Environmental Geochemistry and Health, 1991, Vol. 13(2), page 119
The disposal of flue gas desulphurisation waste: sulphur gas emissions and their control R. Raiswell and S.H. Bottrell Department of Earth Sciences, Universityof Leeds, Leeds LS2 9JT, UK
Abstract Flue gas desulphurisation (FGD) equipment to be fitted to UK coal-fired power stations will produce more than 0.8 Mtonnes of calcium sulphate, as gypsum. Most gypsum should be of commercial quality, but any low grade material disposed as waste has the potential to generate a range of sulphur gases, including H2S, COS, CS2, DMS and DMDS. Literature data from the USA indicates that well-oxidised waste with a high proportion of calcium sulphate (the main UK product of FGD) has relatively low emissions of sulphur gases, which are comparable to background levels from inland soils. However, sulphur gas fluxes are greatly enhanced where reducing conditions become established within the waste, hence disposal strategies should be formulated to prevent the sub-surface consumption of oxygen.
Introduction Electricity in England and Wales is generated from a total of 74 power stations with a capacity of 54 GW (--67% generated by coal, 16% by oil and 9% nuclear). Coal, as the predominant fuel, is combusted at the rate of some 80 Mtonnes yr-1 (Brown and Halstead, 1989), which currently gives rise to the emission of -1-1.2 Mtonnes yr-1 of sulphur as SO2. Emissions from coal-fired power stations have been falling since 1979 and reductions will continue as plants are commissioned using more efficient coal combustion technologies, and as cleaner energy sources are increasingly exploited. There are, in addition, plans to fit flue gas desulphurisation (FGD) equipment on up to 8 GW of existing coal-fired generation capacity, starting with Drax (National Power) where 6 x 660 MW (3.96 GW) units are under construction and will be commissioned between 1993 and 1995 (Brown and Halstead, 1989). It is also planned to fit FGD to a further 2 GW of generating capacity at each of the Powergen operations at Ratcliffe-on-Soar and Ferrybridge C (Butcher, 1991). The FGD process chosen for Drax utilises a ground limestone/water slurry which is sprayed into the flue gases in absorber towers. Sulphur dioxide is absorbed into the slurry, where it reacts with the limestone to form calcium sulphite and/or calcium sulphate. A forced oxidation step is to be incorporated into the Drax system to ensure conversion of all calcium sulphite to sulphate, by which means it is intended that marketable quality gypsum (for the manufacture of wallboard) will be the overwhelming by-product. The 4 GW of FGD at Drax will consume 4).6 Mtonnes of limestone and will produce -0.8 Mtonnes of gypsum. By comparison, UK annual limestone extraction is 85 Mtonnes and gypsum extraction is 3.5 Mtonnes (of which 2.5 Mtonnes is used for wallboard). In the UK, the gypsum output from Drax is of a size
and quality which has allowed commercial outlets to be secured, and disposal should only be necessary for a relatively small fraction of the output. However, in the USA FGD will produce a gypsum surplus by the year 2000 (Butcher, 1991), moreover, many of the older FGD processes have produced a sludge of no commercial value, usually containing about 50% solids (calcium sulphate plus sulphite) which has to be stored in ponds or disposed as landfill, depending on the degree of dewatering carried out. The disposal of this type of waste most obviously causes problems where there is a potential for groundwater pollution. However, FGD wastes also emit sulphur-bearing gases into the atmosphere (Adams, 1979; Adams and Farwell, 1981). In this contribution we describe methods which can be used for the measurement of sulphur gas emissions from FGD waste disposal sites (including a brief description of a simple and rapid method for total reduced sulphur emissions) and, using literature data, attempt to assess the impact of emissions from different types of FGD waste (compared to natural sulphur gas emissions).
Measurement of Sulphur Gas Emissions Two types of method are generally used to measure gas fluxes from the ground to the atmosphere. The dynamic chamber method uses an open-bottom chamber which is placed over the surface to be sampled, and which is used to capture the gases being emitted. A carrier gas is introduced into the chamber to mix with the emissions, and the mixture is sampled and analysed for the gases of interest. Care has to be taken in the choice of chamber-lining material, as wall-loss rates can be significant. Kuster and Goldan (1987) report negligible losses for carbonyl sulphide (COS) and carbon disulphide (CS2), but losses can otherwise be significant for hydrogen sulphide (H2S), dimethyl sulphide (DMS, CH3-S-CH3),
The disposal of flue gas desulphurisation waste
dimethyl disulphide (DMDS, CH3-S-S-CH3) and mercaptans on Pyrex, polycarbonate and TFE Teflon surfaces. Loss rates are strongly influenced by water vapour. In micrometeorological methods, the concenlxation of gas is measured at various altitudes above the source, as are wind speed and direction. This sampling strategy requires the simultaneous measurement of low concentrations of gas at a high precision, which is often difficult to achieve in practice. Only dynamic chamber methods have been used for the measurement of sulphur gases from FGD waste disposal sites. Thus Adams (1979) and Adams and Farwell (1981) obtained samples of sulphur gases from ponds and wet and dry sludge sites by covering the water or sludge surface with a plastic isolation chamber whose edge provides a chamber-to-surface seal. For sampling pond waters, the chamber was floated using an inflated inner-tube. Air was pumped across the enclosed surface in the chamber at--0.15 m s-1 and a known volume cryogenically concentrated in 'U' traps, which were deactivated to minimise trace sulphur gas losses to the glass surface. Analyses were conducted by wall-coated, open-tubular, capillary column, cryogenic sulphur gas chromatography using a sulphur-selective, flame photometric detector. The main species detected were H2S, COS, CS2, DMS and DMDS. Four other organo-sulphur gases were also detected, two of which were probably propanethiols. Gas chromatography is commonly used to provide data on both the nature and abundance of sulphur gas species, although, since the most common gases are all highly odoriferous and have acidic oxidation products, total sulphur gas flux measurements can often be used to assess environmental impact. In these cases, Bottrell et al. (1991) use a modified dynamic chamber method to measure total reduced sulphur gas fluxes from the ground. A chamber (with an edge-to-surface seal) collects sulphur gases emitted from the ground. Air is pumped out of the chamber and bubbled through an NaOH solution, fixing sulphur gases for analysis in the laboratory using a sensitive fluorometric technique (with a range of 0.5-100 ppb S). This technique quantitatively measures H2S, COS, CS2 and DMS, effectively yielding an integrated flux of reduced sulphur compounds (Bottrell et al., 1991). Reconnaissance measurements have shown that this collection and measurement system is simple and robust, and yet can provide useful data on sulphur gas emissions (with precisions of 5-10%) from a variety of sites, including woodland and marsh, mine spoil and landfill. Sources of Sulphur Gas Emissions The natural cycling of sulphur compounds through different surface environments to the atmosphere is at present a major focus of research, against which the magnitude of anthropogenic sources are being assessed. Recent reviews can be found in Ivanov and Freney (1983), Saltzmann and Cooper (1989), Andreae (1990) and Aneja (1990). Together with other work, this literature suggests that there are several natural sources for each of the main
gases (H2S, COS, CS2, DMS and DMDS) identified in FGD waste by Adams and Farwell (1981). These sources are described below. Hydrogen sulphide Hydrogen sulphide (I-I2S) is most commonly produced either by dissimilatory sulphate reduction or by the decay of biogenic organo-sulphur compounds. In the former process, the sulphate ion acts as an oxidising agent for the dissimilation of organic matter by bacteria which are obligate anaerobes, but which are otherwise tolerant of a wide range of conditions. Sulphate reducing bacteria can tolerate temperatures of-5~ up to 80~ a pH range from 5 to 11, and pressures up to 180 NPa (Postgate, i979; McKinley eta/., 1985; Bath et al., 1987). Sulphate reducers are also found in natural waters of all salinities from zero to saturation. Rates of sulphate reduction are greatly dependent upon the metabolisability of the organic matter substrate (Westrich and Berner, 1984; Middleberg, 1989) with labile organic mailer producing rates of up to 0.5 mol L -1 yr-I (Reeburgh, 1983; Canfield and Raiswell, 1991). Sulphate concentrations also limit rates of sulphate reduction, but only significantly at concentrations of less that 5 mmol L-1 (Westrich, 1983; Boudreau and Westrich, 1984). Rates also increase with temperature (commonly by a factor of -3 for an increase of 10~ although the temperature dependence of the rate is linked in a complex way to the nature of the organic substrate (Westrich and Berner, 1988). Postgate (1979) also points out that H2S can be generated from the sulphite ion by micobial reduction using organic matter. The decay of organo-sulphur compounds can also be an important source of H2S. Plants contain an average sulphur content of 0.25%, mostly as cysteine and methionine which are readily degraded by a range of fungi and bacteria to release H2S (Clarke, 1953; Dunnette, 1989). Sulphur emission rates from decaying leaves are about 10--100 times higher than from living leaves of the same species (Lovelock et al., 1972). Emission rates of H2S have been determined for grasslands (Maynard et al., 1986), various crops (Lamb et al., 1987), lawns and pine forest (Delmas et al., 1980) and humid forest (Delmas and Servant, 1983). Carbonyl sulphide The daily cyclic variations of carbonyl sulphide (COS) in surface seawater have been interpreted as due to production by photochemical reactions (Ferek and Andreae, 1984; Andreae, 1990). Laboratory experiments showed that COS was produced in the presence of dissolved organic sulphur compounds (e.g. cysteine, methionine, glutathiomine, dimethyl sulphonium propionate), dissolved oxygen and light (the UV-B part of the spectrum). The reaction is strongly enhanced by photo-sensitising compounds, such as humic and fulvic acids (Zepp and Andreae, 1989). Taylor et al. (1982) produced COS from the experimental weathering of Fe, Pb, Zn sulphides, apparently in the presence of moisture alone. Carbon dioxide was not however effectively excluded from the vessel (and hence is a possible carbon source for COS and CS2), and organic compounds may have been present in the air samples injected into the reaction vessel to force out
R. Raiswell and S. H. Bottrell Shawnee D
Sulphur g m"2 y r r l i
photochemical oxid_afionin the presence of organic matter. Emissions of COS have additionally been reported from deciduous and coniferous trees (Lamb et al., 1987) and some crops (Guenther et al., 1989). COS may also occur as a product of the photochemical oxidation of CS2 (Andreae, 1990).
Shawnee Widows 5herburne A1 Creek
Carbon disulphide This gas (CS2) is also a product from the oxidative decomposition of sulphides along with COS (Taylor et al., 1982; Stedman et al., 1984; Nicholson et al., 1988). The relatively high fluxes of CS2 observed in coastal marsh environments (Adams et al., 1981; Steudler and Peterson. 1984) may originate either from the fermentation of organic-sulphur compounds, or from the reaction of terrigenous plant debris with polysulphides (formed by dissimilatory sulphate reduction). Otherwise CS2 is emitted by some plants and trees (Aneja eta/., 1979; Westburg and Lamb, 1984; Haines et al., 1989; Guenther et al., 1989).
Carbonyl Sulphide 0.02
Figure 1 Rates of sulphur gas emissions from FGD waste ponds. Data from Adams and Farwell (1981). gaseous reaction products. In a subsequent series of experiments Stedman et al. (1984) demonstrated that organic matter was necessary for the production of COS and CS2 from sulphides. The experimental conditions used by these workers may also have allowed COS formation by
Dimethyl sulphide Dimethyl sulphide (DMS) is evolved from dimethyl sulphonium propionate (DMSP) which is present in significant concentrations in most species of algae (Andreae, 1990). In the oceans, DMS is produced more by planktonic than benthic algae (Challenger, 1951). Keller et al. (1989) found that the highest concentrations of DMSP occurred in dinoflagellates, prymnesiophytes and chrysopytes (in the range 0.2--0.4 tool DMSP L-l of cell volume). The release of DMS from DMSP in algae occurs continuously at a slow rate, which increases when the algae are stressed, for example, by salinity changes, stirring or subaerial exposure (Andreae, 1990). DMSP is itself also released by algae, and can be broken down to DMS in seawater (Turner et al., 1988; 1989), the rate of this reaction being enhanced by the presence of microorganisms (Kiene, 1988). The relative significance of the direct emission of DMS by algae as compared to DMSP release and breakdown to DMS is not yet known. DMS can also be released by some plants, including spartina alterniflora, deciduous and coniferous trees and various crops (Aneja et al., 1979; Aneja and Cooper, 1989). On release DMS is normally oxidised to dimethyl sulphoxide (DMSO), and then to methyl sulphonic acid (MSA) with some SO2, but it may be decomposed anaerobically to H2S (Taylor and Kiene, 1989). DMS is also reported in domestic landfill (Young and Heasman, 1989) where it probably arises from the decay of protein sulphur compounds (Banwart and Bremmer, 1975). Dimethyl disulphide Relatively little is known about the origin of dimethyl disulphide (DMDS), although it can be produced by the decomposition of sulphur-bearing proteins (mainly methionine), and it is emitted by the aerobic and anaerobic decay of manures, sewage and plant material (Banwart and Bremner, 1975; 1976). DMDS is also found as a minor constituent of landfill gas (Young and Parker, 1983; Young and Heasman, 1989), and it is also an oxidation product of methyl mercaptan (Adewuyi, 1989).
The disposal of flue gas desulphurisation waste
g S m-2yr-.i
CT-121 Shawnee C
Sulphur Gas Emissions from FGD Waste
Hydrogen Sulphide 0,2
Adams (1979) and Adams and Farwell (1981) provide the only published data on sulphur gas emissions from FGD waste at a variety of disposal sites in the USA. Different FGD techniques were used to remove flue gas SO2 in scrubbers, using either lime, limestone or alkaline fly ash. The end-product was generally a calcium sulphate/sulphite slurry, which was transported to disposal ponds where the solids settled out. Even after prolonged settling and continued removal of the supernatant, the residual sludge is often thixotropic, and further treatment is necessary to stabilise or 'fix' the sludge. The main alternative to ponding is disposal to landfill, and the data reported by Adams and Farwell (1981) refer both to sludge ponds (sampled wet and dry, in fixed and unfixed states) and to landfill sites. The composition and concentration of sulphur gas emissions were concluded to depend mainly on the age and type of sludge (water content, sulphate/sulphite ratio) and the method of disposal. Here these conclusions are re-examined in the light of more recent studies on the formation of H2S, COS, CS2, DMS and DMDS, as discussed above.
N Carbonyl Sulphide
0.01 1 0 Dlmethyl Sulphide 0.03 ,-'5,,
Disposal by ponding Four of the sites studied by Adams and Farweii (1981) contained ponded wastes (Figure 1), two of which were large scale, operating ponds (Widows Creek and Sherbume) whilst the other two were demonstration-sized, water-covered storage sites (Shawnee D and A1). Waste was still being dumped at the first two sites but the other two had been filled several years before measurements of emissions were made. The total sulphur gas emissions (in g S m-2 yr-1) were 0.212 (Shawnee D). 0.103 (Shawnee A1), 0.197 (Widows Creek) and 0.048 (Sherbume). Except at Sherburne, H2S is usually a significant fraction of the total flux. All three sites which emit H2S also contain wet, relatively reduced sludge (with a comparatively large ratio of calcium sulphite/calcium sulphate). In contrast, the FGD technology used at Sherburne is designed to convert sulphite to sulphate, and hence produces wet but highly oxidised waste. This site is therefore a possible analogue for the disposal of FGD waste from Drax. Adams and Farwell (1981) suggest that H2S is probably produced by microbiological sulphate reduction. This process requires anaerobic conditions (see above), which could be established through the removal of oxygen by reaction with calcium sulphite: 2Ca SO3 + 02 ~ 2CASO4 This suggestion is supported by the lack of H2S in the highly oxidised (low sulphite) sludge at Sherburne. Nevertheless, for H2S production, an organic substrate is required. Such a substrate is also required for the formation of COS and CS2 (see above) one or both of which accompany H2S, albeit in relatively low concentrations
Figure 2 Rates of sulphur gas emission under hybrid disposal conditions. Diagonally lined blocks indicate wet disposal and clear blocks indicate dry conditions. Broken lines represent intermediate conditions. Data from Adams and Farwell (1981).
R. Raiswell and S. H. Bottrell Shawnee B g S m-2yr "1
T o t a l sulphur
~ I Hydrogen sulphide
o,1 o Carbonyl
J~r ,, s
0.02 ,,,J, s~
s a J~s
0 (Figure 1). Possible organic matter sources include wind-blown dust or biogenic activity in the ponds. Algae occur in the Shawnee A1 pond, but no biogenic activity was noted elsewhere, Burial of algae in anaerobic conditions beneath the sludge surface would certainly
provide a labile organic matter substrate for sulphate reduction. The decay of sulphur-bearing organic matter is another possible H2S source, but this is likewise dependent on the presence of biogenic material beneath the sludge surface. Adams and Farwell (1981) suggest that rates of sulphate reduction are controlled by nutrient trace element levels, but the lack of an obvious organic matter source for sulphate reduction makes it more probable that rates are limited by organic matter. Sulphate concentrations are almost certainly too high to limit rates of sulphate reduction
(see above). The gases COS and CS2 tend to be highest where H2S concentrations are low or absent (Figure 1). These gases can form by the oxidation of sulphide.s, and their abundance may therefore reflect kinetic factors. Thus, low rates of sulphate reduction at Widows Creek and Sherburne may result in H2S being generated at rates which match its oxidation to COS and CS2, hence the tendency of H2S to be lowest where COS and CS2 are highest. Seasonal temperature variations may also influence rates of H2S supply (by sulphate reduction) relative to rates of oxidation to COS and CS2. Sampling at Widows Creek occurred after three consecutive months with a temperature range of 24~ to 26~ compared to -12~ to 18~ at Sherburne. Thus it is not clear whether low HzS concentrations at Sherburne reflect low rates of sulphate reduction due to production (i.e. temperature effects) or destruction (in a relatively oxidising environment). Finally, we note that organo-sulphur gases (mainly DMS) are highest at Widows Creek and Sherbume, which both contain the youngest sludge. A similar age trend in organo-sulphur emissions has been noted from domestic waste disposal sites (Young and Heasman, 1989) due to the degradation of sulphur-bearing proteins. DMS in the sludge sites may arise similarly or might also be emitted from algae riving in the pond water. In either case, the presence of organo-sulphur gases provides conf'trrnation that an organic substrate must be present within the buried FGD waste.
Hybrid disposal Drainage conditions varied sufficiently at some sites (termed hybrid) for a comparison to be made of total sulphur gas emissions from the same sludge under both wet and dry conditions (Figure 2). Adams and Farwell (1981) note from this data that total emissions are mostly greater in the wet areas, but there are also other factors in operation. The sludge composition is well-oxidised (90--97% sulphate) at the Scholz CT121, CT101 and CEA/ADL sites, but the sulphate content is low at Shawnee C. Additionally, and crucially, there is an unusually large concentration of organo-sulphur gases in CT121 (the average flux of 0.174 g S m"2 yr-1 over the wet and dry areas is comprised of 0.028 g S m-2 yr-1 or 16% DMS plus 1 0.119 g S m-2 yr- 1 or 68% from two unident'fied organo-sulphur gases). Aneja (1990) suggests that these unidentified gases are non-sulphur compounds to which the flame photometric detector is sensitive. No age is given for the sludge at this site, but the other Scholz sites are all fairly old (2-4 years) and organo-sulphur species are accordingly
The disposal of flue gas desulphurisation waste Table 1 Range of emissions of H2S, COS, CS2, DMS and DMDS from well-oxidised FGD waste (g S m -2 yr-1).
H2S COS CS2 DMS DMDS
Inland soils (USA)
Salt marshes (USA)
0.0003-0.15 0.0001-0.017 0.00004-0.017 0.0003-0.0032