Steam reactivation of 16 bed and fly ashes from industrial-scale coal-fired fluidized bed combustors

Steam reactivation of 16 bed and fly ashes from industrial-scale coal-fired fluidized bed combustors

Fuel 85 (2006) 94–106 www.fuelfirst.com Steam reactivation of 16 bed and fly ashes from industrial-scale coal-fired fluidized bed combustors D. Go´ra...

227KB Sizes 0 Downloads 15 Views

Fuel 85 (2006) 94–106 www.fuelfirst.com

Steam reactivation of 16 bed and fly ashes from industrial-scale coal-fired fluidized bed combustors D. Go´raa, E.J. Anthonya,*, E.M. Bulewiczb, L. Jiaa a CETC-O, Natural Resources Canada, 1 Haanel Drive, Ottawa, Ontario, Canada K1A 1M1 Institute of Inorganic Chemistry, Cracow University of Technology, ul. Warszawska 24, 31-155 Kracow, Poland

b

Accepted 31 May 2005 Available online 5 July 2005

Abstract The hydration behaviour of sixteen ashes, obtained from different commercial-scale fluidized bed combustors, has been investigated. Hydration is important for both ash disposal and reactivation of excess lime present in the ashes for further use in flue gas desulphurization. The techniques used were instrumental and conventional chemical analysis, thermogravimetry and X-ray diffraction. The ashes comprised both fly ash and bottom ash, with particle size less than 2 mm. The ashes were heat treated in air to oxidize free carbon and then hydrated with pressurized steam at about 170 8C, alone and with addition of pure CaO. It has been shown that steam hydration is effective in quantitatively converting CaO to Ca(OH)2, but in most cases the free lime content (i.e. CaOCCa(OH)2), expressed as CaO, decreases and added CaO enters into pozzolanic reactions with coal ash components, in part or even completely. Both the chemical evidence and X-ray phase analyses indicate that hydrated silicates and silicoaluminates are formed. The hydrated ashes are all able to take up additional SO2 and it appears that the presence of amounts of Ca(OH)2 detectable by phase analysis is not necessary for such capture. Crown Copyright q 2005 Published by Elsevier Ltd. All rights reserved. Keywords: Fluidized bed combustion; Bed ash; Fly ash; Hydration; Reactivation; Sulphation

1. Introduction One of the well-known advantages of fluidized bed combustion (FBC) is in situ flue gas desulphurization. At 850–900 8C, typical of FBC installations, when fuel and limestone sorbent are fed together, SO2 capture takes place in the combustor itself. The process can be represented by two global reactions: CaCO3 Z CaO C CO2

(1)

CaO C SO2 C ð1=2ÞO2 Z CaSO4

(2)

On thermodynamic grounds, reaction (2) should go to completion, with a Ca:S molar ratio of 1 or more, but sorbent utilization usually reaches only 20–45%. To achieve a satisfactory degree of SO2 removal, a Ca/S molar ratio of * Corresponding author. Tel.: C613 996 2868; fax: C613 992 9335. E-mail address: [email protected] (E.J. Anthony).

2–2.5 is required [1]. Thus the basic simplicity of the process carries a penalty of a considerable increase in the amount of solid wastes in the form of ashes, with a high content of unreacted CaO, causing problems with both surface deposition and utilization [2,3]. Various methods have been proposed, of either increasing sorbent utilization or reactivating excess sorbent in the ash for further SO2 capture. Controlled ash hydration is a method of both reducing the chemical activity of FBC ashes in landfill and of reactivation and the subject has recently been reviewed [4]. Liquid water, steam and pressurized steam can be used to treat the ashes [5–7] and the basic reactions are: CaO C H2 O Z CaðOHÞ2

(3)

CaSO4 C 2H2 O Z CaSO4 $2H2 O

(4)

However, gypsum formation (4) is slow and with reaction times up to one hour, can be ignored [1]. During the seasoning of FBC ashCwater systems or ash hydration, components derived from coal can interact with those from

0016-2361/$ - see front matter Crown Copyright q 2005 Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.fuel.2005.05.024

D. Go´ra et al. / Fuel 85 (2006) 94–106

the sorbent [8–12]. Literature accounts of ash behaviour during hydration are conflicting [4]. However, it is generally agreed that hydration leads to the reactivation of most bed ashes and that the hydration conditions are important, but one of the questions raised is whether fly ashes can be effectively reactivated to take up more SO2 [4,13]. In this work an attempt is made to compare the behaviour of a number of FBC ashes, both bed and fly ashes, hydrated in the laboratory under the same conditions.

2. Materials The ashes came from commercial North American companies (Pennsylvania and New Jersey) operating coalfired circulating FBC (CFBC) boilers, by courtesy of the Anthracite Region Independent Power Producers Association (ARIPPA) [14] and all these CFBC boilers fired coalmining wastes (i.e. high in ash components). Rock fragments with particle size over 2 mm were removed from the bed ashes. The ash origin and type (bed, BA, or fly ash, FA) are given in Table 1 for all samples examined. Table 1 FBC ashes from North American power plants No.

Code

Description

Origin

1

S1

2

S2

3

S3

4

S4

5 6

S5 S6

Bed Ash Cooler Bottom Ash Fluoseal Bottom Ash Multicyclone Fly Ash Baghouse Fly Ash Bottom Ash Fly Ash

7 8

S7 S8

Bottom Ash Fly Ash

C

9 10

S9 S10

Fly Ash Bottom Ash

D

11 12

S11 S12

Fly Ash Bottom Ash

E

13 14 15

S13 S14 S15

F

16

S16

Boiler#2 Fly Ash Boiler#1 Fly Ash Boiler#2 Bottom Ash Boiler#1 Bottom Ash

Symbol

Companies

A

Wheelabrator Frackville Energy Co., 475 Morea Road, Frackville, PA 17931

B

Piney Creek, 428 Power Line, Clarion, PA Tyco Valves and Controls, 505 Sharptown Road, Bridgeport, NJ 08014 Tractebel Power Inc., Northeastern Power Congregation Facility, McAdoo, PA 18237 Broad Mountain Partners, 40 Eleanor Ave, Frackville, PA 17931 Panther Creek Energy Facility, 4 Dennison Road, Nesquehoning, PA 18240

95

3. Experimental methods, results and discussion 3.1. Reashing and chemical analysis of the samples To avoid the reduction of CaSO4 by free C during thermogravimetric analysis (TGA) under N2 and to decompose any CaCO3 and Ca(OH)2 formed during ash handling and transport, samples were heated in air to 850 8C, for 1 h (furnace: Lindberg LBF794C). Reashed materials were analyzed for their free lime content and were used in the hydration experiments. Some samples were also phase-analyzed by X-ray diffraction (XRD). The chemical composition of the untreated ashes was determined at CETC-O, by the ICP (Inductively Coupled Plasma) method and standard chemical methods. Table 2 gives the main analytical components of the samples and the CO2 (carbonate) content. In seven cases the SO3 level was below the limit of the method, i.e. under 0.46%. Only in samples S5, S7 and S8 did the SO3 exceed 9%. These ashes were also rich in CaO. In the remaining ashes CaO was much lower, 2–9 mass %, and in one case, S4, under 1%. The CaO level was usually higher in the bed ash than in the fly ash from the same installation. In all ashes, SiO2 was the dominant component, which came from the coal ash (25–66%), followed by Al2O3 (9–23%) and Fe2O3 (3–8%). Table 3 gives some calculated quantities and ratios, which reveal a number of facts relevant to the interpretation of the results. The mass ratios of SiO2/Al2O3 and Al2O3/K2O vary much less than, and Al2O3/Fe2O3 to about the same extent as the individual figures. This is illustrated in Fig. 1A and B. The concentrations of SiO2, Al2O3 and K2O are clearly connected, while there is no such link between Fe2O3 and Al2O3. It should be noted that, in the BA samples, there was appreciably more SiO2 than in the FA ones. Roughly constant Al2O3/K2O corresponds to a molar ratio Al/K of about 7.2, without any significant difference between bed and fly ash. This must be connected with the nature of the mineral matter in coal. It is well known that in coal ash some of the SiO2 appears as quartz, but Al2O3 does not occur as alumina and is combined in clay minerals, e.g. kaolinite, Al4(OH)8Si4O10 [15]. These results thus indicate that, in the ashes examined, much of the analytical Al2O3 is associated with both SiO2 and K2O, and that the proportion of SiO2 in the form of quartz is probably higher in bed materials. Common minerals like muscovite and orthoclase contain the elements Si, Al and K in the ratio 3/3/1 and 3/1/1, respectively. The molar ratio CaO/SO3 ranges from 2.4 to over 30, but when SO3 is very low, it becomes meaningless. Assuming all SO3 to be combined in CaSO4, all ashes contained an ‘excess’ of CaO, as free CaO and/or other Ca compounds, OCC [1]. With respect to this overestimated excess, the amount of analytical SiO2 is very large. Hence, during hydration, the formation of not only Ca2SiO4 but of

D. Go´ra et al. / Fuel 85 (2006) 94–106

96 Table 2 Important chemical components of the ashes (mass %) Ash

Ash type

Component, in oxide form (wt. %) CaO

SiO2 S1, A S2, A S3, A S4, A S5, B

S6, B S7, C

S8, C S9, D S10, D S11, E S12, E

S13, F S14, F S15, F S16, F

BA BA 1 2 # FA FA BA 1 2 # FA BA 1 2 # FA FA BA FA BA 1 2 # FA FA BA BA

66.50 2.20 65.61 3.12 57.94 6.40 61.78 4.76 48.64 4.13 47.32 0.79 27.59 34.88 25.82 41.17 26.71 38.03 29.93 9.00 38.03 29.84 48.39 20.83 43.21 25.33 39.96 22.05 42.72 2.63 No analytical data available 56.44 2.26 61.02 7.05 59.74 4.95 60.38 # 6.00 47.26 6.59 47.09 6.07 65.75 5.96 65.00 5.95

Al2O3

Fe2O3

K2O

SO3

CO2

21.18 19.86 18.22 19.04 21.02 23.23 9.35 8.95 9.15 13.39 11.66 16.67 14.17 16.23 20.49

3.31 3.75 6.07 4.91 5.08 7.48 5.23 5.28 5.26 7.78 3.11 3.22 3.17 5.15 4.03

2.97 2.75 2.68 2.72 3.03 2.87 1.14 1.17 1.16 1.66 1.31 1.86 1.59 1.96 2.69

!0.46 1.64 4.02 2.83 !0.46 !0.46 17.60 11.35 14.48 1.69 11.03 3.69 7.36 9.15 !0.46

0.22 0.53 0.20 0.36 0.12 0.51 0.81 2.92 1.86 0.37 0.88 0.98 0.93 0.69 1.21

20.06 17.22 19.35 18.29 22.13 22.06 16.67 16.85

8.13 4.16 3.71 3.94 5.95 5.99 3.39 3.40

2.68 2.37 2.89 2.63 2.65 2.70 2.02 2.22

0.65 0.93 !0.46 O0.46 !0.46 !0.46 2.15 1.92

0.45 0.29 0.18 0.23 0.05 0.34 0.21 0.34

BA, bed ash; FA, fly ash; #, mean for two size fractions. Size fractions analyzed: BA-1, 0.150–0.425 mm; BA-2, 0.425–0.841 mm.

other silicates and silicoaluminates is possible in pozzolanic reactions [16]. If all the SiO2 were reactive, the availability of CaO could be the limiting factor in their formation.

3.2. Steam hydration To investigate the reactions and phase changes when ashes interact with water and to observe trends, hydration by

Table 3 Selected ash components and important mass and molar ratios Sample

S1, A S2, A S3, A S4, A S5, B S6, B S7, C S8, C S9, D S10, D S11, E S12, E S13, F S14, F S15, F S16, F

Ash type

BA BA 1 2 FA FA BA 1 2 FA BA 1 2 FA FA BA FA BA 1 2 FA FA BA BA

Notation as for Table 2.

SiO2/Al2O3

Al2O3/K2O

Al2O3/Fe2O3

CaO/SO3

SO3

CaO excess for SO3

SiO2/CaO

SiO2/CaO excess for SO3

Mass ratio

Mass ratio

Mass ratio

Molar ratio

Mass %

Mass %

Molar ratio

Molar ratio

6.40 5.30 3.00 4.14 3.11 1.79 1.70 1.72 3.75 5.18 3.15 5.08

O6.8 2.72 2.27 O12.9 O2.4 2.83 5.19 7.62 3.88 8.07 3.35 O8.2

!0.46 1.46 4.02 !0.46 !0.46 17.60 11.35 1.69 11.03 3.69 9.15 !0.46

O1.9 1.97 3.58 O3.8 O0.5 22.56 33.22 7.82 22.13 18.25 15.63 O2.3

30.2 19.62 8.45 10.99 44.2 0.74 0.58 3.10 1.19 2.17 1.69 15.16

!32.7 33.3 !15.1 !8.7 !88.3 1.14 0.73 3.57 1.60 2.47 2.39 !17.3

2.47 4.14 5.22 3.72 3.68 4.92 4.96

4.99 10.85 O15.4 O31 O19 3.96 4.43

0.65 0.93 !0.46 !0.46 !0.46 2.15 1.92

1.81 6.4 O4.6 O6.3 O5.8 4.45 4.61

23.31 8.08 11.26 6.69 7.24 10.29 10.18

29.01 8.90 !12.1 !7.0 !7.6. 13.8 13.2

3.15 7.13 3.30 7.22 3.18 6.80 2.21 6.94 2.04 8.09 2.95 8.20 2.88 7.65 2.23 8.07 3.26 8.90 2.90 8.96 2.46 8.28 2.08 7.62 No analytical data available 2.81 7.49 3.54 7.27 3.09 6.69 2.14 8.35 2.13 8.17 3.94 8.25 3.85 7.59

D. Go´ra et al. / Fuel 85 (2006) 94–106 80

A

70

y = 3.4713x - 2.6813 R2 = 0.8682

SiO2 [mass %]

60 50 40 30

y = 1.7707x + 9.8143

20

R2 = 0.5882 BA FA

10 0 0 3.5

5

20

25

B

3 K2O [mass %]

10 15 Al2O3 [mass %]

y = 0.1585x - 0.4099 R2 = 0.9296

2.5 2 1.5

y = 0.1305x - 0.0576 R2 = 0.8652

1 BA FA

0.5 0 0

5

10 15 Al 2O3 [mass %]

20

25

Fig. 1. Relationships between selected components of the FBC ashes examined. (A) Relationship between the analytical concentrations of SiO2 and Al2O3. (B) Relationship between the analytical concentrations of K2O and Al2O3.

steam was used. Hydration with steam at an elevated pressure had been investigated before [17]. Reashed samples, 6–6.6 g per charge, were ground (if necessary) and in some tests, mixed with pure CaO. Addition of CaO was used to test if any of it would be consumed during hydration. Deionized water, 20% wt., was then added to the prepared samples, and the slurry was transferred to a special basket and placed in a pressure bomb (Parr 4522M). The bomb was held for 1 h at about 165 8C (the corresponding overpressure being about 6.5 bar). The hydrated solids were filtered off (if necessary) under reduced pressure, to eliminate any excess water and placed in a vacuum oven to dry at 45 8C for 24 h, with the gases being changed four times. The low drying temperature was intended to minimize loss of water from any labile, hydrated compounds formed. ‘Free CaO’ present in the samples was determined, both before and after hydration, and hydrated ashes were also analyzed by TGA. Ashes S6-S9 and samples of these hydrated alone and with the addition of pure CaO, were also examined by XRD.

97

CaO and/or Ca(OH)2 to a soluble calcium compound that can be readily separated from the solids. When the Ca ions pass into solution, their concentration can be determined by any suitable analytical technique. The sucrose method (ASTM C25 Standard) was used here. The results are collected in Table 4, with the results for Ca(OH)2 derived from TGA included for comparison. For samples S5-S9, the XRD phase analysis results for lime and portlandite are also available (see later). At first sight, it is difficult to see regularities in the FL results, but except with FA samples S9 and S11, in the unhydrated samples FL was lower than the estimated excess CaO (Table 3). When the CaO excess was expected to be high, the FL results were also higher, but this was apparent only for bed ashes S5 and S7 and fly ash S8 (S7 and S8 came from the same source). These ashes were highest in both analytical CaO and SO3, so most of the Ca present was clearly sorbent derived. In other ashes, with less than 8% CaO, an appreciable proportion or even most of the analytical Ca was probably combined in OCC, inaccessible or only partly accessible to the sucrose method. On closer examination, the behaviour of the BA samples shows some trends, clearer on considering changes rather than the absolute figures. On hydration, the FL either fell by up to about 3% (ashes S5, S7, S12 and S15) or remained about the same, to within about 1% (S1, S2, S10 and S16). With the addition of 5 and 10% of CaO, after hydration FL rose with respect to the result without the addition, on average by about 2 and 5.4%, respectively, i.e. much less than expected, with all added CaO remaining free. With FA ashes the changes were less systematic. After hydration, FL fell for most samples and increased slightly only for S15. The largest drop, of over 5%, was seen for S8. In most cases (except for S5 and S6) there was some correspondence between FA and BA samples from the same source, Table 1. Hydration with added CaO led to increases in FL, which though somewhat erratic, were clearly considerably smaller than for the bed materials. In some cases the added CaO practically disappeared. For S8, after hydration FL dropped drastically and even after the addition of 10% CaO did not reach the level found for the unhydrated ash. The behaviour of the BA and FA samples is illustrated in Figs. 2A and 3A, respectively. The same results can also be considered in terms of relative effects, by dividing the differences by the FL values for the unhydrated materials. The relative effects are illustrated in Figs. 2B and 3B. Bed ash samples S1, S2 and S10 stand out, but for these the original FL levels were very low indeed. With the fly ashes, only S4 stands out, but again the unhydrated ash contained very little CaO.

3.3. Free lime analysis 3.4. Thermogravimetric analysis All wet chemical methods of determining ‘free lime’ (FL), i.e. the sum of Ca(OH)2 and CaO (expressed as CaO), are based on the use of a suitable reagent mixture to convert all

To investigate reactions associated with the loss of water and CO2 when hydrated samples are heated, TGA was

D. Go´ra et al. / Fuel 85 (2006) 94–106

98

Table 4 Mean ‘free CaO’ in the FBC bed ash samples, alone and with added CaO. TGA results included for comparison Bed ashes

Fly ashes

Sample

Analysis method

Sample

Analysis method

Ash No

Code CaO%

Free lime by FL as CaO [dry, % wt.]

Free lime by TG as CaO [dry, % wt.]

Ash No

Code CaO%

Free lime by FL as CaO [dry, % wt.]

Free lime by TG as CaO [dry, % wt.]

1A

S1 S1CH 0 S1CH 5 S1CH 0 S2 S2CH 0 S2CH 5 S2CH 10 S5 S5CH 0 S5CH 5 S5CH 10 S7 S7CH 0 S7CH 5 S7CH 10 S10 S10CH 0 S10CH 5 S10CH 10 S12 S12CH 0 S12CH 5 S12CH 10 S15 S15CH 0 S15CH 5 S15CH 10 S16 S16CH 0 S16CH 5 S16CH 10

0.66G0.02 0.48G0.03 3.43G0.43 6.15G0.35 0.96G0.05 0.56G0.04 2.62G0.28 6.39G0.60 20.37G2.42 17.28G1.04 19.23G0.38 22.05G1.46 14.91G1.00 11.62G0.99 14.74G1.03 16.08G0.37 0.84G0.08 0.58G0.12 2.93G0.57 4.79G0.49 5.40G0.91 2.78G0.19 6.23G0.88 9.27G1.58 3.11G0.48 1.86G0.19 4.49G0.16 8.37G0.12 3.63G0.51 4.24G0.98 5.79G0.46 8.72G0.14

– 1.0? 4.1 9.3 – 1.0? 3.5 8.7 – 35.0 29.0 31.0 – 20.4 19.8 22.3 – 1.0? 1.4? 4.2 – 1.9? 4.2 8.5 – 1.9? 3.9 7.1 – ? 3.7 7.4

3A

S3 S3CH 0 S3CH 5 S3CH 10 S4 S4CH 0 S4CH 5 S4CH 10 S6 S6CH 0 S6CH 5 S6CH 10 S8 S8CH 0 S8CH 5 S8CH 10 S9 S9CH 0 S9CH 5 S9CH 10 S11 S11CH 0 S11CH 5 S11CH 10 S13 S13CH 0 S13CH 5 S13CH 10 S14 S14CH 0 S14CH 5 S14CH 10

1.08G0.13 1.08G0.09 1.57G0.23 3.82G0.21 0.39G0.03 1.72G0.12 11.30G2.27 8.53G1.65 1.67G0.31 0.99G0.17 3.50G1.51 3.60G0.26 13.82G1.30 6.27G0.54 8.32G1.33 7.76G0.53 4.11G0.84 3.27G1.11 5.16G0.28 3.43G0.83 3.51G0.49 0.94G0.06 4.58G0.88 3.94G0.34 3.13G0.44 1.33G0.12 5.95G0.67 3.55G0.08 2.68G0.20 2.84G0.16 4.10G0.26 4.30G0.55

– 0 1.1? 2.2 – 0 0 0 – 0 0 4.0? – 7.2 7.8 7.0 – 0 0 0 – 0 0 0 – 0 1.5? 2.0? – 0 1.1? 1.8?

2A

*5 B

*7 C

10 D

12 E

15 F

16 F

4A

*6 B

*8 C

*9 D

11 E

13 F

14 F

Ash sources as for Table 2. CH, steam hydrated; *XRD results available.

employed. The instrument used was a Cahn 1100 with computer, so that TG and differential thermogravimetric (DTG) curves could be plotted. Reashed materials, hydrated alone and with CaO addition, were used, dried for 24 h in a vacuum oven, at 45 8C. Samples of 10–16 mg were employed and were examined under N2. The control program was split into 3 steps: from room temperature to 165 8C in 15 min, from 165 to 900 8C over 105 min (heating rate: 7 8C/min) and, finally, holding the temperature constant at 900 8C. Qualitatively, the TG and DTG curves obtained fell between two extremes. At one extreme, the effect clearly associated with Ca(OH)2 decomposition was clear, strong and after CaO addition, increased as expected, Fig. 4A. (S1). At the other extreme were the curves with no clear effects at all, except slow, continuous mass loss, Fig. 4B. (S14). The intermediate cases bring out the general problems with the interpretation of all TG curves, Fig. 5A and B (S8, S9). These problems are:

† Slow, continuous mass loss is always observed, throughout the whole temperature range † Ca(OH)2 decomposition may be superimposed on other effects, which are difficult to identify and to make allowance for † With some samples without added CaO, an effect is observed at just above 500 8C, disappearing after hydration with added CaO. Slow mass loss had been seen in earlier work with both FBC ash and related systems [12], with small effects below the temperature at which the decomposition of Ca(OH)2 could be expected. Some effects could appear to be superimposed and the mass loss ended at about 600 8C, well before the calcination of any CaCO3. The total mass loss, which could not be attributed to the presence of Ca(OH)2 alone, could amount to several per cent and was larger after hydration under pressure [12,18]. Similar effects had also been observed with autoclaved SiO2CCaOCH2O

D. Go´ra et al. / Fuel 85 (2006) 94–106

A 5 4 3 2 1 0 –1

S1

S2

S5

S7

S10

S12

S15

S16

–2 –3

FL change on hydration [mass %]

A 6

FLchange on hydration [mass %]

99

–4

15

10

5

0 S3

S4

S6

9

Relative FL change on hydration

8

Ash + 10% CaO

B Relative FL change on hydration

B

Ash + 5% CaO

7 6 5 4 3 2 1

S2

S5

Ash alone

S7

S10

Ash + 5% CaO

S12

S15

S16

Ash + 10% CaO

Fig. 2. Changes in the free lime content, FL, on the hydration of bed ash samples, alone and with CaO added. (A) Differences between FL with and without added CaO. (B) Differences between FL with and without added CaO, with respect to FL without hydration or CaO addition.

systems [19] and for seasoned ash-water mixtures [8]. The effect can probably be attributed to the loss of water from non-crystalline hydrated silicates and silicoaluminates [11, 12]. The present problem is that the apparent mass loss, continuing right up to 900 8C, could be partly an instrumental effect. If so, it could be about 2% over the temperature range used. This makes it impossible to assess the true magnitude of the effect of slow water loss. Its presence can only be inferred, from a slight change in the slope of some of the TG curves appearing around 550 to 600 8C. With the second group the situation was more complicated. For S8, in Fig. 5A, the decomposition of Ca(OH)2 is clearly superimposed over other, possibly noncharacteristic effects, the rate of mass loss gradually increasing from about 300 8C (the DTG curves are distorted) [20]. The FL obtained from such TG curves can only be approximate and is possibly overestimated. Finally, the curves for fly ash S9, Fig. 5B, show the disappearance of a small effect at about 500 8C after hydration with the addition of CaO. This effect was also seen with fly ash S6 and with bed ashes S15 and S16. It could be due to the presence of a minor component, probably containing OH groups, reacting

S13

S14

Ash + 5%CaO

Ash + 10% CaO

30 25 20 15 10 5 0 S3

S4

S6

S8

S9

S11

S13

S14

–5 Ash alone

S1

S11

–10

0 –1

S9

–5

Ash alone Ash alone

S8

Ash + 5%CaO

Ash + 10% CaO

Fig. 3. Changes in the free lime content, FL, on the hydration of fly ash samples alone and with CaO added. (A) Differences between FL with and without added CaO. (B) Differences between FL with and without added CaO, with respect to FL without hydration or CaO addition.

with added CaO. These complications make it difficult to obtain truly quantitative information for ashes; agreement was fair for S3 and S14 and tolerable for S6 and S13. For S4, S9 and S11 the TG curves did not give any indication of the presence of Ca(OH)2, even after CaO had been added. The sucrose method, however, matched the TG curves. An attempt has, therefore, been made to make the best possible semi-qualitative assessment of the mass loss associated with the decomposition of Ca(OH)2, and this led to the figures given in Table 4. For the bed ashes, agreement between FL from the sucrose method and the TG results is fair, except for bed ashes S5 and S7, for which the TG results were much higher than the FL figures and apparently hardly affected by CaO addition. The fly ash did give results with a tendency to increase after CaO addition. 3.5. XRD analysis The phase composition of five of the ashes, reashed and hydrated alone and with CaO addition, was determined by XRD (at Cracow University of Technology, using a Philips X’pert, instrument, with Cu anode). Two bed ashes were examined, S5 and S7, both relatively rich in CaO and SO3, and fly ashes S6, S8 and S9. The fly ashes exhibited a range of CaO and SO3 contents (Table 2). Tables 5–9 summarize

D. Go´ra et al. / Fuel 85 (2006) 94–106

100

A 100

A 100

99

98 96 S1CH

97

94 Weight [%]

Weight [%]

98 96

S1CH 5%CaO

95 94

S1CH 10%CaO

90

S8CH5%CaO

88 86

93

S8 CH

84

92

82

91 0

200

400 600 Temperature [˚C]

800

S8 CH10%CaO

80

1000

B 100

0

200

400 600 Temperature [˚C]

800

1000

B 100 S14CH

95

98

S14CH5%CaO

90

96 Weight [%]

Weight [%]

92

85 80

92 S9 CH10%CaO

90

S14CH10%CaO

75

S9 CH5%CaO 94

88

70 0

200

400 600 Temperature [˚C]

800

1000

Fig. 4. Examples of TG curves for hydrated ashes, with no interpretation problems. (A) Decomposition of Ca(OH)2 stands out clearly and changes in the expected manner after hydration with added CaO-bed ash S1. (B) No clear effects at all, except slow, continuous mass loss-fly ash S14.

the XRD results on a relative intensity scale. When possible, two lines were taken into account for each phase, and the strongest line in the XRD spectrum for a given ash was arbitrarily taken as having an intensity of 5. Quartz lines dominated all the XRD spectra, with no apparent relationship between their approximate absolute intensities and the analytical SiO2 content (Table 2). After hydration, there was a general tendency for the intensities of the quartz lines to fall. This effect was most marked for bed ashes S5 and S7, with the highest CaO content, and relatively low SiO2/CaO ratio (Table 3) and it is probably relevant to note that, with these ashes, FL dropped strongly on hydration. This could indicate a reaction between SiO2 and CaO. With the fly ashes, changes in the intensities of quartz lines were slight or even dubious. Anhydrite lines were present for all the samples and the line intensities varied in the order S5O S7OS6wS8OS9. Except for fly ash S6, this is the order of the analytical SO3 content of the samples (Table 2). There was some tendency for the peak intensities of anhydrite lines to increase after hydration, which could be due to improved crystallinity. Lines due to lime were present in the XRD spectra for S5, S7 and S8, with intensities in the order of the analytical CaO concentrations, i.e. S5OS7OS8. For the corresponding hydrated samples, XRD lines due to lime were replaced by those for portlandite. Their intensity increased after CaO addition. Lime lines were absent from the XRD spectra for ashes S6 and S9 and there was no

S9

86 0

200

400

600

800

1000

Temperature [˚C]

Fig. 5. Examples of TG curves for hydrated ashes, with interpretation problems. (A) The decomposition of Ca(OH)2 combined with other effectsfly ash S8. (B) No Ca(OH)2 present, but an unidentified effect at ca 500 8Cfly ash S9.

evidence for the formation of portlandite after hydration of these samples, even after the addition of up to 10% CaO. There was also clear evidence of the presence of some dehydroxylated muscovite, K2Al3Si3O11, in all five ashes, consistent with their relatively high potassium content. The XRD spectra of hydrated ashes contained katoite (Ca2.93Al1.97(Si0.64O2.65)(OH)9.44) lines. The formation of this phase had been observed before [11] and this is significant, since katoite is a hydrated calcium aluminosilicate [16]. In addition, the XRD spectra contained various weak lines and the computer program matched them to calcium (ortho)silicate, calcite, periclase, or cesanite. All of these were feasible as traces, except cesanite, Ca1.31Na4.32(OH)0.94 (SO4)3, which would require the presence of sodium. The lines in question were, therefore, attributed to an unidentified crystalline phase, ‘X’. Table 10 summarizes the most important information derived from XRD phase analysis for ashes S5-S9 (Tables 5–9) and brings it together with FL results from the sucrose and TG methods. It is clear that portlandite lines appear only if CaO is present in the reashed samples and that there is some correspondence between FL from the sucrose method, TG results and XRD line intensities. TG results appear to be least affected by the addition of CaO. When the CaO content of the ash is high, as with bed ashes S5 and S7, the sucrose and XRD results appear reasonable, but those from TG are

D. Go´ra et al. / Fuel 85 (2006) 94–106

101

Table 5 Summary of phase analysis results for S5 ashes alone and with CaO added Nr 1 2 3 4 5 6

Phase

7

Lime L Portlandite P Anhydrite A Quartz Q Hematite H Muscovite Dehydroxylated M Katoite K

8 9

Ca Silicate CS X

Formula

Angle, 2Q

Intensity, relative scale 0–5 S5

S5CH

S5CH 5%CaO

S5CH 10%CaO

CaO Ca(OH)2 CaSO4 SiO2 Fe2O3 KAl3Si3O11

37.34, 53.85 18.06, 34.10 25.44, 31.35 26.61, 20.82 24.14, 33.14 8.77, 19.67

1.74, 0.17 – 1.75, 0.46 4.68, 0.51 0.063, 0.18 0.087, 0.15

– 1.17, 1.20 2.88, 0.94 1.27, 0.21 K0.10 K0.066

– 3.57, 2.38 2.33, 0.82 1.09, 0.17 K0.097 –?

– 5.00, 2.70 1.99, 0.70 0.80, 0.20 K0.096 K0.060

Ca2.93Al1.97(Si0.64O2.65)(OH)9.44 Ca2SiO4 Unidentified

17.52, 20.25



0.21, 0.081

0.23, 0.075

0.21, 0.084

29.37, 32.67 10.79, 27.94

– –

0.17, 0.34 0.056, 0.078

0.077, 0.34 0.080, 0.11

K0.25 0.046, 0.067

Relative intensity scale 0–5; CH-hydrated ash.

Table 6 Summary of phase analysis results for S6 ashes alone and with CaO added Nr

Phase

Formula

Angle 2Q

Intensity, relative scale 0–5 S6

S6CH

S6CH 5%CaO

S6CH 10%CaO

1 2 3 4 5 6

CaO Ca(OH)2 CaSO4 SiO2 Fe2O3 KAl3Si3O11

37.34, 53.85 18.06, 34.10 25.4431.35 26.61, 20.82 24.14, 33.14 8.77, 19.67

– – 2.23, 0.64 4.85, 0.83 0.26, 0.98 0.074, 0.26

– – 2.25, 0.68 4.66, 0.86 0,23, 1.03 0.073, 0.26

– – 2.29, 0.73 5.00, 1.00 0.35, 1.00 0.088, 0.25

– – 1.71, 0.57 4.00, 0.75 0.28, 0.81 ?, 0.22

7

Lime L Portlandite P Anhydrite A Quartz Q Hematite H Muscovite Dehydroxylated M Katoite K

17.52, 20.25





0.36, 0.15

0.53, 0.17

8 9

Ca Silicate CS Calcite C

Ca2.93Al1.97(Si0.64O2.65)(OH)9.44 Ca2SiO4 CaCO3

29.37, 32.67 29.31, 47.12

– –

– –

– 0.47, 0.10

– –

Relative intensity scale 0–5; CH-hydrated ash.

Table 7 Summary of phase analysis results for S7 ashes alone and with CaO added Nr

Phase

Formula

Angle 2Q

Intensity, relative scale 0–5 S7

S7CH

S7CH 5%CaO

S7CH 10%CaO

1 2 3 4 5 6

CaO Ca(OH)2 CaSO4 SiO2 Fe2O3 Kal3Si3O11

37.34, 53.85 18.06, 34.10 25.44, 31.35 26.61, 20.82 24.14, 33.14 8.77, 19.67

3.17, 1.40 – 1.95, 0.62 5.00, 0.91 K0.12 –

– 1.23, 1.40 2.81, 0.72 4.43, 0.98 – –

– 2.20, 1.97 2.57, 0.99 3.65, 0.57 – –

– 2.32, 2.82 1.50, 0.68 1.83, 0.70 – –

7

Lime L Portlandite P Anhydrite A Quartz Q Hematite H Muscovite Dehydroxylated M Katoite K

17.52, 20.25



0.11, 0.050

0.16, 0.081

0.170.061

8 9 10 11 12

Ca Silicate CS Calcite C Periclase Pe Magnesite Ma X

Ca2.93Al1.97(Si0.64O2.65)(OH)9.44 Ca2SiO4 CaCO3 MgO MgCO3 Unidentified

29.37, 32.67 29.31, 47.12 42.91 32.65 10.79, 27.94

– – 0.13 0.098 –

0.057, 0.37 – – – –?

K0.37 0.084, 0.72 – – 0.046, 0.10

K0.33 – – – ?, 0.10

Relative intensity scale 0–5; CH-hydrated ash.

D. Go´ra et al. / Fuel 85 (2006) 94–106

102

Table 8 Summary of phase analysis results for S8 ashes alone and with CaO added Nr 1 2 3 4 5 6

7 8 9 10 11

Phase Lime L Portlandite P Anhydrite A Quartz Q Hematite H Muscovite Dehydroxylated M Katoite K Calcium Silicate CS Calcite C Periclase Pe X

Formula

Angle 2Q

Intensity, relative scale 0–5 S8

S8CH

S8CH 5%CaO

S8CH 10%CaO

CaO Ca(OH)2 CaSO4 SiO2 Fe2O3 KAl3Si3O11

37.34, 53.85 18.06, 34.10 25.44, 31.35 26.61, 20.82 24.14, 33.14 8.77, 19.67

1.02, 0.44 – 1.32, 0.14 3.66, 0.93 0.055, 0.17 K0.15

– 0.12, 0.19 1.81, 0.84 5.00, 0.73 0.08, 2, 0.17 0.063, 0.098

– 0.46, 0.56 0.91, 1.30 2.95, 0.65 K0.16 K0.071

– 0.14, 0.25 0.96, 0.86 2.90, 0.60 K0.13 –?

Ca2.93Al1.97(Si0.64O2.65)(OH)9.44 Ca2SiO4

17.52, 20.25



0.29, 0.13

0.38, 0.17

0.30, 0.11

29.37, 32.67







0.075, 0.53

CaCO3 MgO Unidentified

29.31, 47.12 42.91 10.79, 27.94

0.11 –

0.11, 0.10 – 0.11, 0.10

0.057, 0.20 – 0.10, 0.14

– – 0.10, 0.11

Relative intensity scale 0–5; CH-hydrated ash.

improbably high. On the other hand, when neither TG curves nor phase analysis provided any evidence for the presence of Ca(OH)2, the sucrose method yielded FL of about 3–5%. If this was due to the presence of portlandite, the Ca(OH)2 should give detectable XRD lines and mass loss on TG curves of about 1–1.6%, which should certainly be observable. Since there is no ready explanation for these observations, the methods of measuring FL need further verification. It is possible that FL determination by the sucrose method includes CaO from some labile silicates or aluminates formed during the hydration process [20]. Another feature of the results in Table 10 is that lines attributable to katoite and the unidentified phase ‘X’ appear in the XRD spectra of hydrated ashes, but only when CaO is present in the original sample or added before hydration. This suggests that both katoite and phase ‘X’ are formed in interactions between CaO or Ca(OH)2 and other ash components, such as silica or silicates and aluminosilicates,

i.e. in pozzolanic reactions [16]. Katoite contains OH groups and X is also likely to be hydrated, consistent with the slow loss of water evident in the TG curves and distortion of the effect by portlandite. It is worth noting that, compared with portlandite, in relation to Ca, katoite contains about 1.5 times more OH groups. When all added CaO is apparently consumed, the extra water combined during hydration could thus amount to as much as 3–5%. This is of the right order of magnitude to account for some of the observations. 3.6. Sulphation of hydrated ashes To test the effectiveness of hydration in reactivating the excess CaO in FBC ashes, sulphation of hydrated bed and fly ash samples S3-S14 was attempted in the TG apparatus. At the end of the usual heating program, as for portlandite determination, the temperature was held at 900 8C and the

Table 9 Summary of phase analysis results for S9 ashes alone and with CaO added Nr 1 2 3 4 5 6

7 8 9

Phase Lime L Portlandite P Anhydrite A Quartz Q Hematite H Muscovite Dehydroxylated M Katoite K Calcium Silicate CS Gehlenite G

Formula

Angle 2Q

Intensity, relative scale 0–5 S9

S9CH

S9CH 5%CaO

S9CH 10%CaO

CaO Ca(OH)2 CaSO4 SiO2 Fe2O3 Kal3Si3O11

37.34, 53.85 18.06, 34.10 25.44, 31.35 26.61, 20.82 24.14, 33.14 8.77, 19.67

– – 0.52, 0.18 4.96, 0.95 K0.19 0.22, 0.60

– – 0.51, 0.16 4.80, 0.90 0.080, 0.20 0.24, 0.68

– – 0.40, 0.13 5.00, 0.87 K0.18 0.22, 0.51

– – 0.19, 0.14 3.56, 0.42 K0.15 0.14, 0.42

Ca2.93Al1.97(Si0.64O2.65)(OH)9.44 Ca2SiO4

17.52, 20.25





0.28, –

0.25, 0.11

29.37, 32.67









Ca2Al(Al1.22Si0.78O6.78)(OH)0.22

42.44, 54.26





0.21, 0.073



Relative intensity scale 0–5; CH-hydrated ash.

D. Go´ra et al. / Fuel 85 (2006) 94–106

103

Table 10 Comparison of results from the sucrose method, TG and XRD phase analysis, for ashes S5-S9, alone and with CaO added Sample Ash No

5, BA, B

6, FA, B

7, BA, C

8, FA, C

9, FA, D

Analytical results

XRD information

Code CaO %

Free lime by FL as CaO [dry, %wt.]

Free lime by TG as CaO [dry, wt.%]

CaO Signal Intensity

Ca(OH)2 Signal intensity

Other crystalline phases present (other than quartz)

S5 S5CH 0 S5CH 5 S5CH 10 S6 S6CH 0 S6CH 5 S6CH 10 S7 S7CH 0 S7CH 5 S7CH 10 S8 S8CH S8CH 5 S8CH 10 S9 S9CH S9CH 5 S9CH 10

20.4 17.3 19.2 22.1 1.7 1.0 3.5 3.6 14.9 11.6 14.7 16.1 13.8 6.3 8.3 7.8 4.1 3.3 5.2 3.4

– 35.0 29.0 31.0 – 0 0 4.0? – 20.4 19.8 22.3 – 7.2 7.8 7.0 – 0 0 0

4650 – – – 0 – – – 7090 – – – 2300 – – – 0 – – –

– 4720 10000 14200 – 0 0 0 – 3140 4600 5030 – 230 970 280 – 0 0 0

A, H, M A, H, M, Kt, CS, X A, H,-Kt, CS, X A, H, M, Kt, CS, X A, H, M A, H, M A, H, M, Kt, C A, H, M, Kt,-X? A, H, Pe A,-Kt, CS-X? A,-Kt, CS, C, X A,-Kt, CS, C, X A, H, M, Pe A, H, M,-C, Kt, X A, H, M,-C, Kt, X A, H, M?- Kt, X A, H, M A, H, M,A, H, M, Kt A, H, M, Kt

Sample codes as for Table 4. Phase symbols as for Tables 5–9. Bold-strongest lines (apart from quartz), X-an unidentified phase (weak lines). Note: The error limits (given in Table 4) are now omitted. The qualitative correlations are important here, particularly the fact that FL could be determined when portlandite was not detected, by either XRD or TGA.

atmosphere was changed to one containing 5000 ppm of SO2. The mass of the sample rose, the reaction rate falling off with time. After 40 min the sample mass was still slowly increasing (suggesting that the increases in Table 11 are not the final ones). Fig. 6 shows examples of resulting TG curves for bed ash S5 (Fig. 6A) and fly ash S3 (Fig. 6B). To assess the degree of sulphation, the minimum sample mass after all water had been lost had to be accurately determined. This was difficult, because of the suspected instrumental effect mentioned above, but the results showed some systematic trends. For comparison, Table 11 includes collected data on the CaO content of the samples. As expected, hydrated samples containing Ca(OH)2 and with high FL, e.g. bed ashes S5 and S7, appeared to capture a great deal of SO2. Fly ashes high in analytical CaO (S8 and to a lesser extent S6 and S13) also appeared highly reactive towards SO2. Analytically, S8CH contained about 15% of excess non-sulphate CaO, with FL at 6.3%. The interpretation of the TG curves (Fig. 5A) was problematic, but the estimated Ca(OH)2 content, 7.2%, was close to the FL value, although in the XRD spectrum portlandite lines were very weak (Tables 8 and 10). In fly ashes S6 and S13 there was less analytical CaO and S6 contained little and S13 practically no SO3 (Table 2), so most of the analytical CaO present probably came from the coal ash. XRD detected no lime or portlandite in S6 before or after hydration (Table 6). FL was very low and TG confirmed the absence of portlandite. Yet after hydration these ashes appeared to capture appreciable quantities of

SO2 (Table 11). Even S4CH, lowest in non-sulphate analytical CaO gave some reaction and a non-negligible result for FL. In addition, for ashes with high excess CaO and portlandite present after hydration, the extent of SO2 capture appeared much too high, even with all excess analytical CaO reacting. With FA S8CH, if all CaO reacted, sample mass would increase by w22% and if only that suggested by FL and TGA methods, w9%, while the observed increase approached 40%. In an attempt to obtain more evidence, additional tests were carried out at a later date. A sample of unhydrated and hydrated bed ash S5 and S5CH, respectively, and hydrated samples of bed ash S7CH and fly ash S8CH were resulphated, by the same method as before, but the products were analyzed chemically by X-ray fluorescence, XRF, and examined by quantitative X-ray diffraction, QXRD. In all four cases, after the sulphation, the concentrations of all main analytical components fell in roughly the same proportions, except for that of SO3, which strongly increased. These results were thus consistent with the earlier ICP analysis (Table 2) and demonstrated that more SO2 had been captured. The molar ratios of S/Ca (i.e. SO3/ CaO) in the original ash (first stage of sulphation) and after the re-sulphation were calculated and were found to have increased from 0.20–0.29 to 0.71–0.82. The results, and for comparison, relevant information from Tables 2 and 11 are collected in Table 12. It can be seen that the XRF results indicate that in all four cases less than the full analytical amount of CaO reacted, as would be expected. Incomplete

D. Go´ra et al. / Fuel 85 (2006) 94–106

104

Table 11 Sulphation of hydrated and then calcined ash samples; comparison with ‘free CaO’ obtained by different methods ASH From

A A B C D E F F A A B C D E F F

Sulphation, sample mass, % Sample

Bed ashes S1CH S2CH S5CH S7CH S10CH S12CH S15CH S16CH Fly ashes S3CH S4CH S6CH S8CH S9CH S11CH S13CH S14CH

‘Free CaO’, mass %

With respect to original (dry) sample mass

With respect to min. mass

Excess, calc.

FL

TG

Initial

Final

Increase

Increase C

96.5 96.0 77.4 83.3 95.9 92.7 93.0 91.5

– – 137.0 115.0 99.6 97.4 – –

– – 59.6 31.7 3.7 4.7 – –

– – 77.0 38.1 3.9 5.1 – –

O1.9 2.8 27.9 20.2 – O5.5 4.5 4.6

0.5 0.6 17.3 11.6 0.6 2.8 1.9 4.2

1.0 1.0 35.0 20.4 1.0 1.9 1.9 ?

91.7 96.0 90.4 83.1 86.7 92.7 92.7 92.9

99.2 100.7 106.1 122.1 92.5 97.4 105.8 101.0

7.5 4.7 15.7 39.0 5.8 4.7 13.1 8.1

8.6 4.9 17.4 46.9 6.7 5.1 14.1 8.7

O3.8 O0.5 7.8 15.6 O2.3 1.8 O6.3 O5.8

1.1 1.7 1.0 6.3 3.3 0.9 1.3 2.8

0 0 0 7.2 0 0 0 0

Notation as for Table 4. Mass: Initial, as % of original mass of the dry sample, at the start of sulphation; Final, at the end of sulphation, % of original (dry) sample mass; Increase, increase due to sulphation, % of original (dry) sample mass; Increase, C, increase due to sulphation, % of sample mass at 900 8C, prior to sulphation. ‘Free CaO’: calculated, as % of original, unhydrated ash. FL, TG-referred to dry, hydrated samples.

hydration can reactivate fly ashes for further SO2 capture [4, 13]. The results seem to suggest that hydration can bring about an increase in the reactivity of fly ashes, but this may not be via the same mechanism as with bed ashes, S5CHSO2 A 140

120 110 100 90 80 70 0

2000

4000

6000 Time [s]

8000

S3CHSO2

95 90 85

All these observations indicate that after steam hydration a considerable proportion of the analytical CaO may be able to react, though it is still unclear whether this includes any of that present as OCC, in hydrated aluminosilicates such as katoite. Previous workers have shown that Ca silicates and OCC can capture SO2 under FBC conditions [21,22]. The results obtained may also have a bearing on whether

1000 900 800 700 600 500 400 300 200 100 0 10000

Temperature[˚C]

B 100

Weight [%]

† On re-sulphation of dry or hydrated ashes the additional CaO utilized is less than the analytical excess and may correspond to the FL, as determined by the sucrose method. † The earlier TG results may be in error. This had been suspected because of drifting curves and sometimes also poor curve stability. † In the additional experiments, hydration of lime-rich BA S5 gave no benefit, but with just one ash tested with and without hydration, it is impossible to generalize.

1000 900 800 700 600 500 400 300 200 100 0 10000 12000

Temperature [˚C]

130

Weight [%]

reaction and the presence of ‘other Ca compounds’ could account for the unreacted CaO. The QXRD method, however, suggested that the final anhydrite content of the re-sulphated samples was much higher than that calculated from the XRF results, as if all excess analytical CaO (and even some other components) could react, as implied by the earlier TGA observations. However, the problem may lie in the instrument settings or in the operation of the computer program. Bearing all these considerations in mind, it can probably be concluded that:

80 0

2000

4000

6000

8000

Time[s]

Fig. 6. Examples of TG curves for hydrated ashes, with calcination of hydrated samples followed by sulphation at 900 8C. (A) Calcination of Ca(OH)2, very large mass rise on sulphation-bed ash S5. No evidence of Ca(OH)2, clear mass increase on sulphation-fly ash S3.

D. Go´ra et al. / Fuel 85 (2006) 94–106

105

Table 12 Comparison of CaO to CaSO4 conversion after the re-sulphation of a raw dry ash and three hydrated ones with that in the original ashes Ash sample

BA S5 BA S5CH BA S7CH FA S8CH

Ash after first sulphation

Ash after re-sulphation

In dry ash, ICP

QXRD

In dry ash, XRF

SO3 CaO

calc. CaSO4

*calc. CaO

Molar S/Ca

Determined FL CaO

TG CaO

CaSO4

SO3 CaO

14.48, 38.03 14.48, 38.03 7.36, 25. 34 9.15, 22. 05

24.62

28.1

0.27

20.4



69.9

24.62

28.1

0.27

17.3

11.6

65.2

12.51

20.2

0.20

14.9

20.4

43.2

15.56

15.7

0.29

6.3

7.2

39.6

31.76, 27.17 29.47, 26.69 20.33, 19.90 20.46, 17.69

calc. CaSO4

*calc. CaO

Molar S/Ca

reac. CaO

54.0

4.95

0.82

21.0

50.1

6.06

0.77

19.4

344.6

5.67

0.71

12.9

34.8

3.37

0.81

11.5

Combined results from different methods, all concentrations in mass %. *calc., calculated excess analytical CaO; reac., CaO sulphated.

containing sorbent particles sulphated following the shellcore pattern [5]. Increased reactivity of hydrated fly ash may be associated with the conversion of the Ca present to reactive silicates and silicoaluminates. But even if all CaO could react, there is no ready explanation why in the first set of experiments the mass gains associated with SO2 capture by the ashes appeared so large. Drift and instability of TG curves should not lead to such huge errors. To elucidate these problems, all reaction products should have been analyzed, but this could not be done with the very small samples used in the TG apparatus. In the additional tests, with four selected ashes the products were indeed analyzed chemically by XRF and phase analyzed by QXRD. Although the results were not fully consistent, the tentative conclusions given above seem reasonable.

5.

6.

7.

4. Conclusions 1. Hydration by saturated steam at 165 8C is efficient in converting the CaO present in FBC ash to Ca(OH)2, but due to pozzolanic reactions with the coal-derived ash components, some of the CaO may be consumed in side reactions, with the formation of hydrated aluminates, silicates and aluminosilicates, such as katoite. 2. Steam hydration does not result in the formation of any gypsum. 3. The crystalline reaction products give rise to a number of weak and very weak lines in the XRD spectrum, but only katoite, Ca2.93Al1.97(Si0.64O2.65)(OH)9.44, could be identified with any confidence. Full identification was impossible, because of the small amounts of such compounds, possibly containing sulphate ions. In addition, some of the reaction products could be amorphous. 4. Different methods of assessing the amount of ‘free lime’ or reactive CaO in hydrated FBC ashes yield results that agree only at the qualitative level. In the sucrose method, the sugar solution might affect some Ca compounds other

8.

9.

than Ca(OH)2 and the TG curves may be distorted, making the determination of Ca(OH)2 unreliable. However, the ‘free lime’ from the sucrose method can probably be used for comparative purposes. Independent of the presence of any Ca(OH)2, the TG curves indicate that the hydrated ash samples can lose water practically continuously, with few, if any, characteristic effects. It is clear that the mass change of hydrated samples cannot be used as a simple measure of the extent of hydration. An appreciable proportion of the water taken up is combined in phases such as hydrated aluminosilicates. For hydration to be complete, free CaO must no longer be present, but its disappearance does not necessarily mean that it has been fully converted to Ca(OH)2. It can be consumed in pozzolanic reactions and in some cases as much as 10% of added CaO can react. Hence, in systems like FBC ashes, the extent of CaO to Ca(OH)2 conversion can only be defined unambiguously if both compounds are determined independently. Hydration by pressurized steam reactivates FBC ashes for further SO2 capture, but the simple reaction involving CaO and SO2 may not be the only process that occurs. It appears that some SO2 can be taken up even when neither CaO nor Ca(OH)2 are detectable. The difficulties encountered in this work indicate that the results of experiments on real ashes may be difficult to interpret, since the behaviour of such complex material can be influenced by many factors. Whenever possible, more than one experimental method should be used and the results carefully cross-checked; modern analytical methods should be used critically and with care.

Acknowledgements The authors acknowledge David Martin of ARIPPA for the supply of the ashes used in this work.

106

D. Go´ra et al. / Fuel 85 (2006) 94–106

References [1] Anthony EJ, Granatstein DL. Sulfation phenomena in fluidized bed combustion systems. Prog Energy Combust Sci 2001;27:215–36. [2] Smith I. Management of AFBC Residues, IEA Coal Research Report IEACR/21; 1990. [3] Mulder E, van Duin PJ, Grootenboer GJ. Solid residues. In: Valk M, editor. Atmospheric fluidized bed combustion. Amsterdam: Elsevier; 1995. p. 105–37. [4] Anthony E.J., Bulewicz E.M., Jia L. Problems and Issues Concerning Limestone-derived Sorbent Reactivation and the Use of Additives to Enhance Sulphur Capture Processes, Submitted to Progress in Energy and Combustion Science. [5] Laursen K, Duo W, Grace JR, Lim J. Sulfation and reactivation characteristics of nine limestones. Fuel 2000;79:153–63. [6] Wu Y, Anthony EJ, Jia L. Hydration of Ashes from a CFBC Boiler. Combust Sci Technol 2002;174:171–82. [7] Wu Y., Anthony E.J., Jia L. Steam Hydration of Partially Sulphated CFBC Ash, FBC 2003-060,17th International Conference on FBC, Jacksonville, FL; 2003. [8] Bulewicz E.M., Dudek K., Zoldani E., Mazurkiewicz M., Piotrowski Z. Application of AWDS Technology for Circulating FBC and LIFAC Ashes (PERD), 1994, Final Report to CANMET, DSS Contract No. 23440-2-9113/01-SQ. [9] Bulewicz E.M., Brydak Z., Kozak A. Fundamental Studies on Problems Associated with the AWDS Process for Circulating Fluidized Bed Combustion (CFBC) and Pulverized Fuel (PF) Combustion Ashes (PERD), 1995, Final Report to CANMET, DSS No: 23440-3-9075/01-SQ. [10] Bulewicz EM, Dudek K, Go´ra D, Taraszka J. Possible causes of the variability of FBC ash behaviour. Combust Sci Technol 2000;(153): 141–55.

[11] Go´ra D. Interactions between silica, aluminosilicates and the products of flue gas desulphurisation in FBC ashes; 2002, PhD Thesis, Cracow University of Technology, Krako´w, Poland (in Polish). [12] Bulewicz E.M., Go´ra D., Anthony E.J. The Behavior of Free Lime in FBC Ashes, FBC2003-152, 17th International Conference on FBC, Jacksonville, FL, ; 2003. [13] Wu Y, Charland JP, Anthony EJ, Jia L. A Study on the Reactivation of Five Fly Ashes from Commercial Scale FBC Boilers. Energy Fuels 2004;18:830–4. [14] ARIPPA, 2004, http://www.arippa.org/. [15] Smoot DL, editor. Fundamentals of coal combustion: for clean and efficient use. Coal science and technology, vol. 20. Amsterdam: Elsevier; 1992. p. 299–373. [16] Taylor HFW. Cement chemistry. 2nd ed. London: Thomas Telford; 1997. [17] CERCHAR process: 1991, French patent (WO91/10628. [18] Anthony EJ, Bulewicz EM, Dudek K, Kozak A. The long term behaviour of CFBC ash–water systems. Waste Manage 2002;(22): 99–111. [19] Paluch E, Gawlicki M, Biernacka-Zydek K. An assessment of the applicability of chemical methods for determining unreacted Ca(OH)2 in lime-sand materials (in Polish). Cement, Wapno, Gips 1971;11: 341–8. [20] Anthony EJ, Berry EE, Blondin J, Bulewicz EM, Burwell S. LIFAC ash-strategies for management, Waste Manage [accepted for publication]. [21] Dennis JS, Hayhurst AN. Alternative sorbents for flue-gas desulphurisation, especially in fluidised-bed combustors. J Inst Energy 1989;62: 202–7. [22] Anthony EJ, Iribarne AP, Iribarne JV, Jia L. Re-utilization of Land Filled FBC Ashes. Fuel 1997;76:603–6.