Yang Meng,Peng Jiang,Yuxin Yan,Yuxin Pan,Xinyun Wu,Haitao Zhao,Nusrat Sharmin,Edward Lester,Tao Wu,5,Cheng Heng Pang,5,

1 Department of Chemical and Environmental Engineering,The University of Nottingham Ningbo China,Ningbo 315100,China

2 Ningbo New Materials Institute,The University of Nottingham,Ningbo 315042,China

3 Department of Mechanical Engineering,Massachusetts Institute of Technology,Cambridge 02139,USA

4 Department of Chemical and Environmental Engineering,The University of Nottingham,Nottingham NG7 2RD,UK

5 Key Laboratory for Carbonaceous Wastes Processing and Process Intensification Research of Zhejiang Province,The University of Nottingham Ningbo China,Ningbo 315100,China

Keywords:Oil shale Coal Image-based ash fusion test Co-gasification Mineral transformation

ABSTRACT This study investigates the potential of solid fuel blending as an effective approach to manipulate ash melting behaviour to alleviate ash-related problems during gasification,thus improving design,operability and safety.The ash fusion characteristics of Qinghai bituminous coal together with Fushun,Xinghua and Laoheishan oil shales(and their respective blends)were quantified using a novel picture analysis and graphing method,which incorporates conventional ash fusion study,dilatometry and sintering strength test,in a CO/CO2 atmosphere.This image-based characterisation method was used to monitor and quantify the complete melting behaviour of ash samples from room temperature to 1520 °C.The impacts of blending on compositional changes during heating were determined experimentally via X-ray diffraction and validated computationally using FactSage.Results showed that the melting point of Qinghai coal ash to be the lowest at 1116°C,but would increase up to 1208°C,1161°C and 1160°C with the addition of 30%–50% of Laoheishan,Fushun,and Xinghua oil shales,respectively.The formation of high-melting anorthite and mullite structures inhibits the formation of low-melting hercynite.However,the sintering point of Qinghai coal ash was seen to decrease from 1005 °C to 855 °C,834 °C,and 819 °C in the same blends due to the formation of low-melting aluminosilicate.Results also showed that blending directly influences the sintering strength during the various stages of melting.The key finding from this study is that it is possible to mitigate against the severe ash slagging and fouling issue arising from high calcium and iron coals by co-gasification with a high silica-alumina oil shale.Moreover,blending coals with oil shales can also modify the ash melting behaviour of fuels to create the optimal ash chemistry that meets the design specification of the gasifier,without adversely affecting thermal performance.

1.Introduction

Coal is an abundant and inexpensive energy resource in China and has accounted for approximately 70% of its primary energy supply in the last few decades [1].However,increasing environmental and health concerns have arisen from this large-scale consumption of coal.The emergence of clean coal conversion technologies (CCTs) is seen as a potential means to mitigate the environmental burden of coal.Some of the more established CCTs include advanced combustion and liquefaction,integrated gasification combined cycle (IGCC) and carbon capture,storage and utilization (CCUS) [2].Specifically,coal gasification promises a high potential due to its high carbon conversion efficiency,low emissions and valuable producer gas [3].In general,coal gasifiers can be categorized into three main categories based on the reactor type,namely fixed/moving bed,fluidized bed and entrained-flow gasifiers[4–6].Table 1 summarizes a list of major gasification systems together with their operating conditions and ash related issues.The large variation in design and operating conditionamongst these gasifiers result in substantial differences in ash/slag formation,deposition and removal methods [7].Therefore,understanding ash-related problems and the solution to these problems is crucial in effectively operating a gasification system.

Table 1 Major gasifier types in worldwide and summary of ash-related problems

During gasification processes,a large majority of organic matter is converted into syngas whilst inorganic minerals are turned into fused or semi-fused slag and fly ash[13].Ash fusibility is an important parameter,in both gasifier design and operation,as it plays a major role in influencing ash deposition and slag formation[14].To date,the development of experimental techniques for investigating ash fusibility has largely focused on methods such as ash fusion testing [15],thermo-mechanical analysis [16] and heating stage microscopy [17].Amongst these,the conventional ash fusion test is the most commonly used approach to understand ash fusibility as it simplifies the ash melting and slag formation process into four characteristic ash fusion temperatures (AFTs),i.e.deformation temperature (DT),softening temperature (ST),hemispherical temperature (HT) and flow temperature (FT) [18] as described below:

DT:The temperature at which the first sign of rounding of the pellet edges due to melting

ST:The temperature at which the ash softens and becomes plastic;

HT:The temperature at which the ash pellet forms a hemispherical droplet where the height of the pellet is half of the width;

FT:The temperature at which the ash is spread across the sample holder in a thin film where the height the film is one third the height of the pellet at hemispherical temperature;

However,the conventional ash fusion methods are heavily reliant on visual observations rather than physical measurements,thus resulting in low repeatability and reproducibility.Moreover,the lack of information between characteristic temperatures could pose a challenge in accurately predicting slag formation.Recognizing the limitations of conventional ash fusion test,Pang and coworkers [19] have developed an advanced ash fusion test based on computational imaging,known as the picture analysis and graphing (PAnG) method.The PAnG method is an automated ash characterisation method that incorporates conventional ash fusion test,dilatometry and sintering strength test by means of image analysis.It continuously studies the physical shape change of ash pellets at every degree interval from room temperature to 1520°C.For comparison purposes,three characteristic temperatures can be determined from the ash profiles including the sintering point (SP),expanding point (EP) and excessive melting point(EMP) as described below [19]:

SP:The temperature at which sintering starts,i.e.the onset of Phase 2 on the ash profile;

EP:The temperature at which the ash pellet starts to expand and increase in height,i.e.the onset of Phase 3 on the ash profile;

EMP:The temperature at which the ash pellet becomes highly fluid and excessive bubbling starts,i.e.the onset of Phase 4 on the ash profile.

The proposed method produces characteristic ash profiles that reflect the continuous monitoring,analysis and interpretation of ash melting behaviour from room temperature to above 1500 °C.The characteristic ash profiles are unique to each sample and thus form the‘fingerprint’for solid fuels[19].In addition to determining the key ash fusion temperatures,the more subjective and reproducible PAnG method also allows a detailed investigation into the events leading to each ash fusion temperature,as reflected in characteristically different types of ash profiles.However,the potential of PAnG method in characterizing ash fusion in gasifiers has not been fully explored,yet.

The ash properties,including ash fusibility,slagging/fouling tendency and slag viscosity of most of the coals in China are unlikely to directly meet the optimum requirements for a large variety of gasifiers.Thus,blending coal with other fuels is seen as a straightforward,yet effective,approach to modify ash fusion behaviour and also to expand the varieties of gasification feedstock[20].Oil shale,a combustible sedimentary rock,has great potential for exploration and utilization due to its abundant reserves in China[21,22].However,large-scale utilization of oil shale is limited due to its high ash content,low heating value and varied AFTs,in particular the silica-alumina rich oil shale [23–25].The addition of oil shale could avoid expensive refractory agents whilst improving overall gasifier operation [26].However,the variable mineral composition of fuels can lead to complex melting behaviours and detrimental inconsistency in operation.Therefore,a more detailed investigation on the transformation of mineral matters and the ash behaviour of coal and oil shale at elevated temperatures is needed in order to effectively explore the feasibility of blending fuels for specific gasification processes.

Qinghai bituminous coal(QH),which has high iron and calcium contents,has been reported to be problematic during gasification processes [27,28].This is largely because ashes high in Fe2O3and CaO contain eutectics which could enhance slag formation and strongly affect ash fusion behaviour [29,30].Moreover,Fe-bearing minerals in ash are known to experience complex mineral transformation and the AFTs of Fe-rich ash could vary considerably in different atmospheres due to melting points of iron of varying oxidation states,e.g.Fe(1538°C)[31],FeO(1369°C)[31]and Fe2O3(1565°C)[32].In an oxidising atmosphere,Fe-bearing species lead to high melting behaviour due to the increased Fe2O3content at elevated temperatures.On the other hand,a weak reducing atmosphere results in a lower melting point for Fe-rich ash as the iron oxides are reduced to FeO and form low melting-point eutectic mixtures like ferrous sulfide oxide in sulfur-rich ash[29,31,33–35].Similarly,CaO is commonly used as a fluxing agent to decrease the AFTs of high-melting fuels [36].Li and co-workers [29] have combined experimental studies with thermodynamic simulations to investigate the effects of CaO on petroleum coke ash fusibility.The results showed that the AFTs decreased with increasing CaO content due to the formation of low-melting eutectics and the decomposition of high-melting minerals.However,Liu and co-workers [37] further reported that excessive CaO could exist in the form of CaO monomer and eventually increase the AFTs.Thus,the presence and/or absence of certain mineral groups have the potential to shift the system towards or away from optimal operation.

From a process point of view,ash fusion plays a crucial role in influencing gasification processes and,thus,a comprehensive understanding of ash fusion characteristics is essential[7,20,38,39].Coal and oil shale have different ash composition which,upon blending at the right ratio,may complement ash chemistry and lead to desirable ash fusion temperatures [26,40].Previous workers [41] have investigated the effects of co-firing oil shale semi-coke with Zhundong lignite on morphological and mineralogical characteristics of ash deposits in a lab-scale circulated fluidized bed reactor.It was reported that high SiO2and Al2O3content in semi-coke could effectively capture the sodium species in Zhundong lignite via the formation of sodium aluminosilicates.Similarly,Lu and co-workers [26] also found that the sodium in Zhundong coal was transformed from water-soluble phase to aluminosilicate phase during co-combustion with oil shale due to the effects of SiO2and Al2O3.The results also showed that the sodium retention rate and FT of blended ash increased with the addition of oil shale,which implies that the slag properties could be moderated with different blending ratios.Whilst previous works mainly focused on the addition of oil shale addition in combustion systems,there is a need to further understand the effects of oil shale addition on coal melting behaviours under gasification conditions.

This study aims to design and manipulate ash fusion behaviour of solid fuels,via blending of coal and oil shale,to fit specific gasification systems,but not at the expense of calorific values.The ash fusion behaviours of three high-silica-alumina Chinese oil shale samples,QH coal and their respective blends at various blending ratios are investigated under a weak reducing atmosphere.The PAnG (picture analysis and graphing) advanced ash fusion test published elsewhere [19,42] is adopted to determine the characteristic ash fusion profiles of all samples.The mineral transformation in each ash sample is analysed experimentally via XRD at four specific key temperatures,whilst the eutectic processes are thermodynamically examined using FactSage 6.3 software.This paper presents valuable information about how blending can be used effectively in co-gasification technologies,particularly whilst using coal and oil shale.The proposed image-based characterization method is a potential tool for evaluation and prediction of ashrelated problems of fuels in specific gasification systems with known temperature profile.

2.Materials and Methods

2.1.Raw material and ash preparation

The oil shale samples investigated in this study were collected from Fushun (FS),Xinghua (XH) and Laoheishan (LHS) deposits in China,whilst Qinghai (QH) bituminous coal sample was collected from a domestic power plant.Following the ISO 1171:2010 standards[43],the as-received oil shale and coal samples were pulverized using a DY500 lab-scale pulverizer (Xinyuan DY500,China)and the particles sieved to pass a standard 212 μm mesh [44].Proximate analysis was conducted using a thermogravimetric analyser (Netzsch STA449F3,German) following ISO 17246:2010[45],whilst the ultimate analysis was conducted using an elemental analyzer(Euro Vector EA3000,Italy) following ISO 17247:2013[46].The high heating value (HHV) of oil shale and coal samples was calculated from their respective proximate components using Eq.(1)as described in the works of Majumder and co-workers[47].Mineral composition and major mineral phases of oil shale,coal and their respective ashes were analysed using X-ray fluorescence analyser (Bruker S4-Explorer,German) and X-ray diffraction spectrometer (Bruker D8 Advance,German).

The ash samples in this study were prepared using a laboratory muffle furnace.Both oil shale and coal samples were ashed at(815 ± 15)°C following the ISO 1171:2010 standard [43].Briefly,the samples were heated up from room temperature to 500 °C at a heating rate of 10 °C﹒min-1and were kept isothermal at 500 °C for 1 h.This was followed by a 10 °C﹒min-1ramp to 815 °C before maintaining at constant temperature for 3 h.Thereafter,the prepared ash samples were ground to pass through a standard 74 μm sieve.

The blended ashes were prepared by mixing the ashes of QH coal with that of FS,XH,LHS,respectively,at two different blending ratios each,i.e.FS:QH=30:70 (denoted as FS30QH70),FS:QH=50:50 (denoted as FS50QH50),XH:QH=30:70 (denoted as XH30QH70),XH:QH=50:50 (denoted as XH50QH50),LHS:QH=30:70(denoted as LHS30QH70)and LHS:QH=50:50(denoted as LHS50QH50).

2.2.PAnG advanced ash fusion test

The PAnG method is an automated advanced ash fusion test that continuously monitors and characterizes the melting behaviour of ash from room temperature to above 1500 °C via computational imaging.As compared to conventional ash fusion tests,the PAnG method employs a unique set of preparation,characterization and analytical procedures,leading to a more automated,subjective and reproducible investigation.

2.2.1.Pelleting method

Cylindrical ash pellets were formed by compressing 0.75 g of ashes in a custom-made,10 mm diameter die.A standard laboratory tablet press was used to compress each sample following an identical compression profile with a 30 s holding time at 2000 kPa.The coherent cylindrical pellets were used for the PAnG method and sintering strength test.As for the ASTM method,0.5 g of ash were mixed with dextrin (binder) and shaped into pyramidal ash cones following the ASTM standard [48].

2.2.2.Image-based ash fusion temperature test

The PAnG ash fusion test was conducted using a SDAF2000d Ash Fusion Analyser (SDAF2000d,China) equipped with a black and white closed circuit digital camera (CCDC).In this study,the ash fusion test was carried out under a weak reducing atmosphere(also known as a carbon atmosphere)using the carbon seal method[37] to prevent complete oxidation.The cylindrical ash pellets were individually heated to 1520°C at a rate of 10°C﹒min-1,whilst the camera captured the images of pellets at 1 frame per degree increments.For each sample in every run,more than 1500 images were captured and analysed using a self-coded image analysis program developed by Pang and co-workers [19].The height of the pellet in each image (and hence each temperature) was continuously measured via image analysis.The edge detection method used was developed from the gradient method[42].An error tolerance level of 15 in grayscale was set to accommodate noises in low-resolution images.Fig.S1 shows example images of a pellet(at two different temperatures) that were automatically cropped into 160 × 160 pixels frames and converted into binary images with detected edges.The height measurements were carried out on the binary images and the data were used to plot the PAnG characteristic ash profiles consisting of more than 1500 points.

2.3.Sintering strength test

The cylindrical ash pellets were sintered in a closed tube furnace system at key temperatures of 850°C,1000°C,1150°C and 1300°C,and subsequently tested for sintering strength via compression using the MTS Electronic Universal Testing Machine (MTS E45.105,China).The ash pellets were gently heated and sintered at a rate of 7.5°C﹒min-1in order to avoid sudden thermal expansion which could detrimentally affect pellet strength.Upon sintering at the designated temperatures,the pellets were removed and quenched in an ice-water bath to preserve the lattice structure created at high temperatures[49].The sintered ash pellets were individually compressed at a constant loading rate of 2 mm﹒min-1until failure.The point of failure is the peak on the stress–strain profile prior to the drop rate reaching 80% per minute.Three ash pellets were prepared,sintered and strength-tested for each sample.

2.4.FactSageTM thermodynamic modelling and slagging/fouling tendency

Thermodynamic modelling was carried out using the FactSage(version 6.3) software package.The FToxid and FactPS databases were selected to represent the phase formation,combination and transformation of metal oxides[39,50,51].The FToxid calculations were carried out under a weak reducing atmosphere (CO/CO2=6:4),and the changes in the solid and liquid phases from 800 °C to 1520 °C were calculated based on Gibbs’ energy minimization principal.

The slagging and fouling property of an ash sample would vary in different atmospheres as it involves complex eutectic processes which are dependent on its chemical composition.However,empirical indices are able to predict the slagging/fouling propensity of ash samples [52–55].

The empirical indices used in this study are described below:

Base to acid ratio:

where SiO2,Al2O3,TiO2,Fe2O3,CaO,MgO,Na2O and K2O represent the mass fraction of each component in the corresponding ash sample.

Each ash sample was evaluated using the RB/A,Fuand SRindices which could indicate the slagging/fouling tendency as severe/high,medium or low according to the classification published elsewhere[56,57].Samples with RB/A≤0.5 are known to have low slagging propensity,whilst those with 0.5 1 have high tendency for slagging.Similarly,Fu≤0.6 indicates a low fouling inclination,whilst 0.6 40.Additionally,slag viscosity index(SR)is a common parameter to describe the slagging inclination.Samples with SR>72 have a low slagging tendency,whilst those with 65

3.Results and discussion

3.1.Fuel and ash sample characteristics

The proximate and ultimate analyses of three oil shale and coal samples are summarized in Table 2.QH bituminous coal is characterized by its high volatile content (28.68%),medium fixed carbon content (57.71 wt%) and low ash content (11.38 wt%).In comparison with coal,the ash contents of oil shale samples are relatively higher(49.35 wt%–86.34 wt%),whilst the fixed carbon contents are relatively lower (1.30 wt%–15.75 wt%).The remarkable difference in the chemical properties of raw samples leads to distinctly varied high heating values (HHV),which range from 1.12 MJ﹒kg-1to 29.08 MJ﹒kg-1.Additionally,it is worth noting that oil shale samples have higher aromatic H/C ratios (1.04–5.54) compared to QH coal (0.63).Therefore,the addition of oil shale could increase the H/C ratio of blended feedstock which,in turn,could enhance the gasification process by improving the producer gas yield [58].The results given in Table 3 present the chemical composition of ash samples in terms of major elements on an oxide basis.It is apparent from Table 3 that oil shale ash samples are relatively richer in acidic oxides (SiO2and Al2O3),whereas QH coal ash is mainly comprised of basic oxides.Acidic oxides are known to improve the AFTs,whilst basic oxides are seen as fluxing agents to reduce the AFTs[59].More specifically,sodium and magnesium contents in QH coal ash are comparable with those in oil shale ash,whilst iron is the dominant element in QH coal ash,followed by calcium.Previous work[29,37,60]has demonstrated that the melting behaviour of solid fuels varies significantly with CaO and Fe2O3contents.Under a CO/CO2atmosphere,the iron could be present as either Fe2+or Fe3+,and the effects of iron valence on AFT have been reported by Mysen and co-workers[61].It was concluded that iron in form of Fe2+would lower the AFTs,whilst Fe3+has variable effects on AFTs.That being said,Shi and co-workers [60] reported that the CaO in blended ash could inhibit the transformation of Fe3+to Fe2+,and the Fe2+could be stabilised by aluminosilicates in molten slag.Consequently,CaO and Fe2O3in QH coal may lead to ash slagging/fouling problems due to the eutectic processes associated with these two elements [29,37].

The characteristics of the ash components can also be described by random network theory derived from glass melts[59],where a[SiO4]4-tetrahedron network is interrupted by alkali and alkaline earth metal (AEEM).The stability of this network highly depends on the network formers to network modifiers ratio,which can be expressed as a base(oxygen donor)to acid(oxygen accepter)ratio(RB/A) according to Eq.(2).Thus,the RB/Aratio is used to evaluate the ash fusion temperature and slagging/fouling potential.As shown in Table 3,there is a distinct difference in RB/Abetween coal and oil shale samples:QH coal has a higher slagging/fouling tendency (RB/A>1),whilst all oil shale samples exhibit relatively stable ash networks (RB/A<0.5) thus implying lower ash melting potential and,consequently,higher AFTs [62].Owing to the high silicon content and low alkali and alkaline earth metal(AEEM)content,the slagging viscosity indices of all oil shale samples are morethan twice as large as QH coal.Although high slagging viscosity is associated with low slagging tendency,the poor slag mobility is problematic in certain gasification systems such as the entrained-flow gasifier [7].Additionally,the fouling tendencies(Fu) of all tested samples are consistent with the other two calculated indices (RB/Aand SR),except for XH oil shale as it is comparatively richer in AEEMs (Na,Mg and Ca).The alkali species could react with other minerals to form eutectics which,in turn,facilitate the decomposition of high-melting minerals [29,60].As shown in Fig.1,the addition of oil shale could shift the calculated indices towards the medium range.Hence,oil shale/coal blends could be a feasible method for modifying the AFTs of solid fuels to meet different requirements during co-gasification.

Table 2 Proximate and ultimate analysis of the employed fuels

3.2.Ash fusion characteristic profiles

The PAnG method was used to evaluate the ash fusion behaviour of coal,oil shale and their blends.A characteristic ash profile was generated for each ash sample which records the shape change of pellets over a wide range of temperature,i.e.relative height of pellets plotted against temperature.Fig.S2 shows an example of the PAnG characteristic ash profile of FS oil shale.As apparent from the Fig.S2,a typical ash profile has four distinct regions,including dust stage,sintering stage,expanding stage and excessive melting/bubbling stage,whilst three characteristic temperatures could be determined from the ash profile,including the sintering point(SP),expanding point (EP) and excessive melting point (EMP).The variation in geometry of ash pellet (i.e.relative height) measured via image analysis in this work is consistent with previously published works [19,42,63,64].

The ash profiles of oil shale,coal and oil shale/coal blends are shown in Fig.2.It can be seen that every ash profile is unique and no two profiles are identical.This is because the ash samples experience different mineral transformation as a result of different ash composition.In general,only FS and LHS have ash profiles clearly showing the expanding stage and excessive melting stage.However,these stages are lacking in the ash profiles of XH and QH coal which had seen a sudden drop upon heating to above 1000°C.Compared with FS and LHS,QH and XH ash samples have relatively low slag viscosity (Table 2),and thus the ash pellets showed a limited resistance to gravity-driven slumping or toppling[63].Moreover,the EMPs of FS and LHS are relatively higher at 1475 °C and 1277 °C,respectively,compared to XH and QH coal at 1180 °C and 1116 °C,respectively.The results are consistent with prediction by empirical indices shown in Fig.1.However,the sintering points of FS (889 °C) and LHS (937 °C) are comparatively low compared to other samples.This is due to the formation of low-temperature eutectic components (e.g.sodium disilicates)from the interaction between alkali species and quartz in FS and LHS oil shale,thus leading to early sign of sintering [65,66].

Table 3 Chemical composition of ash samples/wt%

Fig.3 shows the effects of blending on ash fusion temperatures obtained from the PAnG method.In general,the expanding and excessive melting points for all samples investigated would gradually increase with the increase in blending ratio(i.e.the increase in oil shale percentage).However,the rate of increase would differ in different samples(e.g.the blending of FS with QH has seen a more significant increase in excessive melting point above the blending ratio of 0.5 as compared to other samples).The increasing trend has been observed because the expanding and excessive melting points of oil shale samples are higher than those of QH coal,in this study.On the contrary,the sintering points of the tested oil shale samples are lower than that of QH coal.This has led to an initial decrease in sintering points of blended samples as observed in Fig.3.Specifically,the decrease is due to three potential transformation mechanisms:(1) sodium species in QH coal react with silica in oil shale ash to form sodium disilicates with a relatively low melting point of 850 °C [65,66];(2) partially oxidized iron species(Fe-S-O/Fe-oxide) together with calcium oxide interact with high temperature clay minerals in oil shale ash to form molten iron and calcium aluminosilicate [67];and (3) sodium-bearing species and calcium oxide in QH coal interact with clay minerals in oil shale ash to facilitate the formation of low-melting aluminosilicate,e.g.sodium calcium aluminosilicate with a melting point of900 °C [40].These mechanisms are discussed in more details in Section 3.4.

Fig.1.Calculated indices under different oil shale and coal blending ratios:(a)Base-to-acid ratio (RB/A);(b) Fouling index (Fu) and (c) Slag viscosity index (SR).

However,further increase in blending oil shale(above blending ratio of 0.3 in this case)would lead to gradual increase in shrinking points,similar to the increasing trend in expanding and excessive melting points.Compared with QH coal,oil shale has higher content of SiO2and Al2O3,which could promote the interactions with calcium and iron to produce molten slag phase [68].This would eventually facilitate the sintering of blended ash,leading to an early sign of contraction.On the other hand,the remaining SiO2and Al2O3may exist in the form of anorthite and mullite which could enhance the EMPs.Consequently,blending coal with oil shale would lead to lower shrinking points whilst increasing expanding and excessive melting points.This implies that the sintering stage of blended samples is significantly larger compared to unblended samples of either coal or oil shale.This effect is reflected in the larger vertical distance (wider temperature range) between shrinking point and expanding point in Fig.3.The wider sintering temperature range is favourable to gasification as it could alleviate the adverse effect on gasification rate caused by the aggregation and spread of molten minerals on char surfaces [69].

Table 4 summarizes the ash fusion temperatures of all blended and unblended samples determined from both the ASTM and PAnG methods.The SP is of key importance in ash fusion studies as it signifies the first sign of melting [70].However,as apparent from Table 4,the SPs of all samples are significantly lower compared to the DT,which is the first melting event detected via the ASTM method.This is because the sintering of ashes (indicated by the SP via PAnG method),and thus shrinking of pellets,may have occurred considerably at lower temperatures before the rounding of ash pellet tips occur (detected via ASTM method).Hence,the SP of PAnG method is more likely to indicate the first sign of melting compared to other methods [19].

Similarly,the excessive melting point determined via the PAnG method is significantly lower than the flow temperature of ASTM method.Both excessive melting point and flow temperature indicate the formation of flowing slag,which is the final stage of ash deposition/melting.Ash deposits above this characteristic temperature are in molten state with progressively lower viscosity.However,compared to the EMP of PAnG method which is governed by viscosity,the detection of flow temperature in the ASTM method is also dependent on the wetting propensity between the molten ash and sample holder(in order to achieve the 1/3r criteria).Hence,the EMP of PAnG method is able to accurately detect the onset of flowing slag based on the change in viscosity and the bubbling effect[71].

3.3.Sintering strength of ash pellets

The ash pellets were sintered at 850 °C,1000 °C,1150 °C and 1300 °C under a weak reducing atmosphere and the sintering strength of each pellet was determined and is presented in Fig.2 on the secondary Y-axis.However,the sintering strength for XH,QH,FS30QH70,XH30QH70 and XH50QH50 was not studied as these pellets have completely melted at 1300 °C.In general,the sintering strength of ash pellets would increase with the decrease in relative height,and vice versa.This is reflected as a negative correlation with the PAnG ash profile as see in Fig.2.Table 5 also shows that the sintering strength increased with temperature within the sintering stage.However,upon reaching the expanding point,the sintering strength decreased.This is because the ashes experience different melting events in different phases as reflected on the ash profile.

3.3.1.Region I– Dust phase

This is the initial horizontal section of the PAnG ash profile where the ashes have not undergone melting.In this phase,the ash particles remain as loose particles and the corresponding ash pellets have relatively weak sintering strength,as shown by the ash pellets of QH and XH sintered at 850 °C and 1000 °C (still in dust phase at these two temperatures).

3.3.2.Region II– Sintering phase

Fig.2.Characteristic profiles (obtained from PAnG method) of every sample and mean sintering strength (labelled as *) at 850 °C,1000 °C,1150 °C and 1300 °C (obtained from sintering strength test).(a) FS;(b) XH;(c) LHS;(d) QH;(e) FS30QH70;(f) FS50QH50;(g) XH30QH70;(h) XH50QH50;(i) LHS30QH70;(j) LHS50QH50.

This is the first depression after Region I where the relative height of ash pellets starts to decrease due to the sintering effect.Sintering is the softening of ash particles below melting point.The minerals become sticky within this phase thus enhancing the interconnection between ash particles which,in turn,facilitates coalescence.Similar sintering behaviour has also been reported by Valenze and co-workers [72] whilst investigating the sintering of refractory tiles.More specifically,the ash chemistry of blended ash shown in Table 2 suggests that the molten slag in initial stage(<1100 °C) is mainly comprised of low-melting eutectics,such as sodium disilicates,hercynite and sodium calcium aluminosilicates as discussed in Section 3.2.As temperature increases,the minerals in blended ash (e.g.anorthite and gehlenite) start to melt and dissolve into molten slag,which subsequently enhances the contraction [63].The consolidation and densification of the ash pellet would reduce the porosity,thus leading to an increase in sintering strength[19].This may explain why most of the pellets in Region II(approximately 1150°C)show a relatively high sintering strength,as listed in Table 5.

Table 4 AFTs of blended and unblended samples determined from both the ASTM and PAnG method

Table 5 Sintering strength at four characteristic temperatures/kPa

3.3.3.Region III– Expanding phase

This region is found as the increase in relative height right after Region II.As shown in Fig.2,a reduction in sintering strength was recorded for most of the ash pellets in this region(e.g.the strength of FS reduced 46%in the expanding phase compared to that in sintering phase).The weakening in structural integrity is caused by increase in pore size and void degree as the melted ash traps the gases being released,as demonstrated in our previous study [19].This implies that if ash deposition occurs at temperatures within the expanding phase in a gasifier,the deposited ash would have a lower sintering strength and be much easier to remove.

3.3.4.Region IV– Flowing slag/excessive bubbling phase

Ash pellets in the flowing slag/excessive bubbling stage are commonly associated with a rapid decrease in height with or without significant oscillations in the plot.This phase would occur at relatively high temperatures and is usually the final stage of melting.The ashes are completely melted in this phase and would form low viscosity molten slag.The viscosity of the molten ash becomes progressively lower with increasing temperature.This is indicated by the active formation of bubbling activities,and thus reflected as oscillations in the ash profile.Such oscillations are usually accompanied by an overall and abrupt decrease in relative height due to the inability of the highly fluid ash to withstand its structure.Depending on the gasification system,some gasifiers require the ashes to be in this phase.

3.4.Mineral transformation behaviour and thermodynamic equilibrium modelling.

3.4.1.XRD analysis

XRD patterns of coal,oil shale and oil shale/coal blends at elevated temperatures under a weak reducing atmosphere are shown in Fig.4.The results show that the main minerals in oil shale samples are clay and quartz along with several iron-bearing minerals.At 850 °C,the dominating minerals are hematite (Fe2O3),quartz(SiO2),and anorthite (CaAl2Si2O8).Upon heating from 1000 °C to 1300 °C,the diffraction peaks of hematite disappeared,thus indicating a reduction from hermatite (Fe2O3) to wustite (FeO),whilst mullite (Al6Si2O13) was generated through the reaction below:

On the other hand,the appearance of cristobalite(SiO2)at 1300°C implies a transformation from amorphous to crystalline silica.Hence,the presence of high-melting point minerals in FS ash samples has led to the extremely high ash fusion temperature.A similar mineral transformation pathway for LHS ash sample is shown in Fig.4(c).In contrary to FS and LHS ash samples,the mineral composition in XH ash sample remained unchanged from 850 °C to 1150 °C,and no peak was observed at 1300 °C.As discussed in Section 3.3,XH ash sample was completely melted at 1300 °C,which suggests the formation of low-melting eutectic mixture at higher temperatures.

As shown in Fig.4(d),QH ash sample is rich in anhydrite(CaSO4),quartz,and hematite at 850 °C.The anhydrite diffraction peaks gradually weakened as the temperature increased from 850 °C to 1000 °C,whilst the diffraction peaks of anorthite and hedenbergite (CaFeSi2O6) appeared.This phenomenon has also been reported by Wang and co-workers [73] whilst investigating the effects of Al2O3/CaO on mineral transformation of Ningdong Coal.As the temperature further increased to 1150 °C,the peaks of anorthite and hedenbergite completely disappeared,whilst the intensity of gehlenite (Ca2Al2SiO7) diffraction became significant.The excessive CaO in QH coal ash (>20 wt%) would continuously consume anorthite to produce gehlenite as shown in Eq.(9).Hence,the potential mineral transformation pathway is summarized as below:

Fig.4(e)–(j)show the mineral transformation of the ashes of QH coal blended with FS,XH and LHS,respectively at two blending ratios.With the addition of FS and LHS oil shale,the ash fusion temperature increased with the increase in oil shale percentage.As seen in Fig.4(e) and (i),the magnetite (Fe3O4) in blended ash is likely to have partially oxidized to form hematite at 1000 °C when the blending ratio of oil shale and coal was 30:70.Interestingly,the addition of XH to QH coal has seen a limited enhancement on ash fusion temperature,as observed in Table 4.This is largely because the high-melting gehlenite formed in Eq.(9) has reacted with wustite to form the low-melting hercynite (FeAl2O4,melting point of 808 °C) in the reduction atmosphere as shown in Eq.(10).Moreover,the addition of oil shale would also lower the CaO/Fe2O3ratio compared to raw coal,which could promote the formation of FeO [60].Consequently,the yield of hercynite is further increased via the interaction between FeO and Al2O3,as shown in Eq.(11).

3.4.2.Thermodynamic modelling

Figs.5 and 6 illustrate the change in ash composition and slag percentage of coal,oil shale and their blends with increasing temperature.In general,the simulated results from FactSage are consistent with the XRD analysis discussed in Section 3.4.1.At a relatively low temperature of approximately 850°C(with minimal melting),oil shale ash is mainly composed of quartz (SiO2),albite (NaAlSi3O8),corderite (Mg2Al4Si5O18) and ferro-corderite(Fe2Al4Si5O18),whilst the QH coal ash is comprised of pyrite(FeS),ferrogehlenite(Ca2FeSi2O7),akermanite(Ca2MgSi2O7),albite(NaAlSi3O8)and hercynite(FeAl2O4).As temperature increases,the existing minerals undergo transformation and re-organisation to form new minerals,thus affecting the ash fusion behaviour[28,69].However,the degree of mineral transformation is highly influenced by the original ash composition,particularly the AAEM species,which is widely varied amongst different species of fuel.This has been observed in this particular set of samples where,upon heating,the QH coal would characteristically form gehlenite and hercynite,whilst oil shale would mainly produce anorthite with mullite found only in FS and LHS oil shale samples.Similar results have also been reported by previous workers[27,39]whilst investigating mineral transformation behaviours of coal and coal blends under gasification condition.However,high CaO and FeO contents could lead to low melting eutectic in QH coal ash as observed in Fig.5(d).This is because CaO reacts with anorthite to produce gehlenite (Eq.(9)) [74],whilst the hercyinite is produced from the reaction between gehlnite and wustite (Eq.(10)) [27,74].

The formation and presence of certain mineral groups have crucial influence on the formation process of liquid slag,i.e.melting.As listed in Table 4,the excessive melting point of QH coal ash is relatively lower compared to oil shale samples.This is attributed to its high calcium (more than 21 wt%) and iron (more than 27 wt%) content.Calcium (CaO) is known to react with SiO2and Al2O3to form anorthite (CaAl2Si2O8) which,upon further reaction with remaining CaO,will form gehlenite.The formation of gehlenite would competitively consume the SiO2and Al2O3thus leading to reduced anorthite and mullite.At the same time,strong interactions between iron,Al2O3and gehlenite would facilitate the formation of hercynite (FeAl2O4),a low melting mineral.A similar observation was reported by Shen and co-workers [74] (whilst studying blended coals) and Li and co-workers [27] (whilst investigating coal and industrial sludge),where Fe-rich samples would lead to the formation of hercynite thus reducing overall ash fusion temperatures.Hence,the enhanced formation of hercynite and the reduced formation of anorthite and mullite have led to a low EMP observed in Ca-and Fe-rich QH coal.

Fig.4.XRD patterns of oil shale,coal and oil shale/coal blends at different temperature.(a) FS;(b) XH;(c) LHS;(d) QH;(e) FS30QH70;(f) FS50QH50;(g) XH30QH70;(h)XH50QH50;(i) LHS30QH70 (j) LHS50QH50.A-Anhydrite;Al-Albite;Ar-anorthite;Cr-Cristobalite;H-Hematite;He-Hedenbergite;M-Magnetite;Mu-Mullite;Q-Quartz;KKaolinite;C-Cordierite.

Fig.5.Calculated slag compositions at equilibrium stage:(a) FS;(b) XH;(c) LHS;(d) QH.

Fig.6.Calculated slag compositions at equilibrium stage:(a) FS30QH70;(b) FS50QH70;(c) XH30QH70;(d) XH50QH50;(e) LHS30QH70;(f) LHS50QH50.

Conversely,oil shale has a significantly different set of ash chemistry which,upon blending with coal,is able to modify the overall ash fusion composition and,hence,ash fusion behaviour.The addition of high silica-alumina oil shale samples (FS,LHS and XH,respectively) would significantly increase the mass fraction of SiO2and Al2O3whilst reducing the concentration of FeO and CaO.This would lead to a substantial increase in the formation of anorthite,as seen in Fig.6.That being said,the mass fraction of anorthite first increased (900–950 °C) before leveling off (950–1150 °C) and followed by an eventual decrease (>1150 °C).Such a trend in anorthite has also been reported by Zhang and coworkers [28] whilst investigating the mineral transformation of gasified semi-char containing a SiO2-to-Al2O3mass ratio of 2.79(note that the samples tested in this study have ratios between 2 to 3).Moreover,the high silica and alumina content in oil shale samples have profound inhibiting effect on the formation of hercynite [27,74],thus explaining the lack of low-melting hercynite in blended ash samples,as observed in Fig.6.This has led to an increase in EMPs of all blended ash samples as compared to QH coal,as shown in Fig.3.However,above the EMP blended samples containing XH do not follow the general trend.As apparent in Fig.6(d),XH50QH50 would form complete liquid slag at 1200°C,which is lower than that of unblended samples(i.e.XH at 1280°C and QH at 1240 °C).Although not severely huge,such a drop in melting temperature is not common in blended samples and this is largely because XH ash is rich in AAEM species.The silica and alumina in oil shale ash are competitively consumed by the AAEM species to produce low-melting aluminosilicate which contributes to the eutectic processes[74].This is reflected in the rapid increase in liquid slag during the sintering stage (as shown in Fig.6(c) and (d))and the lowering of sintering point of blended ash (XH-QH) compared to unblended coal.Therefore,blending of oil shale and coal has the potential to adjust the overall ash fusion temperatures by adjusting the ash chemistry.

3.5.Application of PAnG method in ash characterisation for gasification systems

It is apparent from Table 1 that each gasifier has its own specifications for feed,operating and ash conditions.All of these factors will have to be taken into considerations during fuel characterisation and selection,particularly the highly varied ash behaviour.Hence,a complete understanding of the ash melting event across a wide temperature range (rather than certain fixed temperatures and conditions) is essential.

One of the most relevant examples is the use of indices and its generic interpretation.In this particular study,the Rb/aand Fuindices have predicted the QH coal ash to have severely high slagging and fouling tendencies.Whilst this may be true for a particular system,such indication may not be applicable to other systems.On the other hand,in addition to indicating the slagging and fouling tendencies,the PAnG method analyses the complete ash behaviour of the samples from room temperature to 1520 °C,and identifies ‘events’ at specific points across temperature range.

For instance,if the QH coal is gasified in a system operating at temperatures within its Region I (<1005 °C for QH coal),the QH coal ash will remain as loose particles.If the operating temperature is slightly higher and within the initial stage of Region II,the ash deposits will start shrinking leading to enhanced bed permeability and slag mobility,particularly in BGL and fluidized bed gasifiers.However,if the operating temperatures are within the later part of Region II,i.e.temperatures approaching expanding point,severe coalescence and densification may occur resulting in blockage of slag tap nozzle [75] and defluidization [9].As the operating temperatures reach Region III,porous slag deposits are formed with low sintering strength and are easily removed.However,low strengths may increase the amount of fly ash and lead to blockages in downstream units.In region IV,the ashes form molten liquid slag,which is desirable in some gasifiers.For example,the entrained-flow gasifiers are most optimum when the ashes are in molten slag and with a viscosity of 2.5–25 Pa﹒s [76].Hence,based on the PAnG ash characteristic profile,it is possible to identify the most optimum operating temperature for a specific fuel,or to identify the most optimum fuel for a specific operating temperature.

It is worth noting that not all gasification technologies favour high ash fusion temperatures,e.g.the entrained-flow TEXACO-GE Energy gasifier which favours operations in slagging condition.In addition,blending coal with oil shale may not necessarily lead to higher ash fusion temperatures.However,the blending of fuels does provide a practical strategy to manipulate the ash fusion temperatures to meet specific requirements of different gasification systems.Moreover,the PAnG method,which produces characteristic ash profiles based on continuous monitoring of ash melting properties,allows straightforward and accurate identification of fuels with the right kind of ash behaviour.

4.Conclusions

In this study,the melting characteristics of QH coal,FS,XH,LHS oil shales and their blends were investigated under a weak reducing atmosphere (CO/CO2).The PAnG advanced ash fusion test was used to monitor and quantify the complete melting behaviour of ash samples from ambient to 1520 °C.Mineral transformations were also determined experimentally via XRD and modelled via a FactSage thermodynamic simulation.Compared to oil shale samples,the EMP of QH coal ash is lower due to its high calcium and iron content which facilitate the transition from anorthite to gehlenite to form low-melting hercynite.However,the EMP of QH coal increased from 1116 °C up to 1208 °C,1161 °C and 1160°C with the blending of 30%–50%LHS,FS and XH oil shales,respectively.The addition of high silica-alumina oil shale promotes the formation of high melting anorthite and mullite.Conversely,the SP of QH coal ash decreased from 1005 °C to 855 °C,834 °C,and 819 °C in the same blends,due to the formation of low-melting eutectic components created by the interaction of alkali species in the QH coal with SiO2/Al2O3in the oil shales.

The sintering strength of ash pellets was examined at 850 °C,1000 °C,1150 °C and 1300 °C,respectively.The sintering strength of ash pellets varied with temperature correlating with the trend seen in the ash fusion profiles.During the sintering phase (Region II–850–1150°C),the sintering strength increased up to 300%with increasing temperature,followed by a decrease of up to 46%during the expanding phase (Region III– 1150–1300 °C).The ash fusion profiles are therefore useful when devising ash blending strategies.In effect,co-gasifying coal with oil shale significantly alters the mineral transformation and final composition of the ash,thus affecting the overall melting behaviour.Whilst a gasifier requires both EMP and SP to be within specific ranges to operate efficiently,this can create limitations with fuel selection.However,blending strategies can be determined via the PAnG test that can allow optimal co-gasification,but not at the expense of the thermal performance of the gasifier.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The authors gratefully express gratitude to all parties which have contributed towards the success of this project,both financially and technically,especially the S&T Innovation 2025 Major Special Programme (grant number 2018B10022) and the Ningbo Natural Science Foundation Programme (grant number 2018A610069) funded by the Ningbo Science and Technology Bureau,China,as well as the Industrial Technology Innovation and Industrialization of Science and Technology Project,China(grant number 2014A35001-2)and the UNNC FoSE Faculty Inspiration Grant,China.The Zhejiang Provincial Department of Science and Technology is also acknowledged for this research under its Provincial Key Laboratory Programme (2020E10018).

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2020.10.011.