English

• 

  The properties in relation to SiO₂-Al₂O₃ containing materials at temperature beyond melting point are of vast interest to all relevant industries and sciences. To an extent, the knowledge on viscosity of melting ash is a core work in order to characterize the flow behavior of slag in gasification furnace and provide insights into the structure and dynamics of multiphase melts. For gasification, the viscous properties of the melting ash are decisive for the workability of the process. The chemical composition of coal ash resembles four principle inorganic components, silica (SiO₂), alumina (Al₂O₃), iron oxides and calcium oxide (CaO), which comprise more than 90% of the total compositions^[1-4]. Except these components mentioned above, magnesium oxide (MgO), sodium oxide (Na₂O), and titanium oxide (TiO₂) are also contained but in relatively low contents. Therefore, the quaternary system of SiO₂-Al₂O₃-CaO-FeO_x could be used to simulate the coal ash. Iron oxides and calcium oxide in coal ash and slag exert strong influence on the viscosity^[5-11]. Moreover, Fe-oxidizing morphology is very sensitive to the atmosphere during combustion and gasification^{[5, 7, 12-14]}. Actually, ash is exposed in the oxide-rich circumstance in combustion stage, therefore, Fe₂O₃ would be the dominating form^[15]. During the combustion of char, iron may be released into the flame in form of fused hematite (Fe₂O₃) or magnetite (Fe₃O₄), depending on the local oxygen concentration^[16]. Besides, carbon, carbon monoxide and even hydrogen acting as the reductive matters can lead to the reduction of Fe₂O₃ to the FeO and even to the metallic Fe. Reactions between iron compounds and alumina-silicate minerals may form low temperature iron-bearing spinel^[17]. Thus, fusion temperatures in reducing atmosphere are lower than that in oxidizing atmosphere^{[18, 19]}. Similarly, calcium oxide was also considered as a common and efficient fluxing agent for decreasing the viscosity of aluminosilicate materials^[9]. A deeper insight into the interactions among iron oxides and calcium oxide at melting state is vital and inevitable to the comprehensive understanding of the melting behavior and the evaluation of the viscosity of the material involving those components.

  Accurate prediction of viscosity values requires high-quality measurements on molten samples under conditions closing to gasification^[20]. General method conducted in oxidizing or reducing atmosphere with real ash and slag provides valuable information^{[6, 8, 14]}. However, even the gas flow in the furnace is constant and continuous, the contact surface of melting slag and gas is narrow, which, to a large extent, restricts the simulation of the actual gasifier condition.

  In this study, quaternary system of SiO₂-Al₂O₃-CaO-FeO_x with different iron valences was employed to simulate the different atmospheres of coal gasification process. The melting and reaction behavior was investigated by means of thermo-mechanical analyzer (TMA), differential scanning calorimetry (DSC), X-ray diffraction (XRD) as well as the thermodynamic simulation modeling by “FactSage”. Viscosity was measured by a high temperature rotary viscometer with temperature ranging from 1 700 ℃ to re-solidification temperature. The reactions of the major compositions in such system, especially the interactions between iron oxides and calcium oxide were tentatively investigated.

  1    Experimental

  1.1    Sample preparation

  Samples were prepared using the pure bulk chemical compositions and numbered as “S” series. Characteristics of chemicals used were silicon dioxide (SiO₂), CAS: 60676-86-0, fused granular, AR (analytical reagent) grade; aluminum oxide (Al₂O₃), CAS: 1344-28-1, powder; AR grade; calcium oxide (CaO) CAS: 1305-78-8, powder, AR grade; iron oxide (Fe₂O₃), CAS:1309-37-1, powder, AR grade; magnetite (Fe₃O₄), CAS: 1317-61-9, powder, AR grade and iron (II) oxide (FeO), CAS: 1345-25-1, powder, AR grade. All chemicals were vacuum sealed prior to use. Accurate amounts of chemical agents were well mixed with desired proportions, and then subjected to alumina ball-mill for 5 h followed by being sieved to under 45 μm.

  1.2    Experimental conception

  Based on the real coal ash compositions, six analytical compounds, SiO₂, Al₂O₃, CaO, FeO (note: as FeO_10-1.1), Fe₂O₃ and Fe₃O₄ were selected, and their amounts were normalized to 100%. Eleven samples were prepared and proceed to the analyses and measurement. As shown in Table 1, all the samples have a fixed SiO₂/Al₂O₃ ratio being at 1.89. S01-S04 and S05-S08 were separated according to iron valence. Samples S01-S03 were adjusted using binary composition of SiO₂-Al₂O₃ and ternary composition of SiO₂-Al₂O₃-Fe₂O₃ or SiO₂-Al₂O₃-CaO to investigate the reciprocity among single components. Though samples S04 and S08 were chemically similar to S07 and S11 respectively, they experienced different viscosity measurement processes. Specifically, the viscosities of S07 and S11 were measured by graphite crucible instead of alumina crucible, in which iron oxides was completely reduced to the metallic iron.

  Table1. Composition in SiO₂-Al₂O₃-CaO-FeO_x quaternary system (as equivalent oxide)

  表选项

  导出Excel

  下载全尺寸表图像

  Sample	Composition w/%									Mol ratio
  	SiO2	Al2O3	CaO	Al2O3-crucible				C-cruciblea		SiO2/Al2O3	B/Ab
  				Fe2O3	Fe3O4	FeO		Fe2O3
  S01	65.36	34.64	-	-	-	-		-		1.89	0.00
  S02	49.29	26.09	-	24.62	-	-		-		1.89	0.33
  S03	49.18	25.98	24.83	-	-	-		-		1.89	0.33
  S04	39.52	20.89	19.94	19.65	-	-		-		1.89	0.66
  S05	39.78	21.02	20.07	-	19.13	-		-		1.89	0.66
  S06	40.31	21.31	20.34	-	-	18.04		-		1.89	0.66
  S07	39.52	20.89	19.94	-	-	-		19.65		1.89	0.66
  S08	39.78	21.02	8.25	30.94	-	-		-		1.89	0.64
  S09	40.19	21.24	8.34	-	30.22	-		-		1.89	0.64
  S10	41.05	21.70	8.52	-	-	28.74		-		1.89	0.64
  S11	39.78	21.02	8.25	-	-	-		30.94		1.89	0.64
  a C-crucible was used to simulate the reduction atmosphere, Fe2O3 was reduced to iron metallic particles during measurement; b base/acid ratio (base=w(Fe2O3)+w(CaO), acid=w(SiO2)+w(Al2O3)), where only four analytical compounds were calculated, and the FeO, Fe3O4 were normalized to the Fe2O3

  Table1. Composition in SiO₂-Al₂O₃-CaO-FeO_x quaternary system (as equivalent oxide)

  1.3    Viscosity measurement

  Viscosity was measured under N₂ atmosphere with temperature rising to 1 700 ℃ by a home-made high temperature rotary viscometer, as described in Figure 1. Samples of ca. 100 g were used to measure the viscosity with the temperatures interval of 50 ℃. The viscometer was calibrated using the reference material 717a (borosilicate glass) certificated by the American National Institute of Standards & Technology with the deviation being less than 5%.

  

  Figure 1. Schematic diagram of high temperature rotary viscometer for viscosity measurement

  图选项

  下载全尺寸图像

  下载幻灯片

  1.4    Analysis methodologies

  TMA experiments were conducted in a SII EXSTAR TMA/6300 (SII Instrument ltd) thermo-mechanical analyzer to study the potential correlation between heating shrinkage and fusibility characteristics. Approximate 50 mg samples was placed into a high purity alumina crucible, and then compacted with a force about 5 N. A penetrating ram was inserted into the crucible and pressured at the interface between the penetrating ram and sample surface by loading the weight of 100 g. The samples were heated up to 1 500 ℃ with the step of 10 ℃/min under 100 mL/min N₂ gas flow. The expansion and/or shrinkage of probe and ram were calibrated and reset prior to analysis.

  DSC experiments were performed in an EXSTAR DSC-6300 (SII Instrument ltd). All the samples were heated with a rate of 10 ℃/min under N₂ atmosphere and the weight of samples was constant at 15 mg with the deviation of 0.5 mg.

  XRD analysis was carried out by Rigaku RINT ultimate-Ⅲ powder diffractometer using a graphite monochromator, NaI (TI) detector and Cu Kα radiation. The XRD scans were performed between 5° and 90° with a step of 0.02(°)/s.

  1.5    “FactSage” modeling

  The current study characterized the phase equilibrium and thermodynamic properties in the quaternary system SiO₂-A1₂O₃-CaO-FeO_1-x by computer-aided thermodynamic modeling using “FactSage 7.0”. In the thermodynamic “optimization” of a system, all available thermodynamic and phase equilibrium data for the system were evaluated simultaneously in order to obtain one set of model equations for the Gibbs energies of all phases as functions of temperature and composition. Consequently, all of the thermodynamic properties and the phase diagrams can be back-calculated. In this way, all the data were rendered self-consistently and consistent with thermodynamic principles. Thermodynamic property data, such as activity data, were helpful in the evaluation of the phase diagram, and the phase diagram measurements can be applied to deduce thermodynamic properties. The models (quasi-chemical, sub-lattice) were used to predict the thermodynamic properties of a multi-component solution from the optimized parameters of its binary, ternary and (where available) quaternary sub-systems.

  2    Results and discussion

  2.1    Melting behavior traced by TMA

  The melting behavior of samples resulted from the reaction among minerals during thermal treatment was analyzed by the TMA and its traces for samples were presented in Figure 2. It must be noted that the total amount and height of samples compacted in the crucible sometimes had deviation. The shrinkage ratio was defined as the shrinkage (ΔH) divided by the original height (H) of samples.

  

  Figure 2. TMA analyses of mineral transformations in SiO₂-Al₂O₃-CaO-FeO_x quaternary system

  图选项

  下载全尺寸图像

  下载幻灯片

  The peaks were correlated to structural changes of the sample (sintering or melting) and sudden acceleration of the penetration ram. Effort was done to identify the species that caused melting at given temperatures in the samples. The TMA traces showed very similar penetration values for all samples. For each group, two graphs were depicted including the penetration height and the shrinkage rate versus temperature. The scale on the Y-axes of the shrinkage rate graphs varied for different groups.

  As shown in Figure 2, the percentages of penetration expressed by the TMA trace of sample started to change when temperature was higher than 850 ℃, indicating decomposition occurred around this temperature. The likely process was the transformation of crystal. As shown in Figure 2(a), the S01 with only SiO₂ and Al₂O₃ showed a smooth pattern with less shrinkage and without transition point, indicating that it was still in the un-melting state. In contrast, the penetration of S02 started at around 1 382 ℃, while, that of S03 at ca. 1 306 ℃. Thus, CaO was concluded to exert more significant impact than F₂O₃ on the melting behavior.

  From Figure 2(b), it can be seen that the TMA peaks of S04, S05 and S06 presented different transition temperatures. CaO with same proportion was supposed to impose the similar influences on these samples. For TMA curves, the deformation temperatures were still within a temperature region (around 850 ℃) of little shrinkage. In fact, deformation temperature had extremely poor reproducibility. However, the fusion temperatures exhibited high reliability with apparently rapid shrinkage. The fusion temperatures of S04, S05 and S06 detected by TMA analysis were 1 307, 1 192 and 1 160 ℃, respectively. Sample S04 showed three peaks. The first one at 1 251 ℃ was considered to be the softening temperature caused by the transformation of hematite to magnetite and/or wüstite. The second peak at around 1 307 ℃ was attributed to the melt and reaction of hematite with other components to generate the eutectic. It is important to note that the atmosphere of TMA analysis was N₂ that was not exactly consistent with the pure oxidizing and reducing conditions. The transition temperatures of three iron oxides with various valences were considered to be independent under inert condition. The last peak at 1 347 ℃ might be caused by the unstable penetration of rod after the melting of sample species, indicating the inhomogeneous fusion of minerals and/or agglomeration of solid particles. Sample S05 with Fe₃O₄ addition presented lower fusion temperature than that of S04 with Fe₂O₃ addition. Fe²⁺ in Fe₃O₄ was considered to lead to structural distortion more easily than Fe³⁺ in Fe₂O₃, and therefore, the liquid phase was formed more easily in S05. Similarly, S06 manifested a lower fusion temperature at 1 160 ℃ due largely to the formation of lower temperature eutectics by wüstite phase, that is, the Fe²⁺ exerted significant effect on the generation of low melting point phase than that of Fe³⁺ under the same condition (in N₂ atmosphere).

  The S08-S10 with lower CaO proportion compared with that of S04-S06 were used to investigate the effect of CaO interaction on iron oxides. With less CaO interaction, sample S08 with addition of Fe₂O₃ demonstrated higher shrinkage temperature, at ca. 1 316 ℃, while S09 and S10 consisting of Fe₃O₄ and FeO showed lower shrinkage temperature, at ca. 1 200 ℃ and 1 231 ℃, respectively (Figure 2 (c)). Summarily, compared with Fe³⁺, Fe²⁺ imposed more positive influence on decreasing melting temperature. However, samples S08-S10 without Ca interaction presented higher melting temperatures than S04-S06 (Figure 2 (b)), that is, Ca accompanied by iron oxide is supposedly accelerate the formation of lower temperature eutectics in comparison with the independent iron oxide and/or calcium oxide.

  Despite TMA analysis revealed remarkable deformation during ash melting, it had revealed partially unfeasibly since mixtures of species contributed to multi peaks in TMA. The correlation between morphological transformation and mineral reaction is quite necessary for the prediction of the sample properties at melting stage.

  2.2    Qualitative DSC analyses on mineral reaction

  The thermal analysis methods such as DSC could reveal whether the peaks are corresponding or not with melting^[21]. Precise determination on the onset of fusion can also be provided by thermal analysis as an endothermic phenomenon. Nevertheless, these techniques are based on compounds that is no representative to accurately predict the slagging potential. The research efforts have been paid to enhance slagging prediction taking into account the different components, as well as the impact of reducing environment^{[12, 16, 19]}.

  For these experiments, high heating rate in DSC furnace was favored as it resulted in better resolution of peaks. The evolution of sample during thermal treatment is shown in Figure 3. Various behaviors for all the samples were determined by DSC analysis due to the different components they contained.

  

  Figure 3. DSC analyses of mineral reactions in SiO₂-Al₂O₃-CaO-Fe₂O₃ quaternary system

  图选项

  下载全尺寸图像

  下载幻灯片

  S01 with only SiO₂ and Al₂O₃ showed a small exothermal peak at ca. 1 240 ℃ that was considered to be the transition of Al₂O₃ and SiO₂ to α-Al₂O₃ and cristobalite, respectively^[22]. However, the melting and reaction in S01 are not observed, as shown in Figure 3 (a). For S02, a small exothermal peak is found at 918 ℃ that might derive from the partial transition of iron oxide^[23] while the broad endothermic peak at around 1 284 ℃ should be the reaction between alumina and iron oxide to spinel. Significant distinctions are observed in S03, as shown in Figure 3 (a) with three sharp endothermic peaks being detected at 1 234, 1 355 and 1 484 ℃. The peak at about 1 234 ℃ is thought to be the reaction of CaO with silica and alumina to form the low temperature eutectic (e.g. anorthite). The eutectic then accelerates the dissolution of calcium oxide to generate other Ca-bearing aluminosilicates (e.g. wollastonite).

  Samples S04-S06 prepared with similar amount of iron oxides and calcium oxide are used to investigate the interactions among calcium oxide and iron oxides, as shown in Figure 3(b). A large peak in S04 at ca. 1 318 ℃ is supposed to primarily be derived from the reaction of calcium oxide with silica and alumina and/or partial decomposition of Fe₂O₃. Similar curves are obtained for S05 and S06, revealing the identical effect of Fe₃O₄ and FeO in this system. However, as mentioned above, Fe₃O₄ demonstrates higher activity than that of FeO because of their different crystal structures. The reaction in S05 begins at 1 182 ℃, which is lower than that in S06 at 1 221 ℃, meaning that the reactions of ferric oxide were restrained by the addition of Ca-bearing material in this quaternary system. On the contrary, the calcium oxide may be able to accelerate the interactions of ferrous oxide in this system to form lower temperature eutectics.

  The phase transformation predicted by “FactSage” for the SiO₂-Al₂O₃-CaO-Fe₂O₃-O system in equilibrium with 0.8/0.2 of SiO₂/Al₂O₃ mole ratio is represent in Figure 4. The phase transformations primarily occur at 1 200-1 700 ℃. The region below the liquidus temperatures stands for the homogeneous metastable liquids. Apparently, the anorthite and wollastonite dominate the phase transformation. The composition with low Ca and Fe₂O₃ contents are located in the mullite primary phase field, and thus, the liquidus temperatures are predicted to be above 1 600 ℃. In contrast, Fe₂O₃ is able to accelerate the formation of low melting composition in the lower liquidus temperature region. CaO demonstrates similar effect on the system. Nevertheless, the excessive amount of CaO tends to generate monoxide composition (i.e., CaO residue) with high melting temperature.

  

  Figure 4. Phase diagram of the Al₂O₃+SiO₂+CaO+“Fe₂O₃” system with SiO₂/Al₂O₃(mol ratio)=0.8/0.2 showing the phase compositions and equilibrium of the minerals used in this investigation

  图选项

  下载全尺寸图像

  下载幻灯片

  The proportion of calcium oxide in S08-S10 is decreased to ca. 8.0%, while the iron oxides are dominant in these samples. With less interaction of calcium oxide, the main endothermic peaks are almost consistent, except a larger endothermic peak at 1 350 ℃ (i.e. melting point of hematite) for S08. The reaction occurred here is thought mainly to be the decomposition of ferric oxide to ferrous oxide and the following reaction with silica and/or alumina to generate iron-bearing spinels, as shown in Figure 3(c).

  2.3    Effects of CaO and FeO_x on the viscosity in the system

  The minerals in S01 are not molten, and therefore, its viscosity measurement is failed, while the viscosities of S02 and S03 are shown in Figure 5(a). Though some solid particles (mullite and/or cristobalite) still exist in S02, its viscosities are still lower than that of S03. The molten S02 is suddenly re-solidified at about 1 550 ℃ due to the rapid formation of high temperature spinel crystal. In contrast, the S03 is continuously in the melting phase until 1 400 ℃. It can be thus concluded that the Fe-bearing crystal prefers to generate higher temperature spinel than Ca-bearing minerals.

  

  Figure 5. Apparent viscosities vs. temperature of samples at cooling cycle

  图选项

  下载全尺寸图像

  下载幻灯片

  The addition of ferrous oxide in S04 results in lower viscosity than that of ferric oxide in S05, as shown in Figure 5(b). After the reduction and separation of iron oxides from system in S07, Ca-bearing mineral plays the dominant role in keeping molten phase, nevertheless, its viscosity increases subsequently.

  It is consistent with “FactSage” simulation (Figure 4) that the iron oxides performe high efficiency to lower the viscosities in S08-S10, as illustrated in Figure 5(c). Additionally, ferrous oxide tends to decrease the viscosity more remarkably than ferric oxide, and therefore, the viscosity of S10 is lower than that of S08. Besides, ferroferric oxide presents similar effect on the viscosity with ferric oxide, that is, Fe (Ⅲ) in the molten sample imposes the dominated impact on viscosity. The viscosity of S11 increases significantly after strong reduction of iron oxide to metallic iron due to the re-generation of high temperature minerals, such as mullite.

  Though both of the Ca-bearing and Fe-bearing minerals are able to form eutectic, the minerals formed by Ca-bearing eutectic are considered to present higher viscosity. The basic structural unit of all Fe (Ⅲ) oxides is an octahedron where each Fe atom is surrounded by six O atoms. O ions form layers which are either approximately hexagonally close-packed, as are in goethite and hematite, or approximately cubic close-packed^{[24, 25]}. Fe³⁺ in the octahedral position may be partly replaced by other trivalent metal cations of similar size, such as A1³⁺, without modifying the structure (isomorphous substitution)^[26]. Hereby, CaO leads to lower transformation temperature and more eutectic, and also it is able to accelerate the dissolution of iron oxides, while iron oxide can mutually substitute with alumina and loosen the large framework structure of alumina to small fragments, resulting in lower viscosity. Moreover, Ca ion exhibits higher reaction activity than iron ions, and thus iron oxides can be driven out of the molten phase while coexists with CaO, as seen in Figure 4.

  2.4    Crystal transformation correlating with components

  XRD analysis presents the primary crystalline structure after viscosity measurement and slows re-solidification until room temperature. As depicted in Figure 6, crystalline structure in S01 are mullite, cristobalite and α-alumina, indicating the silica and alumina transition and their reaction with mullite. All of these compositions in S01 are high-melting point minerals; , and therefore, melting cannot occur at this temperature. The Fe₂O₃ addition in the S02 results in the similar crystal with that in S01. Moreover, the peaks of alumina are disappeared while that of aluminum iron oxide was found (Figure 6), which reveals that the iron oxide could primarily react with alumina to form spinel. However, this reaction was incomplete, because some hematite peaks are observable. Still, high temperature mullite and cristobalite stably exist in the solid phase, and calcium oxide strongly reacts with silica and alumina to generate eutectics. As a result, the peaks in Figure 6 primary represent Ca-bearing minerals formed at high temperature.

  

  Figure 6. XRD patterns of S01-S03 after viscosity measurement

  图选项

  下载全尺寸图像

  下载幻灯片

  The peaks in S04, S05 and S06 with addition of iron oxides and calcium oxide are similar, as shown in Figure 7, in which anorthite is the major crystalline structure. And interestingly, maghemite is once again found to be isolated component, which might be caused by its partial separation from molten sample due to its polarization^[27]. The quaternary system after measurement in carbon crucible experiences serious reduction. The iron oxide in S07 is considered to be reduced to the metallic iron. Therefore, its crystal structures were thus destroyed to X-ray amorphous structure, as illustrated in Figure 7.

  

  Figure 7. XRD patterns of S04-S07 after viscosity measurement

  图选项

  下载全尺寸图像

  下载幻灯片

  Figure 8 shows XRD patterns of S08-S11 with higher proportion of iron oxides. Addition of Fe₂O₃ and Fe₃O₄ only forms the crystalline structure of iron silicate, while a part of magnetite is surplus. Some weak peaks of mullite are also detected in S08 and S09, indicating the silica and alumina partially prioring to forming mullite as well as Fe-bearing materials. FeO leads to the generation of X-ray amorphous mineral in S10，which means that the Fe-bearing crystal experiences molten and re-solidification, and then becomes instable due to the distortion of iron oxide lattice. Iron oxides in S11 are totally reduced to metallic iron, and subsequently, silica and alumina re-combined to form mullite, as shown in Figure 7.

  

  Figure 8. XRD patterns of S08-S11 after viscosity measurement

  图选项

  下载全尺寸图像

  下载幻灯片

  Comparing Figure 7 with Figure 8, the calcium oxide exhibits higher activity to form crystal than that of iron oxides. Calcium oxide can firstly generate low temperature eutectic and then induce the reaction with iron oxides. However, higher calcium oxide content can also drive ferric oxide, especially, out of the liquid system, causing the isolation of iron oxides.

  3    Conclusions

  The synergistic effects of the components on mineral behaviors in SiO₂-Al₂O₃-CaO-FeO_x quaternary system were evaluated. The melting and reaction behaviors were investigated using TMA, DSC, XRD and “FactSage” modeling etc., and viscosities were measured from liquid to re-solidification phases. The result showed that during heat treatment silica and alumina preferred to be transformed to high temperature cristobalite and α-alumina, respectively, which featured in relatively stable structure and high viscosity. Calcium and iron oxides were valuable fluxing agents for accelerating the reaction and transformation of silica and alumina. Synthetic slags with addition of ferrous oxide showed lower viscosity than those with ferric oxide and ferriferous oxide. Furthermore, the reduction of iron oxides to metallic iron could increase the viscosity quite significantly. Ferric oxide may take part in the random glass network in a manner similar to alumina. However, iron in Fe²⁺ oxidation state may also act as a modifier under weak reducing condition and higher temperature. The sensitivity of viscosity of mineral matters to temperature excursion decreased with the increasing of the content, as calcium oxide was considered to be able to enhance the solution ability of iron oxides in the quaternary SiO₂-Al₂O₃-CaO-FeO_x system.

• 1. [1]

     ELLIOTT L. Dissolution of lime into synthetic coal ash slags[J]. Fuel Process Technol, 1998, 56(56):  45-53.

  2. [2]

     NOWOK J W, BENSON S A, STEADMAN E N, BREKKE D W. The effect of surface-tension viscosity ratio of melts on the sintering propensity of amorphous coal ash slags[J]. Fuel, 1993, 72(7):  1055-1061. doi: 10.1016/0016-2361(93)90308-O

  3. [3]

     HURST H J, NOVAK F, PATTERSON J H. Viscosity measurements and empirical predictions for fluxed Australian bituminous coal ashes[J]. Fuel, 1999, 78(15):  1831-1840. doi: 10.1016/S0016-2361(99)00094-0

  4. [4]

     CHAKRABORTY S, SARKAR S, GUPTA S, RAY A. Damage monitoring of refractory wall in a generic entrained-bed slagging gasification system[J]. Proc Inst Mech Eng Part A, 2008, 222(8):  791-807. doi: 10.1243/09576509JPE638

  5. [5]

     BRYANT G W, MCLENNAN A R, BAILEY C W, STANMORE B R, WALL T F. Index for iron-based slagging for pulverized coal firing in oxidizing and reducing conditions[J]. Energy Fuels, 2000, 14(2):  349-354. doi: 10.1021/ef990127d

  6. [6]

     HURST H J, NOVAK F, PATTERSON J H. Viscosity measurements and empirical predictions for some model gasifier slags[J]. Fuel, 1999, 78(4):  439-444. doi: 10.1016/S0016-2361(98)00162-8

  7. [7]

     HUFFMAN G P, HUGGINS F E, DUNMYRE G R. Investigation of the high-temperature behavior of coal ash in reducing and oxidizing atmospheres[J]. Fuel, 1981, 60(7):  585-597. doi: 10.1016/0016-2361(81)90158-7

  8. [8]

     WU H W, WEE H L, ZHANG D K, FRENCH D. The effect of combustion conditions on mineral matter transformation and ash deposition in a utility boiler fired with a sub-bituminous coal[J]. Proc Combust Inst, 2005, 30(2):  2981-2989. doi: 10.1016/j.proci.2004.08.059

  9. [9]

     RUPRECHT P, KONKOL W, CORNILS B, LANGHOFF J, BRUNKE W, SCHAFER W. Slagging gasification of coal under addition of flux agents[J]. Fuel Process Technol, 1985, 11(1):  1-12. doi: 10.1016/0378-3820(85)90011-6

  10. [10]

      GROEN J C, BROOKER D D, WELCH P J, OH M S. Gasification slag rheology and crystallization in titanium-rich, iron-calcium-aluminosilicate glasses[J]. Fuel Process Technol, 1998, 56(1/2):  103-127.

  11. [11]

      LIN X C, IDETA K, MIYAWAKI J, TAKEBE H, YOON S H, MOCHIDA I. Correlation between fluidity properties and local structures of three typical asian coal ashes[J]. Energy Fuels, 2012, 26(4):  2136-2144. doi: 10.1021/ef201771f

  12. [12]

      BAI J, LI W, LI B Q. Characterization of low-temperature coal ash behaviors at high temperatures under reducing atmosphere[J]. Fuel, 2008, 87(4):  583-591.

  13. [13]

      AIKIN T L H, CASHION J D, OTTREY A L. Mossbauer analysis of iron phases in brown coal ash and fireside deposits[J]. Fuel, 1984, 63(9):  1269-1275. doi: 10.1016/0016-2361(84)90436-8

  14. [14]

      MCLENNAN A R, BRYANT G W, STANMORE B R, WALL T F. Ash formation mechanisms during of combustion in reducing conditions[J]. Energy Fuels, 2000, 14(1):  150-159. doi: 10.1021/ef990095u

  15. [15]

      SCHOTTE E, LORENZ H, RAU H. Investigating gasification processes of solid fuels by in situ gaspotentiometric oxygen probes[J]. Chem Eng Technol, 2003, 26(7):  774-778. doi: 10.1002/ceat.200301667

  16. [16]

      SONG W J, TANG L H, ZHU X D, WU Y Q, ZHU Z B, KOYAMA S. Prediction of Chinese coal ash fusion temperatures in Ar and H₂ atmospheres[J]. Energy Fuels, 2009, 23(4):  1990-1997. doi: 10.1021/ef800974d

  17. [17]

      WALL T F, LOWE A, WIBBERLEY L J, STEWART I M. Mineral matter in coal and the thermal performance of large boilers[J]. Prog Energy Combust Sci, 1979, 5(1):  1-29. doi: 10.1016/0360-1285(79)90017-0

  18. [18]

      LLOYD W G, RILEY J T, ZHOU S, RISEN M A, TIBBITTS R L. Ash fusion temperatures under oxidizing conditions[J]. Energy Fuels, 1993, 7(4):  490-494. doi: 10.1021/ef00040a009

  19. [19]

      BRYANT G W, MCLENNAN A R, BAILEY C W, STANMORE B R, WALL T F. An experimental comparison of the ash formed from coals containing pyrite and siderite mineral in oxidizing and reducing conditions[J]. Energy Fuels, 2000, 14(2):  308-315. doi: 10.1021/ef990092h

  20. [20]

      URBAIN G, BOTTINGA Y, RICHET P. Viscosity of liquid silica, silicates and alumino-silicates[J]. Geochim Cosmochim Acta, 1982, 46(6):  1061-1072.

  21. [21]

      ARVELAKIS S, JENSEN P A, DAM-JOHANSEN M. Simultaneous thermal analysis (STA) on ash from high-alkali biomass[J]. Energy Fuels, 2004, 18(4):  1066-1076. doi: 10.1021/ef034065+

  22. [22]

      PINTO R R C, VALLE M L M, SOUSA-AGUIAR E F. The influence of silica to alumina ratio on Y zeolite activity by simultaneous TG and DSC analysis[J]. J Thermal Anal Calorim, 2002, 67(2):  439-443(5). doi: 10.1023/A:1013976512286

  23. [23]

      GUO X H, ZHONG S L, ZHANG J, WANG W, MAO J J, XIE G. Synthesis, phase transition, and magnetic property of iron oxide materials: Effect of sodium hydroxide concentrations[J]. J Mater Sci, 2010, 45(23):  6467-6473. doi: 10.1007/s10853-010-4733-8

  24. [24]

      EBSWORTH E A V, WEIL J A. Paramagnetic resonance absorption in peroxydicobalt complexes[J]. J Phys Chem, 2002, 63(11):  1890-1900.

  25. [25]

      RISTIĆ M, POPOVIĆ S, MUSIĆ S. Investigation of crystalline phases in the alpha-Fe₂O₃/alpha-Al₂O₃ system[J]. Croat Chem Acta, 2009, 82(2):  397-404.

  26. [26]

      NOWOK J W, HURLEY J P, STANLEY D C. Local structure of a lignitic coal ash slag and its effect on viscosity[J]. Energy Fuels, 1993, 7(6):  1135-1140. doi: 10.1021/ef00042a063

  27. [27]

      EBNER A D, RITTER J A. Retention of iron oxide particles by stainless steel and magnetite magnetic matrix elements in high-gradient magnetic separation[J]. Sep Sci Technol, 2004, 39(12):  2863-2890. doi: 10.1081/SS-200028811

• 

  Figure 1  Schematic diagram of high temperature rotary viscometer for viscosity measurement

  下载: 全尺寸图片 幻灯片
  

  Figure 2  TMA analyses of mineral transformations in SiO₂-Al₂O₃-CaO-FeO_x quaternary system

  下载: 全尺寸图片 幻灯片
  

  Figure 3  DSC analyses of mineral reactions in SiO₂-Al₂O₃-CaO-Fe₂O₃ quaternary system

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Phase diagram of the Al₂O₃+SiO₂+CaO+“Fe₂O₃” system with SiO₂/Al₂O₃(mol ratio)=0.8/0.2 showing the phase compositions and equilibrium of the minerals used in this investigation

  下载: 全尺寸图片 幻灯片
  

  Figure 5  Apparent viscosities vs. temperature of samples at cooling cycle

  下载: 全尺寸图片 幻灯片
  

  Figure 6  XRD patterns of S01-S03 after viscosity measurement

  C: 74-1081# corundum-Al₂O₃; A: 82-1584# aluminum iron oxide-Al₃Fe₅O₁₂; H: 24-0072# hematite-Fe₂O₃; Cr: 85-0621# cristobalite, syn-SiO₂; M: 15-0776# mullite-Al₆Si₂O₁₃; Ca: 72-0767# calcium aluminum oxide-CaAl₄O₇; P: 74-0874# pseudowollastonite-Ca₃(Si₃O₉)

  下载: 全尺寸图片 幻灯片
  

  Figure 7  XRD patterns of S04-S07 after viscosity measurement

  A: 41-1486# anorthite-CaAl₂Si₂O; M: 39-1346# maghemite-Fe₂O₃; H: 70-1876# hedenbergite-Fe_1.3CaO_0.7(Si₂O₆); I: 77-2355# FeO-iron oxide

  下载: 全尺寸图片 幻灯片
  

  Figure 8  XRD patterns of S08-S11 after viscosity measurement

  M: 15-0776# mullite-Al₆Si₂O₁₃; Ma: 75-0449# magnetite-Fe₃O₄; I: 80-1625# iron silicate-Fe₂(SiO₄)

  下载: 全尺寸图片 幻灯片

  Table 1.  Composition in SiO₂-Al₂O₃-CaO-FeO_x quaternary system (as equivalent oxide)

  Sample	Composition w/%									Mol ratio
  	SiO2	Al2O3	CaO	Al2O3-crucible				C-cruciblea		SiO2/Al2O3	B/Ab
  				Fe2O3	Fe3O4	FeO		Fe2O3
  S01	65.36	34.64	-	-	-	-		-		1.89	0.00
  S02	49.29	26.09	-	24.62	-	-		-		1.89	0.33
  S03	49.18	25.98	24.83	-	-	-		-		1.89	0.33
  S04	39.52	20.89	19.94	19.65	-	-		-		1.89	0.66
  S05	39.78	21.02	20.07	-	19.13	-		-		1.89	0.66
  S06	40.31	21.31	20.34	-	-	18.04		-		1.89	0.66
  S07	39.52	20.89	19.94	-	-	-		19.65		1.89	0.66
  S08	39.78	21.02	8.25	30.94	-	-		-		1.89	0.64
  S09	40.19	21.24	8.34	-	30.22	-		-		1.89	0.64
  S10	41.05	21.70	8.52	-	-	28.74		-		1.89	0.64
  S11	39.78	21.02	8.25	-	-	-		30.94		1.89	0.64
  a C-crucible was used to simulate the reduction atmosphere, Fe2O3 was reduced to iron metallic particles during measurement; b base/acid ratio (base=w(Fe2O3)+w(CaO), acid=w(SiO2)+w(Al2O3)), where only four analytical compounds were calculated, and the FeO, Fe3O4 were normalized to the Fe2O3

  下载: 导出CSV
•