Temperature-Dependent Fractionation of Particulate Matter and


Temperature-Dependent Fractionation of Particulate Matter and...

0 downloads 105 Views 834KB Size

806

Energy & Fuels 2000, 14, 806-815

Temperature-Dependent Fractionation of Particulate Matter and Sulfates from a Hot Flue Gas in Pressurized Pulverized Coal Combustion (PPCC) M. Enders,*,† A. Putnis,† and J. Albrecht‡ Institut fu¨ r Mineralogie, WWUsMu¨ nster, Corrensstr. 24, D-48149 Mu¨ nster, Germany, and Lurgi Umwelt GmbH, Lurgiallee 5, D-60295 Frankfurt/Main, Germany Received August 18, 1999

The development of new highly efficient combined-cycle power plants is fundamentally linked to the development of new hot flue gas purification systems. Hot flue gas purification systems aim to remove particles and newly formed particulate matter from condensation reactions from the flue gas to prevent any damage to the gas turbine. This study focuses on the characterization of particulate matter which was deposited onto the inner walls of flue gas probes in a 1 MW pilot plant for pressurized pulverized coal combustion under slagging conditions (PPCC). The particulate matter in the flue gas stream originates from incompletely separated fly ash particles and particles formed during condensation processes from the hot flue gas. The phase analysis of the materials was done by high-resolution scanning electron microscopy. The combination of wavelength dispersive electron microprobe with a knowledge of the approximate temperature profile through the flue gas probe allowed the definition of dew points of single crystalline phases under the actual pressure conditions from the hot flue gas in the pilot plant. At high temperatures (>1200 °C) the deposits on the inner wall of the flue gas probe are mainly fly ash particles which were deposited in the liquid state. With decreasing temperature, CaSO4, Na2SO4, K2Ca2(SO4)3, and K3Na(SO4)2 condense from the flue gas. The ubiquitous occurrence of crystalline Cr2O3 can be traced to high-temperature refractory materials in the boiler and the subsequent particle separator of the pilot plant. At low temperatures some sulfates may decompose and form highly corrosive pyrosulfate melts. The changes in the phase composition of the deposits correlate well with chemical profiles along the flue gas probe.

Introduction The future development of power plants is driven by economic and ecologic considerations. The major targets of research are to reduce CO2 and other gaseous emissions, solid residues, and fuel consumption. In recent years the combination of modern gas turbine processes and the traditional steam process has led to commercially available highly efficient combined-cycle power plants based on clean hydrocarbon fuels (oil or gas). The maximum efficiencies attained in commercial combined-cycle power plants are close to 60%.1 Coal is an easily available and cheap fuel. A goal of current research in development of combustion power plants is the transfer of combined-cycle techniques to coal combustion power plants. Several models of coal burning combined-cycle power plants are under discussion, such as coal gasification, pressurized fluidized bed systems, and pressurized coal combustion under slagging conditions,2 where a coal fired turbine is implemented prior to a conventional steam process. From the viewpoint of efficiency high-temperature combustion concepts are favored, as the efficiency of a gas turbine is correlated to the gas entrance temperature.1 The * Author to whom correspondence should be addressed. Phone: 0049/251/8333500. Fax: 0049/251/8338397. E-mail: enders@ nwz.uni-muenster.de. † Institut fu ¨ r Mineralogie, WWUsMu¨nster. ‡ Lurgi Umwelt GmbH.

maximum attainable efficiencies for coal fired combinedcycle power plants are close to 55% for pressurized coal combustion under slagging conditions.1-3 However, this technology requires gas entrance temperatures into the turbine in the range of 1200-1300 °C.1-3 The very high gas entrance temperature is a challenge for the development of materials for the gas turbine. In contrast to hydrocarbon fuels such as oil and gas, coals are always contaminated with mineral matter. During the combustion process in pulverized coal combustion the mineral fraction of the raw coal appears as particulate and potentially volatilized material in the flue gas. This material may cause corrosion, erosion, or mechanical instability of the gas turbine. Therefore the particulate fractions have to be removed at high temperatures from the hot flue gas prior to the gas turbine. Techniques for hot flue gas purification are still in the stage of development.3,4 The fate of the mineral fraction of the coal during combustion processes is complex. The original mineral (1) Strauβ, K. Kraftwerkstechnik; Springer-Verlag: Berlin, Heidelberg, New York, Barcelona, Budapest, Hong Kong, London, Mailand, Paris, Santa Clara, Singapore, Tokyo, 1997; p 494. (2) Romey, I.; Rode, H.; Schuknecht, M. Erstes Statuseminar Druckflamm, 17.Nov. 1998; Hannes, K., Ed.; Zeche Zollverein Essen, Germany, 1999; pp 163-173. (3) Hannes, K. W. Proceedings of the EU Seminar .Status of Development and Market Penetration of Clean Coal Technologies (CCT) for Power Generation,, Du¨sseldorf, Germany, 5/6 Nov. 1998; Europ. Commmision/BEO, 1998; pp 1-15.

10.1021/ef9901793 CCC: $19.00 © 2000 American Chemical Society Published on Web 05/05/2000

Particulate Matter and Sulfates from a Hot Flue Gas

matter in the coal may be dehydrated, decomposed, physically altered, volatilized, or fused.5-7 The type and products of mineral reactions are determined by the composition of the ash fraction in the coal, the combustion temperature, and the combustion technique.5,6 The mineral matter in conventional power plants with a pulverized coal combustion technology is carried with the flue gas stream (fly ash). The flue gas heats a steam boiler via heater tubes. At temperatures in the range of 3-5 µm from the flue gas stream,4 it is obvious that small silicate particles in the lower micron and submicron range were deposited on the wall of the probe in liquid state and then coalesced and lengthened under gravity to form the observed large melt droplets. Smaller crystals resting on the surface of the melt droplets were determined from energy dispersive microanalysis to be Na2SO4 (needlelike, Figure 3B right side) and CaSO4 crystals (Figure 3B left side). At a distance of 35 cm from the hot end of the flue gas probe the lengthened melt droplets are still the most prominent larger scale feature. In contrast to the former observations, the surfaces of the droplets are densely covered with Na2SO4 crystals (Figure 3C,D). At 70 cm distance from the entrance aperture of the flue gas probe, the total number of melt droplets decreases (Figure 3E). The occurrence of well rounded fly ash spheres suggests that the melt temperature dropped below flow point. The surface of the inner wall of the probe is covered with spherical particles of agglomerated microcrystals. Spherical fly ash particles stick to these agglomerates (Figure 3F). At a distance of 100 cm from the hot end of the probe the amount of deposited molten mineral matter is reduced (Figure 3G) and the color of the quartz glass tube changes from reddish to greenish colors. In this area dust particles are attached to the top of a regular framework of tabular crystals (Figure 3G,H). From energy dispersive microanalysis and X-ray diffraction these tabular crystals were determined to be Cr2O3 (eskolaite). Chromium is only a trace element in the residual ash fraction of the Spitzbergen coal, and the observed enrichment in the deposit from the flue gas can only be explained by a partial mobilization of chromium as a volatile species from high-temperature insulation bricks in the boiler area. This effect is wellknown in high-temperature oxidizing atmospheres. The vaporized chromium later condenses as Cr2O3 (eskolaite) from the flue gas. At 135 cm from the flue gas entry into the probe the surface of the inner wall of the probe is covered with

Energy & Fuels, Vol. 14, No. 4, 2000 809

some dust consisting of fine grained spheres and agglomerated material (Figure 3I). Part of the chromium oxide surface is covered with small microcrystals (Figure 3J). According to energy dispersive microanalysis they are composed of sodium, potassium, sulfur, and oxygen. By comparison with X-ray results of flue gas filters from the same plant they are assumed to consist of K3Na(SO4)2 (aphthithalite).8,9 The regular crystal morphology indicates an origin as a condensate from the passing flue gas stream, rather than from a melt. At about 170 cm from the entrance aperture the most prominent feature is still a chromium oxide framework. The chromium oxide crystals are rather sharp edged. In some areas the framework is covered with a sulfurrich melt (Figure 3K, upper left corner). The local change in fabric cannot be explained by an overall change in chemical composition, but indicates a reaction of the previously formed condensates with the SO2-rich flue gas atmosphere. Agglomerated particles in this part of the flue gas probe are in the micrometer range. The source particles of the agglomerated particles are in the range of 100 nm (Figure 3L). Particles in this size range are assumed to form during condensation processes from flue gas.12 At the cool end of the probe (200 cm) the shape of the chromium oxide crystals has changed. The formerly clearly defined tabular crystals with sharp edges now have rounded edges and they appear to be covered with a continuous melt (Figure 3M, right top). On top of the chromium oxide crystals radially arranged CaSO4 crystals occur (Figure 3N). The shape of the crystals has changed compared to those observed in the entrance region of the flue gas probe (Figure 3B). The fibrous growth is an indication for a rapid growth from a melt. Electronprobe Microanalysis The chemical composition of the deposits on the flue gas probe was determined with wavelength dispersive electron microprobe analysis. Although these values cannot be interpreted quantitatively the relative changes of the chemical composition in relation to distance and temperature monitor the deposition of molten silicate particles and reactions from the hot flue gas to solidstate material (Figure 4). The accumulated data are compiled in Table 2. Broadly the analyzed elements can be divided in two groups: those enriched in the hot part of the flue gas probe (Na, Mg, Al, Si, Ca, Fe) and those enriched at the cool end of the probe (S, K, Cr). The binary distribution indicates a successive occurrence of different temperature-dependent processes leading to the inner wall deposits. Elements enriched at the hot end of the probe are refractory elements such as Mg and Al. Si can only be volatilized as SiO in highly reducing environments.7,13 The condensation of Cr2O3 from the flue gas confirms oxidizing conditions in the flue gas probe. Ca and Na are fixed in alumino-silicate phases and glass. The similar distribution pattern of Ca and Na compared to (12) Baumbach, G. Luftreinhaltung, 3rd ed., Springer-Verlag: Berlin, Heidelberg, New York, London, Paris, Tokyo, Hong Kong, Barcelona, Budapest, 1994; 461 pp. (13) Gumz, W.; Kirsch, H.; Mackowsky, M.-Th. Schlackenkunde; Springer-Verlag: Berlin, Go¨ttingen, Heidelberg, 1958; 422 pp.

810

Energy & Fuels, Vol. 14, No. 4, 2000

Enders et al.

Particulate Matter and Sulfates from a Hot Flue Gas

Energy & Fuels, Vol. 14, No. 4, 2000 811

Figure 3. Electron micrographs of the inner walls of the flue gas deposits. (A,B) At the entrance aperture, molten ash droplets with sulfate cover; (C,D) at 30 cm from the entrance aperture: molten ash droplets with sulfate overgrowth of small needlelike crystals; (E,F) at 70 cm from the entrance aperture: molten ash droplets and close up of a agglomerate and a fly ash sphere; (G,H) at 100 cm from the entrance aperture: chromium oxide substrate with dust and close-up with microcrystal; (I,J) at 130 cm from the entrance aperture: chromium oxide substrate with microcrystal of aphthithalite; (K,L) at 170 cm from the entrance aperture: chromium oxide with pyrosulfate melt and microparticle; (M,N) at 200 cm from the entrance aperture: chromium oxide with sulfate melt and newly formed calcium sulfate crystal.

Al, Si, and Fe indicates their primary association with molten silicate particles (Figure 3). The high sodium content at the very hot end of the flue gas probe exceeds the sodium content of the residual ash fraction from the raw coal. This enrichment of Na compared to raw coal can be explained by a condensation of Na2SO4 from the flue gas. Electron microscopy independently confirmed sodium sulfate crystals on top of the molten silicate droplet deposits.

With decreasing temperature the major components of the flue gas deposits change (Figure 3), with a reduction of Al, Si, and Fe and an increase of K, Ca, and S. The similar distribution pattern of K, Ca, and S in the flue gas probe suggests a common crystalline phase for these elements. Previous work described K2Ca2(SO4)3 (Ca-langbeinite) as a major sulfate component in flue gas deposits from the PPCC pilot plant.9 Although K2Ca2(SO4)3 has not been positively identified

812

Energy & Fuels, Vol. 14, No. 4, 2000

Enders et al.

Figure 4. Temperaturesconcentration relationship along the flue gas probe. The temperature had been calculated from the known temperature distance profile.

during electron microscopy the common distribution pattern of these elements is strong evidence for the formation of this phase from the flue gas. Chromium is highly enriched in the central part of the flue gas probe. Chromium is rarely bound into siliceous slag.9 The observed high enrichment of chromium in the flue gas deposits can be traced to hightemperature insulation materials in the boiler part of the plant and the subsequent particle and alkali separator (Figure 1). At very low temperatures close to the end of the flue gas probe an increase in the sulfate content of the flue

gas deposit can be observed. This increase in the distribution pattern is not linked to other elements which would be expected to form crystalline sulfates (K, Na, Ca) (Figure 4). This increase of SO3 content is evidence for the formation of a high sulfur phase at rather low temperatures. Discussion Flue gases are complex mixtures of gaseous components from the combustion together with solid or liquid residues of the mineral matter in the coal and vaporized species originating from both the mineral matter and

Particulate Matter and Sulfates from a Hot Flue Gas

Energy & Fuels, Vol. 14, No. 4, 2000 813

Table 2. Chemical Profiles for Different Sample Series through the Flue Gas Probe sample

0204V101

0204V107

0204V114

0204V121

0204V127

0204V134

distance [cm] temperature [°C] Na2O MgO Al2O3 SiO2 P2O5 SO3 K2O CaO TiO2 Cr2O3 MnO FeO sum

0 1200 5.0 2.0 14.5 33.3 2.0 12.4 3.9 6.8 0.9 13.0 0.1 6.1 100.0

30 1087 4.0 1.6 9.2 22.6 1.0 19.9 11.3 6.8 0.5 17.3 0.0 5.7 100.0

Flue Gas Probe 0204V1 65 946 3.5 0.5 2.3 8.8 0.7 15.3 8.0 2.3 0.1 56.7 0.1 1.7 100.0

100 841 2.6 0.6 1.8 4.4 0.7 15.2 7.9 2.7 0.1 62.7 0.1 1.3 100.0

130 733 3.1 0.8 2.1 8.4 0.9 18.6 9.4 4.0 0.1 50.9 0.1 1.6 100.0

165 607 2.9 0.7 1.8 10.5 1.5 19.5 6.7 3.6 0.1 51.4 0.1 1.4 100.0

sample

0205V201

0205V207

distance [cm] temperature [°C] Na2O MgO Al2O3 SiO2 P2O5 SO3 K2O CaO TiO2 Cr2O3 MnO FeO sum

0 1250 5.3 1.7 11.7 29.3 4.3 8.5 4.1 5.6 0.8 22.5 0.1 6.1 100.0

35 1087 7.5 0.9 3.5 11.2 1.8 21.6 5.7 5.0 0.2 38.6 0.1 3.9 100.0

sample

0206V101

0206V107

distance [cm] temperature [°C] Na2O MgO Al2O3 SiO2 P2O5 SO3 K2O CaO TiO2 Cr2O3 MnO FeO sum

0 1250 3.7 1.8 12.9 34.6 1.9 19.4 3.0 10.2 0.5 2.9 0.1 9.0 100.0

35 1087 7.0 1.3 4.9 49.9 1.0 18.8 5.9 6.9 0.3 1.0 0.0 3.0 100.0

0205V214

0205V221

Flue Gas Probe 0205V2 75 105 946 841 3.5 2.1 0.3 0.4 1.4 1.1 8.5 3.9 0.7 0.6 21.3 20.1 13.3 12.2 1.8 1.9 0.1 0.1 47.6 56.3 0.1 0.1 1.4 1.2 100.0 100.0 0206V114

0206V121

Flue Gas Probe 0206V1 75 105 946 841 6.7 4.2 2.9 1.5 7.1 3.1 14.0 38.5 1.1 0.7 28.0 17.6 8.5 4.5 9.7 5.5 0.5 0.1 11.3 21.5 0.1 0.0 10.0 2.9 100.0 100.0

organically bonded elements. During the cooling of the flue gas some components might stick to surrounding surfaces, while others attain supersaturation and start to sublime on materials surfaces or fly ash particles. In the following we describe the reactions from the flue gas to form solid state materials on the inner walls of the flue gas probe. However, the formulation of reaction equations is limited due to a lack of knowledge of the speciation of certain elements in the hot flue gas, especially for alkalies and alkali earths. At the entrance aperture of the flue gas probe, silicate melt droplets and films are a characteristic feature of the deposits. The size of single droplets is well above the upper size limit of particles passing the molten ash separator (3-5 µm).5 The form of the droplets confirm that they have been lengthened by gravity suggesting that a melt film is still fluid at this point inside the flue gas probe. This observation correlates with the flow point determined from physical analysis (Table 1). The temperature of the flow point also verifies the accuracy of the temperature profile along the probe.

0205V227

0205V234

0205V240

140 733 2.6 0.7 1.5 9.4 0.8 28.8 15.3 4.5 0.1 35.3 0.0 1.0 100.0

175 607 3.0 0.4 2.0 9.4 1.6 26.5 8.6 1.9 0.1 45.0 0.0 1.6 100.0

210 500 1.4 0.1 0.8 21.8 2.3 9.6 3.3 0.8 0.0 59.2 0.0 0.5 100.0

0206V127

0206V134

0206V140

140 733 4.5 0.9 1.5 34.1 0.6 17.3 4.6 3.3 0.0 31.7 0.1 1.3 100.0

175 607 1.8 0.5 1.3 56.2 0.5 10.2 2.8 2.4 0.0 22.6 0.0 1.6 100.0

210 500 0.8 0.4 1.0 79.7 0.3 3.0 0.9 0.9 0.1 11.9 0.0 1.1 100.0

Chemically the Al and Si are the major constituents of the deposit. The common distribution pattern of both Al and Si is evidence that they are bound to the same phase in the inner deposits and that they are fractionated during the same process. Furthermore a major vaporization and recondensation of Si during the combustionswhich is possible under highly reducing conditions7,13scan be excluded. At the low-temperature end of the probe the microprobe analyses for SiO2 are less reliable, due to a probable contribution of the silicon glass substrate to the Si signal. The concentration of Na is higher in the deposits than the Na concentration in the mineral matter of the raw coal. If sodium only partitioned into the silicate melt droplets, its concentration should be close to the concentration in the mineral matter of the raw coal (Table 1). The enrichment of Na to double the value compared to residual ash fraction of the coal (Table 1) can only be explained by additional input of Na as a volatile species which then condenses on the surface. The speciation of sodium in gases under high temperatures is still in

814

Energy & Fuels, Vol. 14, No. 4, 2000

Enders et al.

discussion. From a thermodynamic viewpoint Na can occur as Na-metal, NaOH and NaCl.7 Recent experiments by Chadwick and co-workers14 confirmed NaOH in the combustion gas of Australian coals. Using this information possible reaction paths leading to the formation of calcium sulfate and sodium sulfate are (eq 1 - eq 4):

CaO + SO2 + 0.5O2 ) CaSO4

(1)

Na2O + SO2 + 0.5O2 ) Na2SO4

(2)

4NaOH + 2SO2 + O2 ) 2Na2SO4 + 2H2O

(3)

2Na0 + SO2 + O2 ) Na2SO4

(4)

The proposed reaction schemes exclude the formation of CaSO3 and Na2SO3. Neither sulfite phase was found in filter residues from the same plant.9 The lack of sulfite phases in filter residues makes their occurrence as deposits from hot flue gas improbable. In hot flue gases (>600 °C) the concentration of SO3 as a gaseous species is very low. SO3 forms at temperatures less than 600 °C in catalytic reactions gas with oxygen.12 From this an indirect formation via the proposed reaction schemes (eq 1, eq 2, eq 3) of the sulfate phases is very probable. Na2SO4 occurs as small lathlike crystals which can be observed in the electron microscope. The occurrence of these crystals corresponds to a Na-enrichment on the surfaces of the silicate melt droplets. The occurrence of the sodium sulfate crystals is restricted to the part of the probe which experienced temperatures in the region of 1200 °C. This is in good correlation with previous results from a glass melt, where an increased partial pressure of SO2 lead to a raised upper thermal stability limit of Na2SO4.15 The small needles (Na2SO4) probably formed in condensation reactions from the gas atmosphere. Any other process leading to the formation of idiomorphic crystals would require a melt or fluid. If a crystal forms from a melt or a fluid the medium would only be consumed totally if its composition is identical to that of the newly formed crystal. In any other case residues of the melt or at least the occurrence of a second crystalline species should be observed. This is not the case. We therefore assume a heterogeneous crystallization of the given crystalline phases from the flue gas to the substrate. Calcium oxide is a highly refractory component which should not be vaporized during combustion. The formation of CaSO4 in gas-solid reactions is still unclear.7 The results of this study (Figure 3B) suggest the formation of CaSO4 via the reaction of highly reactive calcium oxide with the flue gas similar to those used for desulfurization of fluidized bed combustion systems.16 (14) Chadwick, B. L.; Ashman, R. A.; Campisi, A.; Crofts, G. J.; Godrey, P. D.; Griffin, P. G.; Ottrey, A. L.; Morrison, R. J. S. Int. J. Coal Pet. 1996, 32, 241-253. (15) Williams, R. O.; Pasto, A. E. J. Am. Ceram. Soc. 1982, 65, 602606. (16) Atimay, A. T.; Harrison, D. O. Desulfurization of hot coal gas; NATO ASI Series G, Vol. 42; Springer-Verlag: Berlin, Heidelberg, New York, Barcelona, Budapest, Hong Kong, London, Milan, Paris, Singapore, Tokyo, 1998; 408 pp.

The upper stability limit of crystalline Na2SO4 is reported to be at 884 °C.13,17 The temperatures at this point of the probe of 1200 °C should induce either melting of Na2SO414 or the decomposition of the mineral phase. However, the observed Na2SO4 is crystalline (Figure 3D). This observation can be explained by a considerable extension of the stability limit of this phase toward higher temperature due to a high partial pressure of SO2 in a coal combustion system with a total pressure of about 10 bar.3,4 With decreasing temperature the silicate melt droplets are still the dominating microtopographic feature on the flue gas probe walls. However, the surface of the melt droplets is increasingly covered with characteristic needleshaped microcrystals of Na2SO4 and also K2Ca2(SO4)3. The latter is evident from chemical profiles (Figure 4), where a significant increase in the potassium and sulfur concentration on the inner wall deposits is observed. The onset of K2Ca2(SO4)3 formation is at temperatures of about 1100 °C (Table 2). This temperature of formation is higher than temperatures reported by Trojer.18 A possible mechanism of formation is

K2O + 2CaO + 3SO2 + 1.5O2 ) K2Ca2(SO4)3 (5) Electron microscopy shows that sulfate phases (CaSO4, Na2SO4, K2Ca2(SO4)3) are enriched on the surface of the melt droplets in the hot part of the flue gas probe. With decreasing temperature the amount of deposited siliceous melt droplets on the inner walls decreases (Figure 3E). In this region the silicate particles become more viscous and sticky. This can be seen from the occurrence of well-rounded fly ash spheres in the submicron range (Figure 3F). The next characteristic change in chemical composition at temperatures of around 1000 °C is a major increase in the chromium content. Chromium condenses in the form of the Cr3+ oxide, mainly as small plates. The source of the chromium oxide crystals are hightemperature ceramics in the boiler. The hot oxidizing atmosphere leads to a major chromium mobilization as volatile species. In a future commercial PPCC plant a slag fur on top of the refractory bricks will prevent chromium emission. During the cooling of the flue gas in the pilot plant, this chromium condenses at temperatures less than 1000 °C. Figure 3G,H shows the formation of a dense fabric of tabular chromium oxide crystals. This central region of the probe is also very rich with other condensation products containing potassium and sulfur. Elements bound to silicate melt droplets such as aluminum and iron decrease. In some areas new sulfate crystals form; from the EDX spectra these crystals can be identified as K3Na(SO4)2 (aphthithalite). This phase has been previously described from a PPCC pilot plant8 and cement furnaces.18 Toward the end of the flue gas probe a melt-like overgrowth on chromium oxide crystals as well as a sulfate melt appear as new features in combination with a different morphology of newly formed CaSO4 crystals (17) Blachnik, R. Taschenbuch fu¨ r Chemiker und Physiker; SpringerVerlag: Berlin, Heidelberg, New York, Barcelona, Budapest, Hong Kong, London, Milan, Paris, Santa Clara, Singapore, Tokyo, 1998; 1463 pp. (18) Trojer, F. Die oxidischen Kristallphasen der anorganischen Industrieprodukte; E. Schweizerbart’sche Verlagsbuchhandlung: Stuttgart, 1963; 428 pp.

Particulate Matter and Sulfates from a Hot Flue Gas

Energy & Fuels, Vol. 14, No. 4, 2000 815

Table 3. Tabulated Results of Electron Microscopy and Microanalysis with Temperatures of Formation, Probable Mechanism of Formation, and the Source of Evidence newly formed phase

temperature of formation

reaction

evidence

fly ash CaSO4 Na2SO4 K2Ca2(SO4)3 Cr2O3 K3Na(SO4)2 K2S2O7

.1087 °C >1200 °C (1200 °C