially take the place ofsulphates in ettringite: up to 10% of sulphates for AsO4

2− and 100% for SeO42−. These

anions are then linked to the ettringite structure and are no longer solubilized. Anotherassumption sets that As, Se and Cr in cationic form can partially substitute to Al(Leroy 1994), so PCL ash should be partially stabilized when wetted. Moreover, thisphenomenon, together with the low permeability of solidified blocks (Jozewicz et al.1991), should cut down pollutant release in leaching water.

CFBC ash did not show the same properties because the CaO content is too low toform enough ettringite: As concentration remained nearly constant in all samples,whatever the water:ash ratio and the age may be. Chromium concentration slightlydecreased in the early ages but increases later. Only Se concentration showed a decreaseby an order of magnitude (136 lg l−1 to 12 lg l−1) in the course of 2 months: selenateions may, more easily than CrO4

2− ions, substitute for SO42− ions in the ettringite phase

(Kumarathasan et al. 1990).

4. Conclusion

The aim of this work was to completely characterize two types of desulphurizationashes: one from a lime injection process in a PC furnace and the other from a CFBCunit using limestone, and to study their physico-chemical properties with moisture.

The following conclusions were drawn:

(1) The PCL ash morphology is very similar to, though finer than, that of PC ash, andshows lots of microspheres. Conversely, CFBC ash shows only irregularly-shapedparticles. By addition of lime into the furnace, sulphates appear in the fly ashcontent.

(2) The addition of sorbent during combustion has a dilution effect on most traceelement concentrations in fly ash. Moreover, it seems that it can help to trap volatileelements (such as As, Cd) in the ash during combustion.

(3) Most of the trace elements show an inverse concentration–particle size dependencein PCL ash. This trend does not exist for CFBC ash: trace element concentrationsare constant whatever the particle size may be.

(4) Leachates obtained from PCL ash and CFBC ash are both rich in soluble salts,mainly CaSO4 and Ca(OH)2. On the other hand, As and Se are prevented fromsolubilizing by high lime content. Thus, Cr is the only measured trace element thatis appreciably solubilized (up to 2.3% of the total Cr content in ash).

(5) When wetted, PCL ash hardens and forms compact blocks: the formation ofettringite crystals stabilizes Ca2+ and SO4

2− ions. Simultaneously, arsenate, selenateand chromate anions are trapped in the crystal, possibly thanks to the substitutionfor sulphate ions (Kumarathasan et al. 1990). CFBC ash does not show thesame properties, because the lime content is too low. However, the leached Seconcentration is less in wetted CFBC ash samples.

This property of desulphurization fly ash is particularly interesting because it limitsthe impact of the leachates and can allow many utilizations.

I. Lecuyer et al.28

The high pH and Ca concentration should allow the use of this fly ash for acidicsoil stabilization or for agricultural improvement.

The hydraulic binder property (hardening with water and decrease of permeability)could be used to treat degraded soils or for the construction of landfill capping ofwaste disposal sites and as components in the manufacture of low strength buildingmaterials (R. T. Johnston, personal communication).

Finally, the property to fix some chemical compounds should allow the use of thisfly ash for the stabilization of industrial wastes, and even for the decontamination ofpolluted waters and soils (Vangronsveld et al. 1991).

References

Behr-Andres, C. B., Douglas McDowell, S. & Hutzler, N. J. (1993) Quantitative mineral de-termination of industrial coal ash. Journal of the Air Waste Management Association 43,1245–1251.

Carles-Gibergues, A. (1978) Proprietes hydrauliques des cendres volantes sulfocalciques deGardanne (Hydraulic properties of the sulfo-calcic fly ashes of Gardanne). Bulletin de Liaisondu Laboratoire de Physique et Chimie 93, 5–8.

Carles-Gibergues, A. & Delsol, C. (1994) Annales de l’ITBTP. 525, Serie Materiaux 85, p. 46.Davison, R. L., Natusch, D. F. S., Wallace, J. R. & Evans Jr., C. A. (1974) Trace elements in fly

ash: dependence of concentration on particle size. Environmental Science & Technology 13,1107–1112.

Delsol, C. Thesis in progress at “Laboratoire des Materiaux et des Durabilites des Constructions”,INSA Toulouse, France (unpublished).

EPRI (1994) Chemical Attenuation Reactions of Selenium. Report No. 103535. EPRI, U.S.A.Gay, A. J. (1989) Characterization of fly ash from Fluidised Bed Combustors with regard to its

utilisation and safe disposal. TNO Environmental and Energy Research, 17pp.Gutierrez, B., Pazos, C. & Coca, J. (1993) Characterization and leaching of coal fly ash. Waste

Management & Research 11, 279–286.Jozewicz, W., Gullett, B. K. & Hueckel, T. (1991) Permeability properties of fly ash from furnace

sorbent injection process, Processing and Utilization of High-Sulphur Coals IV (P. R. Dugan,D. R. Quigley and Y. A. Attia, eds). Amsterdam, The Netherlands: Elsevier Science Publishers,685–699.

Kumarathasan, P., MacCarthy, G. J., Hassett, D. J. & Plfughoeft-Hassett, D. F. (1990) Flyash and Coal-Conversion By-Products: Characterization, Utilization and Disposal V. InMaterials. Research Society Symposium Proceedings (R. L. Day & F. P. Glasser, eds). 178,83–104.

Leroy, J. B. (1994) Les dechets et leur Traitement (Wastes and their Treatment). Que sais-je 946,54–55. Paris, France: PUF.

Moretti, J. C., Henke, K. R. & Wentz, C. A. (1988) Immobilization of arsenic in coal combustionfly ash leachates. Particulate Science and Technology 6, 393–404.

Sato, K. (1986) Leaching Characteristics of Various Elements from Coal Fly Ash. CRIEPITechnical Report ET 86001, 37pp.

Smith, R. D. (1980) The trace element chemistry of coal during combustion and the emissionsfrom coal-fired plants. Progress in Energy and Combustion Science, 6, 53–119.

Soreau, S. & Jacob, D. (1993) Analyses de Charbons, Cendres et Lixiviats: Etude Comparativede 4 Laboratoires (Analysis of Coals, Ashes and their Leachates: Comparative Study at 4Laboratories). Rapport EDF-DER HE-35/93/24, 60pp.

Theis, T. L. & Richter, R. O. (1980) Chemical speciation of heavy metals in power plant ashpond leachate. Environmental Science and Technology 14, 219–224.

Theis, T. L. & Wirth, J. L. (1977) Sorptive behaviour of trace metals on fly ash in aqueoussystems. Environmental Science and Technology 11, 1096–1100.

Vangronsveld, J., Cammaer, C., Van Assche, F. & Clijsters, H. (1991) Reclamation of a “desertlike” site in the north east of Belgium: evolution of the metal pollution and experiments insitu. International Conference of Heavy Metals in the Environment, Edinburgh, U.K. 1, 58–61.