Xinjiang lignite ash slagging and flow under the weak reducing environment at 1300 °C – Release of sodium out of slag and its modelling from the mass transfer perspective

Xinjiang lignite ash slagging and flow under the weak reducing environment at 1300 °C – Release of sodium out of slag and its modelling from the mass transfer perspective

Fuel Processing Technology 170 (2018) 32–43 Contents lists available at ScienceDirect Fuel Processing Technology journal homepage: www.elsevier.com/...

1MB Sizes 0 Downloads 6 Views

Fuel Processing Technology 170 (2018) 32–43

Contents lists available at ScienceDirect

Fuel Processing Technology journal homepage: www.elsevier.com/locate/fuproc

Research article

Xinjiang lignite ash slagging and flow under the weak reducing environment at 1300 °C – Release of sodium out of slag and its modelling from the mass transfer perspective

MARK

Hengsong Jia,b, Xiaojiang Wua,c, Baiqian Daia, Lian Zhanga,⁎ a b c

Department of Chemical Engineering, Monash University, Victoria 3800, Australia Department of New Energy and Engineering, Jiangsu University, Zhenjiang, Jiangsu, China Shanghai Boiler Works Co Ltd, 250 Huaning Road, Minhang, Shanghai 200245, China

A R T I C L E I N F O

A B S T R A C T

Keywords: Xinjiang lignite Ash slagging Sodium loss Mass transfer coefficient

This study examined the Na loss upon the slagging of five typical Xinjiang lignite ashes and their blends with clay at different ratios, under the simulated cyclone combustion conditions. The ash samples were laden on a 25° – inclined plate and exposed in 1% CO (nitrogen balanced) at 1000–1300 °C for maximum 2 h at each temperature. Apart from experimental approach, modelling based on the mass transfer mechanism was also attempted. The results show that, the Na loss ratios achieved are lower than that has been reported for the combustion of the same coals in the pulverised – coal fired plants, varying from 40 to 70% of the entire Na that largely depends on the basicity of ash or ash – clay blend. The reaction temperatures examined here exert a marginal influence on the Na loss. The addition of clay to raw coal ash and its dosage amount are crucial. At a mild combustion temperature of 1300 °C, the addition of 8–10wt% clay, on the ash mass basis to a basic lignite ash is the optimum, simultaneously accelerating both slag flow rate and Na retention in slag matrix. Additionally, the addition of clay alleviated the penetration and corrosion of Na vapour into the refractory corundum, thus extended its lifespan. The Na loss rate is dominated by the internal mass transfer rate of Na vapour inside slag film, in particular at the elevated temperatures from 1300 °C when ash is partially/mostly molten, as well as when the clay is added into the original Xinjiang ash. For the original ash which is rarely molten before 1200 °C, the gassolid reaction prevailed, leading to the capture of Na vapour by Al and Fe – bearing solid species and even the refractory corundum plate. In contrast, at 1300 °C, the formation of slag film decreased the Na vapour mass transfer coefficient drastically. The resistance was further enlarged by the formation of more molten species and a denser slag film upon the addition of clay. The original Na content in ash has proven insignificant in the mass diffusion control regime.

1. Introduction Vapourisation of alkali elements including sodium (Na) and potassium (K) causes severe deposition and fouling in the coal/biomassfired boilers [1–2]. It also fouls the syngas cooler in the downstream of a gasifier [3]. Understanding the emission and speciation of Na in a hot gas environment, either oxidising or reducing is essential in terms of boiler/gasifier design and operation. This is especially significant for the use of high - sodium lignite. Compared to the bituminous coal and anthracite with a decreased reserve, lignite such as Xinjiang lignite from China has much larger reserve and can be used for hundred years based on the current usage rate [4]. One of the critical issues related to the use of Xinjiang lignite is the large content of sodium within it,



which can go up to 5–10 wt% on the total ash-forming element basis. That is far more than the normal sodium content level (< 2%) [5]. Such a high sodium content has been found to possess the strongest slagging and fouling propensity compared to the other lignites in the rest of the world [2,6]. Since Na is a key trigger for the ash fouling and slagging inside a furnace, plenty of researches have been conducted on the Na vapourisation, especially for high sodium Xinjiang coal. The ash deposition of coal with high Na/Ca content was studied in full-scale and lab-scale furnaces [7]. The effect of alumina - silica additives on ash fusion and transformation of Xinjiang high-sodium coal was revealed [8]. The emission behaviour for organically bound elements in Xinjiang lignite during air and oxy-fuel combustion was also studied and compared with

Corresponding author. E-mail address: [email protected] (L. Zhang).

http://dx.doi.org/10.1016/j.fuproc.2017.10.016 Received 13 August 2017; Received in revised form 23 October 2017; Accepted 24 October 2017 0378-3820/ © 2017 Elsevier B.V. All rights reserved.

33

4.27 1.5–5.8

< 74 mm Ultrafine

7.28

20.55

< 74 mm

180–355 μm

4.64–12.21

< 74 μm

Na-rich Xinjiang lignite [13] Australian Loy Yang [13] Yidong [13] Ash slags from gasifier [14]

High-Na Zhundong coal [16]

2.1

Raw slag mixed with different additives

Slag from a PPCC pilot plant [12]

0–10

4.46

Not mentioned

Ultrafine

8.53

< 2.5 mm, mean diameter of 0.77 mm

High-sodium Zhundong coal [10] Zhundong coal [11]

Synthetic high-Na ashes [15]

Na2O wt% in ash

Size

Coal type

Lab-scale quartz - made fluidised bed reactor for pyrolysis and gasification

High-temperature horizontal tube furnace

KEMS

Lab-scale coal steam gasifier (40% steam in 60% N2)

Knudsen effusion mass spectrometer (KEMS)

Computer modelling

CFB gasifier

Experimental rig

850–950

Up to 1700

1300–1600

900–1000

1250–1450

1000–1300

850–1000

Temperature range, °C

Table 1 Literature review and summary on the release/loss of sodium from slag, in particular from Xinjiang lignite ash [10–16].

Na loss: 38.8% at 850 °C, 40% at 900 °C and 76.2% at 950 °C in total. For the char gasification stage, Na loss: 24.1% at 850 °C, 25.6% at 900 °C and 64.4% at 950 °C.

Not mentioned

20% at 1000 °C Not mentioned

25% @ 900 °C, 50% at 1000 °C, and

Na loss decrease by 30% upon adding 5% SiO2/TiO2 at1400 °C Na loss decreases by 50% upon adding 10% SiO2 at1400 °C 5–15% at1000 °C

Not mentioned

Not mentioned

% Na lost

50

30

N/A

20–60

% Na captured in solid/slag

Na loss correlates well with water-soluble Na in coal Na retention into NaAlSiO4 High SiO2/Al2O3 slag favours the capture of Na vapour Na captured in anorthite (CaAlSiO4) The draining slag rich in SiO2 can capture Na vapour Addition of Na2O reduces ash fusion temp by max 200 °C, down to minimum 950 °C Minerals albite (NaAlSi3O8) and nepheline (NaAlSiO4) react with anorthite (CaAl2Si2O8) to form low-temp eutectics Water-soluble and organic - Na released during pyrolysis stage

Na release augmented by chlorine (Cl) Na retention into NaAlSiO4 at high temperature Addition of kaolin increases the ash fusion temperatures by 250 °C maximum TiO2 is stronger than SiO2 for Na retention and network formation; Addition of MgO increases the activity (and loss) of Na out of slag Na loss correlates well with Cl in coal

Notes

H. Ji et al.

Fuel Processing Technology 170 (2018) 32–43

Fuel Processing Technology 170 (2018) 32–43

H. Ji et al.

Australian brown coal [9]. Table 1 briefly summarises the up-to-date researches related to emission of Na from the combustion/gasification of Xinjiang lignite and others, based on the recently published references [10–16]. In general, the emission of Na is augmented by the existence of chlorine (Cl) in the original coal, same as that has been found for the biomass [17]. The presence of alumina-silicate, either as an inherent mineral in the original fuel or as an external additives such as clay is beneficial in the immobilisation of Na into refractory minerals to form albite (NaAlSiO8) and nepheline (NaAlSiO4) [10,14]. However, most of the studies related to Xinjiang lignite were conducted at relatively low temperatures, such as 800–1000 °C that is in the operation window for CFB boilers/gasifiers [10,11,13,16]. The emission extent of Na also varies broadly between different Xinjiang lignites tested, demonstrating the complexity of the underpinning mechanisms for the Na release. For instance, a lab-scale steam gasifier of 900–1000 °C confirmed that maximum 10% of Na was lost from a Xinjiang lignite tested, relative to an Na loss of 50% and 20% from an Australian lignite, Loy Yang and an Inner Mongolia lignite Yidong, respectively [13]. In another study under the similar conditions [10], the Na loss was found to range from 20% to 60% for the Xinjiang lignite tested. With respect to the high temperatures such as 1200–1600 °C when the majority of the ash-forming elements are melted into slag [12], the Na loss is estimated to account for around 30% for a Germany slag. The addition of external SiO2 and TiO2 reduced the Na loss extent, whereas the addition of magnesium oxide (MgO) enhanced the activity and loss of Na out of the molten slag. Na was also partially bound into anorthite (CaAlSiO4) forming a complex structure in slag [15]. Additionally, the draining slag, which is rich in quartz (SiO2) was also able to capture the Na vapour upon the contact with the Na-bearing syngas. Once having Na embedded inside, the flow fluidity of a slag will also be enhanced greatly [15]. As far as the authors are aware, the knowledge related to the emission of Na out of Xinjiang slag has yet to be touched in the past. In particular, the knowledge is missing for the emission of Na out of a cyclone furnace where the ash is slagged under the mild conditions relative to that in an entrained-flow gasifier. Obtaining such knowledge is crucial for a proper design and operation of a cyclone furnace which aims to melt the majority of ash – forming elements including Na into the bottom slag [17]. This is expected to be superior over the existing pulverised – coal fired boiler in which > 80% of Na converts into sulphates that deposit firmly on and corrode the heat – exchange tube surfaces [2,18,19]. In parallel to another study examining the slag flow rate of Xinjiang ashes and their blends with clay under the mild conditions [20], this paper aims to quantitatively reveal the partitioning of Na during the ash slagging process, as a function of ash basicity (base/ acid ratio), temperature, clay additive percentage and exposure time. Efforts were also made to correlate the Na loss extent with slag flowability, thereby optimising the addition percentage of clay. Furthermore, the vapourisation rate of Na was interpreted from the mass transfer perspective, in which both the thermodynamic equilibrium and the internal resistance of slag liquid layer against the diffusion of Na vapour are taken into account. The derived overall mass transfer coefficient from the mass transfer modelling is expected to be used in assisting the design of the bottom slag – tapping section of the furnace, so as to maximise the retention of Na inside the furnace. The modelling approach here is also expected to be simpler and even superior compared to the kinetic modelling which is affected by a large number of unknown chemical and physical variables.

Table 2 Ash compositions of five Xinjiang lignite and clay, wt%.

Na2O MgO Al2O3 SiO2 P2O3 SO3 K2O CaO TiO2 Mn3O4 Fe2O3 Others

Ash A

Ash B

Ash C

Ash D

Ash E

Clay

3.76 7.05 12.39 15.67 0.13 19.51 0.4 38.12 0.2 0.18 2.34 0.25

3.63 5.4 13.94 27.66 0 13.95 0.96 31.27 0.18 0 2.46 0.55

3.28 7.98 10.63 15.83 0 20.22 0.32 37.32 0.19 0 3.77 0.46

3.21 2.92 19.37 46.64 0.6 10.22 1.44 6.83 0.2 0.03 8.54 0

3.09 2.59 23.91 49.74 0.34 5.62 1.84 6.38 0.3 0 6.08 0.11

0.01 1.62 21.97 71.93 0 0.19 1.75 0.22 0 0 1.16 1.15

original ash samples with the basicity variation from 0.27 to 1.98. The narrow range of Na2O content is close to the Xinjiang ash samples tested elsewhere [10,11,13,16], and also falls in the range of the Na2O contents tested for the Germany slags [12,14]. Note that, an extra 5–10 wt% Na2O was added into the basic ash A in this study, which increases the maximum Na2O content in the ash to 13.76 wt%. That is in line with the upper limit 12.21 wt% Na2O that has been found in a Xinjiang lignite ash previously [13]. Apart from the five original ashes, namely ashes A through to E with the basicity varying from 1.98 to 0.27 [20], each ash was also blended with clay at 10–40 wt% to assess its influence on both slag flow and Na retention. The clay used here is the overburden from a Xinjiang opencut mine, which is on top of the lignite layer and rich in SiO2 and Al2O3. The composition for the clay is also listed in Table 2. The content of Na2O is negligible in the clay additive, and hence, the blending of clay with each of the five Xinjiang ashes is insignificant in affecting the content of Na2O. Same as the high-purity corundum substrate that was used for ash slagging in this study, in which Na2O is also negligible, as confirmed by the SEM-EDX quantification (data not shown). For the blending of clay, it was prior ground to < 63 μm, which is the same size as the original ashes. Regarding the addition of extra 5–10 wt% Na2O, the reagent – grade NaOH was used as the precursor and prepared at the respective percentage to mix with ash A. 2.2. Quantification of Na2O in slag To reiterate, around 200 mg of an ash sample was loaded on a 25° – inclined, high-purity corundum plate and exposed in a weak reducing environment, i.e. 1% CO in nitrogen at flow rate of 3 L/min for maximum 2 h [20]. After the tests at 1000 °C and 1100 °C, the ash samples still remained un-molten and were easily removed from the plate (as evident in Fig. 3). They were subsequently dispersed uniformly on a double – side carbon tape, and carbon – coated prior to being analysed by SEM–EDX. In contrast, for another two high – temperatures 1200 °C and 1300 °C tests, the ash samples, termed as slag hereafter were mostly molten and stuck firmly on the corundum plate. The whole slag – laden plate was embedded into epoxy resin, and then halved along the vertical centreline of the plate by a Stuers low – speed saw. The resultant cross – section, bearing a maximum width of 2 mm was further polished by SiC papers with the grits of 1200 and 2400, subsequently carbon coated and subjected to the SEM–EDX analysis. Three standard fly ash samples, same as those used for the ICP-OES calibration elsewhere [21], were measured under the same operating conditions to validate the accuracy of the SEM-EDX quantification results. As evident in Table 3, all the SEM-EDX measured results for Na and other elements are quite consistent with the respective certified/ recommended values, demonstrating the high accuracy of the quantification method used throughout this study. The percentage of Na lost due to vapourisation was calculated by using silicon in ash/slag as the tracer, based on the assumption that

2. Experimental section 2.1. Properties of Xinjiang lignite ash samples and clay As shown in the Table 2, the ash composition for the test five ash samples are provided (which is the same as in the companion paper [20]). The contents of Na2O vary from 3.09 to 3.76 wt% in the five 34

Fuel Processing Technology 170 (2018) 32–43

H. Ji et al.

where the symbol KL on the right hand side stands for the liquid slag – side mass transfer coefficient based on the molar fraction of Na (mol/ m2·s·mol fraction). The symbols, Na(g)_L, and Na⁎(g)_L refer to the molar fraction of Na vapour in bulk slag, and the equilibrated molar fraction of Na vapour on the slag/gas interface in the slag side, respectively. These two molar fractions can be further visualised in panel b, where the molar fraction Na(g)_L in bulk slag is quantified by the SEM-EDS, whereas the Na⁎(g)_G is thermodynamically equilibrated with the molar fraction of Na vapour in the bulk gas by Henry's law, as shown in Eq. (3) below.

Table 3 Validation of the SEM–EDX quantification accuracy by using three standard ash samples. NIST2690 fly ash

Na Mg AI Si P S Cl K Ca Ti Mn Fe

Certified 0.24 1.53 12.35 25.85 0.52 0.15 1.04 5.71 0.52 0.03 3.57

BCR38 fly ash

SRM19 HTA Measured 0.30 1.65 12.28 23.3 0.75 0.21 0 1.16 6.67 0.71 0.12 4.54

Certified 0.79 0.44 15.59 19.29 0.07 1.8 0.73 3.65 0.42 0.06 4.5

Measured 0.69 0.33 13.39 18.28 0.08 1.89 0 0.39 2.74 0.44 0.19 2.33

Certified 0.37 0.89 15.57 27.15 0.23 1.1 3.07 1.32 0.58 0.05 3.38

Measured 0.45 0.91 15.16 25.23 0.21 0.96 0.04 4.22 1.67 0.65 0 4.02

Na∗(g) L = Na (g) G/H′

where the symbol Na(g)_G refers to the concentration of Na in the bulk gas that can be estimated based on the original Na content and its loss fraction, whereas H′ is Henry's law constant that can be estimated by the thermodynamic equilibrium calculation. Here the ‘equilib’ module in FactSage 6.4 was employed to estimate this constant for each ash. The input for the calculation includes the absolute amounts of individual elements in an ash in which the content of Na2O was manipulated to vary from 3 wt% through to 24 wt%, the amount of CO calculated by bulk gas flow rate and exposure time and temperature which was fixed at 1000–1300 °C. The solid solution of Slag A in FactSage was used here, considering that it has been used for the ash slagging estimation and it also converged for most of the cases. Fig. 2 demonstrates the calculation for ash A with different Na2O contents. Irrespective of the content of Na2O in the solid ash, its vapourisation loss remains constantly at around 99.8%, suggestive of the high volatility of this element, as expected. By further calculating the molar fraction of Na vapour in the gas phase (molar amount of Na over the molar amount of total bulk gas) and plotting against the concentration of Na2O in the original ash, a linear correlation trend was derived and its slope, 2.778 × 10− 5 with a unit of mol_gas/weight_solid was consequently read off as the Henry's law constant. Back to Eq. (2), the left hand side for molar flux can be determined by the Eq. (4) below:

silicon has no loss upon the high-temperature exposure test. Such an assumption holds true, since the ash powders remain static during the test. With regards the calculation method based on the change of the absolute mass of Na upon a test, its error is larger, considering that the content of carbon, namely loss on ignition (LOI) cannot be quantified by SEM–EDX analysis. Calculation of Na loss was based on the following Eq. (1):

( ) −( ) Na lost% = ( ) Na Si b

Na Si a

Na Si b

× 100 (1)

where the term Na/Si refers to the mass ratio of Na and Si in a sample quantified by SEM-EDX, whereas the two subscripts, b and a stand for the original and used ash before and after a test, respectively. 2.3. Modelling of Na2O vapourisation out of slag Although the kinetic regarding the vapourisation of Na has been approached from an Arrhenius equation perspective [22], the genericity of a kinetic model is highly sensible to the properties of the ash matrix hosting Na, including the physical properties of the ash powder such as size, porosity and even tortuosity, mode of occurrence of Na and the other ash – forming elements in the matrix as well. This is evident by the large discrepancy between the Na loss extents summarised in Table 1. Additionally, compared to the reaction kinetic, the diffusion control is more dominant at a very high temperature [23]. In particular, the slag film formed on the wall surface provides a considerable resistance against the release of any volatile elements including Na out of it [12,14,24]. As has been confirmed in a similar study on the use of 30 mg ash/slag in a melting test, the overall Na loss rate is fully controlled by mass transfer from 3500 s onward [25]. Given the fact that a larger amount of ash (i.e. ~200 mg) was used in each test here and the exposure time was also relatively long, it is reasonable to assume that the mass transfer far overwhelms the reaction kinetics. Based on these considerations, the vapourisation of Na out of Xinjiang lignite slag was modelled from the mass transfer perspective. As depicted in Fig. 1a, once being disassociated, the Na vapour is expected to diffuse out of the un-molten ash matrix via the inter-particle pores at low temperatures. Upon the progress of ash exposure over the increase of either temperature or time, the inter-particle pores are gradually blocked by the molten slag until the whole ash is fully molten. Consequently, the Na vapour needs to diffuse out of the liquid slag before it is swept away by the bulk flue gas, as demonstrated in Fig. 1b. On the assumptions of a steady state and that Henry's law for dilute system holds on the slag/gas interface, the emission molar flux, NNa of Na vapour can be expressed by the classic overall mass transfer mechanism shown in Eq. (2) below, which is based on the liquid slag side.

NNa = KL (Na (g) L −Na∗(g)L )

(3)

NNa =

∆Na L A × ∆t

(4)

where ΔNa_L refers to the loss of Na in moles after a test, Δt for the exposure duration, and the symbol A for the surface area of the liquid slag which is a product of the width and length of the slag based on the assumption that slag is in a rectangle shape (see Fig. 3). Both the width and length of slag were photographed and measured after each test [20]. It is noteworthy that, for the calculation here, the bulk motion effect was ignored, considering that Na is a minor element and its concentration in either bulk slag or gas is also extremely low. Moreover, the Henry's law constant calculated here is not a pure, physical partitioning coefficient of Na vapour based on its solubility. Instead, it should also include any possible reactions between Na vapour and other elements such as silica and alumina in the ash, and even the corundum plate. These reactions have been taken into account during the FactSage calculation, except the reaction between slag and corundum plate the extent of which is unknown. However, the slag – plate reaction will be examined and its effect on the change on value of KL will be discussed qualitatively later. 3. Results and discussion 3.1. Na release from basic ash A with a base/acid ratio of 1.85 Effort was first made to examine the influences of exposure temperature and the addition of 10 wt% clay on the release of Na out of the basic ash A which is a typical lignite ash that is rich in alkali and alkaline earth metals. Fig. 3 depicts the temperature - dependent slagging behaviour of this ash. The exposure time is 2 h and the gas composition is 1% CO in nitrogen, with a total flow rate of 3 L/min for all the four

(2) 35

Fuel Processing Technology 170 (2018) 32–43

H. Ji et al.

interface Liquid/Solid

(a)

interface

Gas

Liquid/Solid Na(g)_L

Na(g) Flue gas

Na(g)_iL

Fig. 1. Mass transfer mechanism for sodium vapour out of slag (a), proposed concentration profile of sodium vapour across the slag layer and in bulk gas (b), and proposed vapourisation and release of Na(g) out of ash/slag matrix (c).

(b)

Gas H’ Na*(g)_G Na(g)_iG Na(g)_G

Corundum plate

Slag layer

(c) Na(g)

Na(g)

Solid species

Na(g)

Molten species Pores Temperature: low (1000-1100oC) Time: short (~ 10 min)

medium (1200oC) medium (~ 30 min)

high (1300oC) long (>60 min)

slagging of ash matrix by clay led to the formation of a dense liquid film retarding the diffusion of Na vapour out of it. Due to the mass transfer resistance, the measured Na losses are far from those predicted based on the thermodynamic equilibrium, as shown in dashed curves.

temperatures. As can be seen, the original ash sintered and shrank before 1200 °C, whereas it spread a little bit at 1300 °C. The addition of 10 wt% clay caused little change before 1200 °C, but led to a rapid motion and even over - flowing at 1300 °C. With regard the Na loss, the experimental measurements in solid lines confirmed a loss of approximately half of the entire Na at 1000 °C for the raw ash, which was increased slowly to around 60% at 1200 °C and 1300 °C. The addition of clay caused little change before 1200 °C. This further suggests that the kinetic of Na capture is even not a limiting factor before 1200 °C. The vapourised Na escaped the ash matrix rapidly, leaving little chance for the mobile Na ion to be scavenged by the static clay particles. The influence of clay addition is considerable at 1300 °C. However, it should not be due to the chemical scavenging of Na vapour by clay according to the reaction (5). The fixation of Na by aluminosilicate is only favoured at the temperature below 1000 °C, above which the surface of clay is gradually sintered and deactivated [12]. The thermodynamic equilibrium prediction confirmed a 100% loss of Na upon the addition of clay (see the open triangle symbols in Fig. 3). Clearly, the accelerated

100.0

0.0007

(a)

Na2O + 2SiO2 + Al2 O3 = 2NaAiSiO4

The calculated slag – side overall mass transfer coefficients were plotted in Fig. 4 for ash A with and without 10% clay. This quantitatively confirmed a remarkable change on the diffusion resistance with the rise on temperature and clay addition. For ash A only, the mass transfer coefficient of Na vapour out of it is the largest at 1000 °C, suggestive of a least resistance from the ash matrix which is made up of loose solid particles, as evident in Fig. 3. The overall mass transfer coefficient decreases and reaches its minimum at 1100 °C, implying a complete blockage of inter – particle passages for Na vapour due to the great sintering and shrinkage of the ash matrix. Upon the further increase of the temperature, the overall mass transfer coefficient increases slightly, indicating that the change on the ash shape due to spreading reduced the thickness of the ash/slag layer, and thus accelerated the Fig. 2. Vapourisation extent of Na as a function of the content of Na2O in Xinjiang ash A (a), and the molar concentration of Na(g) vapour versus content of Na2O in Xinjiang ash (b).

(b)

99.8 0.0006

99.6

Na(g) Vapor, mol%

99.4

Na% Lost

99.2 99.0 98.8 98.6 98.4

0.0005

y = (2.778E-5)x, 2 R = 0.9958

0.0004

0.0003

0.0002

98.2 98.0

(5)

0.0001 4 6 8 10 12 14 16 18 20 22 24

4 6 8 10 12 14 16 18 20 22 24

Na2O wt% in Ash

Na2O wt% in Ash 36

Fuel Processing Technology 170 (2018) 32–43

H. Ji et al.

Ash A

Ash A + 10 wt% clay 100

1000oC

1100 oC

% Na Lost

80

1200oC

60

Ash A only Ash A + 10% clay Predicted_ash A only Predicted_ash A + 10% clay

40

1300oC

20 1000

1050

1100

1150

1200

1250

1300

Temperature, oC 20 mm Fig. 3. Sodium release during the slagging of Xinjiang ash A, with and without the addition of 10 wt% clay at four different temperatures, with a residence time of 2 h.

exposed at 1300 °C for 2 h. It can be clearly seen that a considerable amount of Na crossed the interface and penetrated inside the plate, with a penetration thickness of around 10 μm. Such a phenomenon has also been observed elsewhere [3,26], which is however believed to shorten the life-time of the refractory wall. More interestingly, for the exposure of ash A blended with 10% clay, cross-section of the resultant slag is distinctive from ash A only. As depicted in Fig. 6, the crystal species disappeared. Except Mg, all the other elements including Na evenly dispersed in the liquid slag. This indicates that the penetration of Na ion inside the corundum plate is unable to take place, or to a less extent upon the addition of clay. Instead, it was wrapped by the liquid slag and rapidly flew away from the plate. Apart from the entrainment of Na, such a positive effect is clearly also beneficial in the protection of the refractory wall.

outward diffusion of Na, by a limited extent though. The mass transfer coefficients for the ash A blended with 10 wt% clay are generally less than those obtained for the original ash A only. In particular, the gap is larger at the two end temperatures, 1000 °C and 1300 °C. For the former end temperature, the capture of Na vapour by clay via gas-solid reaction should be predominant, whereas at the latter end temperature, the resistance from the liquid slag should prevail. Compared to the raw ash having only 30 wt% molten at 1300 °C, the liquid fraction is increased to around 80 wt% once the clay is added in the ash [20]. This is also evident in Fig. 3 for an over-flowing of this slag. Apart from the mass transfer resistance, the capture of Na ion by the corundum substrate is also a possible factor in alleviating the Na loss, as visualised in Fig. 1a. Fig. 5 depicts the cross-section of ash A after

Fig. 4. Mass transfer coefficient for the release of Na(g) vapour out of ash A as a function of temperature in 1% CO in argon.

-5

1.0x10

Ash A only Ash A + 10 wt% silica

-6

9.0x10

-6

.

.

Mass transfer coefficient, 2 mol(m s mol frac)

8.0x10

-6

7.0x10

-6

6.0x10

-6

5.0x10

-6

4.0x10

-6

3.0x10

-6

2.0x10

-6

1.0x10

0.0 1000

1050

1100

1150

1200

1250 o

Reaction Temperature, C 37

1300

Fuel Processing Technology 170 (2018) 32–43

H. Ji et al.

Al

25 µm

Ca

S

Slag

Plate

100 µm

O

Na

Si

Fe

Mg

Fig. 5. Elemental mapping for the cross-sectional surface of corundum plate loaded with ash A at 1300 °C, 2 h. The dashed curve refers to the interface/boundary between liquid slag and corundum substrate.

around 10 min, implying that the slagging of ash matrix is completed in the first 10 min. Afterwards, the emission of Na is simply dependent on its thermodynamic equilibrium nature, i.e. the saturation vapour pressure on the slag surface, and the mass transfer resistance. This is consistent with the kinetics of Na2O loss from the CaO-SiO2-Al2O3 slag studied before [25]. With respect to the influence of clay addition, it is more effective in reducing the Na loss at the initial stage up to 20 min, for both basic ashes. However, from 30 min onwards, the Na loss is

3.2. Time-resolved release of Na from basic ashes The second effort was made to examine the time - resolved release of Na from two basic ash samples, A and C with a basicity of 1.85 and 1.97, respectively. Each ash sample was also blended with 10 wt% clay and tested for comparison. The exposure temperature was fixed at 1300 °C here. As depicted in Fig. 7, for all the cases except the case of ash A blended with 10 wt% clay, the Na loss is quickly levelled off from

Al Coal slag

O

S

Slag

Plate

Corundum plate 25 µm

100 µm

Ca

Si

Na

Mg

Fe

Fig. 6. Elemental mapping for the cross-sectional surface of corundum plate loaded with ash A +10 wt% clay at 1300 °C, 2 h. The dashed curve refers to the interface/boundary between liquid slag and corundum substrate.

38

Fuel Processing Technology 170 (2018) 32–43

H. Ji et al.

100

80

% Na Lost

maturation of the slag matrix. However, for the two basic ashes in particular ash A, the mass transfer coefficient changes much more slowly. It decreases rapidly in the first 20 min, then slowly until converges to the same value for the two basic ashes after 2 h. This echoes a slow change on the colour and shape of the slag formed from ash A only in Fig. 8. The extra information related to the cross-section of ash A slag formed at various times, shown in Fig. 10 supports the slow change on the mass transfer coefficient of Na vapour. As it can be seen, the ash matrix remains highly porous in the first 10 and 20 min. It is only slightly sintered after 30 min and even 60 min. On the one hand, such a high porosity is in favour of the diffusion of Na vapour via inter-particle gaps. On the other hand, the abundance of solid species in ash matrix can promote the gas-solid reaction for the capture of Na vapour, in particular from 30 min onwards. This is evident by the mapping of Na, Al and Fe elements in Fig. 10 that proves the co - existence of these three elements as crystal species in the ash matrix. Such a negative effect in turn counter balances the positive role of ash porosity, leading to a comparable Na loss rate compared to the other cases in Fig. 8. Back to Fig. 9, for all the blends made up of raw ash with clay or Na2O, their mass transfer coefficient profile is rather identical, showing a descending trend upon the increase on the exposure time. This suggests that the two additives, clay and Na2O are comparable in terms of promoting the basic ash slagging behaviour. Moreover, irrespective of the exposure time, their mass transfer coefficients are far less than the two raw ashes. Their decrease upon the exposure time is also rather marginally. All echo a rapid melting of ash matrix upon the blending with clay or Na2O. The resultant slag films are also rather similar in structure, thereby exerting a similar resistance against the diffusion of Na ion.

Ash A only Ash A + 10 wt% Clay

60

40

20

0 0

20

40

60

80

100

120

Residence time, min 100

Ash C only Ash C + 10 wt% Clay 80

% Na Lost

60

40

3.3. Influence of ash basicity on the Na loss

20

The final effort was made to evaluate the effect of ash basicity (i.e. base/acid) ratio on the Na loss. This was conducted by testing five different original ash samples, namely ash A through to ash E with the basicity being 1.83, 1.05, 1.98, 0.35 and 0.27, respectively. Each ash was also blended with 10 wt% clay and tested. In addition, ash A blended with up to 40 wt% clay was further tested, with its basicity decreasing from 1.83 to 1.35, 1.02, 0.78 and 0.60 for ash A only, ash A + 10 wt% clay, ash A + 20 wt% clay, ash A +30 wt% clay, and ash + 40 wt% clay, respectively. All the tests were conducted at 1300 °C for a duration of 2 h. As shown in Fig. 11(a) for the basic ash A blended with clay at different ratios, it is clear that the Na loss is decreased slightly and linearly upon the rise of the clay addition ratio. For instance, relative to around 70% Na loss for the raw ash A only, the addition of 40 wt% clay, on the mass basis of ash A reduced the Na loss down to 60%. Such an observation is qualitatively in agreement with those summarised in Table 1, strengthening the statement that Na is stabilised into the alumina-silicate matrix. However, the extent of Na loss reduction is only 14% upon the addition of 40 wt% clay, which is far lower than the ratio of 30–50% reduction for the addition of only 5–10 wt% SiO2 at 1400 °C [12,14]. Such a difference should be attributed to a lower extent of polymerisation of the slag formed for the lignite ash A tested here. The Germany slag tested in [12,14] is made up of 55.152–59.366 wt% SiO2, whereas the ash A blended with 40% clay here only increased the SiO2 content to 38 wt%. For all the slag-forming elements, SiO2 has the largest effect on the activity of alkali metals, acting as a strong network former [12,14,27]. The results for the five different ashes in panel b show the similar phenomena. That is, the most acidic ash E has the lowest Na loss ratio, followed by acidic ash D, neutral ash B and then two basic ashes A and E. It further confirms that the addition of 10 wt% clay is insignificant, except the acidic ash D. A good correlation was also confirmed between the basicity of ash and the respective Na loss ratio in Fig. 12, where all the samples including original five ashes and their blends with different ratios of clay

0 0

20

40

60

80

100

120

Residence time, min Fig. 7. Time-resolved loss of Na from Xinjiang ashes A and C at 1300 °C with and without the addition of 10 wt% clay.

rather independent on the addition of clay. In light of this, it is referable that the mobile Na ion should be in portion bound chemically with the alumina-silicate matrix that is formed in the first 20 min. Afterwards, the alumina-silicate slag matrix is depolymerised by the accumulated Na and other basic ions within it, and consequently, the chemically bound Na is re–mobilised and the resultant free Na vapour is released out. The depolymerisation effect of Na2O to reduce the slag viscosity has been reported elsewhere [15,27]. For the CaO-SiO2 – dominant slag, the introduction of Na can progressively depolymerise the CaSiO3 chain, cutting them into shorter and non-uniform ones. It can even replace CaO to form SiO4 tetrahedrons and Na-bearing crystals [14,15,27]. This is also proven in Fig. 8 upon the test on the addition of extra 5–10 wt% Na2O into ash A. As shown on the left hand side, the addition of 5 wt% Na2O led to an over-flowing slag in 30 min, whereas the addition of 10 wt% extra Na2O even caused a rapid over-flowing in 10 min. Moreover, regarding the Na loss rate profile, it remains identical and independent on the addition of extra Na2O into the slag. This substantiates the statement that the Na loss rate is independent on the original concentration of Na2O in ash matrix. The overall Na loss should not be governed by a first-order chemical reaction either. Instead, it merely depends on the thermodynamic equilibrium nature of Na which determines the saturation pressure of Na vapour on the slag surface. Fig. 9 shows the calculated time – dependent mass transfer coefficients for Na vapour in the different cases. For all the cases, the mass transfer coefficient decreases exponentially and then levels off, due to a gradual 39

Fuel Processing Technology 170 (2018) 32–43

H. Ji et al.

10min

20min

30min

40min Na loss percentage, % 100

Raw Ash A Raw Ash A + 5 wt% NaOH Raw Ash A + 10 wt% NaOH

Ash A only

% Na Lost

80

Ash A+ 5 wt% NaOH

60

40

20

Ash A+ 10 wt% NaOH

0 0

5

10

15

20

25

30

35

40

45

Time, min

Fig. 8. Effect of additional Na2O on ash A slag flow and the loss of added Na2O at 1300 °C, 1% CO in N2.

8% clay addition is the most beneficial in slag flow rate. At such a low clay addition ratio, the slag formed can entrain around 40% Na (as shown in Fig. 12) and also have the potential to quickly flow out of the furnace. As substantiated in Fig. 14, in parallel to the acceleration of slag flow rate, both the Na loss ratio and the overall mass transfer coefficient decrease simultaneously and proportionally. In particular, the overall mass transfer coefficient decreases more steeply. This in turn alleviates the Na – related fouling issues in the heat – exchanger zones. Instead, the use of excess clay increases the viscosity of the ash/slag matrix and thus difficulty for the discharge of slag out of the furnace, although it can reduce the Na loss down to 40%.

were included. As shown in panel a, the Na loss ratio initially increases rapidly, and then levels off from the basicity of around 0.8–1.0 for an equivalent content between the base and acid components in the ash. By plotting the thermodynamically predicted Na loss in panel b, one can further see that the acidic ashes or ash – blends with a basicity close to 0.4 and less show a good match between the measured and predicted Na loss results. However, beyond the basicity of 0.4, the gap is enlarged and stabilised between the measured and predicted Na loss results. Again, this proves the controlling of internal mass transfer resistance for the diffusion of Na ions at 1300 °C. From the Na capture perspective, the use of excess clay such as 40 wt% on the mass basis of a Xinjiang basic ash is essential, as substantiated in Fig. 12. However, as discussed in the companion paper [20] and further shown in Fig. 13 for ash A blended with different ratios of clay, it is obvious that a slightly acidic ash – clay blend, referring to

3.4. Practical implications Although a similar study has yet to be conducted in the literature,

-5

5.0x10

Ash A Ash A + 10wt% silica Ash C Ash C + 10% silica Additional 5 wt% Na2O Aditional 10 wt% Na2O

Ash A

-5

3.0x10

.

.

Mass transfer coefficient, 2 mol/(m s mol frac)

-5

4.0x10

Ash C

-5

2.0x10

-5

1.0x10

0.0 0

20

40

60

80

100

Residence time, min 40

120

Fig. 9. Time-resolved mass transfer coefficients for the release of Na(g) vapour out of Xinjiang ash slag at 1300 °C, in 1% CO in argon.

Fuel Processing Technology 170 (2018) 32–43

H. Ji et al.

Corundum Ash/Slag

Al

Na

Fe

10 min

20 min

30 min

60 min 25µm

100µm Fig. 10. Micro-structure of the cross-sections for ash A at 1300 °C with different residence time, and the elemental mapping for Al, Na and Fe on the slags formed at 30 min and 60 min.

collected from a PPCC pilot plant in which around 30% Na is stabilised in the slag at 1400 °C. They are far better than the loss of Na in the typical pulverised coal – fired plant burning the same coal in which only 20% Na was captured in the bottom ash/slag [18–21]. Clearly, the slagging – type cyclone combustion at a mild temperature such as 1300 °C is suitable for Xinjiang lignite. In addition, it is confirmed that

the 40–70% Na loss ratio based on a slag-tapping combustion process, as confirmed in this study is comparable to the results achieved at 850–1000 °C in the lab-scale circulated fluidised bed reactor, suggesting that the liquid film resistance is comparable with the kinetic restrictions encountered at the low temperatures, such as solid-solid/gas - solid reactions. Moreover, the results achieved are comparable with the slags

% Na Lost

100

100

(a)

80

80

60

60

40

40

20

20

0

(b)

No silica 10 wt% silica

0 0

10

20

30

40

Wt% of silica added to ash A

A

B

C

D

Ash Samples 41

E

Fig. 11. Sodium release for the addition of 10-40 wt% clay to ash A (a), and the five different ash samples with and without the addition of 10 wt% clay (b), at 1300 °C and 2 h.

Fuel Processing Technology 170 (2018) 32–43

H. Ji et al.

100

100

(a) Ash only Ash + 10% silica Ash A + >10 wt% silica

90

Measured Predicted

80

% Na Lost

Fig. 12. Sodium release as a function of base/ratio for five Xinjiang ashes with and without clay additive, at 1300 °C and 2 h. Panel (a) is experiment results for ash A and ash A with clay additive; panel (b) is the comparison between the measured and thermodynamic equilibrium predicted values.

(b)

80

70

60

60

50

40

40

30 0.0

0.4

0.8

1.2

1.6

2.0

2.4

0.0

0.4

0.8

1.2

1.6

2.0

2.4

(Na2O + K2O + CaO + MgO +Fe2O3)/(Al2O3+SiO2+TiO2)

Xinjiang lignites. 2) The addition of clay to raw coal ash is essential to promote the entire ash slagging, slag flow rate as well as capture/entrainment of Na vapour inside the slag liquids. However, the clay addition percentage is crucial. At a mild combustion temperature of 1300 °C, the addition of 8–10wt% clay, on the ash mass basis to a basic lignite ash is the optimum, simultaneously accelerating both slag flow rate and the Na retention in slag matrix. Additionally, the addition of clay alleviated the penetration and corrosion of Na vapour and other ash – forming elements into the refractory corundum, which can extend its lifespan. 3) The Na loss rate is dominated by the mass transfer inside the slag film, in particular at the elevated temperatures such as 1300 °C when ash is partially/mostly molten, as well as when the clay is added into the original Xinjiang ash. For the original ash which is rarely molten before 1200 °C, the solid – gas reaction prevail, leading to the capture of Na vapour by Al and Fe – bearing species and even refractory wall. In contrast, at 1300 °C, the formation of slag film decreases the Na vapour mass transfer coefficient drastically. This is further accelerated by the formation of more molten species and a denser slag film upon the addition of clay. The original Na content in ash has proven insignificant in the mass diffusion control regime.

the addition of 8–10 wt% clay, corresponding to a basicity of 1.4 is optimum, which accelerates the slag flow rate remarkably, and simultaneously reduces the mass transfer coefficient for the Na vapour. A quick discharge of the molten slag out of the furnace can also help entrain the Na vapour to flow out the furnace together. Moreover, the corrosion of refractory wall by the Na vapour is alleviated upon the addition of clay. 4. Conclusions This study has examined the Na loss upon the slagging of five typical Xinjiang lignite ashes and their blends with clay at different ratios, under the simulated cyclone combustion conditions. The ash samples were laden on a 25° – inclined plate and exposed in 1% CO + nitrogen at 1000–1300 °C for maximum 2 h at each temperature. Apart from experimental approach, modelling based on the mass transfer mechanism was also attempted. The major conclusions achieved are summarised as follows: 1) The Na loss varies from 40 to 70% of the entire Na, depending largely on the basicity of ash or ash – clay blend. The reaction temperatures examined here exert a marginal influence on the Na loss. The Na loss ratios achieved are lower than that has been reported for the combustion of the same coals in the pulverised – coal fired plants, proving that cyclone combustion is suitable for the

Clay

0%

2%

Base/Acid Ratio

1.83

1.72

4% 1.62

6% 1.52

8%

10%

20%

30%

40%

1.43

1.35

1.02

0.78

0.60

Ash A

Fig. 13. Slag flow rate as a function of clay addition percentage to ash A in 1% CO in argon, at 1300 °C. The ratios from 0% to 10% was conducted in 40 min, whereas the other three ratios were conducted in 2 h.

42

Fuel Processing Technology 170 (2018) 32–43

H. Ji et al. -5

70

1.2x10

(a)

(b)

60

-5

Overall Transfer Coefficient, 2 [mol/)m .s. mol frac]

1.0x10

% Na Lost

50

40

30

20

10

0 0.00

0.02

0.04

0.06

0.08

Slag flow rate, mm/min

0.10

Fig. 14. Na loss ratio and overall transfer coefficient of Na vapour as a function of slag flow rate. The experimental points listed here include five original ashes with and without 10 wt% clay (2 h exposure time), and ash A blended with different ratios (same as in Fig. 13) at 1300 °C.

-6

8.0x10

-6

6.0x10

-6

4.0x10

-6

2.0x10

0.0 0.00

0.02

0.04

0.06

0.08

0.10

Slag Flow Rate, mm/min

Acknowledgment [12]

This work was supported by Australian Research Council (ARC) under its Industrial Research Training Hub (150100006) scheme for the joint project between Monash and Shanghai Boiler Works Co Ltd. The author Xiaojiang Wu is grateful to Science and Technology Commission of Shanghai Municipality (15dz1206500). The author Hensong Ji is also grateful to China Scholarship Council (CSC) for the support to his one – year visit in Monash from June 2016 to July 2017. Bai-Qian Dai is partially supported by Brown Coal Innovation Australia (BCIA) for the PhD top-up scholarship.

[13]

[14]

[15]

[16]

References [17] [18]

[1] Q. Zhu, Update on Lignite Firing, IEA Clean Coal Centre, (June 2012) (London). [2] B.Q. Dai, F. Low, A. De Girolamo, X. Wu, L. Zhang, Characteristics of ash deposits in a pulverized lignite coal – fired boiler and the mass flow of major ash – forming inorganic elements, Energy Fuel 27 (10) (2013) 6198–6211. [3] P. Wang, M. Massoudi, Slag behaviour in gasifiers. Part 1: influence of coal properties and gasification conditions, Energies 6 (2013) 784–806. [4] J. Li, X. Zhuang, X. Querol, O. Font, N. Moreno, J. Zhou, Environmental geochemistry of the feed coals and their combustion by-products from two coal-fired power plants in Xinjiang Province, Northwest China, Fuel 95 (2012) 446–456. [5] Y. Gao, L. Ding, X. Li, W. Wang, Y. Xue, X. Zhu, H. Hu, G. Luo, I. Naruse, Z. Bai, H. Yao, Na & Ca removal from Zhundong coal by a novel CO2-water leaching method and the ashing behavior of the leached coal, Fuel 210 (2017) 8–14. [6] X.J. Wu, X. Zhang, K. Yan, N. Chen, J. Zhang, X. Xu, B.Q. Dai, J. Zhang, L. Zhang, Ash deposition and slagging behaviour of Chinese Xinjiang high-alkali coal in 3MWth pilot-scale combustion test, Fuel 181 (2016) 1191–1202. [7] X. Wang, Z. Xu, B. Wei, L. Zhang, H. Tan, T. Yang, H. Mikulčić, N. Duić, The ash deposition mechanism in boilers burning Zhundong coal with high contents of sodium and calcium: a study from ash evaporating to condensing, Appl. Therm. Eng. 80 (2015) 150–159. [8] B. Wei, X. Wang, H. Tan, L. Zhang, Y. Wang, Z. Wang, Effect of silicon–aluminum additives on ash fusion and ash mineral conversion of Xinjiang high-sodium coal, Fuel 181 (2016) 1224–1229. [9] F. Low, A. De Girolamo, B.-Q. Dai, L. Zhang, Emission of organically bound elements during the pyrolysis and char oxidation of lignites in air and oxyfuel combustion modes, Energy Fuel 28 (2014) 4167–4176. [10] Y. Zhang, H. Zhang, Gasification characteristics and sodium transformation behaviour of high-sodium Zhundong coal, Energy Fuel (2017), http://dx.doi.org/10. 1021/acs.energyfuels.6b03456. [11] M. Li, Z.X. Zhang, X. Wu, J. Fan, Experiment and mechanism study on the effect of

[19]

[20]

[21] [22]

[23] [24]

[25] [26]

[27]

43

kaolin on melting characteristics of Zhundong coal ash, Energy Fuel 30 (2016) 7763–7769. W. Willenborg, M. Müller, K. Hilpert, Alkali removal at about 1400 °C for the pressurized pulverized coal combustion combined cycle. 1. Thermodynamics and concept, Energy Fuel 20 (2006) 2593–2598. S. Guo, Y. Jiang, J. Li, Z. Yu, J. Zhao, Y. Fang, Correlations between coal compositions and sodium release during steam gasification of sodium – rich coals, Energy Fuel (2017), http://dx.doi.org/10.1021/acs.energyfuels.7b00674. L. Bencze, M. Ryś- Matejczuk, E. Yazhenskikh, M. Ziegner, M. Müller, Investigation of vaporization of alkali metals from solidified gasifier slags by knusen effusion mass spectrometry, Energy Fuel 30 (2016) 657–665. Y. Wang, Y. Xiang, D.X. Wang, C.Q. Dong, Y.P. Yang, X.B. Xiao, Q. Lu, Y. Zhao, Effect of sodium oxides in ash composition on ash fusibility, Energy Fuel 30 (2016) 1437–1444. R.B. Li, Q. Chen, H.X. Zhang, Detailed investigation on sodium (Na) species release and transformation mechanism during pyrolysis and char gasification of high – Na Zhundong coal, Energy Fuel (2017), http://dx.doi.org/10.1021/acs.energyfuels. 7b00410. G.R. Couch, Power Generation From Lignite, IEA Coal Research, (1989) (London). B.Q. Dai, X. Wu, A. De Girolamo, L. Zhang, Inhibition of lignite ash slagging and fouling upon the use of a silica-based additive in an industrial pulverised coal-fired boiler. Part 1: changes on the properties of ash deposits along the furnace, Fuel 139 (2015) 720–732. B.Q. Dai, X. Wu, A. De Girolamo, L. Zhang, Inhibition of lignite ash slagging and fouling upon the use of a silica-based additive in an industrial pulverised coal-fired boiler. Part 2: speciation of iron in ash deposits and separation of magnetite and ferrite, Fuel 139 (2015) 733–745. X. Wu, H.S. Ji, B.Q. Dai, L. Zhang, Xinjiang lignite ash slagging and flow under the weak reducing environment at 1300 °C – establishment of a new slag flow index for basic ashes, submitted to, Fuel Process. Technol. (July 2017). F. Low, L. Zhang, Microwave digestion for the quantification of inorganic elements in coal and coal ash using ICP-OES, Talanta 101 (2012) 346–352. P.J. van Eyk, P.J. Ashman, G.J. Nathan, Mechanism and kinetics of sodium release from brown coal char particles during combustion, Combus. Flame 158 (2011) 2512–2523. S. Niksa, Coal Combustion Modelling, IEA Coal Research, (1996) (London). P. Pednekar, D. Bhattacharyya, J.S. Kasule, R. Turton, R. Rengaswamy, Dynamic model of a slagging entrained-flow gasifier including models of slag transport, deposition, and slag layer, Ind. Eng. Chem. Res. 55 (2015) 279–292. M. Li, T. Utigard, M. Barati, Kinetics of Na2O and B2O3 loss from CaO-SiO2-Al2O3 slags, Metall. Mater. Trans. B Process Metall. Mater. Process. Sci. 46 (2015) 74–82. G. Zhang, M. Reinmöller, M. Klinger, B. Meyer, Ash melting behaviour and slag infiltration into alumina refractory simulating co – gasification of coal and biomass, Fuel 139 (2015) 457–465. I. Escobar, M. Müller, Alkali removal at about 1400 °C for the pressurized pulverized coal combustion combined cycle. 2. Sorbents and sorption mechanisms, Energy Fuel 21 (2007) 735–743.