Formation and use of coal combustion residues from three types of power plants burning Illinois coals

Fuel 80 (2001) 1659±1673

www.fuel®rst.com

Formation and use of coal combustion residues from three types of power plants burning Illinois coals Ilham Demir*, Randall E. Hughes, Philip J. DeMaris Illinois State Geological Survey, 615 E. Peabody Drive, Champaign, IL 61820, USA Received 13 December 1999; revised 31 January 2001; accepted 8 February 2001

Abstract Coal, ash, and limestone samples from a ¯uidized bed combustion (FBC) plant, a pulverized coal combustion (PC) plant, and a cyclone (CYC) plant in Illinois were analyzed to determine the combustion behavior of mineral matter, and to propose bene®cial uses for the power plant ashes. Pyrite and marcasite in coal were converted during combustion to glass, hematite and magnetite. Calcite was converted to lime and anhydrite. The clay minerals were altered to mullite and glass. Quartz was partially altered to glass. Trace elements in coal were partially mobilized during combustion and, as a result, emitted into the atmosphere or adsorbed on ¯y ash or on hardware on the cool side of the power plants. Overall, the mobilities of 15 trace elements investigated were lower at the FBC plant than at the other plants. Only F and Mn at the FBC plant, F, Hg, and Se at the PC plant and Be, F, Hg, and Se at the CYC plant had over 50% of their concentrations mobilized. Se and Ge could be commercially recovered from some of the combustion ashes. The FBC ashes could be used as acid neutralizing agents in agriculture and waste treatment, and to produce sulfate fertilizers, gypsum wall boards, concrete, and cement. The PC and CYC ¯y ashes can potentially be used in the production of cement, concrete, ceramics, and zeolites. The PC and CYC bottom ashes could be used in stabilized road bases, as frits in roof shingles, and perhaps in manufacturing amber glass. q 2001 Elsevier Science Ltd. All rights reserved. Keywords: Coal combustion residues; Combustion at power plants; Illinois coals; Mineral matter

1. Introduction Inorganic matter in coal is converted to coal combustion residues (CCRs) at coal-®red power plants. Virtually every economical use of coal depends in part on the amount and variety of its inorganic matter. Inorganic matter in coal can be the source of deleterious pollutants and corrosive elements, but it can also be a source of useful by-products. Sixteen elements in coal (As, Be, Cd, Cl, Co, Cr, F, Hg, Mn, Ni, P, Pb, Sb, Se, Th, U) are among the 189 hazardous air pollutants (HAPs) mentioned in the 1990 Clean Air Act Amendments (CAAA) [1]. Partitioning of these and other elements among the CCRs and ¯ue gas is highly variable [2±5] because of the variations in the types and operational conditions of combustion units, the characteristics of coal, and the modes of occurrences of the elements in the coal. Dif®culties in obtaining representative samples, as well as analytical errors, contribute to the variability of data on the retention or emission of potentially toxic elements from the power * Corresponding author. Tel.: 11-217-244-0836; fax: 11-217-333-2830. E-mail address: [email protected] (I. Demir).

plants. The HAPs provisions of the CAAA presently focus on municipal incinerators and petrochemical and metal industries. A risk analysis by the US Environmental Protection Agency concluded that, at present, only Hg emission from coal-®red electrical utilities requires further investigation [6]. A ®nal regulation on Hg emission is expected by the end of 2004. US utilities annually generate 107 million tons of CCRs, 59% of which is ¯y ash (Fig. 1). Bene®cial uses of CCRs include soil and mine waste treatments, admixtures in cement and concrete, making bricks and other ceramic products, ®ll materials in civil engineering projects, and extraction of valuable materials [8±17]. However, only about 31% of all the CCR generated in the US is used commercially (Fig. 1); the remainder is discarded in land®lls or in coal mines. This study evaluated chemical, mineralogical, and microscopic characteristics of feed coals and CCRs from three types of electrical power plants in Illinois to determine (1) the combustion behavior of minerals, (2) the fate of 15 elements (As, Be, Cd, Co, Cr, F, Hg, Mn, Ni, Pb, P, Sb, Se, Th, and U) of environmental concern, and (3) the potential economic value of the CCRs from the power plants.

0016-2361/01/$ - see front matter q 2001 Elsevier Science Ltd. All rights reserved. PII: S 0016-236 1(01)00028-X

1660

I. Demir et al. / Fuel 80 (2001) 1659±1673

Fig. 1. US production and use of coal combustion residues for the year 1999 (plotted using data from ACAA [7]).

2. Experimental 2.1. Sample collection and preparation Samples of feed coals and CCRs (¯y ash and bottom ash) were collected from a ¯uidized bed combustion (FBC) plant, a cyclone (CYC) plant, and a pulverized coal combustion (PC) plant; all burned Illinois coals (Table 1). A sample of limestone that is used in the FBC plant to capture sulfur dioxide was also collected. At the PC and CYC plants Table 1 Quantities and descriptions of the samples of coal, limestone, and CCRs collected from a FBC plant, a PC plant, and a CYC plant. All of the plants were located in Illinois and burned Illinois coals Plant type Sample type Amount (kg) Sample description a FBC

coal ¯y ash

11.34 8.16

bottom ash

10.89

limestone CYC

PC

a

11.34

coal ¯y ash

4.08 3.63

bottom ash

5.44

coal ¯y ash

5.44 3.63

bottom ash

6.80

, 0.95 cm (,3/8 00 ) size coal Very ®ne particle size, light gray color , 0.64 cm (,1/4 00 ) particle size, mostly yellowish and grayish particles, and small number of black particles. Crushed and off-white color Crushed coal Fine particle size, dark gray color , 0.95 cm (,3/8 00 ) particle size, black color, and high moisture content because of quenching in water Crushed coal b Fine particle size, light gray color , 0.95 cm (,3/8 00 ) particle size, dark gray to black color, and high moisture content because of quenching in water

The size ranges are semiquantitative values based on visual examination. The sample of crushed coal was taken before the coal went through the pulverizer. The pulverizer did not reject pyrite; it ground and injected the entire feed coal into the burner. b

automatic samplers were used to collect the samples of coal and CCRs, as part of the routine sampling. For the FBC plant, a sampling shovel was used to collect 15 increments of the coal from the automatic belt carrying coal to the boiler. To sample the FBC ¯y ash, the ash stream was directed to a sampling bag for about 30 s through an opening at the particulate collection system. The FBC bottom ash and crushed limestone was sampled by collecting 15 to 20 increments of the materials from widely spaced locations of their stock piles using a sampling shovel. Representative splits of the coal and coarse-grained bottom ash samples were separated by rif¯ing, splitting, and then grinding to ,250 mm (260 mesh) particle size. The ¯y ash samples were already ,250 mm in particle size and, therefore, did not need to be ground. Small splits of the ground coal and bottom ash samples and of ¯y ash samples were prepared for chemical, mineralogical, and microscopic analyses by rif¯ing and splitting. 2.2. Chemical analysis The coal and CCR samples were chemically analyzed using well-established analytical methods [18,19]. These methods and parameters (shown in parentheses) determined by each method were: wavelength dispersive X-ray ¯uorescence spectrometry (Ge, Mn, and oxides including SiO2, Al2O3, Fe2O3, CaO, MgO, K2O, Na2O, Ti2O, P2O5, MnO, and SO3), cold-vapor atomic absorption spectrometry (Hg), instrumental neutron activation analysis (As, Co, Cr, Ni, Sb, Se, Th, U, Zn), optical emission spectrometry (Be), pyrohydrolysis-ion chromatography (F), inductively coupled plasma-mass spectrometry (Cd, Pb), and standard ASTM methods of coal analysis (ash, moisture). 2.3. Mineralogical and microscopic analyses The ®rst step of mineralogical analysis was the generation of the low-temperature ash (LTA) of each coal sample at about 1508C in an activated oxygen±plasma atmosphere created by passing oxygen into a radio-frequency ®eld [20]. The minerals in the LTA were essentially unaltered from their original state in the coal. The LTA samples were washed with water to remove elements adsorbed on minerals and thus improve the accuracy of mineralogical analysis [10,21,22]. The water-extracted LTA samples were dried, weighed, and micronized with Na-dithionite buffered water in a McCrone w grinder for 15 min. The samples were then centrifuged in 50-ml tubes for 20 min at 2000 rpm. After discarding the clear supernatant, the paste was mixed thoroughly, and a portion of it was smeared on glass slides. The remaining paste was dried overnight at 708C, mixed with a mortar and pestle, and then packed into an end-loading sample holder (a random bulk pack). The random bulk pack samples were examined by X-ray diffraction (XRD). The smear slides were X-rayed after drying in air, after at least two days exposure to ethylene glycol, after

I. Demir et al. / Fuel 80 (2001) 1659±1673

heating for 1 h at 3508C, and after heating to 5508C, which produced the strong 1.4 nm XRD peak for detection of small amounts of chlorite. The peak areas of the XRD spectra were calculated with a software package. The bulk pack XRD analysis was the preferred method for determination of nonclay minerals. The XRD analysis of smear slides was used for clay mineral analysis. The XRD detection limit for calcite in the LTAs was about 5 wt%. Therefore, the calcite content in the LTA determined by the XRD analysis was corrected using the chemically determined CaO content of the coal. This approach was judged to be reasonable because no other Ca minerals were detected in the samples and Ca associated with clay minerals and plagioclase was insignificant relative to Ca associated with calcite. The steps for calculating the mineralogical composition of the coal samples from the XRD spectra and chemical analysis were as follows: 1. The XRD patterns were deconvoluted, and the data were transferred electronically to a computer spreadsheet. 2. From the chemically-determined wt% pyritic S (%Spy) content of the LTA, the Fe-sul®de mineral content was calculated as wt%(pyrite 1 marcasite) in LTA ˆ %Spy in LTA £ 1.871. 3. From the bulk pack XRD peak areas, the percentages of quartz, K-feldspar, plagioclase, calcite, pyrite, and marcasite were calculated on a 100% nonclay mineral basis. 4. Dividing wt% (pyrite 1 marcasite) from step 2 by wt% (pyrite 1 marcasite) from step 3 yielded decimal nonclay minerals (DNC) in the LTA. 5. Percentage of each nonclay mineral in the LTA was calculated by multiplying its wt% in the nonclay fraction from step 3 by DNC from step 4. 6. From smear slide XRD peak areas, the percentages of clay minerals (mixed-layered illite/smectite, illite, kaolinite, chlorite) were calculated on a 100% clay basis; 7. Percentage of each clay mineral in the LTA was calculated by multiplying its wt% in the clay fraction from step 6 with (1-DNC). 8. Using chemically-determined CaO contents of coal, wt% calcite was calculated on LTA basis, this chemical-based calcite content was replaced for the XRDbased calcite content from step 5, and the percentages of other minerals from step 5 and 7 were adjusted accordingly. 9. Percentage of each mineral in whole coal was calculated by multiplying (wt%LTA 4 100) by its percentage in the LTA from step 8. Most of the CCR samples from the power plants were dominated by an amorphous (glass) fraction. These ashes do not contain much water-soluble inorganic elements loosely adsorbed on the ash particles. However, they do contain substantial amounts of lime and anhydrite that react with water. Therefore, a somewhat different procedure of miner-

1661

alogical analysis was followed for the ¯y and bottom ashes to obtain the results: water extraction step was omitted, 8 wt% dolomite was added as an internal standard, and the samples were micronized in propanol. The limestone sample was analyzed using the same XRD procedures used in the analysis of the CCRs. For electron microscope analysis, the ash samples were mounted on aluminum stubs and coated with Au±Pd; the bottom ash samples were ground in a ceramic mortar and pestle prior to mounting on the aluminum stubs. The samples were examined under a scanning electron microscope (SEM) equipped with an energydispersive X-ray analyzer (EDX). The EDX spectra of the ash particles were produced for sample spots with 5±15 nm in diameter. 3. Results and discussion 3.1. Combustion transformation of minerals Major minerals in the feed coals were quartz, kaolinite, illite, mixed-layered illite/smectite, pyrite, and calcite (Table 2). Small amounts of chlorite, marcasite, and feldspars were also present. The total clay mineral content of coal was generally high relative to the other minerals, and among the clay minerals, kaolinite was the most abundant (Fig. 2). Combustion in the power plant furnaces converted the coal minerals to the solids that made up the bottom and ¯y ashes. The details of these conversion reactions are dif®cult to establish because of the random collision, coalescing, and chemical interactions of a large number of solid, liquid, and gas phases at very high temperatures in a matter of seconds. However, a comparison of the minerals in the coals with those in the CCRs (Table 2), along with published data [23±30], suggests that the coal minerals decomposed at certain temperatures and then, during further heating or cooling, yielded certain combustion products as follows: 2 pyrite or marcasite …2FeS2 † 250±8508C

!

2 pyrrhotite …2FeS† 1 S2 .8508C

5FeS 1 S2 1 10:5O2 ! hematite …Fe2 O3 † 1 magnetite …FeO´Fe2 O3 † 1 7SO2 hematite and magnetite

1000±16508C

!

partial conversion

to amorphous Fe oxides in glass phase of CCRs .8508C

calcite …CaCO3 † ! lime …CaO† 1 CO2 .8508C

CaO 1 SO2 1 0:5O2 ! anhydrite …CaSO4 †

coal limestone coal coal

FBC

¯y ash bottom ash

CYC

b

26 (2.7) 1.5 34 (3.5) 23 (2.1)

Kaolinite 20 (2.1) 3.4 23 (2.3) 16 (1.5)

Illite 17 (1.8) nd b 3.4 (0.35) 4.9 (0.46)

Illite/Smectite 1.5 (0.16) 1.5 1.9 (0.19) 0.43 (0.04)

Chlorite

1.9 nd b

7.6 1.3

0.008 1.0

8.4 8.2

nd b nd b

5.1 nd b

Quartz

Mullite

3.0 nd b

2.1 2.0

nd b 0.4 0.8 nd b

3.0 nd b

Hematite

,0.1 1.9

Calcite

4.0 0.4

4.7 4.9

0.69 nd b

Magnetite

Minerals and amorphous material in the coal combustion residues (wt%)

16 (1.7) 5.0 16 (1.6) 32 (3.0)

Quartz

Minerals in the low-temperature ashes of coal and in limestone (wt%) a

nd b nd b 5.0 0.5

nd b nd b

nd b 2.0

Gypsum

2.9 (0.3) nd b 0.66 (0.07) , 0.1 (,0.01)

Marcasite

0.7 nd b

17 16

Anhydrite

8.5 (0.89) nd b 12 (1.3) 12 (1.1)

Pyrite

nd b nd b

nd b nd b

2.3 40

Lime

8.2 (0.86) 86 9.7 (1.0) 11 (1.0)

Calcite

nd b nd b

nd b nd b

11 13

78 98

87 92

57 19

Amorphous material (glass)

, 0.1 (,0.01) nd b , 0.1 (,0.01) , 0.1 (,0.01)

nd b 4.0 nd b nd b

Portlandite

K-feldspar

Dolomite

0.04 (,0.01) nd b 0.24 (0.03) 1.4 (0.2)

Plagioclase

Values on whole coal basis are given in parentheses next to the values on LTA basis. The water-extracted LTA contents of the FBC, PC, and CYC coals were 10.47, 10.31 and 9.40 wt%, respectively. Values below detection limits were indicated as (nd).

¯y ash bottom ash

PC

a

¯y ash bottom ash

FBC

PC CYC

Sample type

Plant type

Table 2 Mineralogical composition of the samples of coal, limestone, and CCRs from the three types of power plants

1662 I. Demir et al. / Fuel 80 (2001) 1659±1673

I. Demir et al. / Fuel 80 (2001) 1659±1673

1663

Fig. 2. Abundances of minerals in samples of feed coals combusted at the FBC, PC, and CYC plants. Keys: Kaol, kaolinite; I, illite; I/S, illite/smectite; Chl, chlorite; Qtz, quartz; Pyr, pyrite; Marc, marcasite; Calc, calcite; Plag, plagioclase; K-f, potassium feldspar.

clay minerals 850±10008C

!

1000±11008C

!

50±6008C

!

dehydration and dehydroxylation

glass

glass 1 mullite …3…Al2 O3 †´2…SiO2 ††

1 cristobalite …SiO2 † .11008C

!

glass 1 mullite …3…Al2 O3 †´2…SiO2 ††

1 cristobalite …SiO2 † 1 Na2 O…gas† 1 K2 O…gas† quartz

.12008C

!

partial melting and glass formation

Pyrite, marcasite, and pyrrhotite particles most likely melted to varying degrees prior to their decomposition and oxidation [27,28]. A study by Hatt and Bull [31] suggested that rapid quenching of iron in the air promotes the formation of hematite while exposure of iron to hot ¯ue gas for extended periods of time promotes the formation of magnetite. Once formed, hematite is stable until a temperature of 14008C is reached [32]. Therefore, hematite melted in the CYC furnace, where the temperature reaches 16508C, and then recrystallized upon cooling. Magnetite, a hightemperature Fe-spinel, probably formed between 10008C and 12008C [33]. Although the average FBC temperature is about 9008C, the ¯ame temperature on particle surfaces could be ,10008C; this would explain the presence of a small amount of magnetite in FBC ashes. Magnetite has a melting point of ,16008C suggesting that, like hematite, magnetite in the CYC ashes formed largely as the melt cooled. The combination of melting at 16508C and quenching in water reduced the recrystallization of magnetite and hematite and, as a result, incorporated most of the Fe-oxide phase into glass in the CYC bottom ash.

Calcite was partially converted to lime (CaO); some calcite was still present in all the CCRs (Table 2), even though the combustion temperature was above the calcination temperature of this mineral in all three power plants. The likely reasons were: (1) interiors of large calcite particles were not heated enough to decompose during their short residence time in the furnace, and (2) CaO in the ashes was partially converted back to calcite after leaving the furnace through reaction with ¯ue gas CO2 or with atmospheric CO2. Portlandite, a hydrated form of lime, formed as a result of post-combustion exposure of lime to moisture in the ¯ue gas and air. Similarly, gypsum formed from anhydrite. Clay minerals and feldspars were converted primarily to aluminosilicate glass and mullite. The alteration of clay minerals and feldspars might have also generated a small amount of cristobalite, a high-temperature polymorph of quartz. Cristobalite was reported to form from kaolinite heated to 900±13008C [24]. However, the amount of cristabolite that might have formed in the combustion ashes of the power plants was less than the detection limit of the XRD technique. Also, mullite was less than the detection limit in the FBC ashes, indicating that the FBC ¯ame temperature probably did not reach the mullite formation temperature of .10008C. The melting of clay minerals and glass formation at the relatively low temperatures of the FBC resulted most likely from the ¯uxing actions of K2O from illite [26] and of CaO from calcining of limestone. Gibbon [34] observed that mullite crystals were embedded in glass particles of ¯y ashes. Mullite normally forms through solid-state reactions primarily from kaolinite, and to a lesser extent from high-Al illite and illite/smectite. However, the difference between the amount of mullite in the ¯y ash and that in the bottom ash from the CYC and PC plants suggests the crystallization of some mullite from the melt. Apparently the mullite crystals melted along their margins in the combustion zones of the PC and CYC furnaces. The outermost layers of the melted margins

1664

I. Demir et al. / Fuel 80 (2001) 1659±1673

Fig. 3. (a) SEM micrograph of CYC bottom ash and (b) EDX spectrum of one of the glass (G) particles shown on the SEM image. The bottom ash consists almost entirely of aluminosilicate glass with varying amounts of Fe and Ca, and sometimes also K and Na, as indicated by the EDX spectrum. The bottom ash was ground for preparing the SEM/EDX sample. Therefore, the actual particle sizes in the sample were larger than indicated on the SEM micrograph.

attained lower liquidus temperature (temperature of complete melting) than the rest of the molten phase as a result of interaction with the ¯ux compounds FeO and CaO generated in the combustion zones of the PC and CYC furnaces. This was probably why substantial amounts of Fe and Ca occurred in the aluminosilicate glass, as determined by SEM/EDX analysis (Fig. 3). Thermodynamic modeling by Evgueni et al. [33] suggests that mullite and its molten phase coexisted between ,11008C and ,15008C in the PC and CYC furnaces with the mullite/melt ratio decreasing with increased temperature. As the particles cooled, mullite recrystallized from its molten phase except in the ¯uxed outermost layers; eventual rapid cooling below 10008C in the ¯ue gas stream converted the ¯uxed outer layers to glass. Quenching in water increased glass formation at the expense of mullite formation in the PC and CYC bottom ashes. Mitchell and Gluskoter [24] reported the conversion of

quartz to glass as a result of exposure to temperatures of ,12008C to ,13008C for 30 min under oxidizing conditions. Then, not surprisingly, quartz survived the combustion process at the FBC plant where the combustion temperature was less than 10008C. Even at the PC and CYC plants where the combustion temperatures exceeded 13008C, the decomposition of quartz was limited to partial melting and subsequent conversion to glass because of the short residence time of the particles in the ¯ame. 3.2. Combustion mobility of HAPs Upon combustion, trace elements in coal were retained in the CCRs in varying amounts depending on their volatilization±condensation characteristics under the operating conditions of the power plants. The term ªcombustion mobilityº was used in this study to express the portions of the trace elements that were not retained in

I. Demir et al. / Fuel 80 (2001) 1659±1673

the CCRs: that is, the mobility of an element is the difference between the amount in the feed (coal or, in the case of the FBC unit, 0.75 parts coal 1 0.25 parts limestone) and the amount retained in the bottom and ¯y ashes. The retention and mobility calculations took into consideration the mass ratios of ¯y ash to bottom ash, as well as the measured concentrations of high-temperature ash and elements in the feed and CCR samples. The ¯y-ash to bottom-ash weight fraction ratios were 0.8:0.2 for the FBC plant, 0.75:0.25 for the PC plant, and 0.25:0.75 for the CYC plant. The steps for calculating the retention or mobility of elements at the FBC plant were:

1665

Fba ˆ 1 2 Ffa where Ffa and Fba are loss-on-ignition (LOI)-free fractions of ¯y ash and bottom ash, respectively, in the total of FBC CCRs, and Afa and Aba are the percent ash yields of ¯y ash and bottom ash, respectively, as determined in the laboratory by burning their residual carbons at 10008C. Cel-output ˆ Ffa …Cfa-ash † 1 Fba …Cba-ash † where Cel-output is the concentration of an element in the total of FBC CCRs, and Cfa-ash and Cba-ash are the concentrations of an element in the LOI-free ¯y ash and bottom ash, respectively. Unlike at the FBC plant, limestone was not added to the feed coal at the PC and CYC plants. Therefore, Cel-input for the PC and CYC plants was the concentration of the element in the feed coal on an ash basis. Calculation of the Cel-output for the PC and CYC plants was similar to the calculation used for the FBC plant except that the 0.8 ¯y ash and 0.2 bottom ash fractions were replaced with 0.75 and 0.25, respectively, for the PC plant and with 0.25 and 0.75, respectively, for the CYC plant. Once Cel-input and Cel-output were known, the %retention and %mobility were calculated as follows:

Flst ˆ …0:25Alst †=…0:25A1st 1 0:75Acoal † Fcoal ˆ 1 2 Flst where Flst and Fcoal are fractions of limestone ash and coal ash, respectively, in the ash of the FBC feed, and Alst and Acoal are the percent ash yields of limestone and coal, respectively, as determined at 10008C in the laboratory. Cel-input ˆ Flst …Clst-ash † 1 Fcoal …Ccoal-ash † where Cel-input is the concentration of an element in the FBC feed on an ash basis, and Clst-ash and Ccoal-ash are the concentrations of an element in the limestone and coal components of the FBC feed on an ash basis, respectively.

%retention ˆ ……Cel-output =Cel-input †=…CAl-output =CAl-input †† £ 100 %mobility ˆ 100 2 %retention

Ffa ˆ …0:8Afa †=…0:8Afa 1 0:2Aba †

Table 3 Retentions and mobilities of Al and 15 HAPs calculated from their concentrations in coal, CCRs, and limestone from the three power plants Location and sample

Al

HAP elements As

FBC plant Concentration (mg/kg), dry basis in coal (ash basis) in limestone (ash basis) in LOI-free ¯y ash in LOI-free bottom ash Retention % Mobility a, %

93197 10645 34805 11181 76 24

Be

Cd

Co

Cr

F

Hg

Mn

Ni

P

Pb

Sb

16.8 15.8 5.6 2.0 6.3 6.2 4.1 0.5 81 97 19 3

9.32 0.90 2.55 0.64 73 27

42.9 233 1277 5.9 19 8 12.6 77 54 4.4 36 49 76 96 15 24 4 85

0.93 746 168 1221 , 0.01 2349 16 443 0.26 637 68 582 0.01 318 14 492 83 42 108 103 17 58 0 0

PC plant Concentration (mg/kg), dry basis in coal (ash basis) 76900 20.6 9.4 in LOI-free ¯y ash 86900 34.9 9.2 in LOI-free bottom ash 69100 3.0 9.0 Retention % 107 121 91 Mobility a, % 0 0 9

3.94 5.02 0.47 92 8

32.8 159 20.7 180 22.5 136 60 99 40 1

591 69 5 8 92

0.75 0.10 , 0.01 9 91

750 103 472 79 460 85 58 73 42 27

413 30.0 538 50.2 262 10.2 106 125 0 0

5.6 8.1 2.2 110 0

38.2 1.86 16.0 13.14 5.5 0.48 20 170 80 0

38.2 162 27.7 259 20.1 112 54 85 46 15

601 263 15 11 89

0.76 0.83 0.02 25 75

763 76 354 198 540 70 62 124 38 0

420 31.9 848 102 522 2.6 135 75 0 25

6.7 15.5 1.1 62 38

CYC plant Concentration (mg/kg), dry basis in coal (ash basis) 76300 in LOI-free ¯y ash 80200 in LOI-free bottom ash 79800 Retention % 105 Mobility a, % 0 a

25.0 57.8 1.2 53 47

40.4 3.7 8.5 0.3 14.5 1.3 4.92 0.8 84 105 16 0

Se

51.3 , 0.5 14.8 , 1.0 87 13

27.2 7.6 2.2 21 79

Th

U

16.8 17.7 1.7 4.7 5.8 7.3 1.9 4.1 94 94 6 6

13.1 12.9 10.4 87 13

21.6 17.2 12.5 69 31

25.8 11.5 49.5 16.7 2.1 11.3 47 104 53 0

19.1 24.8 11.1 71 29

%Mobility ˆ 100±retention. Negative mobility values resulting from greater than 100% retention were assumed to indicate no mobility (see text). For concentrations below detection limits, 1/2 of the detection limits were used in the calculations of retention and mobility.

1666

I. Demir et al. / Fuel 80 (2001) 1659±1673

where CAl-output and CAl-input are the concentration of aluminum (Al) in the combination of ¯y and bottom ashes on an LOI-free basis and in the feed coal on an ash basis, respectively, of a power plant. This normalization of the retention of a trace element to the retention of Al was done to minimize the effect of sampling and analytical errors. Aluminum is a refractory element that occurs in large concentrations in coal, and is expected to have retention values similar to those of combustion ashes. Less than 1% of the ¯y ashes were expected to escape the particulate collection systems with ultra-®ne, air-borne ¯y ash particles [3]. Mass balances (retentions) of close to 100% for Al at the CYC and PC plants (Table 3) indicated the reliability of the sampling, analytical techniques, and retention calculations performed in this study. The Al mass balance for the FBC plant (76%) was less than the Al mass balance for the CYC and PC plants. This was likely the result of the dif®culty in obtaining representative samples from the FBC plant. Both the limestone and the coal used in the FBC plant were blends of products from many different quarries and mines in Illinois. Therefore, in future studies, several sets of samples should be collected over a period of several months of the FBC operation, and the average analytical data for these samples should lead to less variation. If the amount of an element retained in the CCRs accounted for 100% of its amount in the feed, then the mobility of the element was assumed to be zero. If the retention was less than 100%, then the difference was considered to be the percentage of the element mobilized during combustion. Negative mobility values resulting from excess retention values (.100% retention) (Table 3) probably resulted partly from analytical error and partly from

contamination of the CCRs by erosion of boiler hardware. For convenience, the negative mobility values were assumed to indicate zero mobility in this study. The mobilized portions of the elements were not necessarily emitted into the atmosphere, except in the cases of elements with low condensation temperatures, such as F, Hg, Se, and As. These elements were probably emitted into the atmosphere through the gas phase or partially through the gas phase and partially through condensation on ultra-®ne, air-borne ¯y ash particles. The mobilized portions of the elements with higher condensation temperatures may have condensed partially on the ultra-®ne, air-borne ¯y ash particles and partially on the power plant hardware after leaving the furnace. For convenience, the mobility values of 15 elements investigated were divided into three categories; low (,25%), moderate (25±50%), and high (.50%). For the FBC plant, the mobility of all elements, except F (85%), Mn (58%), and Cd (27%) was low (Fig. 4). Querol et al. [35] reported that Mn has an af®nity for Fe-oxide in combustion residues. The concentration of Mn in the CYC and PC feeds (coal) was less than the Mn concentration in the FBC feed (coal 1 limestone) (Table 3). Furthermore, CCRs from the CYC and PC plants contained more magnetite than the CCRs from the FBC plant (Table 2). This may have been why Mn had a greater mobility at the FBC plant than at the CYC and PC plants. The low mobility of the highly volatile elements Hg (17%) and Se (13%) at the FBC plant was somewhat surprising. The relatively low combustion temperature or the chemical environment created by the addition of limestone in the FBC seem to have reduced elemental mobilities. The effect of combustion temperature

Fig. 4. Combustion mobilities of 15 HAP elements at the FBC, PC, and CYC plants.

I. Demir et al. / Fuel 80 (2001) 1659±1673

was evident from the general increase in the mobility of the elements from the lowest-temperature plant to the highest one (Fig. 4). It was likely that the effect of the limestone addition was superimposed on the effect of temperature. For example, several authors [3,35±39] reported that lime, limestone, or Ca has the ability to capture substantial amounts of As, Hg, Sb, and Se during combustion. Suarez-Fernandez et al. [40], on the other hand, did not ®nd any major difference between the combustion behavior of trace elements in a laboratory-scale FBC unit with and without the addition of limestone. High mobilities were observed for the highly volatile elements F (92%, 89%), Hg (91%, 75%), and Se (79%, 53%) at both the PC and CYC plants (Fig. 4). Fluorine was likely emitted into the atmosphere as HF. Mercury was emitted as Hg 0, HgO, HgCl2, or CH3HgCl, and Se was emitted as SeO2 [3,41,42]. Arsenic is vaporized as As2O3 above 14008C [3], explaining the greater As mobility at the CYC plant than at the other plants. Beryllium, a moderately volatile element [3], showed high mobility at the CYC plant. It is possible that a large portion of Be was in the form of BeS in the coals or converted to BeS in the furnace. The BeS has a volatilisation temperature of ,15498C [43] and, thus, was mobilized at the CYC plant but not at the other two plants. Another possibility is that ultra-®ne, air-borne particles carrying Be and other elements into the atmosphere from the CYC plant were a relatively large portion of the total ash output. The ¯y ash-to-bottom ash ratio was 25:75 at the CYC plant (reverse of the ratio at the other two plants), resulting in smaller ash density and, accordingly, less agglomeration of the small particles through collision in the CYC ¯ue gas stream. However, a good Al mass balance (retention) for the CYC plant indicated that such mechanism did not take place to a signi®cant degree. Moderate mobilities were observed for Co (40%), Mn (42%), Ni (27%), and U (31%) at the PC plant and for As (47%), Co (46%), Mn (38%), Pb (25%), Sb (38%), and U (29%) at the CYC plant. The mobilities of other elements from the PC and CYC plants were low (0±15%).

1667

two to three times greater than the highest Ge concentration found in the ¯y ashes. However, Ge in the ¯y ashes could be concentrated easily by simple magnetic or particle size separation [49]. Past studies [35,40,50±60] suggest that As, Cd, Zn, and perhaps some other elements could also be concentrated to levels of economic importance by sieving and magnetic separation of the ¯y ashes. The PC bottom and ¯y ashes and the CYC ¯y ash contained 4±5 wt% magnetite (Figs. 5±7) which can potentially be recovered and used for metallic Fe production or for preparing dense ¯otation media for use in coal cleaning processes. The magnetite-rich fractions of the PC and CYC ¯y ashes that were separated crudely using a hand-held

3.3. Economic value of coal combustion residues 3.3.1. Extraction of valuable elements and minerals CCRs from the three power plants were enriched in some elements and depleted in others relative to the earth's crust, shales, or soils (Table 4). A number of elements have enrichment factors of greater than 10 relative to the three geologic materials. Particularly, Se and Ge were highly enriched in ¯y ashes from all three power plants. Selenium is produced industrially as a by-product of Cu re®ning [47], and a Se concentration of ,8 mg/kg reported for an Arizona Cu ore deposit [47] is within the range of the 7.4 to 43.3 mg/ kg Se found in the ¯y ashes of this study. Germanium is produced as a by-product of processing Cu±Pb±Zn sul®des. The Ge concentration (400±610 mg/kg) of a Cu±Zn ore in a Utah mine opened to produce primarily Ge and Ga [48] is

Fig. 5. Mineralogical compositions of ashes from the FBC, PC, and CYC plants. Keys: Mul, mullite; Calc, calcite; Qtz, quartz; Magn, magnetite; Hemt, hematite; Anhyd, anhydrite; Lime, lime, Portl, portlandite; Gyps, gypsym; Glass, amorphous phase.

4.0 0.3 0.6

34 2.6 4.9

3.0 0.2 0.4

51 3.9 7.2

1.2 0.1 0.2

FBC bottom ash with respect to crust to shale to soil

PC ¯y ash with respect to crust to shale to soil

PC bottom ash with respect to crust to shale to soil

CYC ¯y ash with respect to crust to shale to soil

CYC bottom ash with respect to crust to shale to soil

b

a

2.0 2.0 6.7

6.1 0.5 0.9

2.4 1.6 0.8

58 38 19

2.5 1.7 0.8

25 16 8.2

3.0 2.0 1.0

13 8.3 4.2

2.5 0.6 4.9 0.5 11.5 0.48

0.2 0.3 0.6

Cd

0.8 1.1 2.2

1.0 1.3 2.7

0.9 1.2 2.4

0.8 1.1 2.2

0.2 0.2 0.5

0.5 0.6 1.4

12.3 4.3 20 22 24.2 20.2

25 19 9

Co

1.1 1.2 2.0

2.3 2.5 4.1

1.4 1.5 2.5

1.8 1.9 3.2

0.4 0.4 0.6

0.8 0.8 1.4

75 35 175 136 227 112

100 90 55

Cr

0.0 0.0 0.1

0.4 0.3 0.8

0.0 0.0 0.0

0.1 0.1 0.2

0.1 0.1 0.2

0.1 0.1 0.2

53 48 67 5 230 15

544 800 300

F

0.3 0.1 0.2

9.1 4.1 7.3

0.1 0.0 0.1

1.3 0.6 1.0

0.1 0.1 0.1

3.1 1.4 2.5

0.25 0.01 0.10 0.005 0.73 0.02

0.08 0.18 0.10

Hg

0.6 0.6 1.0

0.3 0.4 0.6

0.5 0.6 0.9

0.5 0.6 0.9

0.3 0.4 0.6

0.7 0.7 1.1

620 310 468 468 310 542

950 850 550

Mn

0.9 1.0 3.5

2.3 2.5 8.7

1.1 1.3 4.3

1.0 1.1 3.9

0.2 0.2 0.7

0.9 1.0 3.3

66 14 77 85 173 70

75 68 20

Ni

0.5 0.7 1.0

0.7 1.1 1.5

0.3 0.4 0.5

0.5 0.7 1.0

0.5 0.7 1.0

0.6 0.8 1.1

567 480 524 262 742 524

1000 700 500

P

0.2 0.1 0.1

6.9 3.6 4.7

0.8 0.4 0.5

3.8 2.0 2.6

0.4 0.2 0.3

1.1 0.6 0.7

14.1 4.8 49 10 89.7 2.7

13 25 19

Pb

5.5 0.7 1.6

68 9.1 19

11 1.5 3.1

40 5.3 11

4.0 0.5 1.1

6.5 0.9 1.9

1.3 0.8 7.9 2.2 13.6 1.1

0.2 1.5 0.7

Sb

42 4.2 5.3

866 87 108

44 4.4 5.5

148 15 19

10 1.0 1.3

288 29 36

14.4 0.5 7.4 2.2 43.3 2.1

0.05 0.5 0.4

Se

The values for Earth's crust, shale, and soils were compiled from Swaine [2], Clarke and Sloss [3], Drever [44], Gluskoter et al. [45], and Krauskopf [46]. Enrichment factors of greater than 10 are shown in bold.

1.8 1.8 6.1

4.7 4.7 16

3.0 3.0 10

3.0 3.0 10

0.2 0.2 0.6

6 0.5 9 9 14 5.5

3 3 0.9

6.1 4 34 3 51 1.2

1 13 7

Be

FBC Ð ¯y ash Ð bottom ash PC Ð ¯y ash Ð bottom ash CYC Ð ¯y ash Ð bottom ash Enrichment factors b FBC ¯y ash with respect to crust to shale to soil

Concentrations (mg/kg) a Crustal average Shale average Soil average

As

Table 4 Concentrations of 15 HAPs, Ge, Zn, and Ti in earth's crust, shales, soils, and CCRs, and the enrichments factors for the same elements in the CCRs

1.6 0.9 1.3

2.0 1.2 1.6

1.4 0.9 1.2

1.8 1.1 1.4

0.3 0.2 0.2

0.8 0.5 0.6

5.6 1.9 12.6 10.4 14.6 11.3

7.2 12 9

Th

6.2 3.0 4.1

12 5.9 8.0

6.9 3.4 4.6

9.3 4.5 6.2

2.2 1.1 1.5

3.9 1.9 2.6

7.1 4.0 16.8 12.5 21.7 11.1

1.8 3.7 2.7

U

6.0 4.5 9.0

133 100 200

13 10 20

76 57 114

2.7 2.0 4.0

8.7 6.5 13

13 4 114 20 200 9

1.5 2.0 1.0

Ge

1.4 0.8 1.4

14 8.0 14

1.9 1.1 1.9

4.4 2.6 4.4

1.0 0.6 1.0

2.8 1.6 2.8

193 73 306 133 962 97

70 120 70

Zn

0.9 0.9 1.3

1.4 1.3 2.0

0.9 0.8 1.3

1.2 1.2 1.8

0.2 0.1 0.2

0.5 0.4 0.7

1980 660 5460 3900 6114 4016

4400 4600 3000

Ti

1668 I. Demir et al. / Fuel 80 (2001) 1659±1673

I. Demir et al. / Fuel 80 (2001) 1659±1673

1669

Fig. 6. SEM micrograph of (a) magnetite-rich fraction of PC ¯y ash and (b) magnetite-rich fraction of CYC ¯y ash. Particles of crystalline iron oxide (IO), glass (G), quartz (Q), anhydrite (A), lime (L), and mixed phases of quartz±glass (Q±G) and of iron oxide±glass (IO±G) were identi®ed with the help of EDX spectra of the particles. The EDX spectra of the IO particles in the PC ¯y ash contained mostly Fe and O peaks; the peaks of other elements, if they occurred, were relatively small. The EDX spectra of the IO particles in the CYC bottom ash generally contained relatively large peaks of Si, Al, and other elements, in addition to the Fe and O peaks, indicating that the magnetic and glass phases were fused together at high temperatures.

magnet still contained a large number of nonmagnetic particles (Fig. 6). In a potential industrial process, one might try to wash away the nonmagnetic particles while holding magnetic particles back with the help of a strong magnetic ®eld. As a result, ore grade magnetite fractions containing over 50 wt% Fe could be separated from the ashes, provided the magnetic particles consist of relatively pure Fe-oxide minerals (magnetite 1 hematite). This appears to be the case for the PC ¯y ash (Fig. 6a) and bottom ash (Fig. 7), but not for the CYC ¯y ash (Fig. 6b). Or in the case of PC bottom ash, the magnetic particles that are much smaller than the other particles (Fig. 7) could be separated using a combination of sieving and applied magnetic ®eld. 3.3.2. Other commercial applications CCRs are considered nonhazardous according to the

Resource Conservation and Recovery Act and their leachates generally meet the requirements of existing environmental regulations [61±69]. Some CCRs could even be used as adsorbents to remove certain metals from solutions [70]. The FBC ashes were low in LOI (Table 5) and high in alkaline and sulfate-rich minerals (Figs. 5 and 8). Their high alkalinity make them suitable for acid-mine drainage abatement, as well as for use in agriculture and in treating pyrite-containing coal cleaning wastes [9,13,71,72]. Anhydrite (or gypsum) in the FBC ashes can be concentrated through the use of hydrocyclones and ¯otation [73] for manufacturing a highly enriched sulfate fertilizer or gypsum wall boards. The FBC ¯y ash had high lime plus portlandite content (Fig. 5) and enough silica, alumina, and Feoxides (Table 5) to a be a suitable pozzolan and cementitious material in concrete and cement once their sulfate minerals

1670

I. Demir et al. / Fuel 80 (2001) 1659±1673

Fig. 7. SEM micrograph of PC bottom ash. Some particles of glass (G), Feoxide (IO), and Fe-oxide containing Al and Si peaks (IO-AlSi) were identi®ed with the help of EDX spectra of the particles. The bottom ash was ground for preparing the SEM/EDX sample. Therefore, the actual particle sizes were larger than indicated on the SEM micrograph.

are removed. The ASTM speci®cations [74] that at least a third of a ¯y ash should have particles smaller than 45 mm in size in order to be used as pozzolan in cement and concrete applications is easily met by the FBC ¯y ash (Fig. 8). The SO3 content and LOI of CCRs from the PC power plant were both less than 3.0 wt% (Table 5) which is about the maximum amount allowed in using them as pozzolan in cement applications. The CYC ¯y ash can also be used in cement application providing that its high LOI (unburnt carbon) content (Table 5) is lowered. The unburnt carbon may be recovered through ¯otation [76], electrostatic separation [77], or centrifugation [78], and then used as adsorbent carbon [12] or as clean fuel. The PC and CYC ¯y ashes would be a better pozzolan than the FBC ¯y ash because their particles are more spherical (Fig. 9) which would increase pumping ef®ciency, and provide a smoother ®nished surface with less permeability in construction when

Fig. 8. SEM micrograph of (a) ¯y ash and (b) bottom ash from the FBC plant. Some particles of glass (G), quartz (Q), glass±lime mixture (G±L), quartz±lime mixture (Q±L), lime (L), lime±anhydrite mixture (L±A), and anhydrite (A) were identi®ed with the help of EDX spectra of the particles. The bottom ash was ground for preparing the SEM/EDX sample. Therefore, the actual particle sizes in the bottom ash were larger than indicated on the SEM micrograph.

Table 5 Oxide compositions of the coal, ¯y ash, bottom ash, and limestone samples from the three types of electrical power plants burning Illinois coals. All values are in wt% on a dry basis SiO2

Al2O3

Fe2O3

CaO

MgO

K2O

Na2O

Ti2O

P2O5

MnO

SO3

LOI a

10.73 97.40 97.49 59.18

53.03 20.66 13.12 7.74

17.43 6.41 2.06 2.01

16.87 6.40 1.51 1.71

4.47 32.32 62.51 84.40

1.12 9.86 1.57 2.04

2.14 0.93 0.24 0.35

1.40 0.52 0.15 0.19

0.84 0.33 0.11 0.10

0.28 0.13 0.11 0.10

0.09 0.08 0.04 0.30

2.05 19.41 15.72 0.64

89.27 2.60 2.51 40.82

coal b ¯y ash bottom ash

10.66 97.48 99.96

51.13 52.06 46.56

14.45 16.01 13.05

20.73 17.00 30.63

5.25 5.10 5.49

0.84 0.85 0.76

1.69 1.86 1.33

1.31 1.54 0.80

0.75 0.91 0.65

0.09 0.12 0.08

0.09 0.06 0.06

3.28 1.57 0.20

89.34 2.52 0.04

coal b ¯y ash bottom ash

10.48 87.52 100.4

50.95 46.51 52.25

14.15 13.26 15.13

20.71 15.93 19.63

5.34 3.45 8.84

0.76 0.82 0.87

1.62 2.19 1.34

1.34 1.97 0.86

0.76 1.02 0.67

0.10 0.17 0.12

0.10 0.04 0.07

3.34 1.82 0.21

89.52 12.48 0.00

Plant type

Sample type

FBC

coal b ¯y ash bottom ash limestone b

PC

CYC

a b

10008C ash

LOI ˆ 100 2 %ash at 10008C. Oxide concentrations in coal and limestone are given on an ash basis.

I. Demir et al. / Fuel 80 (2001) 1659±1673

1671

4. Conclusions CCRs from a FBC plant, a PC plant, and a CYC plant in Illinois were analyzed chemically and mineralogically in order to understand the combustion behavior of the inorganic matter in the coal, and to propose potential bene®cial uses of the CCRs. Upon combustion, the mineral matter in coal decomposes to form the crystalline and amorphous phases that make up the ¯y and bottom ashes. Pyrite and marcasite were mostly converted to glass in the CYC bottom ash (because of melting at 16508C and subsequent quenching in water), and to hematite, magnetite, and glass in all other ashes. Calcite was converted to lime, and some of this lime reacted with SO2 and CO2 to form anhydrite and calcite, respectively, in all three plants. Clay minerals and feldspars were melted and converted to glass at the FBC plant with the help of the ¯uxing action of K2O, FeO, and CaO. In the PC and CYC ¯ue gas-¯y ash streams, kaolinite was converted to mullite, glass, and probably a small amount of cristabolite while other clay minerals and feldspars were converted to glass. Quenching in water suppressed mullite formation in favor of glass formation in the PC and CYC bottom ashes. Quartz survived the combustion process at the FBC plant but was partially converted to glass at the PC and CYC plants. Combustion mobilities of 15 elements of environmental concern in the feed coals were:

Location

FBC plant Fig. 9. SEM micrograph of (a) PC ¯y ash and (b) CYC ¯y ash. Some particles of glass (G), quartz (Q), anhydrite (A), Fe-oxide (IO), and organic matter±glass mixture (OM±G) were identi®ed with the help of EDX spectra of the particles.

added to cement and concrete. Like the FBC ¯y ash, the particle-size distribution of the PC and CYC ¯y ashes (Fig. 9) meets the ASTM speci®cation [74] for cement and concrete applications. Zeolites, high-capacity adsorbents used for chemical puri®cation and separation, could be synthesized from the PC and CYC ¯y ashes by reacting the glass fraction of the ash with alkali hydroxides. Probably converting the ash into zeolites can be done best at the boiler while the ash is still hot. Or, the activation process for cold ash could be accelerated to achieve zeolite synthesis through the use of microwaves [75]. Both the PC and CYC bottom ashes were composed mostly of glass (Fig. 5) that formed relatively large particles (Table 1, Figs. 3 and 7); they, therefore, can be used in stabilized road bases, as frits in roof shingles, and perhaps in manufacturing amber glass, especially after eliminating the magnetic fraction.

Low mobility (,25%)

As, Be, Co, Cr, Hg, Ni, P, Pb, Sb, Se, Th, U PC plant As, Be, Cd, Cr, P, Pb, Sb, Th CYC plant Cd, Cr, Ni, P, Th

Moderate mobility (25±50%)

High mobility (.50%)

Cd

F, Mn

Co, Mn, Ni, U As, Co, Mn, Pb, Sb, U

F, Hg, Se Be, F, Hg, Se

The mobilized portions of F, Hg, Se, and As, were emitted into the atmosphere through the gas phase or partly through the gas phase and partly through condensation on the ultra-®ne, air-borne ¯y ash particles. The mobilized portions of other elements probably condensed partially on the power plant hardware and partially on the ultra®ne, air-borne ¯y ash particles upon cooling. Relatively low mobilities at the FBC plant resulted from the cooler operating temperature (9008C) when compared with the PC (13508C) and CYC (16508C) plants and/or from the effect of limestone added to the coal. Size fractionation of the ¯y ashes through sieving could concentrate Ge and possibly other elements to levels of economic importance because the trace elements tend to

1672

I. Demir et al. / Fuel 80 (2001) 1659±1673

be enriched in smaller ¯y ash particles. The PC bottom and ¯y ashes and the CYC ¯y ash contained 4±5 wt% magnetite and up to 2±3 wt% hematite. The magnetic particles in the PC ashes were composed of relatively pure, ®ne-grained Fe-oxides, probably fused mixtures of magnetite and hematite. Therefore, an ore-grade magnetic fraction containing over 50 wt% Fe could be separated from the PC ashes using various combinations of applied magnetic ®eld, sieving, and water-washing. Most of the magnetic particles in the CYC ¯y ash were fused together with the glass phase and, thus, would be dif®cult to separate. The alkaline and sulfate-rich FBC ashes could be used in the neutralization of acidic soils, coal cleaning wastes, acidmine drainage, and other acidic industrial wastes. Commercial materials that could be produced from the FBC ashes include sulfate fertilizers, gypsum wall boards, concrete, and cement. The low LOI (resulting from the low organic carbon content), low SO3 content, and small, spherical glass particles in the PC ¯y ash are the characteristics desired in the production of cement, concrete, and ceramics. The CYC ¯y ash is not suitable for similar applications because its organic carbon content is too high (12.5%). However, this unburnt carbon could be separated and used as adsorbent carbon or as clean fuel, making the remaining fraction of the CYC ¯y ash attractive for cement, concrete, and ceramic applications. Zeolites are the other commercial materials that could be produced from both PC and CYC ¯y ashes. The PC and CYC bottom ashes are composed of mostly large, glassy particles and, therefore, could be used in stabilized road bases, as frits in roof shingles, and perhaps in manufacturing amber glass. Acknowledgements We thank R.A. Cahill, Y. Zhang, and J.D. Steele for chemical analysis of the samples, and A.T. Sanders for assisting with the SEM/EDX work. This study was supported, in part, by grants from the Illinois Department of Commerce and Community Affairs (IDCCA) through the Illinois Coal Development Board (ICDB) and the Illinois Clean Coal Institute (ICCI). Disclaimer Neither the authors, Illinois Department of Commerce and Community Affairs (IDCCA), Illinois Coal Development Board (ICDB), Illinois Clean Coal Institute (ICCI), nor any person acting on behalf of either: (a) makes any warranty of representation, express or implied, with respect to the accuracy, completeness, or usefulness of the information contained in this paper, or that the use of any information, apparatus, method, or process disclosed in this

paper may not infringe privately owned rights; or (b) assumes any liabilities with respect to the use of, or for damages resulting from the use of, any information, apparatus, method, or process disclosed in this paper. Reference herein to any speci®c commercial product, process or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring; nor do the views and opinions of the authors expressed herein necessarily state or re¯ect those of the IDCCA, ICDB, or ICCI. References [1] U.S. Public Law 101-549. Clean Air Act Amendments, Title 3. 104 Stat 2531±2535, 1990. [2] Swaine DJ. Trace elements in coal. London: Butterworths, 1990. [3] Clarke LB, Sloss LL. Trace elements Ð emission from coal combustion and gasi®cation. London: IEA Coal Research, 1992 (IEACR/49). [4] Wesnor JD. ABB's investigation into the utility air toxic problems. Paper presented at EPRI/EPA SO2 Conference, Boston, MA, 1993. [5] Davidson RM, Clarke LB. Trace elements in coal. London: IEA Coal Research, 1996 (IEAPER/21). [6] USEPA. Study of hazardous air pollutant emissions from electric utility steam generating units Ð ®nal report to Congress, vols. 1, 2. Washington, DC: United States Environmental Protection Agency, 1998 (EPA-453/R-98-004b). [7] ACAA. Coal combustion product production and use for 1999. Alexandria, VA: American Coal Ash Association, 2000. [8] Clarke LB. Application for coal-use residues. London: IEA Coal Research, 1992 (IEACR/50). [9] Dreher GB, Roy WR, Steele JD. Laboratory studies on the codisposal of ¯uidized-bed combustion residue and coal slurry solid, Environmental geology 150. Champaign, IL: Illinois State Geological Survey, 1996. [10] Hughes RE, DeMaris PJ, Dreher GB, Moore DM, Rostam-Abadi M. Brick manufacture with ¯y ash from Illinois coals. Carterville, IL: Illinois Clean Coal Institute, 1996 (®nal report). [11] Wirtz GP, Bukowski JM. Extruded honeycomb construction materials from ¯y ash. Carterville, IL: Illinois Clean Coal Institute, 1996 (®nal report). [12] DeBarr JA, Rapp DM, Rostam-Abadi M, Lytle JM, Rood ML. Valuable products from utility ¯y ash. Carterville, IL: Illinois Clean Coal Institute, 1996 (®nal report). [13] Roy WR, Dreher GB, Steele JD, Darmody RG, Tungate D, Giles WE, Phifer SC. Direct revegetation of coal slurry after amendment with FBC residues. Carterville, IL: Illinois Clean Coal Institute, 1997 (®nal report). [14] Dhir RK, Jones MR. Fuel 1999;78:137. [15] Niewiadomski M, Hupka J, Bokotko R, Miller JD. Fuel 1999;78:161. [16] Sloan JJ, Dowdy RH, Dolan MS, Rehm GW. Fuel 1999;78:169. [17] Barbieri L, Lancelotti I, Manfredini T, Queralt I, Rincon JMa, Romero M. Fuel 1999;78:271. [18] Demir I, Ruch RR, Damberger HH, Harvey RD, Steele JD, Ho KK. Fuel 1998;77:95. [19] Zhang Y, Talbott JL, Wiedenmann L, DeBarr J, Demir I. Adv X-ray Anal 1998;41:879. [20] Gluskoter HJ. Fuel 1965;44:285. [21] Hughes RE, Moore DM, Glass HD. In: Amonette JE, Zelazny LW, editors. Quantitative methods in soil mineralogy, Soil Science Society of America Miscellaneous Publication, 1994. p. 330. [22] Hughes RE, Dreher GB, Fiocchi T, Frost JK, Moore DM, RostamAbadi M, Swartz D. Brick manufacture with ¯y ash from Illinois coals. Carterville, IL: Illinois Clean Coal Institute, 1995 (®nal report). [23] O'Gorman JV, Walker PL. Fuel 1973;52:71.

I. Demir et al. / Fuel 80 (2001) 1659±1673 [24] Mitchell RS, Gluskoter HJ. Fuel 1976;55:90. [25] Reid WT. In: Elliott MA, editor. Chemistry of coal utilization, second supplementary volume. New York: Wiley, 1981. p. 1389. [26] Huffman GP, Huggins FE, Dunmyre GR. Fuel 1981;60:585. [27] Raask E. Mineral impurities in coal combustion. New York: Hemisphere, 1985. [28] Nayak RV, Bauer FW, Tonden TP. J Coal Qual 1987;4:37. [29] Mattigod SV, Rai D, Eary LE, Ainsworth CC. J Environ Qual 1990;19:188. [30] Reifenstein AP, Kahraman H, Coin CDA, Calos NJ, Miller G, Uwins P. Fuel 1999;78:1449. [31] Hatt RM, Bull DL. In: Bryers RW, Vorres KS, editors. Mineral matter and ash deposition from coal. New York: Engineering Foundation, 1990. p. 215. [32] Wertz DL, Collins LW, Froelicher F. Adv X-ray Anal 1988;31:343. [33] Evgueni J, Degterov S, Hayes PC, Petton AD. Fuel 1998;77:77. [34] Gibbon DL. Scanning Electron Microsc 1979;1979/I:501. [35] Querol X, Fernandez-Turiel JL, Lopez-Soler A. Fuel 1995;74:331. [36] Meij R. In: Chow W, Lewin L, editors. Proceedings of Second International Conference on Managing Hazardous Air Pollutants. Palo Alto, CA: Electric Power Research Institute, 1993. p. V83 (EPRI TR-104295). [37] Meij R. Distribution of trace species in power plant streams: a European perspective. Proceedings of 56th Annual Meeting of American Power Conference, vol. 56-I. Chicago, IL, April 25±27, 1994a. p. 458. [38] Gullet BK, Ragnunathan K. Energy and Fuels 1994;8:1068. [39] Bool LE, Helble JJ. Energy and Fuels 1995;9:880. [40] Suarez-Fernandez G, Querol X, Fernandez-Turiel JL, Benito-Fuertes AB, Martinez-Tarazona MR. Am Chem Soc, Fuel Chem Prepr 1996;41:796. [41] Bloom NS. Mercury speciation in ¯ue gases: overcoming the analytical dif®culties. Paper presented at the EPRI conference on ªManaging Hazardous Air Pollutants: State of the Artº. Washington, DC, November 4±6, 1991. [42] Meij R. In: Vourvopolis G, editor. Elemental analysis of coal and its by-products. Singapore: World Scienti®c, 1992. p. 299. [43] Yokoyama T, Asakura K, Seki T. Field study of trace element behavior in coal-®red plants. Komae: Komae Research Laboratory, 1991 (report ET 91002). [44] Drever JI. The geochemistry of natural waters. Upper Saddle River, NJ: Prentice-Hall, 1997. [45] Gluskoter HJ, Ruch RR, Miller WG, Cahill RA, Dreher GB, Kuhn JK. Trace elements in coal: occurrence and distribution. Champaign, IL: Illinois State Geological Survey, 1977 (circular 499). [46] Krauskopf KB. Introduction to geochemistry. New York: McGrawHill, 1967. [47] Rose AW, Hawkes HE, Webb JS. Geochemistry in mineral exploration. 2nd ed. New York: Academic Press, 1979. [48] Dutrizac JE, Jambor JL, Chen TT. Econ Geol 1986;61:946. [49] Hower JC, Rathbone RF, Robertson DJ, Peterson G, Trimble AS. Fuel 1999;78:197. [50] Hullet LD, Weinberger AJ, Northcutt KJ, Ferguson M. Science 1980;210:1356. [51] Eary LE, Rai D, Mattigod SV, Ainsworth CC. J Environ Qual 1990;19:202. [52] Mukhopadhyay PK, Lajeunesse G, Crandlemire AL. Int J Coal Geol 1996;32:279. [53] Davison RL, Natusch DFS, Wallace JR, Evans CA. Environ Sci Technol 1974;8:1107. [54] Meij R. Fuel Process Technol 1994;39:199.

1673

[55] Tumati PR, DeVito MS. Retention of condensed/solid phase trace elements in an electrostatic precipitator, Managing hazardous air pollutants: state of the art. Palo Alto, CA: Electric Power Research Institute, 1991. [56] Tumati PR, DeVito MS. Trace element emissions from coal combustion Ð a comparison of baghouse and ESP collection ef®ciency, Proceedings of Third International Conference on Effects of Coal Quality on Power Plants. Palo Alto, CA: Electric Power Research Institute, 1993 (EPRI TR-1-2280, p. 3±35). [57] Dale LS, Lawrensic SA, Chapman JF. Mineralogical residence of trace elements in coal Ð environmental implications of combustion in power plants, 1992. (CSIRO investigation report CET/IRO58). [58] DeVito MS, Jackson BL. Trace element partitioning and emissions in coal-®red utility systems. Paper presented at the 87th Annual Meeting and Exhibition of Air and Waste Management Association. 94WP73.03, Cincinnati, OH, June 19±24, 1994. [59] Helble JJ. Fuel Process Technol 1994;39:159. [60] Cereda E, Braga MGM, Pedretti M, Grime GW, Baldacci A. Nucl Instrum Methods Phys Res B 1995;99(1/4):414. [61] Akers DJ. Trace elements in coals and coal wastes. Palo Alto, CA: Electric Power Research Institute, 1989 (Interim report EPRI GS6575, projects 1400-6, -11). [62] Hassett DJ. Evaluation of leaching potential of solid coal combustion wastes. Indianapolis, IN: Indiana Coal Council, Inc, 1991 (®nal report). [63] USEPA. Environmental fact sheet, large-volume wastes from coal®red electrical utilities exempt as hazardous waste, EPA530-F-93014. Washington, DC: United States Environmental Protection Agency, 1993. [64] Rogbeck J, Elander P. In: Goumans JJJ, van der Sloot HA, Aalbers ThG, editors. Environmental aspects of construction with waste materials. Amsterdam: Elsevier, 1994. p. 589. [65] Tyson S, Blackstock T. Am Chem Soc, Fuel Chem Prepr 1996;41:587. [66] Hower JC, Robertson JD, Thomas GA, Wong AS, Schram WH, Graham UM, Rathbone RF, Robl TL. Fuel 1996;75:403. [67] Tuttle TE. Removing barriers: the Illinois ash use amendment, International ash utilization symposium volume. Lexington, KY: University of Kentucky, Center for Applied Energy Research, 1997. [68] USGS. Radioactive elements in coal and ¯y ash: abundance, forms, and environmental signi®cance. U.S. Geological Survey Fact Sheet FS-163-97, 1997. [69] Roy WR, Grif®n RA. Environ Sci Technol 1984;18:739. [70] Ayala J, Blanco F, Garcia P, Rodriguez P, Sancho J. Fuel 1998;77:1147. [71] Stout WL, Daily MR, Nickeson TL, Svendsen RL, Thompson GP. Fuel 1997;76:767. [72] Stout WL, Sharpely AN, Gburek WJ, Pionke HB. Fuel 1999;78:175. [73] Shoop KJ, Blystone SS, Kawatra SK. Zeta potential measurements on the components of wet ¯ue gas scrubber sludge. SME Preprint 96±99. SME Annual Meeting. Phoenix, AZ, March 11± 24, 1996. [74] ASTM. Concrete and aggregates, Annual book of ASTM standards, vol. 4.02. Philadelphia, PA: American Society for Testing and Materials, 1996. [75] Querol X, Alastuey A, Lopez-Soler A, Plana F, Andres JM, Juan R, Ferrer P, Ruiz CR. Environ Sci Technol 1997;31:2527. [76] Groppo JG, Robl TL, Graham UM. Mining Eng 1996;48:50. [77] Tondu E, Thompson WG, Whitlock DR, Bittner JD, Vasiliauskas A. Mining Eng 1996;48:47. [78] Maroto-Valer MM, Taulbee DN, Hower JC. Energy and Fuels 1999;13:947.