- Email: [email protected]
Effects of CaCO3 on slag flow properties at high temperatures
Fuel 109 (2013) 76–85
Contents lists available at SciVerse ScienceDirect
Fuel journal homepage: www.elsevier.com/locate/fuel
Effects of CaCO3 on slag flow properties at high temperatures Lingxue Kong a,b, Jin Bai a,⇑, Zongqing Bai a, Zhenxing Guo a, Wen Li a a b
State Key Laboratory of Coal Conversion, Institute of Coal Chemistry, Chinese Academy of Sciences, P.O. Box 165, Taiyuan 030001, PR China University of Chinese Academy of Sciences, Beijing 100039, PR China
h i g h l i g h t s " Effects of limestone on coal samples of different Si/Al were carefully studied. " The essences of influences on AFTs, slag viscosity, temperature of critical temperature and type of slag were discussed. " The prediction methods are raised for guiding the use of limestone. " A new method for predicting critical temperature is presented.
a r t i c l e
i n f o
Article history: Received 10 February 2012 Received in revised form 26 October 2012 Accepted 5 November 2012 Available online 30 November 2012 Keywords: CaCO3 Ash fusion temperatures Slag viscosity Temperature of critical viscosity Type of slag
a b s t r a c t Experiments were conducted on the selected ashes with different additions of CaCO3 for understanding the effects on slag flow properties including ash fusion temperatures, slag viscosity, temperature of critical viscosity and type of slag. ICP-AES, XRD and FTIR analyses were applied to determine the component and structure of the slags. Factsage was used to calculate liquidus temperatures in the SiO2–Al2O3–CaO– FeO system and to predict formed mineral matters and proportion of solid phase as a function of temperatures. The results show that the liquidus temperature calculated by Factsage well predicts the variation of ash fusion temperatures. Slag viscosity behavior changes with increasing addition of CaCO3 because the formation process of solid phase is different. The Fourier transform infrared (FTIR) spectrum indicates that Ca2+ leads to break polymerized Si–O–Si into Si–O, so the increasing Ca2+ in slag results in the decrease of viscosity above liquidus temperature. Below liquidus temperature, solids content decreases with increasing addition of CaCO3 above the temperature of critical viscosity (Tcv). Meanwhile, it is found that the rate of solid formation is related with Tcv and a new prediction method of Tcv based on that was proposed. Moreover, the type of slag changed with addition of CaCO3 was predictable by XRD analysis. The prediction on ash fusion temperature, Tcv and type of slag is expected to serve as a reference for adding flux to regulate coal/ash properties suitable for slag tapping gasification technology. Crown Copyright Ó 2012 Published by Elsevier Ltd. All rights reserved.
1. Introduction The need for higher efficiency of future power generation and chemical production leads to increased interest in IGCC technologies, especially in advanced gasification technology such as entrained-flow gasifier [1–6]. In the gasifier, under high temperatures over 1400 °C and strong turbulence, organic matter in coal is completely combusted and gasified in a short time, and the mineral matter in coal transforms into ash. The ash becomes liquid slag owing to the melting and reactions of its component mineral matter at high temperature [2]. Slag fluid properties are much more important for entrained-flow gasifiers with liquid slag-removal process than the conversion of organic matter in coal.
⇑ Corresponding author. Tel.: +86 351 4040289; fax: +86 351 4050320. E-mail address: [email protected] (J. Bai).
Continuous slag tapping is the key for the successful operation for different entrained flow gasifiers (GE, Shell, Prenflo, GSP, Texaco, Eagle) [3–21], so the flow properties of slag and the influence of additives on them at high temperature are of great importance. Generally, ash fusion temperature and slag viscosity curve are used for describing the slag flow properties. Ash fusion temperatures (AFTs) are referred for slag tapping in GE gasifier. Operating temperature is above the flow temperature (FT) of coal ash, and the FT should not be higher than 1300 °C. Slag viscosity vs. temperature curve is applied for slag tapping in the membrane wall gasifiers, such as Shell and GSP gasifiers. It is generally accepted that the slag viscosity must be in a certain range for not only slag tapping but also membrane wall design [4]. Two aspects were considered from slag viscosity curve including viscosity dependence on temperature and the temperature of critical viscosity (Tcv). For example, the viscosity is required in 2.5–25 Pa s at temperatures from 1300 to 1500 °C. Besides, Tcv should be at least 150 °C lower than
0016-2361/$ - see front matter Crown Copyright Ó 2012 Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.fuel.2012.11.014
77
L. Kong et al. / Fuel 109 (2013) 76–85 Table 2 Ash content and fusion temperatures of samples. MgO
TiO2
K2O
Na2O
Si/Al
YC SL FG
56.98 60.82 61.08
27.33 22.89 18.10
4.92 3.86 5.25
5.19 3.82 7.48
0.98 2.55 2.12
1.05 1.13 0.97
1.06 1.82 1.76
0.43 1.10 0.46
2.08 2.66 3.37
the temperature of 2.5 Pa s. However, slag properties of most coals, especially lignite in Northwest China did not satisfy the requirement for slag tapping. In hence, fluxing agent is used to adjust slag properties and limestone is the widely used one in China for its abundance and low cost. Thus, it is necessary to understand the influence of limestone on slag properties including AFTs, slag viscosity and Tcv, for continuous operation of the various gasifiers at the expected temperature range. Generally, two methods, both experimental and thermodynamic calculation are used to study AFTs of coal ash with addition of limestone [4–9]. Hurst et al. [4] used the ternary equilibrium phase diagrams to study the fluxing effect addition of CaCO3 on Australian coal ashes. Song et al. [5] examined the effect of CaO as pure compounds on the AFTs of coal ash with Factsage. It was found that when the content of CaO was over 35%, the AFTs increased quickly with more CaO. The similar phenomenon is also noticed by other researchers, but the minimum AFTs appears at different content of CaO for various Si/Al coal ashes [9]. Since the AFTs also increased with adding limestone, the proper selection of flux amount is important for gasification operation. Seggiani [6] provided a method from chemical composition to predict the effect of adding minerals (such as CaO) on AFTs. Jak et al. [7] also applied the thermodynamic computer package Factsage to predict AFTs and found that they do correlate with the liquidus temperatures. However, much fewer attempts have been made so far with regards to proper flux ratio based on the AFTs of raw coal aiding with thermodynamic calculation. Viscosity of most slags exponentially decreases with increasing temperature and irregularly decreases with adding limestone. The decrease is mainly caused by the variation of solid phase in melted slag [9–12]. Wu et al. [13] studied the effect of solid particles on liquid viscosity by adding solids into silicate liquid. The amount and type of solid phase precipitated in melt slag showed significantly influence on slag viscosity. Ilyushechkin [11] found that below the liquidus temperature, slag viscosity increased sharply with the increasing amount of solids. The sharp increase of viscosity with relatively low Si/Ai ratio (<2) slag was due to solid level or size of the crystal, and a high Si/Al ratio slag was solid precipitation. The alkaline earth metal oxides (e.g. CaO) mostly acted as network modifiers to reduce the slag viscosity. Song et al. [12] also found that solids amount in melted slag decreased with the increasing CaO content which resulted in viscosity decreasing. The Tcv is another critical factor of slag flow behavior. Tcv indicates a point of abrupt change in the viscosity–temperature curve when solid phase in liquid slag begins to crystallize and to separate out from the liquid phase [14]. It is also assumed that Tcv indicates the boundary between crystal-affected viscosity and viscosity not affected by the presence of crystals [15]. Some slags exhibit the classical behavior of a glass that a continuous increase in viscosity as the temperature decreases and the slag is referred as glassy slag. Others show a rapid increase of viscosity when the temperature is lowered below Tcv, which is referred as plastic slag or crystalline slag [16]. Many researchers [4–9] used CaO or CaCO3 as fluxing agent to study their effect on slag viscosity. Although these results can give good guide in gasification, yet little work has been published regarding essences for the effect of flux on slag viscosity, Tcv and type of slag.
Samples
Ash content/Ad (wt%)
AFTs (°C) IDT
ST
HT
FT
YC SL FG
15.30 14.05 4.57
1404 1306 1180
1427 1353 1206
1443 1357 1224
1496 1372 1318
(a) 1650 T liq
1600
FT
1550 o
CaO
Temperature / C
Fe2O3
1500 1450 1400
CaO=33.89%
1350 1300 1250 1200 1150 -2
0
2
4
6
8
10
12
14
16
18
20
22
Addition of CaCO3 / wt% (coal basis)
(b) 1650
Tliq
1600
FT 1550 1500
o
Al2O3
Temperature / C
SiO2
1450 1400 1350 1300
CaO=31.24%
1250 1200 1150 0
2
4
6
8
10
12
14
16
18
Addition of CaCO3 / wt% (coal basis)
(c) 1450
Tliq
1425
FT
1400 1375 o
Samples
Temperature / C
Table 1 Chemical composition of coal ashes (wt%).
1350 1325
CaO=29.17%
1300 1275 1250 1225 1200 1175 -0.5
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
Addition of CaCO3 / wt% (coal basis) Fig. 1. Flow temperature vs addition of CaCO3 curves. (a) YC; (b) SL; (c) FG.
78
L. Kong et al. / Fuel 109 (2013) 76–85
Fig. 2. Calculated liquidus temperatures in the SiO2–Al2O3–CaO–FeO system on the pseudoternary section with a SiO2/Al2O3 mass ratio of 2.66.
(ICP-AES) was used to characterize ash composition based on ASTM D6349. The chemical compositions of coal ash are given in Table 1.
FT=0.86Tliq+86.86 R=0.9211
1450
o
Flow temperature (FT) / C
1500
2.2. Measurement of AFTs
1400 o
+40 C
1350 1300
YC SL FG
1250 1200 1150 1250
1300
1350
1400
1450
1500
1550
1600
1650
o
Liquidus temperature (Tliq) / C Fig. 3. Correlation temperatures.
between
flow
temperature
(FT)
values
and
liquidus
In this work, samples with three different Si/Al ratios were chosen to investigate the effects of CaCO3 on slag flow properties including AFTs, slag viscosity, Tcv and type of slag. Mineralogical composition and structure of slag were analyzed by XRD and FTIR. The computer software package Factsage has been used to calculate liquidus temperatures in the SiO2–Al2O3–CaO–FeO system and to predict component variation of solid and liquid phase in slag as a function of temperature.
2. Experimental 2.1. Samples Three coals including YC, SL, and FG with different Si/Al ratios were used in the study. The ash samples were prepared in a muffle furnace at 815 °C according to the Chinese standard GB/T15742007. Inductively coupled plasma–atomic emission spectrometry
The ash fusion temperatures auto detecting system developed by KY company, China was used to investigate the AFTs with different additions of CaCO3 (wt/%, coal basis). The measurements were performed following the Chinese standard procedures (GB/T2192008) – under mild reducing atmosphere (CO/CO2 = 6:4). The initial deformational temperature (IDT), softening temperature (ST), hemispherical temperature (HT), and flow temperature (FT) were recognized and recorded by the auto detecting system with the accuracy of 1 °C. Ash content and fusion temperatures of samples are shown in Table 2. 2.3. Viscosity test The viscosity of slag was analyzed with a Theta high-temperature rotating viscometer under a reducing atmosphere as measurement of AFTs. The test started above liquidus temperature for keeping the sample totally melted. The maximum temperature of the viscometer is 1680 °C, which is high enough to melt most ash into slag. The ramp rate was 1 °C/min. The viscosity and temperature were recorded continuously at interval of 0.1 °C, so a viscosity–temperature curve was obtained. The molybdenum rotors and cylinder crucibles were used and the parameters for the rotor crucible combination were calibrated with a standard reference material of 717A glass. The sample temperature was recorded using a type-B platinum thermocouple in an alumina pedestal and corrected using a previously determined temperature calibration experiment with a thermocouple inside the crucible filled with magnesium oxide. 2.4. Slag quenching experiment The quenching experiment was performed in order to obtain slag sample at certain temperature. Slag with different CaCO3
79
L. Kong et al. / Fuel 109 (2013) 76–85
additions were placed in tube furnace. The temperature was kept according to the viscosity test procedure. When the temperature reached the target value, the slag was taken out of furnace and then quenched by water immediately. 2.5. XRD analysis The slag samples from viscosity test were ground to less than 0.074 mm. A RIGAKU D/max-rB X-ray powder diffractometer was applied for XRD patterns using Cu Ka radiation (40 kV, 100 mA, Ka1 = 0.15408 nm). The samples were scanned with a step size of 0.01° at 4°/min over a 2h range of 10–90°.
ture was similar to the trend of FT with addition of CaCO3. Many researchers have found that the linear relation between FT and Tliq is possible as FT ¼ a þ b T liq [8]. Fig. 3 shows the FT as a function of the calculated liquidus temperatures. Good correlation between FT and liquidus temperatures is determined when only CaO concentration is changed in the ash. In addition all of the ash samples exhibit flow temperatures within
(a) 300 250
FTIR spectra were recorded with a Bruker Vertex 70 Fourier transform infrared spectrometer. Slag sample was accurately weighed and then mixed with KBr with the ratio of 1:200. The mixture was finely milled in the agate mortar for 30 min and then pressed into pellets under 12 MPa. The spectral region from 4000 to 400 cm1 was examined.
200
Viscosity / Pa s
2.6. FTIR analysis
Factsage [17] was used to calculate component of solid phase and change of liquid phase at different temperatures with CaO– SiO2–Al2O3–FeO system. Calculations were carried out between the solid temperature and liquidus temperature in mild reducing atmosphere under 0.1 MPa.
150
100
6% CaCO3 5% CaCO3
25 Pa s
0 1200 1250 1300 1350 1400 1450 1500 1550 1600 1650 1700 o
Temperature / C
(b) 300 0% CaCO3
250
Viscosity / Pa s
2.8. Thermodynamic equilibrium calculations
3% CaCO3
50
2.7. SEM A LEO 1450 scanning electron microscope (SEM) was employed to assess the microstructure of the quenched slag. For SEM observations, the slags were mounted in liquid epoxy resin, pelletized, polished and finally sputtered-coated with carbon. Analysis of the quenched sample provides information about the amount of solids in slags. The microstructure of the quenched slags was investigated using SEM in back-scattering mode (BSE), and Energy Dispersive X-ray Spectroscopy (EDX) was carried out to identify the composition of solid in the quenched samples.
0% CaCO3
3% CaCO3 7% CaCO3
200
9% CaCO3
150
100 15% CaCO3
50 25 Pa s
0 1100 1150 1200 1250 1300 1350 1400 1450 1500 1550 1600 1650
3. Results and discussion
o
Temperature / C
3.1. Effects of limestone on AFTs
(c) 300 250
Viscosity / Pa s
Fig. 1a–c shows the variation of flow temperature (FT) against addition of CaCO3. The FT of coal ash samples decreases with increasing addition of CaCO3 until the CaO content reaches about 30–35%. At higher addition of CaCO3, the FT of coal ash samples increases, but slower than the decreasing rate. This trend is similar to that displayed by the liquidus temperature (Tliq) as addition of CaCO3 increases. Fig. 2 illustrates a pseudoternary section construction for the SiO2–Al2O3–CaO–FeO system that can be used to visually express liquidus temperature of sample SL with a SiO2/Al2O3 mass ratio of 2.66 as a function of CaCO3 addition. The lines of the same color represent all compositions having a given liquidus temperature. The initial composition of coal ash lies in the mullite region and has a melting point above 1600 °C. The addition of CaCO3 changes the composition from the mullite region to the anorthite (CaAl2Si2O8) with an initial reduction in liquidus temperatures. However, the composition lies in the ps-wallostonite (CaSiO3 (s2)) and gehlenite (Ca2Al2SiO7) regions at more addition of CaCO3 that causes liquidus temperatures increase. The change of liquidus tempera-
0% CaCO3
200 1% CaCO3
150 3.5% CaCO3
100 4.5% CaCO3
50 25 Pa s
0 1100
1200
1300
1400
1500
1600
o
Temperature / C Fig. 4. Viscosity–temperature curves for samples with different addition of CaCO3. (a) YC; (b) SL; (c) FG.
L. Kong et al. / Fuel 109 (2013) 76–85
(a) 50 Silica Mullite
4.5
Si/Al
4.0
35
3.5
30
3.0
25
2.5
20
2.0
15
1.5
10
1.0
5
0.5 1450
1500
1550
3.5 3.0
40
2.5 30
2.0 1.5
20
1.0 10
0.5
0 1260 1280 1300
1600
1320 1340 1360 1380
o
o
Si/Al CaO/CaO* FeO/FeO*
5
Solids / %
3 30 2 20
Si/Al CaO/CaO* FeO/FeO*
4
70 60
3 50 40
2
30 20
1
5
Liquid phase change
40
80
Liquid phase change
4
10
Wallostonite Anorthite
90
50
0 1150
(d)100
Solids / %
60
0.0 1400 1420 1440
Temperature / C 6
Wollastonite Anorthite
4.5 4.0
Temperature / C
(c) 70
Si/Al CaO/CaO*
50
0.0 1400
Silica Anorthite
Liquid phase change
40
5.0
70 60
Liquid phase change
Solids / %
45
0 1350
(b)
5.0
Solids / %
80
1
10 1200
1250
1300
0
0 1150
0 1400
1350
1200
o
1250
1300
1350
o
Temperature / C
Temperature / C
(e) 100
4.0 Ca3Si2O7
90
Si/Al CaO/CaO*
CaSiO3
80
3.5
Gehlenite 3.0
Solids / %
2.5
60
2.0
50 40
1.5
30
Lquid phase change
70
1.0 20 0.5
10 0 1250
0.0 1260
1270
1280
1290
1300
o
Temperature / C Fig. 5. Solid phase content and liquid phase change of sample SL with selected additions of CaCO3 (FeO⁄ and CaO⁄ are initial concentrations in liquid phase).
experimental errors of ±40 °C, which indicates that the liquidus temperature predicts FT with a good accuracy (R = 0.9211). The FT of the ash samples were expressed as a linear function of the liquidus temperature following the equation:
sage. Linear relation between Tliq and FT can then be used to determine the amount of limestone required at target FT.
T FT ¼ 86:86 þ 0:86T liq
3.2. Effect of limestone on slag viscosity
ð1Þ
where Tliq is the liquidus temperature of the samples in degrees Celsius (°C). With the linear relation, it is possible to guide the used of limestone for coal ash. In other words, the liquidus temperatures with different addition of limestone were first calculated by Fact-
For reverse flow gasifier, the maximum tapping viscosity is 25 Pa s for continuous operation of the entrained-flow gasifier at slag tapping temperature of 1300–1500 °C. However, at 1500 °C the viscosity of ash without flux addition of CaCO3 is higher than
L. Kong et al. / Fuel 109 (2013) 76–85
15% CaCO3
9% CaCO3
7% CaCO3
3% CaCO3 0% CaCO3
4000
3500
3000
2500
2000
Wavenumbers / cm
1500
1000
500
-1
Fig. 6. FTIR for sample SL with different CaCO3 additions.
0.8 0.7
NBO/BO ratio
0.6 0.5 0.4 0.3 0.2 0.1 0.0 -0.1 0
2
4
6
8
10
12
14
16
CaCO3 addition / % Fig. 7. NBO/BO ratios with CaCO3 addition above liquidus temperature.
25 Pa s because of the high proportion of solid particles in liquid slag formed by ash at this temperature. Fig. 4a–c shows the variation of slags viscosities versus temperature at selected addition of CaCO3. The viscosity of slag is dependent on temperature and chemical composition. The dotted lines are drawn at 25 Pa s, which is considered the upper viscosity limit of operability. The viscosity of slag without addition of CaCO3 is higher than the tappability limit of 25 Pa s at 1500 °C. However, after addition of CaCO3, it is lower than the tappability limit of 25 Pa s at 1500 °C. Fig. 4a–c also represents the effect of addition CaCO3 on slag viscosity behavior. Within the measured temperature range (1100– 1630 °C), the viscosity behavior changed with increasing addition of CaCO3. For example, the slags are glassy type with addition of 0%, 3% and 7% for SL. However, when addition increases to 9%
81
and 15%, they become crystalline type. Ilyushechkin has proved that the change of solid and liquid phase is important characteristic of slag affecting the slag’s viscosity behavior [11]. A possible reason for this phenomenon is that formation process of solid phase changes with increasing addition of CaCO3 in slag. The different formation process of solid phase in slag may influence solid phase content and composition of liquid phase as a function of temperature. Fig. 5 shows solid phase content and liquid phase change with different additions of CaCO3 for SL with selected additions of CaCO3, as calculated by Factsage. In Fig. 5a, mullite appears as primary phase below 1625 °C for the raw slag, where the amount of solid and Si/Al ratio gradually increased with decreasing temperature. At 1456 °C, the second solid phase silica starts to precipitate. The solid formation and accompanied change of the Si/Ai ratio have very little effect on viscosity; the amount of solids is still quite low (just about 10% when temperature decreased by about 170 °C) and relative change of Si/Ai ratio is not significant. Considering calcium and iron concentrations, the liquid phase compositional change is not considered, as their contents are low. In the case of composition with 3% addition (Fig. 5b), the solids are anorthite (primary solid phase) and silica as second phase. Solids start to precipitate below 1440 °C. The amount of solids reaches 27% until the temperature decreases to 1315 °C, where silica starts to precipitate. The amount of solid decreases at the same temperature compared with the raw sample. For example, the solids amount is 34% at 1400 °C with addition of 0%, whereas it is 14% for 6% addition. For composition with 7% addition (Fig. 5c), the solids are anorthite (primary solid phase) and wollastonite as the second phase. Solids start to precipitate on cooling below 1400 °C. The amount of solids reaches 25% until the temperature decreases to 1235 °C, where wollastonite starts to precipitate. The change of solid amount is similar with 3% addition, which explains the similar shape of the viscosity curves for 3% and 7% addition. It is also found that FeO concentration increased when wollastonite starts to precipitate and affects the trend of viscosity change. The sharp increase in viscosity below 1400 °C might be due to increasing solids amount and FeO concentration. When addition is 9% (Fig. 5d), the primary solid phase is anorthite and wollastonite exists as the second phase. Solids start to precipitate below 1350 °C. The solid amount increases sharply below 1270 °C where wollastonite starts to precipitate. For example, the solids amount improves 58% from 1272 to 1200 °C, and also significantly improves iron oxide content in the slag’s liquid phase. Fig. 5e presents the case of addition 15%. It is found that solid content increases drastically with formation of CaSiO3 when temperature is below 1268 °C. Meanwhile, concentration of Ca also decreases in the liquid phase due to the formation of CaSiO3. 3.3. Slags characteristics In order to understand the viscosity behavior of slag with addition of CaCO3 above and below the liquidus temperature, sample SL was used to conduct a series of slag quenching experiments. The FITR analysis and SEM were carried out to identify solid and liquid phase compositions of the quenched samples. 3.3.1. FTIR analysis CaO usually is classified as network modifiers and trends to lower viscosity. With the addition of alkaline earth metal oxide CaO in silica, Si–O–Si bonds are broken and the cations are randomly distributed throughout the lattice. In order to understand the effect of Ca2+ on structure in slag with different of additions CaCO3, the slags above liquidus temperature were analyzed with FTIR, as shown in Fig. 6. The peak at 1100–900 cm1 is assigned as Si–O anti-symmetric stretching vibration, which is quite strong.
82
L. Kong et al. / Fuel 109 (2013) 76–85
Fig. 8. SEM-EDX of directly cold water quenched slags at 1300 °C. (a) 0%; (b) 3%; (c) 7%; (d) 9%; (e) 15%.
Peaks at 600–400 cm1 are assigned as Si–O vibrations. The content of Si–O increased with adding CaCO3, which was confirmed by the stronger peaks at 1100 and 460 cm1. The peak at around
700 cm1 is assigned to Si–O–Si bond which decreased with increased additions of CaCO3. This indicates that the content of Si– O–Si in the slags decreased with adding CaCO3.The peak at around
83
L. Kong et al. / Fuel 109 (2013) 76–85
Fig. 8. (continued)
1550 1500
YC 1400
dm(solids)/dT
o
Temperature / C
1450 SL FG
1350 1300 1250 1200 1150 -2
0
2
4
6
8
10
12
14
16
0.0 -0.1 -0.2 -0.3 -0.4 -0.5 -0.6
0% CaCO 3 3% CaCO 3 7% CaCO 3 9% CaCO3
-0.7 -3.1 -3.2 -3.3 -3.4 -3.5 -3.6 -3.7 1150
15% CaCO3
1200
1250 1300
1400
1450
1500 1550
1600
o
Fig. 11. The rate of solid formation with temperature in slag.
Fig. 9. Temperature of critical viscosity of slags with addition of CaCO3.
1550
100 90
1500
80
1450
70
Tcv=0.98Tmax+17.33 R=0.9492
1400
60
o
Tcv / C
Solid amount / %
1350
Temperature / C
Addition of CaCO3 / %
YC
50
SL
40
1350 YC SL FG
1300
FG
o
+50 C
30
1250
20 1200 10 0 0
2
4
6
8
10
12
14
16
Addition of CaCO3 / %
1150 1150
1200
1250
1300
1350
1400
1450
1500
1550
Tmax / o C
Fig. 10. Solid amount at Tcv with addition of CaCO3.
Fig. 12. Correlation between the measured Tcv values and temperatures of maximum rate of solid formation (Tmax).
1000 cm1 moving to high frequency also suggested the addition of Ca2+ [18]. The notation of bridging and non-bridging oxygen bonds (BO and NBO) is used to describe oxygen related to silicon atom in network theory [18]. BO is linked to two silicon atoms (Si–O–Si), while
a NBO is only linked to one silicon atom (Si–O). And the addition of CaO introduces NBO in the network, resulting in a strong negative effect on viscosity [15]. Senior and Srinivasachar defined the ratio of NBO–BO (NBO/BO) to illustrate structure of silicate slag [19].
84
L. Kong et al. / Fuel 109 (2013) 76–85
In Fig. 7, NBO/BO ratios changed with CaCO3 addition above liquidus temperature. The NBO/BO ratio increased with increasing CaCO3, which indicates that the content of Si-O–Si bond decreased while Si–O increased in slags with increasing CaCO3 addition. It also revealed big network was broken with Ca2+, which is consistent with the FTIR results. In hence, when the temperature is above liquidus temperature, Ca2+ lowers the viscosity by breaking the network structure. 3.3.2. Microstructure of slags BSE photomicrographs of slag samples quenched at 1300 °C are shown in Fig. 8. The composition of the solid phase changes with additions of CaCO3 was also given by EDX analysis. The number of particles existed in the slag was greater than other four samples, when addition was 0% (Fig. 8a), and so the slag has higher viscosity. When the addition increased to 3% and 7%, a few crystalline particles appear (Fig. 8b and c), which should be one of the reasons for the slag viscosity decreased. When the addition increased to 9% and 15% (Fig. 8d and e), the number of crystalline particle decreased drastically. When the temperature was lower than liquidus temperature, the slag viscosity decreased with adding limestone because both silicates with lower melting point formed and less crystal particles were formed. However, when the addition of limestone increased, the size of crystal particles increased obviously. Meanwhile, the slag type turned into crystal slag according to the viscosity temperature curve. The change of slag type should be related with the growth of particle size, so the formation of crystals is the essential reason for change of slag type, but the size of crystal particle is critical condition. Nonetheless, it is almost impossible to attain the image information by SEM at high temperatures, so it is difficult to determine the particle size leading to crystal slag.
This equation provides a good way to well predict Tcv through the calculation of Tmax. 3.5. Prediction of slag type by XRD The type of slag is also important information for slag tapping. Generally, slag is divided into glassy, crystalline and plastic slag from viscosity–temperature curves. Normally, type of slag is judged from viscosity–temperature curve. Each type of slag presents different viscosity properties with temperature decreased [22]. The viscosity of glassy slag increased gradually with decreasing temperature. However, viscosity of crystalline slag and plastic slag increased quickly below Tcv. In order to prevent sharp increase of viscosity in gasification process, glassy slag is usually obtained by blend proper ratio of either flux or other coals with suitable composition. Fig. 13 depicts the X-ray diffraction patterns of samples after viscosity test. Take SL for example: XRD pattern of slags showed that the slags have mainly an amorphous character when CaCO3 additions were 0%, 3% and 7%, while it presented a crystalline character when CaCO3 additions were 9% and 15%. From Fig. 4b slags were grouped as glassy type when CaCO3 additions were 0%, 3% and 7%, respectively and crystal minerals will not appear with decreasing temperature. However, slag became crystalline type when CaCO3 additions were 9% and 15%,
(a)
6% CaCO3
3.4. Slag temperature of critical viscosity (Tcv)
5% CaCO3
Fig. 9 shows the Tcv for slag with different addition of CaCO3. Tcv first decreased with CaCO3 additions, while it increased at higher addition. The temperature of critical viscosity indicates a point of abrupt change in the viscosity–temperature relationship because of solid phases in liquid slag beginning to crystallize and separate out from the liquid phase. Fig. 10 presents solid amount at Tcv with different addition of CaCO3. The solid amount at Tcv increased with CaCO3 additions for SL, while it first increased with CaCO3 additions and then decreased at higher addition for YC and FG. These trends are different from that seen in Tcv with addition of CaCO3. The result demonstrates that solid amount in liquid slag have no direct relationship with Tcv. It is currently impossible to make an accurate prediction or estimation of the Tcv of a specific slag based on its chemical composition. Although the relationship between solid amount and Tcv is not yet clear, there is still a close relation. Fig. 11 presents the rate of solid formation with temperature for sample SL, where dm(solids)/dT is the rate of solid formation. It is obvious that the curve shows similar trend with each other. A maximum rate appeared with temperature decreased in slags. In order to find the relation between rate of solid formation and Tcv, the temperature of maximum rate is defined as Tmax. Fig. 12 shows the measured Tcv of slags as a function of the calculated Tmax for the three samples with different CaCO3 additions. Good correlation between the Tcv and Tmax is observed. Moreover, the R value is 0.9492, which indicates that the excellent accuracy of the results is within the experimental error. The Tcv of slags could be expressed as a linear function of the Tmax according to the equation:
T cv ¼ 0:98T max þ 17:33
ð2Þ
3% CaCO3 0% CaCO3
15
30
45
60
2θ /
75
90
o
(b)
15% CaCO3 9% CaCO3 7% CaCO3 3% CaCO3 0% CaCO3
15
30
45
60
2θ /
75
90
o
(c)
4.5% CaCO3 3.5% CaCO3 1% CaCO3 0% CaCO3
15
30
45
60
2θ /
75
90
o
Fig. 13. XRD patterns of slag after viscosity test. (a) YC; (b) SL; (c) FG.
L. Kong et al. / Fuel 109 (2013) 76–85
respectively and crystals appeared with decreasing temperature. These results were consistent well with the XRD analysis. The same results can also be observed for samples YC and FG. This provides a good method to predict the type of slag by XRD. However, there is still not enough evidence to prove the quantitative relationship between type of slag and amount of crystals in slag. This should be studied further in the future work. In order to get a glassy slag for slag tapping, the method of predicting slag type and Tcv can be used to determine the amount of flux required or the ratio of blended coal. The slag type with different added ratios of flux or blended coal was firstly judged by XRD analysis. Tcv of slag blended with flux or coal were then calculated by Factsage. According to the results, the amount of flux required or the ratio of blended coal to get a glassy slag can be determined. 4. Conclusions In this study, influences of additions of CaCO3 on AFTs, slag viscosity, Tcv and slag type were investigated. The conclusions were expected to guide the use of limestone for slag tapping gasifiers. Liquidus surfaces in phase equilibrium diagrams for the pseudoternary systems correlated well with the trends of AFTs and the linear relation was set up between FT and liquidus temperature, TFT = 86.86 + 0.86Tliq. The relation was able to be used for conducting addition of limestone for gasification (R = 0.9211). The slag viscosity behavior changed with increasing addition of CaCO3. Above liquidus temperature, addition of CaCO3 increased NBO and FTIR of slag proved the content of Si–O–Si decreased. Below the liquidus temperature, the addition of CaCO3 not only influenced by solid content but also changed composition of solid phase. A new method to predict the Tcv was established based on the rate of solids formation. An apparent linear correlation and excellent agreement between Tcv and maximum rate of solids formation (Tmax) calculated by Factsage was found, Tcv = 0.98Tmax + 17.33 (R = 0.9492). The slag type also changed with addition of CaCO3. XRD analysis of slag showed a good consistency with result from viscosity–temperature curves. Acknowledgments Financial support from the National Basic Research Program of China (2010CB227005-02), Joint Foundation of Natural Science Foundation of China and Shenhua Group Corporation Ltd. (U1261209), the National Natural Science Foundation of China
85
(21006121), the Youth Foundation of Shanxi Province (2010021008-2), and the State Key Laboratory of Coal Conversion and State Key Laboratory of Coal Combustion (FSKLCC0909) was gratefully appreciated. References [1] Takematsu T, Maude C. Coal gasification for IGCC power generation. London: IEA Coal Research; 1991. [2] Song W, Tang L, Zhu X, Wu Y, Rong Y, Zhu Z, Shuntarou Koyama. Fusibility and flow properties of coal ash and slag. Fuel 2009;88:297–304. [3] Vassilev S, Kitano K, Takeda S, Tsurue T. Influence of mineral and chemical composition of coal ashes on their fusibility. Fuel Process Technol 1995;45: 27–51. [4] Hurst HJ, Novak F, Patterson JH. Phase diagram approach to the fluxing effect of additions of CaCO3 on Australian coal ashes. Energy Fuels 1996;10:1215–9. [5] Song W, Tang L, Zhu X, Wu Y, Rong Y, Zhu Z, et al. Effect of coal ash composition on ash fusion temperatures. Energy Fuels 2010;24:182–9. [6] Seggiani M. Empirical correlations of the ash fusion temperatures and temperature of critical viscosity for coal and biomass ashes. Fuel 1999;78: 1121–5. [7] Jak E, Degterov S, Pelton AD. Thermodynamic modelling of the system Al2O3– SiO2–CaO–FeO–Fe2O3 to predict the flux requirements for coal ash slags. Fuel 1998;77:77–84. [8] Browning GJ, Bryant GW, Hurst HJ, Lucas JA, Wall TF. An empirical method for the prediction of coal ash slag viscosity. Energy Fuels 2003;17:731–7. [9] Hurst HJ, Novak F, Patterson JH. Viscosity measurements and empirical predictions for fluxed Australian bituminous coal ashes. Fuel 1999;78: 1831–40. [10] Patterson JH, Hurst HJ. Ash and slag qualities of Australian bituminous coals for use in slagging gasifiers. Fuel 2000;79:1671–8. [11] Ilyushechkin AY, Hla San S, Roberts DG, Kinaev NN. The effect of solids and phase compositions on viscosity behaviour and Tcv of slags from Australian bituminous coals. J Non Cryst Solids 2011;357:893–902. [12] Song W, Tang L, Zhu X, Wu Y, Rong Y, Zhu Z. Fusibility and flow properties of shell gasifier slag. J Chem Ind Eng 2009;60:1782–6. [13] Wu L, Mattias EK, Song M, Du S. The effect of solid particles on liquid viscosity. Steel Res Int 2011;82:388–97. [14] Mills KC, Broadbent CP. Engineering foundation conference on the impact of ashdeposition on coal fired plants. Solihull (UK); 1993. p. 513–25. [15] Vargas S, Frandsen FJ, Dam-Johansen K. Rheological properties of hightemperature melts of coal ashes and other silicates. Prog Energy Combust Sci 2001;27:237–429. [16] Oh MS, Brooker DD, Paz de EF, Brady JJ, Decker TR. Effect of crystalline phase formation on coal slag viscosity. Fuel Proc Technol 1995;44:191–9. [17] Bale CW, Chartrand P, Degterov SA, Eriksson G, Hack K, Mahfoud RB, et al. FactSage thermochemical software and databases. Calphad 2002;26:189–228. [18] Li H, Qiu X, Tang Y. Ash melting behavior by Fourier transform infrared spectroscopy. J China Univ Min Technol 2008;18:245–9. [19] Senior CL, Srinivasachar S. Viscosity of ash particles in combustion systems for prediction of particle sticking. Energy Fuel 1995;9:277–83. [20] Kong L, Bai J, Li W, Bai Z, Guo Z. Effect of lime addition on slag fluidity of coal ash. J Fuel Chem Technol 2011;39:407–11. [21] Folkedahl BC, Schobert HH. Effects of atmosphere on viscosity of selected bituminous and low-rank coal ash slags. Energy Fuels 2005;19:208–15. [22] Gong D. The high-temperature viscosity model of coal ash. Therm Power Generation 1989;1:28–33.
Contents lists available at SciVerse ScienceDirect
Fuel journal homepage: www.elsevier.com/locate/fuel
Effects of CaCO3 on slag flow properties at high temperatures Lingxue Kong a,b, Jin Bai a,⇑, Zongqing Bai a, Zhenxing Guo a, Wen Li a a b
State Key Laboratory of Coal Conversion, Institute of Coal Chemistry, Chinese Academy of Sciences, P.O. Box 165, Taiyuan 030001, PR China University of Chinese Academy of Sciences, Beijing 100039, PR China
h i g h l i g h t s " Effects of limestone on coal samples of different Si/Al were carefully studied. " The essences of influences on AFTs, slag viscosity, temperature of critical temperature and type of slag were discussed. " The prediction methods are raised for guiding the use of limestone. " A new method for predicting critical temperature is presented.
a r t i c l e
i n f o
Article history: Received 10 February 2012 Received in revised form 26 October 2012 Accepted 5 November 2012 Available online 30 November 2012 Keywords: CaCO3 Ash fusion temperatures Slag viscosity Temperature of critical viscosity Type of slag
a b s t r a c t Experiments were conducted on the selected ashes with different additions of CaCO3 for understanding the effects on slag flow properties including ash fusion temperatures, slag viscosity, temperature of critical viscosity and type of slag. ICP-AES, XRD and FTIR analyses were applied to determine the component and structure of the slags. Factsage was used to calculate liquidus temperatures in the SiO2–Al2O3–CaO– FeO system and to predict formed mineral matters and proportion of solid phase as a function of temperatures. The results show that the liquidus temperature calculated by Factsage well predicts the variation of ash fusion temperatures. Slag viscosity behavior changes with increasing addition of CaCO3 because the formation process of solid phase is different. The Fourier transform infrared (FTIR) spectrum indicates that Ca2+ leads to break polymerized Si–O–Si into Si–O, so the increasing Ca2+ in slag results in the decrease of viscosity above liquidus temperature. Below liquidus temperature, solids content decreases with increasing addition of CaCO3 above the temperature of critical viscosity (Tcv). Meanwhile, it is found that the rate of solid formation is related with Tcv and a new prediction method of Tcv based on that was proposed. Moreover, the type of slag changed with addition of CaCO3 was predictable by XRD analysis. The prediction on ash fusion temperature, Tcv and type of slag is expected to serve as a reference for adding flux to regulate coal/ash properties suitable for slag tapping gasification technology. Crown Copyright Ó 2012 Published by Elsevier Ltd. All rights reserved.
1. Introduction The need for higher efficiency of future power generation and chemical production leads to increased interest in IGCC technologies, especially in advanced gasification technology such as entrained-flow gasifier [1–6]. In the gasifier, under high temperatures over 1400 °C and strong turbulence, organic matter in coal is completely combusted and gasified in a short time, and the mineral matter in coal transforms into ash. The ash becomes liquid slag owing to the melting and reactions of its component mineral matter at high temperature [2]. Slag fluid properties are much more important for entrained-flow gasifiers with liquid slag-removal process than the conversion of organic matter in coal.
⇑ Corresponding author. Tel.: +86 351 4040289; fax: +86 351 4050320. E-mail address: [email protected] (J. Bai).
Continuous slag tapping is the key for the successful operation for different entrained flow gasifiers (GE, Shell, Prenflo, GSP, Texaco, Eagle) [3–21], so the flow properties of slag and the influence of additives on them at high temperature are of great importance. Generally, ash fusion temperature and slag viscosity curve are used for describing the slag flow properties. Ash fusion temperatures (AFTs) are referred for slag tapping in GE gasifier. Operating temperature is above the flow temperature (FT) of coal ash, and the FT should not be higher than 1300 °C. Slag viscosity vs. temperature curve is applied for slag tapping in the membrane wall gasifiers, such as Shell and GSP gasifiers. It is generally accepted that the slag viscosity must be in a certain range for not only slag tapping but also membrane wall design [4]. Two aspects were considered from slag viscosity curve including viscosity dependence on temperature and the temperature of critical viscosity (Tcv). For example, the viscosity is required in 2.5–25 Pa s at temperatures from 1300 to 1500 °C. Besides, Tcv should be at least 150 °C lower than
0016-2361/$ - see front matter Crown Copyright Ó 2012 Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.fuel.2012.11.014
77
L. Kong et al. / Fuel 109 (2013) 76–85 Table 2 Ash content and fusion temperatures of samples. MgO
TiO2
K2O
Na2O
Si/Al
YC SL FG
56.98 60.82 61.08
27.33 22.89 18.10
4.92 3.86 5.25
5.19 3.82 7.48
0.98 2.55 2.12
1.05 1.13 0.97
1.06 1.82 1.76
0.43 1.10 0.46
2.08 2.66 3.37
the temperature of 2.5 Pa s. However, slag properties of most coals, especially lignite in Northwest China did not satisfy the requirement for slag tapping. In hence, fluxing agent is used to adjust slag properties and limestone is the widely used one in China for its abundance and low cost. Thus, it is necessary to understand the influence of limestone on slag properties including AFTs, slag viscosity and Tcv, for continuous operation of the various gasifiers at the expected temperature range. Generally, two methods, both experimental and thermodynamic calculation are used to study AFTs of coal ash with addition of limestone [4–9]. Hurst et al. [4] used the ternary equilibrium phase diagrams to study the fluxing effect addition of CaCO3 on Australian coal ashes. Song et al. [5] examined the effect of CaO as pure compounds on the AFTs of coal ash with Factsage. It was found that when the content of CaO was over 35%, the AFTs increased quickly with more CaO. The similar phenomenon is also noticed by other researchers, but the minimum AFTs appears at different content of CaO for various Si/Al coal ashes [9]. Since the AFTs also increased with adding limestone, the proper selection of flux amount is important for gasification operation. Seggiani [6] provided a method from chemical composition to predict the effect of adding minerals (such as CaO) on AFTs. Jak et al. [7] also applied the thermodynamic computer package Factsage to predict AFTs and found that they do correlate with the liquidus temperatures. However, much fewer attempts have been made so far with regards to proper flux ratio based on the AFTs of raw coal aiding with thermodynamic calculation. Viscosity of most slags exponentially decreases with increasing temperature and irregularly decreases with adding limestone. The decrease is mainly caused by the variation of solid phase in melted slag [9–12]. Wu et al. [13] studied the effect of solid particles on liquid viscosity by adding solids into silicate liquid. The amount and type of solid phase precipitated in melt slag showed significantly influence on slag viscosity. Ilyushechkin [11] found that below the liquidus temperature, slag viscosity increased sharply with the increasing amount of solids. The sharp increase of viscosity with relatively low Si/Ai ratio (<2) slag was due to solid level or size of the crystal, and a high Si/Al ratio slag was solid precipitation. The alkaline earth metal oxides (e.g. CaO) mostly acted as network modifiers to reduce the slag viscosity. Song et al. [12] also found that solids amount in melted slag decreased with the increasing CaO content which resulted in viscosity decreasing. The Tcv is another critical factor of slag flow behavior. Tcv indicates a point of abrupt change in the viscosity–temperature curve when solid phase in liquid slag begins to crystallize and to separate out from the liquid phase [14]. It is also assumed that Tcv indicates the boundary between crystal-affected viscosity and viscosity not affected by the presence of crystals [15]. Some slags exhibit the classical behavior of a glass that a continuous increase in viscosity as the temperature decreases and the slag is referred as glassy slag. Others show a rapid increase of viscosity when the temperature is lowered below Tcv, which is referred as plastic slag or crystalline slag [16]. Many researchers [4–9] used CaO or CaCO3 as fluxing agent to study their effect on slag viscosity. Although these results can give good guide in gasification, yet little work has been published regarding essences for the effect of flux on slag viscosity, Tcv and type of slag.
Samples
Ash content/Ad (wt%)
AFTs (°C) IDT
ST
HT
FT
YC SL FG
15.30 14.05 4.57
1404 1306 1180
1427 1353 1206
1443 1357 1224
1496 1372 1318
(a) 1650 T liq
1600
FT
1550 o
CaO
Temperature / C
Fe2O3
1500 1450 1400
CaO=33.89%
1350 1300 1250 1200 1150 -2
0
2
4
6
8
10
12
14
16
18
20
22
Addition of CaCO3 / wt% (coal basis)
(b) 1650
Tliq
1600
FT 1550 1500
o
Al2O3
Temperature / C
SiO2
1450 1400 1350 1300
CaO=31.24%
1250 1200 1150 0
2
4
6
8
10
12
14
16
18
Addition of CaCO3 / wt% (coal basis)
(c) 1450
Tliq
1425
FT
1400 1375 o
Samples
Temperature / C
Table 1 Chemical composition of coal ashes (wt%).
1350 1325
CaO=29.17%
1300 1275 1250 1225 1200 1175 -0.5
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
Addition of CaCO3 / wt% (coal basis) Fig. 1. Flow temperature vs addition of CaCO3 curves. (a) YC; (b) SL; (c) FG.
78
L. Kong et al. / Fuel 109 (2013) 76–85
Fig. 2. Calculated liquidus temperatures in the SiO2–Al2O3–CaO–FeO system on the pseudoternary section with a SiO2/Al2O3 mass ratio of 2.66.
(ICP-AES) was used to characterize ash composition based on ASTM D6349. The chemical compositions of coal ash are given in Table 1.
FT=0.86Tliq+86.86 R=0.9211
1450
o
Flow temperature (FT) / C
1500
2.2. Measurement of AFTs
1400 o
+40 C
1350 1300
YC SL FG
1250 1200 1150 1250
1300
1350
1400
1450
1500
1550
1600
1650
o
Liquidus temperature (Tliq) / C Fig. 3. Correlation temperatures.
between
flow
temperature
(FT)
values
and
liquidus
In this work, samples with three different Si/Al ratios were chosen to investigate the effects of CaCO3 on slag flow properties including AFTs, slag viscosity, Tcv and type of slag. Mineralogical composition and structure of slag were analyzed by XRD and FTIR. The computer software package Factsage has been used to calculate liquidus temperatures in the SiO2–Al2O3–CaO–FeO system and to predict component variation of solid and liquid phase in slag as a function of temperature.
2. Experimental 2.1. Samples Three coals including YC, SL, and FG with different Si/Al ratios were used in the study. The ash samples were prepared in a muffle furnace at 815 °C according to the Chinese standard GB/T15742007. Inductively coupled plasma–atomic emission spectrometry
The ash fusion temperatures auto detecting system developed by KY company, China was used to investigate the AFTs with different additions of CaCO3 (wt/%, coal basis). The measurements were performed following the Chinese standard procedures (GB/T2192008) – under mild reducing atmosphere (CO/CO2 = 6:4). The initial deformational temperature (IDT), softening temperature (ST), hemispherical temperature (HT), and flow temperature (FT) were recognized and recorded by the auto detecting system with the accuracy of 1 °C. Ash content and fusion temperatures of samples are shown in Table 2. 2.3. Viscosity test The viscosity of slag was analyzed with a Theta high-temperature rotating viscometer under a reducing atmosphere as measurement of AFTs. The test started above liquidus temperature for keeping the sample totally melted. The maximum temperature of the viscometer is 1680 °C, which is high enough to melt most ash into slag. The ramp rate was 1 °C/min. The viscosity and temperature were recorded continuously at interval of 0.1 °C, so a viscosity–temperature curve was obtained. The molybdenum rotors and cylinder crucibles were used and the parameters for the rotor crucible combination were calibrated with a standard reference material of 717A glass. The sample temperature was recorded using a type-B platinum thermocouple in an alumina pedestal and corrected using a previously determined temperature calibration experiment with a thermocouple inside the crucible filled with magnesium oxide. 2.4. Slag quenching experiment The quenching experiment was performed in order to obtain slag sample at certain temperature. Slag with different CaCO3
79
L. Kong et al. / Fuel 109 (2013) 76–85
additions were placed in tube furnace. The temperature was kept according to the viscosity test procedure. When the temperature reached the target value, the slag was taken out of furnace and then quenched by water immediately. 2.5. XRD analysis The slag samples from viscosity test were ground to less than 0.074 mm. A RIGAKU D/max-rB X-ray powder diffractometer was applied for XRD patterns using Cu Ka radiation (40 kV, 100 mA, Ka1 = 0.15408 nm). The samples were scanned with a step size of 0.01° at 4°/min over a 2h range of 10–90°.
ture was similar to the trend of FT with addition of CaCO3. Many researchers have found that the linear relation between FT and Tliq is possible as FT ¼ a þ b T liq [8]. Fig. 3 shows the FT as a function of the calculated liquidus temperatures. Good correlation between FT and liquidus temperatures is determined when only CaO concentration is changed in the ash. In addition all of the ash samples exhibit flow temperatures within
(a) 300 250
FTIR spectra were recorded with a Bruker Vertex 70 Fourier transform infrared spectrometer. Slag sample was accurately weighed and then mixed with KBr with the ratio of 1:200. The mixture was finely milled in the agate mortar for 30 min and then pressed into pellets under 12 MPa. The spectral region from 4000 to 400 cm1 was examined.
200
Viscosity / Pa s
2.6. FTIR analysis
Factsage [17] was used to calculate component of solid phase and change of liquid phase at different temperatures with CaO– SiO2–Al2O3–FeO system. Calculations were carried out between the solid temperature and liquidus temperature in mild reducing atmosphere under 0.1 MPa.
150
100
6% CaCO3 5% CaCO3
25 Pa s
0 1200 1250 1300 1350 1400 1450 1500 1550 1600 1650 1700 o
Temperature / C
(b) 300 0% CaCO3
250
Viscosity / Pa s
2.8. Thermodynamic equilibrium calculations
3% CaCO3
50
2.7. SEM A LEO 1450 scanning electron microscope (SEM) was employed to assess the microstructure of the quenched slag. For SEM observations, the slags were mounted in liquid epoxy resin, pelletized, polished and finally sputtered-coated with carbon. Analysis of the quenched sample provides information about the amount of solids in slags. The microstructure of the quenched slags was investigated using SEM in back-scattering mode (BSE), and Energy Dispersive X-ray Spectroscopy (EDX) was carried out to identify the composition of solid in the quenched samples.
0% CaCO3
3% CaCO3 7% CaCO3
200
9% CaCO3
150
100 15% CaCO3
50 25 Pa s
0 1100 1150 1200 1250 1300 1350 1400 1450 1500 1550 1600 1650
3. Results and discussion
o
Temperature / C
3.1. Effects of limestone on AFTs
(c) 300 250
Viscosity / Pa s
Fig. 1a–c shows the variation of flow temperature (FT) against addition of CaCO3. The FT of coal ash samples decreases with increasing addition of CaCO3 until the CaO content reaches about 30–35%. At higher addition of CaCO3, the FT of coal ash samples increases, but slower than the decreasing rate. This trend is similar to that displayed by the liquidus temperature (Tliq) as addition of CaCO3 increases. Fig. 2 illustrates a pseudoternary section construction for the SiO2–Al2O3–CaO–FeO system that can be used to visually express liquidus temperature of sample SL with a SiO2/Al2O3 mass ratio of 2.66 as a function of CaCO3 addition. The lines of the same color represent all compositions having a given liquidus temperature. The initial composition of coal ash lies in the mullite region and has a melting point above 1600 °C. The addition of CaCO3 changes the composition from the mullite region to the anorthite (CaAl2Si2O8) with an initial reduction in liquidus temperatures. However, the composition lies in the ps-wallostonite (CaSiO3 (s2)) and gehlenite (Ca2Al2SiO7) regions at more addition of CaCO3 that causes liquidus temperatures increase. The change of liquidus tempera-
0% CaCO3
200 1% CaCO3
150 3.5% CaCO3
100 4.5% CaCO3
50 25 Pa s
0 1100
1200
1300
1400
1500
1600
o
Temperature / C Fig. 4. Viscosity–temperature curves for samples with different addition of CaCO3. (a) YC; (b) SL; (c) FG.
L. Kong et al. / Fuel 109 (2013) 76–85
(a) 50 Silica Mullite
4.5
Si/Al
4.0
35
3.5
30
3.0
25
2.5
20
2.0
15
1.5
10
1.0
5
0.5 1450
1500
1550
3.5 3.0
40
2.5 30
2.0 1.5
20
1.0 10
0.5
0 1260 1280 1300
1600
1320 1340 1360 1380
o
o
Si/Al CaO/CaO* FeO/FeO*
5
Solids / %
3 30 2 20
Si/Al CaO/CaO* FeO/FeO*
4
70 60
3 50 40
2
30 20
1
5
Liquid phase change
40
80
Liquid phase change
4
10
Wallostonite Anorthite
90
50
0 1150
(d)100
Solids / %
60
0.0 1400 1420 1440
Temperature / C 6
Wollastonite Anorthite
4.5 4.0
Temperature / C
(c) 70
Si/Al CaO/CaO*
50
0.0 1400
Silica Anorthite
Liquid phase change
40
5.0
70 60
Liquid phase change
Solids / %
45
0 1350
(b)
5.0
Solids / %
80
1
10 1200
1250
1300
0
0 1150
0 1400
1350
1200
o
1250
1300
1350
o
Temperature / C
Temperature / C
(e) 100
4.0 Ca3Si2O7
90
Si/Al CaO/CaO*
CaSiO3
80
3.5
Gehlenite 3.0
Solids / %
2.5
60
2.0
50 40
1.5
30
Lquid phase change
70
1.0 20 0.5
10 0 1250
0.0 1260
1270
1280
1290
1300
o
Temperature / C Fig. 5. Solid phase content and liquid phase change of sample SL with selected additions of CaCO3 (FeO⁄ and CaO⁄ are initial concentrations in liquid phase).
experimental errors of ±40 °C, which indicates that the liquidus temperature predicts FT with a good accuracy (R = 0.9211). The FT of the ash samples were expressed as a linear function of the liquidus temperature following the equation:
sage. Linear relation between Tliq and FT can then be used to determine the amount of limestone required at target FT.
T FT ¼ 86:86 þ 0:86T liq
3.2. Effect of limestone on slag viscosity
ð1Þ
where Tliq is the liquidus temperature of the samples in degrees Celsius (°C). With the linear relation, it is possible to guide the used of limestone for coal ash. In other words, the liquidus temperatures with different addition of limestone were first calculated by Fact-
For reverse flow gasifier, the maximum tapping viscosity is 25 Pa s for continuous operation of the entrained-flow gasifier at slag tapping temperature of 1300–1500 °C. However, at 1500 °C the viscosity of ash without flux addition of CaCO3 is higher than
L. Kong et al. / Fuel 109 (2013) 76–85
15% CaCO3
9% CaCO3
7% CaCO3
3% CaCO3 0% CaCO3
4000
3500
3000
2500
2000
Wavenumbers / cm
1500
1000
500
-1
Fig. 6. FTIR for sample SL with different CaCO3 additions.
0.8 0.7
NBO/BO ratio
0.6 0.5 0.4 0.3 0.2 0.1 0.0 -0.1 0
2
4
6
8
10
12
14
16
CaCO3 addition / % Fig. 7. NBO/BO ratios with CaCO3 addition above liquidus temperature.
25 Pa s because of the high proportion of solid particles in liquid slag formed by ash at this temperature. Fig. 4a–c shows the variation of slags viscosities versus temperature at selected addition of CaCO3. The viscosity of slag is dependent on temperature and chemical composition. The dotted lines are drawn at 25 Pa s, which is considered the upper viscosity limit of operability. The viscosity of slag without addition of CaCO3 is higher than the tappability limit of 25 Pa s at 1500 °C. However, after addition of CaCO3, it is lower than the tappability limit of 25 Pa s at 1500 °C. Fig. 4a–c also represents the effect of addition CaCO3 on slag viscosity behavior. Within the measured temperature range (1100– 1630 °C), the viscosity behavior changed with increasing addition of CaCO3. For example, the slags are glassy type with addition of 0%, 3% and 7% for SL. However, when addition increases to 9%
81
and 15%, they become crystalline type. Ilyushechkin has proved that the change of solid and liquid phase is important characteristic of slag affecting the slag’s viscosity behavior [11]. A possible reason for this phenomenon is that formation process of solid phase changes with increasing addition of CaCO3 in slag. The different formation process of solid phase in slag may influence solid phase content and composition of liquid phase as a function of temperature. Fig. 5 shows solid phase content and liquid phase change with different additions of CaCO3 for SL with selected additions of CaCO3, as calculated by Factsage. In Fig. 5a, mullite appears as primary phase below 1625 °C for the raw slag, where the amount of solid and Si/Al ratio gradually increased with decreasing temperature. At 1456 °C, the second solid phase silica starts to precipitate. The solid formation and accompanied change of the Si/Ai ratio have very little effect on viscosity; the amount of solids is still quite low (just about 10% when temperature decreased by about 170 °C) and relative change of Si/Ai ratio is not significant. Considering calcium and iron concentrations, the liquid phase compositional change is not considered, as their contents are low. In the case of composition with 3% addition (Fig. 5b), the solids are anorthite (primary solid phase) and silica as second phase. Solids start to precipitate below 1440 °C. The amount of solids reaches 27% until the temperature decreases to 1315 °C, where silica starts to precipitate. The amount of solid decreases at the same temperature compared with the raw sample. For example, the solids amount is 34% at 1400 °C with addition of 0%, whereas it is 14% for 6% addition. For composition with 7% addition (Fig. 5c), the solids are anorthite (primary solid phase) and wollastonite as the second phase. Solids start to precipitate on cooling below 1400 °C. The amount of solids reaches 25% until the temperature decreases to 1235 °C, where wollastonite starts to precipitate. The change of solid amount is similar with 3% addition, which explains the similar shape of the viscosity curves for 3% and 7% addition. It is also found that FeO concentration increased when wollastonite starts to precipitate and affects the trend of viscosity change. The sharp increase in viscosity below 1400 °C might be due to increasing solids amount and FeO concentration. When addition is 9% (Fig. 5d), the primary solid phase is anorthite and wollastonite exists as the second phase. Solids start to precipitate below 1350 °C. The solid amount increases sharply below 1270 °C where wollastonite starts to precipitate. For example, the solids amount improves 58% from 1272 to 1200 °C, and also significantly improves iron oxide content in the slag’s liquid phase. Fig. 5e presents the case of addition 15%. It is found that solid content increases drastically with formation of CaSiO3 when temperature is below 1268 °C. Meanwhile, concentration of Ca also decreases in the liquid phase due to the formation of CaSiO3. 3.3. Slags characteristics In order to understand the viscosity behavior of slag with addition of CaCO3 above and below the liquidus temperature, sample SL was used to conduct a series of slag quenching experiments. The FITR analysis and SEM were carried out to identify solid and liquid phase compositions of the quenched samples. 3.3.1. FTIR analysis CaO usually is classified as network modifiers and trends to lower viscosity. With the addition of alkaline earth metal oxide CaO in silica, Si–O–Si bonds are broken and the cations are randomly distributed throughout the lattice. In order to understand the effect of Ca2+ on structure in slag with different of additions CaCO3, the slags above liquidus temperature were analyzed with FTIR, as shown in Fig. 6. The peak at 1100–900 cm1 is assigned as Si–O anti-symmetric stretching vibration, which is quite strong.
82
L. Kong et al. / Fuel 109 (2013) 76–85
Fig. 8. SEM-EDX of directly cold water quenched slags at 1300 °C. (a) 0%; (b) 3%; (c) 7%; (d) 9%; (e) 15%.
Peaks at 600–400 cm1 are assigned as Si–O vibrations. The content of Si–O increased with adding CaCO3, which was confirmed by the stronger peaks at 1100 and 460 cm1. The peak at around
700 cm1 is assigned to Si–O–Si bond which decreased with increased additions of CaCO3. This indicates that the content of Si– O–Si in the slags decreased with adding CaCO3.The peak at around
83
L. Kong et al. / Fuel 109 (2013) 76–85
Fig. 8. (continued)
1550 1500
YC 1400
dm(solids)/dT
o
Temperature / C
1450 SL FG
1350 1300 1250 1200 1150 -2
0
2
4
6
8
10
12
14
16
0.0 -0.1 -0.2 -0.3 -0.4 -0.5 -0.6
0% CaCO 3 3% CaCO 3 7% CaCO 3 9% CaCO3
-0.7 -3.1 -3.2 -3.3 -3.4 -3.5 -3.6 -3.7 1150
15% CaCO3
1200
1250 1300
1400
1450
1500 1550
1600
o
Fig. 11. The rate of solid formation with temperature in slag.
Fig. 9. Temperature of critical viscosity of slags with addition of CaCO3.
1550
100 90
1500
80
1450
70
Tcv=0.98Tmax+17.33 R=0.9492
1400
60
o
Tcv / C
Solid amount / %
1350
Temperature / C
Addition of CaCO3 / %
YC
50
SL
40
1350 YC SL FG
1300
FG
o
+50 C
30
1250
20 1200 10 0 0
2
4
6
8
10
12
14
16
Addition of CaCO3 / %
1150 1150
1200
1250
1300
1350
1400
1450
1500
1550
Tmax / o C
Fig. 10. Solid amount at Tcv with addition of CaCO3.
Fig. 12. Correlation between the measured Tcv values and temperatures of maximum rate of solid formation (Tmax).
1000 cm1 moving to high frequency also suggested the addition of Ca2+ [18]. The notation of bridging and non-bridging oxygen bonds (BO and NBO) is used to describe oxygen related to silicon atom in network theory [18]. BO is linked to two silicon atoms (Si–O–Si), while
a NBO is only linked to one silicon atom (Si–O). And the addition of CaO introduces NBO in the network, resulting in a strong negative effect on viscosity [15]. Senior and Srinivasachar defined the ratio of NBO–BO (NBO/BO) to illustrate structure of silicate slag [19].
84
L. Kong et al. / Fuel 109 (2013) 76–85
In Fig. 7, NBO/BO ratios changed with CaCO3 addition above liquidus temperature. The NBO/BO ratio increased with increasing CaCO3, which indicates that the content of Si-O–Si bond decreased while Si–O increased in slags with increasing CaCO3 addition. It also revealed big network was broken with Ca2+, which is consistent with the FTIR results. In hence, when the temperature is above liquidus temperature, Ca2+ lowers the viscosity by breaking the network structure. 3.3.2. Microstructure of slags BSE photomicrographs of slag samples quenched at 1300 °C are shown in Fig. 8. The composition of the solid phase changes with additions of CaCO3 was also given by EDX analysis. The number of particles existed in the slag was greater than other four samples, when addition was 0% (Fig. 8a), and so the slag has higher viscosity. When the addition increased to 3% and 7%, a few crystalline particles appear (Fig. 8b and c), which should be one of the reasons for the slag viscosity decreased. When the addition increased to 9% and 15% (Fig. 8d and e), the number of crystalline particle decreased drastically. When the temperature was lower than liquidus temperature, the slag viscosity decreased with adding limestone because both silicates with lower melting point formed and less crystal particles were formed. However, when the addition of limestone increased, the size of crystal particles increased obviously. Meanwhile, the slag type turned into crystal slag according to the viscosity temperature curve. The change of slag type should be related with the growth of particle size, so the formation of crystals is the essential reason for change of slag type, but the size of crystal particle is critical condition. Nonetheless, it is almost impossible to attain the image information by SEM at high temperatures, so it is difficult to determine the particle size leading to crystal slag.
This equation provides a good way to well predict Tcv through the calculation of Tmax. 3.5. Prediction of slag type by XRD The type of slag is also important information for slag tapping. Generally, slag is divided into glassy, crystalline and plastic slag from viscosity–temperature curves. Normally, type of slag is judged from viscosity–temperature curve. Each type of slag presents different viscosity properties with temperature decreased [22]. The viscosity of glassy slag increased gradually with decreasing temperature. However, viscosity of crystalline slag and plastic slag increased quickly below Tcv. In order to prevent sharp increase of viscosity in gasification process, glassy slag is usually obtained by blend proper ratio of either flux or other coals with suitable composition. Fig. 13 depicts the X-ray diffraction patterns of samples after viscosity test. Take SL for example: XRD pattern of slags showed that the slags have mainly an amorphous character when CaCO3 additions were 0%, 3% and 7%, while it presented a crystalline character when CaCO3 additions were 9% and 15%. From Fig. 4b slags were grouped as glassy type when CaCO3 additions were 0%, 3% and 7%, respectively and crystal minerals will not appear with decreasing temperature. However, slag became crystalline type when CaCO3 additions were 9% and 15%,
(a)
6% CaCO3
3.4. Slag temperature of critical viscosity (Tcv)
5% CaCO3
Fig. 9 shows the Tcv for slag with different addition of CaCO3. Tcv first decreased with CaCO3 additions, while it increased at higher addition. The temperature of critical viscosity indicates a point of abrupt change in the viscosity–temperature relationship because of solid phases in liquid slag beginning to crystallize and separate out from the liquid phase. Fig. 10 presents solid amount at Tcv with different addition of CaCO3. The solid amount at Tcv increased with CaCO3 additions for SL, while it first increased with CaCO3 additions and then decreased at higher addition for YC and FG. These trends are different from that seen in Tcv with addition of CaCO3. The result demonstrates that solid amount in liquid slag have no direct relationship with Tcv. It is currently impossible to make an accurate prediction or estimation of the Tcv of a specific slag based on its chemical composition. Although the relationship between solid amount and Tcv is not yet clear, there is still a close relation. Fig. 11 presents the rate of solid formation with temperature for sample SL, where dm(solids)/dT is the rate of solid formation. It is obvious that the curve shows similar trend with each other. A maximum rate appeared with temperature decreased in slags. In order to find the relation between rate of solid formation and Tcv, the temperature of maximum rate is defined as Tmax. Fig. 12 shows the measured Tcv of slags as a function of the calculated Tmax for the three samples with different CaCO3 additions. Good correlation between the Tcv and Tmax is observed. Moreover, the R value is 0.9492, which indicates that the excellent accuracy of the results is within the experimental error. The Tcv of slags could be expressed as a linear function of the Tmax according to the equation:
T cv ¼ 0:98T max þ 17:33
ð2Þ
3% CaCO3 0% CaCO3
15
30
45
60
2θ /
75
90
o
(b)
15% CaCO3 9% CaCO3 7% CaCO3 3% CaCO3 0% CaCO3
15
30
45
60
2θ /
75
90
o
(c)
4.5% CaCO3 3.5% CaCO3 1% CaCO3 0% CaCO3
15
30
45
60
2θ /
75
90
o
Fig. 13. XRD patterns of slag after viscosity test. (a) YC; (b) SL; (c) FG.
L. Kong et al. / Fuel 109 (2013) 76–85
respectively and crystals appeared with decreasing temperature. These results were consistent well with the XRD analysis. The same results can also be observed for samples YC and FG. This provides a good method to predict the type of slag by XRD. However, there is still not enough evidence to prove the quantitative relationship between type of slag and amount of crystals in slag. This should be studied further in the future work. In order to get a glassy slag for slag tapping, the method of predicting slag type and Tcv can be used to determine the amount of flux required or the ratio of blended coal. The slag type with different added ratios of flux or blended coal was firstly judged by XRD analysis. Tcv of slag blended with flux or coal were then calculated by Factsage. According to the results, the amount of flux required or the ratio of blended coal to get a glassy slag can be determined. 4. Conclusions In this study, influences of additions of CaCO3 on AFTs, slag viscosity, Tcv and slag type were investigated. The conclusions were expected to guide the use of limestone for slag tapping gasifiers. Liquidus surfaces in phase equilibrium diagrams for the pseudoternary systems correlated well with the trends of AFTs and the linear relation was set up between FT and liquidus temperature, TFT = 86.86 + 0.86Tliq. The relation was able to be used for conducting addition of limestone for gasification (R = 0.9211). The slag viscosity behavior changed with increasing addition of CaCO3. Above liquidus temperature, addition of CaCO3 increased NBO and FTIR of slag proved the content of Si–O–Si decreased. Below the liquidus temperature, the addition of CaCO3 not only influenced by solid content but also changed composition of solid phase. A new method to predict the Tcv was established based on the rate of solids formation. An apparent linear correlation and excellent agreement between Tcv and maximum rate of solids formation (Tmax) calculated by Factsage was found, Tcv = 0.98Tmax + 17.33 (R = 0.9492). The slag type also changed with addition of CaCO3. XRD analysis of slag showed a good consistency with result from viscosity–temperature curves. Acknowledgments Financial support from the National Basic Research Program of China (2010CB227005-02), Joint Foundation of Natural Science Foundation of China and Shenhua Group Corporation Ltd. (U1261209), the National Natural Science Foundation of China
85
(21006121), the Youth Foundation of Shanxi Province (2010021008-2), and the State Key Laboratory of Coal Conversion and State Key Laboratory of Coal Combustion (FSKLCC0909) was gratefully appreciated. References [1] Takematsu T, Maude C. Coal gasification for IGCC power generation. London: IEA Coal Research; 1991. [2] Song W, Tang L, Zhu X, Wu Y, Rong Y, Zhu Z, Shuntarou Koyama. Fusibility and flow properties of coal ash and slag. Fuel 2009;88:297–304. [3] Vassilev S, Kitano K, Takeda S, Tsurue T. Influence of mineral and chemical composition of coal ashes on their fusibility. Fuel Process Technol 1995;45: 27–51. [4] Hurst HJ, Novak F, Patterson JH. Phase diagram approach to the fluxing effect of additions of CaCO3 on Australian coal ashes. Energy Fuels 1996;10:1215–9. [5] Song W, Tang L, Zhu X, Wu Y, Rong Y, Zhu Z, et al. Effect of coal ash composition on ash fusion temperatures. Energy Fuels 2010;24:182–9. [6] Seggiani M. Empirical correlations of the ash fusion temperatures and temperature of critical viscosity for coal and biomass ashes. Fuel 1999;78: 1121–5. [7] Jak E, Degterov S, Pelton AD. Thermodynamic modelling of the system Al2O3– SiO2–CaO–FeO–Fe2O3 to predict the flux requirements for coal ash slags. Fuel 1998;77:77–84. [8] Browning GJ, Bryant GW, Hurst HJ, Lucas JA, Wall TF. An empirical method for the prediction of coal ash slag viscosity. Energy Fuels 2003;17:731–7. [9] Hurst HJ, Novak F, Patterson JH. Viscosity measurements and empirical predictions for fluxed Australian bituminous coal ashes. Fuel 1999;78: 1831–40. [10] Patterson JH, Hurst HJ. Ash and slag qualities of Australian bituminous coals for use in slagging gasifiers. Fuel 2000;79:1671–8. [11] Ilyushechkin AY, Hla San S, Roberts DG, Kinaev NN. The effect of solids and phase compositions on viscosity behaviour and Tcv of slags from Australian bituminous coals. J Non Cryst Solids 2011;357:893–902. [12] Song W, Tang L, Zhu X, Wu Y, Rong Y, Zhu Z. Fusibility and flow properties of shell gasifier slag. J Chem Ind Eng 2009;60:1782–6. [13] Wu L, Mattias EK, Song M, Du S. The effect of solid particles on liquid viscosity. Steel Res Int 2011;82:388–97. [14] Mills KC, Broadbent CP. Engineering foundation conference on the impact of ashdeposition on coal fired plants. Solihull (UK); 1993. p. 513–25. [15] Vargas S, Frandsen FJ, Dam-Johansen K. Rheological properties of hightemperature melts of coal ashes and other silicates. Prog Energy Combust Sci 2001;27:237–429. [16] Oh MS, Brooker DD, Paz de EF, Brady JJ, Decker TR. Effect of crystalline phase formation on coal slag viscosity. Fuel Proc Technol 1995;44:191–9. [17] Bale CW, Chartrand P, Degterov SA, Eriksson G, Hack K, Mahfoud RB, et al. FactSage thermochemical software and databases. Calphad 2002;26:189–228. [18] Li H, Qiu X, Tang Y. Ash melting behavior by Fourier transform infrared spectroscopy. J China Univ Min Technol 2008;18:245–9. [19] Senior CL, Srinivasachar S. Viscosity of ash particles in combustion systems for prediction of particle sticking. Energy Fuel 1995;9:277–83. [20] Kong L, Bai J, Li W, Bai Z, Guo Z. Effect of lime addition on slag fluidity of coal ash. J Fuel Chem Technol 2011;39:407–11. [21] Folkedahl BC, Schobert HH. Effects of atmosphere on viscosity of selected bituminous and low-rank coal ash slags. Energy Fuels 2005;19:208–15. [22] Gong D. The high-temperature viscosity model of coal ash. Therm Power Generation 1989;1:28–33.