Interactions between oxygen carriers used for chemical looping combustion and ash from brown coals

Fuel Processing Technology 147 (2016) 71–82

Contents lists available at ScienceDirect

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

Interactions between oxygen carriers used for chemical looping combustion and ash from brown coals Alexander Y. Ilyushechkin ⁎, Mark Kochanek, Seng Lim CSIRO, Australia

a r t i c l e

i n f o

Article history: Received 27 July 2015 Received in revised form 30 October 2015 Accepted 24 November 2015 Available online 3 December 2015

a b s t r a c t Iron-based oxides are considered as potential oxygen carriers in coal chemical looping combustion. However, their redox or oxidation behaviour can be affected by coal mineral matters. The interactions between oxygen carriers and coal ashes depend strongly on the type of ash, which varies significantly in Australian brown coals. We studied the interactions between two types of coal ash (silica-rich and iron and magnesium-rich) and two types of iron-based oxygen carriers (iron ore and industrial-grade ilmenite) via microstructural analysis of samples processed in reducing and oxidising atmospheres at 900 and 950 °C, respectively. Both types of ash reacted locally with iron-based oxides, forming alumina silicates, spinels, magnesioferrites, and iron silicate phases as solid solutions or through liquid phase formation. These interactions can affect the oxidation and reduction kinetics of oxygen carriers. Our results show that iron-rich ash improves the oxidation and reduction kinetics of iron ore. It does not affect the reduction kinetics of ilmenite, but significantly increases ilmenite oxidation time. Silica-rich ash decreases oxidation rates of iron ore, but has less of an effect on the kinetics of ilmenite processing. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Processing low-rank Australian brown coals by chemical looping combustion (CLC) allows coal to be used in a more environmentally friendly manner, because carbon dioxide (CO2) gas emissions can be easily captured. In CLC, oxygen is delivered for coal combustion by solid oxygen carriers (OCs): typically, metal oxides. Supplying oxygen without air allows CO2 to be captured easily because it not diluted by nitrogen. Mixed valency metal oxides can be used in CLC processes. During circulation between two interconnected fluidised beds, these oxides donate oxygen in the fuel reactor, and when oxidised, in the air reactor [1–2].To be effective, oxides must be highly reactive with the fuel, have high oxygen capacity and good mechanical strength, and be nontoxic. Industrial-grade, iron-based oxides such as iron ore and ilmenite (FeTiO3) are low-cost materials often considered as OCs in CLC [3–6], especially where loss of materials is expected during circulation and separation of ash out of the system. The reactivity of iron ore and ilmenite have been extensively studied and compared with other Fe2O3based OC [5–7]. Iron ore has been also tested in a commercial, coalderived CLC unit with Chinese bituminous coal [8]. In recent studies of the combustion of Victorian brown coals with various OCs in CLC, coals were co-processed with CuO [9–10], Mn2O3 [11], and laboratory and commercial Fe2O3 and NiO powders [11–13].

⁎ Corresponding author. E-mail address: [email protected] (A.Y. Ilyushechkin).

http://dx.doi.org/10.1016/j.fuproc.2015.11.019 0378-3820/© 2016 Elsevier B.V. All rights reserved.

These studies mainly focused on the reactivity of OCs in thermogravimetric analysers (TGAs) or fluidised bed reactors. OC particles may interact with coal ash in CLC fuel and air reactors. This can cause deactivation, agglomeration, or attrition of the OCs. Even with the possible separation of the ash and OC particles due to differences in density, some solid–solid contact in the reactor is still expected [14–15]. The effect of different coal ashes on OCs in CLC depends on the ash content and mineralogy, experimental conditions, and OC composition [14–20]. Some interactions have been observed between ash and Fe2O3 OCs [16], while ilmenite is most resistant to reactions with individual common ash minerals [14]. The composition of coal ashes may also affect redox and oxidation kinetics [17–19] and char conversion rates [20]. Increased gas conversion rates have been seen for iron-based OCs in the presence of Ca-rich and Fe-rich coal ashes [17]. However, silica in ashes of sub-bituminous coal and anthracite reacts with CuFe2O4 in a redox cycle, forming stable Fe2SiO4 iron silicate, which reduces the activity of OCs in CLC [19]. Australian brown coals from the Latrobe valley typically have a low ash content. As the composition of minerals varies greatly in coals from different mines, we expect that the possible interactions of coal ashes with OCs could also vary for different brown coals. For example, when interactions of brown coal minerals with CuO were investigated for single and multi- CLC cycles, a strong interaction was observed for one type of coal during the reduction cycle, but another type of coal ash did not react with CuO [9]. In the present work, we investigate and describe the interactions between iron-based OCs and two types of Australian brown coal – high

72

A.Y. Ilyushechkin et al. / Fuel Processing Technology 147 (2016) 71–82

Table 1 Sample identification (ID) and materials ratio used in this study. Oxygen carrier, OC

Brown coal ash 1 (Bca1) mixture

Brown coal ash 2 (Bca2) mixture

Sample ID

Oxygen carrier

Sample ID

Bc ash1: OC (wt:wt)

Sample ID

Bc ash2: OC (wt:wt)

OC1 OC2

Iron ore Ilmenite

OC1-Bca1 OC2-Bca1

1:5.7 1:4.4

OC1-Bca2 OC2-Bca2

1:2.7 1:2.1

iron and magnesium, and silica-rich – using different experimental techniques and thermodynamic modelling. We then discuss the mechanisms of possible reactions and their potential impact on oxidation and reduction kinetics. 2. Experimental 2.1. Materials preparation Industrial-grade iron ore (OC1) and ilmenite (OC2) were used as raw materials for OCs. Both minerals were crushed to obtain particle sizes of b250 μm, and heat treated (calcined) in air in a muffle furnace at 900 °C. Brown coal ashes were prepared by ashing the coal samples at 780 °C for 20 h in a muffle furnace. Ash and OCs were mixed and pelletised, followed by heat treatment in a tube furnace at 900 °C for 5 h under a neutral atmosphere (N2 flow) and with a graphite lining in order to provide reducing conditions. Subsamples were taken for analysis, and the remainder of the samples were placed in alumina crucibles and heat treated in air at 950 °C for 5 h to recover the OC. Sample IDs and mixing ratios are listed in Table 1. 2.2. Materials analysis Calcined ilmenite and iron ore, as well as their mixtures with brown coal ashes, were tested in a Thermogravimetric analysis (TGA) at 1 bar pressure in CO (at 900 °C) and in air (at 950 °C) to provide reducing and oxidising conditions, respectively. Before analysis, the materials were heated to the desired temperature in flushing nitrogen, and then the mass changes were recorded. The bulk composition of coal ashes and OC composition were determined by X-ray fluorescence (XRF) according to the ASTM D3174-12 standard. Samples processed in the tube furnace were analysed by Xray diffraction (XRD) and scanning electron microscopy (SEM) with electron probe microanalysis (EPMA). XRD analysis was performed on a Bruker D8 Advance Powder Diffractometer with Bragg–Brentano geometry at 40 kV, 30 mA, using Cu Kα radiation in conjunction with a graphite monochromator. Peaks and phases were determined using Bruker DiffracPlus Evaluation Software (Ver. 15.0.0). SEM/EPMA of the processed OC-ash mixtures was performed using JEOL 8200 and JEOL JXA8530F electron probes. While iron is present as Fe2 + and Fe3 + in samples processed at different atmospheres, EPMA was used to obtain information on the total iron concentration only; the oxidation states of iron in various phases were not measured. For ease of presentation of XRF and EPMA results, all of the iron was recalculated to the ferric oxidation state.

Table 2 Chemical composition (wt.%) of oxygen carriers (OC) and brown coal ash (Bca), determined by X-ray fluorescence.

OC1 OC2 Bca1 Bca2

Fe2O3

Al2O3

SiO2

Mn3O4

MgO

CaO

TiO2

Na2O

K2O

SO3

BaO

96 46.8 45.1 3.6

1.3 0.6 1.7 10.3

2.3 0.5 3.2 76.7

0 1.5 0.5 0

0 1 18.5 2.2

0 0 7.3 1.1

0 50.1 0.1 0

0 0 5.2 1.8

0 0 0.4 0.4

0 0 16.7 2.8

0 0 0.9 0.1

2.3. Thermodynamic modelling The thermodynamic phase equilibria package FactSage (version 6.4) was used to evaluate possible interactions between coal ash and OCs [21]. Databases used for calculations of compounds and solutions were FACT53 and FToxid. FactSage calculations were performed according to the scheme shown in Fig. 1, where outputs of calculations in coal combustion conditions (inorganic species only) are used as input for calculations in oxidising conditions. This approach differs from previously published thermodynamic studies on the reaction of coal mineral matter with OCs [14,19], in which only individual minerals were considered in the reactions with OCs, thus ignoring possible interactions between ash minerals.

3. Results 3.1. Characteristics of starting materials The chemical composition of OCs and brown coal ashes are listed in Table 2, and their phase compositions are shown in Fig. 2. According to the XRD results, calcined iron ore mainly contains an iron oxide in hematite form, with some aluminium and silica impurities. Iron ore sample OC1, processed in reducing conditions, consists of a mixture of magnetite (Fe3O4) and hematite (Fe2O3), with a very small fraction of wustite (FeO) (Fig. 2a). Only hematite was detected in the oxidised OC1 sample, as marked in Fig. 2b. Iron and titanium oxides are the main components of ilmenite (Table 2). Small impurities of silica, magnesium and aluminium oxides (~ 2 wt%) were identified by XRF only. XRD analysis revealed that ilmenite, FeTiO3, appears in OC2 after processing under reducing conditions, along with rutile (TiO2) and small amounts of pseudobrookite (Fe2TiO5) (Fig. 2c). Calcined (as starting material) or oxidised (following the redox processing) OC2 contains pseudobrookite, as rutile and small amounts of hematite (Fig. 2d). Fig. 3 shows backscattering images of a polished cross-section of a calcined iron ore sample (a, b) and ilmenite (c, d). Calcined iron ore, OC1, has areas that are more porous than other areas, while the microstructure of ilmenite looks uniform. The chemical compositions of the two brown coal ashes (Table 2) as well as their phase compositions (Fig. 4) are very different. The main content of brown coal ash 1 (Bca1) is magnesioferrite (MgFe2O4), and possibly magnetite (Fe(Mg)3O4) and periclase (MgO), the XRD peaks of which overlapped with MgFe2O4. Other elements exist in ash as

Fig. 1. Scheme of thermodynamic modelling calculations on oxygen carrier-brown coal ash (OC-Bca) mixtures. XRD = X-ray diffraction; XRF = X-ray fluorescence.

A.Y. Ilyushechkin et al. / Fuel Processing Technology 147 (2016) 71–82

73

Fig. 2. X-ray diffraction (XRD) spectra of oxygen carrier 1 (OC1) processed in (a) reducing and (b) oxidising atmospheres, and OC2 processed in (c) reducing and (d) oxidising atmospheres (d).

minor phases: sulphur is mainly represented by anhydrite (CaSO4), and sodium is represented by thenardite (Na2SO4) and/or aegirine (NaFe(SiO3)2). Bca2 is mainly crystalline silica (as quartz, SiO2) with some inclusions of aluminium, calcium, magnesium and sodium oxides; however, only anhydrite and possibly hematite and magnesium sulphate (MgSO4) are clearly identified. 3.2. Iron ore and coal ashes 3.2.1. Interactions between iron ore and iron- and magnesium-rich ash All major and minor crystalline phases detected by XRD analysis of samples processed in reducing and oxidation conditions are listed in Table 3. The crystalline phases observed in coal ashes and OCs are also

presented for comparison with the phases detected in coal ash-OC mixtures. The XRD results reveal that ash-iron ore mixtures have crystalline phases, mainly originating from the starting materials. Mixtures processed in reducing conditions have a magnetite from iron ore and magnesia-ferrite from Bca1. The phase compositions of the OC1-Bca1 mixture are uncertain, because some of the peaks from different iron and magnesium-rich phases overlap. Many more details were obtained by SEM analysis of the ash-iron ore mixtures. Fig. 5 shows cross-sections of the OC1-Bca1 mixture exposed in reduced conditions. Table 4 provides the composition of phases and solid solutions observed in Fig. 5. The composition of solid solutions in phases with fine (typically submicron) dispersed particles varies; the compositions listed in Table 4 and the following tables represent an

74

A.Y. Ilyushechkin et al. / Fuel Processing Technology 147 (2016) 71–82

Fig. 3. Microstructure of oxygen carriers (a, b) OC1 and (c, d) OC2.

average composition of three randomly selected 5-μm spots, corresponding to the indicated area (for example, areas (2), (3), (5), and (6) in Fig. 5c). The standard deviation in the concentration of major elements in these compositions is typically 20–30% of their value. Two types of reaction products were observed in the OC1-Bca1 sample processed in reducing conditions: (1) including only Fe and Mg in ash, and (2) including other elements (mainly Si and Al) from ash. In Fig. 5b, the reaction product is a sinter between fine particles of ash magnetite (Fe(Mg)3O4) and large particles of iron ore magnetite (Fe3O4) with magnesioferrites. In Fig. 5c, the reaction product is a

solid solution of alumina silicates and iron oxides, with the stoichiometry closed to iron cordierite type (2FeO·2Al2O3·5SiO2). Fig. 6 demonstrates a typical image of an oxidised OC1-Bca1 sample (Fig. 6a), previously exposed in reducing conditions, and three types of interactions (Fig. 6b–d) observed at higher magnification. The composition of phases and solid solutions indicated in Fig. 6 is given in Table 4. The first type of interaction is a solid–solid reaction between Naalumina silicates from ash and iron ore hematite, illustrated in Fig. 6b; the second is bonding of silica with hematite by iron silicate and liquid phase (Fig. 6c); and the third is bonding of Na-aluminosilicates with

Fig. 4. X-ray diffraction (XRD) spectra of brown coal ashes (a) Bca1 and (b) Bca2.

A.Y. Ilyushechkin et al. / Fuel Processing Technology 147 (2016) 71–82

75

Table 3 Phase compositions of brown coal ashes (Bca) and mixtures with oxygen carriers (OC) detected by X-ray diffraction. Bca1 Major: magnesioferrite (MgFe2O4) and/or magnetite (Fe(Mg)3O4), periclase (MgO) Minor: anhydrite(CaSO4), thenardite (Na2SO4), aegirine (NaFe(SiO3)2) Bca2 Major: silica (SiO2) Minor: hematite (Fe2O3), magnesium sulphate (MgSO4), anhydrite (CaSO4) OC1

OC1 + Bca1

OC1 + Bca2

OC2

OC2 + Bca1

OC2 + Bca2

Major: magnesioferrite (MgFe2O4) and/or magnetite (Fe(Mg)3O4)

Major: silica (SiO2), magnetite (Fe3O4)

Major: ilmenite (FeTiO3), rutile (TiO2)

Major: silica (SiO2), ilmenite (FeTiO3)

Minor: hematite (Fe2O3), wustite (FeO)

Minor: hematite (Fe2O3) and/or (Fe0.95Mg0.05)2O3

Minor: hematite (Fe2O3), hercynite (FeAl2O4)

Minor: pseudobrookite, (Fe2TiO5)

Major: ilmenite (FeTiO3), magnetite (Fe3O4) and/or magnesioferrite (Mg0.64Fe2.36O4) Minor: hematite (Fe2O3)

After oxidation Major: hematite (Fe2O3)

Major: hematite (Fe2O3)

Major: silica (SiO2), hematite (Fe2O3)

Major: pseudobrookite (Fe2TiO5), rutile (TiO2)

Minor: magnesioferrite (MgFe2O4) and/or magnetite (Fe(Mg)3O4) and/or maghemite (Fe2O3)

Minor: maghemite (Fe2O3) and/or magnetite (Fe3O4)

Minor: hematite (Fe2O3) and/or silica (SiO2)

After reduction Major: magnetite (Fe3O4)

magnesioferrites (Fig. 6d). Since the composition of phases observed in the first and third type of interaction are similar to the compositions of phases observed after processing in reducing conditions, it is likely that these interactions occurred before oxidation treatment. This observation is in general agreement with thermodynamic FactSage calculations, which predict a very small amount of Na-rich liquid phase that consumes some Fe from iron ore and exits at 950 °C. This liquid phase disappears at T = 930 °C, forming Ca(Al,Fe)2O4 and adding

Major: pseudobrookite (Fe2TiO5), magnesioferrite (MgFe2O4), and/or maghemite titanian (Fe(Fe0.92Ti 0.61)O4); hematite (Fe2O3) and/or Fe9TiO15 Minor: Mg1.05Ti1.95O5

Minor: iron titanium oxide (Fe3Ti3O10) Major: silica (SiO2), pseudobrookite (Fe2TiO5), rutile (TiO2)

other metals (including Na) in already existing phases, such as spinel, Na2Ca2Si3O9 and others. According to the FactSage calculations, some decrease of hematite and formation of spinel (mainly magnetite and magnesioferrite (FeMg2O4), where Mg is supplied from coal ash) is expected. These phases were observed by EPMA and are listed in Table 4. FactSage calculations also indicate that sulphur appears in the OC1Bca1 system as sodium sulphate (Na2SO4). This corresponds to the compositions in areas (2) and (3) in Fig. 5 and areas (1–3) in Fig. 6, which

Fig. 5. Oxygen carrier-brown coal ash sample OC1-Bca1 after reduction at 900 °C, indicating phases listed in Table 4.

76

A.Y. Ilyushechkin et al. / Fuel Processing Technology 147 (2016) 71–82

Table 4 Composition (wt.%) of phases and agglomerates in mixtures of oxygen carrier 1 (OC1) and coal ash, determined by EPMA. Al2O3

CaO

Fe2O3

MgO

TiO2

Phase identified

High-iron and magnesia ashes, in reducing atmosphere (Fig. 5) 1 0.2 0.2 0 0.9 2 10.2 5.7 1.8 43.9 3 19.1 0.5 2.7 37.2 4 0 0 0 0.4 5 1.3 0.8 0.2 42.8 6 1.4 1.2 0.3 30.5

0.7 34.3 31.2 0.5 33.7 27.4

0.1 0.2 0.7 0.7 0.2 0.4

90.4 7.6 9.7 93.4 19.9 36.4

7.4 0.1 0.1 5.4 0 0.2

0.1 0.8 0.7 0 1.0 1.7

Fe(Mg)3O4 + MgFe2O4 Na-aluminosilicate Na-aluminosilicate Fe(Mg)3O4 + MgFe2O4 Aluminosilicate solid solution Aluminosilicate solid solution

High-iron and magnesia ashes, in oxidising atmosphere (Fig. 6) 1 6.4 0.2 1.2 46.3 2 10.3 1.7 0.8 29.7 3 11.6 0 0.2 64.2 4 0.2 0 0.1 26.2 5 1.2 0.2 0.2 38.3 6 10.4 0 0.2 64.1 7 7.7 0.2 0.8 55.8

36.8 23.1 23.9 0.3 0.6 22.7 18.7

2.8 6.2 4.8 0.3 2.1 4.8 9.2

3.6 27.5 0.84 69.1 54.0 0.9 4.5

0.5 0.7 0 0.4 6.1 0 0.5

0 0.6 0 0.1 0.1 0 0

Na-aluminosilicate NaSO4 + aluminosilicate Na-aluminosilicate Iron silicate Former liq. Phase +Iron silicate Na-aluminosilicate Na-aluminosilicate

High-silica ashes in reducing and oxidising atmospheres (Fig. 7) 1 5.8 0.1 1.7 75.3 2 0.3 0.1 0.0 4.9 3 1.7 0.0 0.2 78.6

17.0 1.5 3.1

0.7 1.7 0.6

2.5 84.7 7.5

No.

Na2O

SO3

K2O

SiO2

are listed in Table 4 as its fine crystallites dispersed in alumina silicates along with Ca and Fe-containing ash minor minerals. 3.2.2. Interaction between iron ore and silica-rich ash The OC1-Bca2 sample, processed under reducing conditions, consists mainly of magnetite from iron ore and silica from Bca2, detected by XRD. It also has hercynite (FeAl2O4) peaks, detected in the sample after reduction processing, which indicates a possible interaction between aluminium from ash samples and iron ore.

1.1 0.9 0.2

0.2 0.0 0.0

Aluminosilicate Magnetite/aluminosilicate Si-rich solid solutions

The microstructures of the OC1-Bca2 sample processed under reducing conditions (Fig. 7a, b) and then oxidised (Fig. 7c, d) demonstrate two types of interactions. The first type is bonding of ash particles with iron ore without a visible reaction product: the phases detected and listed in Table 4 are alumina-silicate and iron oxide (Fig. 7a), and silica and iron oxide (Fig. 7c). The second type is a diffusion of iron from iron ore to silica and/or aluminosilicate particles, as indicated in Fig. 7d and Table 4, area (3). The agglomerate formed between silica and iron oxide shown in Fig. 7b is likely an iron ore particle with

Fig. 6. Oxygen carrier-brown coal ash sample OC1-Bca1 after oxidation at 950 °C, indicating phases listed in Table 4.

A.Y. Ilyushechkin et al. / Fuel Processing Technology 147 (2016) 71–82

77

Fig. 7. Oxygen carrier-brown coal ash sample OC1-Bca2 (a, b) after reduction at 900 °C and (c, d) after oxidation at 950 °C, indicating phases listed in Table 4.

penetrated small ash particles. SEM/EPMA did not reveal any distinctive hercynite phase, which is probably incorporated in other solids. Thermodynamic calculations reveal that the major phase in OC1Bca2 mixtures processed under reducing conditions is spinel (FeAl2O4, AlFe2O4, FeMg2O4), which consumes some Al and Mg from coal ash and Fe from iron ore. Silica from coal ash is included in minor phases such as andradite (Ca3Fe2Si3O12) and anorthite (CaAl2Si2O8). These phases are likely finely dispersed in iron oxide in area (2) indicated in Fig. 7b and Table 4. According to FactSage, oxidation of OC1-Bca2 mixtures leads to formation of Fe(Al)2O3 as a major phase, and a few aluminosilicates such as clinopyroxene (CaMgSi2O6, Mg2Si2O6, CaFeSi2O6), cordierite (Mg2Al4Si5O18) and CaAl2Si2O8. These phases possibly formed in a silica-rich matrix of ash and can be observed in area (3) in Fig. 7d. 3.3. Ilmenite with coal ashes 3.3.1. Interaction between ilmenite and iron- and magnesium-rich ash XRD analysis of OC2-Bc mixtures (Table 3) shows that co-processing ilmenite with coal ash Bca1 under reducing conditions leads to the appearance of magnesioferrite and the coexistence of magnetite with small amount of hematite. Oxidation of the OC2-Bca1sample results in the decomposition of ilmenite to pseudobrookite and rutile. The presence of coal ash increases the hematite content and possibly forms titanian maghemite (Fe(Fe0.92Ti 0.61)O4), with the disappearance of rutile. Interactions between coal ash and ilmenite are difficult to identify by XRD, because the dominant iron-titania phases and coal ash phases possibly cover the peaks from ash-OC interactions. As was previously observed for samples with iron ore, many more insights are delivered by SEM/EPMA. Detailed microstructures of the OC2-Bca1 sample processed under reducing and then oxidising conditions are given in

Figs. 8 and 9, respectively. Fig. 8 (a) shows general image of OC2-Bca1 sample, revealing two distinctive areas of ilmenite particles and fine ash particles. The interaction of these particles can be detected only in the interfaces and bonding, and are shown in Fig. 8(b)–(e). Table 5 provides the composition of phases and solid solutions observed in Fig. 8. Three different mechanisms of particle interactions are observed. Fig. 8b demonstrates particle physical bonding only: no visible reaction products are observed. Fig. 8c illustrates dissolution of ash elements in ilmenite matrix by solid–solid reactions: a clear reaction layer is observed in the ilmenite–ash agglomerates interface. Fig. 8d shows bonding ilmenite particles by Si-rich liquid phases from molten ash. Fig. 9 shows different types of particle interactions in the oxidised OC2-Bca1 sample previously exposed to reducing conditions. The composition of phases and solid solutions, marked in Fig. 9, are listed in Table 5. Overall observation (Fig. 9a) of the OC2-Bca1 mixture demonstrates an interaction between ash and OCs only on the surfaces and interfaces. Segregation of iron and magnesium in ash particles and formation of magnesioferrite rims on the surface of ash agglomerates is clearly visible in Fig. 9a–b. A similar Fe-enriched layer on the surface of ilmenite particles was observed when ilmenite was co-processed with gasification coal ashes [20]. However, this rim was also observed in ilmenite particles without ash [22]. In our study, the interaction between ash and ilmenite was confirmed by the presence of Mg in the rim, which is one of the major elements of Fe and Mg-rich coal ash. Another type of interaction in OC2-Bca1, illustrated in Fig. 9c–d, is bonding of silica particles with ilmenite by silica-rich liquid phase. According to FactSage calculations, a small consumption of titanium dioxide (TiO2) from calcined ilmenite is expected in OC2-Bca1 mixtures due to the reaction with Fe, Mg, Ca and Na species from coal. This results in the formation of karrooite (MgTi2O5), sphene (CaSiTiO5), perovskite (CaTiO3) and (Na2O)(TiO2)6. There is no consumption of Fe from ferric pseudobrookite (Fe2TiO5), which is a major phase in calcined ilmenite.

78

A.Y. Ilyushechkin et al. / Fuel Processing Technology 147 (2016) 71–82

Fig. 8. Oxygen carrier-brown coal ash sample OC2-Bca1 after reduction at 900 °C, indicating phases listed in Table 5.

All S-containing species are transferred in the gaseous phase during the combustion step. The predicted and experimentally obtained major phases in the mixtures are in good agreement, except for the appearance of magnesioferrite in the experimental samples that coexist with ilmenite, pseudobrookite and rutile. However, the predicted reaction products are very different from the experimentally observed products. Experimental reaction products that appeared were iron and magnesia silicates solid solutions, with the presence of small amounts of liquid phases. These were not predicted by FactSage. 3.3.2. Interaction with silica-rich ash In sample OC2-Bca2, processed under reducing conditions, XRD detects only the crystal phases attributed to the coal ash (quartz) and OCs such as ilmenite, ferric pseudobrookite and small amounts of rutile. When this sample was oxidised, the main phase became ferric pseudobrookite and rutile with quartz from coal ash. The interactions between coal ashes and ilmenite OC were detected only by SEM-EPMA. Fig. 10 a–d shows examples of the interactions between iron and titanium oxides and Si-rich coal ash in the OC2-Bca2 mixtures processed under reducing conditions. The microstructure of the sample subsequently processed in an oxidising atmosphere is shown in Fig. 10 e–f. It is likely that the bonding between particles occurred through a locally formed liquid phase. The exception is area (1) in Fig. 10b, which could be solid solutions formed through diffusion of elements near the interface. The composition of ‘bonding’ phases shown in Fig. 10b–d and f are listed in Table 5. Compositions of these phases vary; however, all phases are silica-based with a significant Ti, Fe and Al content. The liquid phase in the OC2-Bca2 mixture is not predicted by FactSage phase equilibria calculations. According to FactSage, under reducing conditions, the main component of the system along with

phases delivered by OC2 (ilmenite and rutile), and coal ash (quartz) is spinel (FeAl2O4, AlFe2O4, FeMg2O4). Minor phases, corresponding to ash and reaction products, are sphene (CaSiTiO5), or cordierite and anorthite, depending on the ash content in the system. Under oxidising conditions, ferric pseudobrookite, rutile and quartz from coal ash remain major phases, while minor phases associated with coal ash and reaction products are cordierite, anorthite and mullite (Al6Si2O13 and Fe6Si2O13). To understand the appearance of the liquid phase in the OC2-Bca1 and OC2-Bca2 systems, compositions (2), (4) and (5) from Table 4, Fig. 8, (4) and (5) from Table 4, Fig. 9, and (1)–(5) from Table 4, Fig. 10, were used to calculate their phase equilibria in FactSage at 900 °C in reducing atmospheres and 950 °C in oxidising atmospheres. For those compositions, a liquid phase exists in both reducing and oxidising atmospheres, along with solids such as anorthite, cordierite, clinopyroxene, ortho-enstatite (MgSiO3), sphene, silica and Fe, and Tioxides (ilmenite, pseudobrookite, rutile) delivered from ilmenite OC2. Silica and Fe, and Ti-oxides, bonded with the former liquid phase, were observed in the microstructures, while the others were not distinctively identified. It is possible that they coexist in the liquid phase as fine dispersed crystals; however, it is more likely that phase equilibria has not been achieved in these areas. The experiments were conducted at relatively low temperatures, and require significantly longer than 5 h to achieve phase equilibria. The areas with liquid phases also indicate that some local inhomogeneity exists in ash-OC mixtures, and could be expected in a real CLC system. This is likely attributed to variation in coal ash mineralogy, rather than in ilmenite OC composition, which mainly consists of iron and titanium oxides particles. It has been suggested that solid–solid reactions between ash and metal oxides could happen very slowly, and that some of the reaction phases would not appear in the mixtures even when these reactions are thermodynamically feasible [17].

A.Y. Ilyushechkin et al. / Fuel Processing Technology 147 (2016) 71–82

79

Fig. 9. Oxygen carrier-brown coal ash sample OC2-Bca1 after oxidation at 950 °C, indicating phases listed in Table 5.

Fig. 11a–b shows iron ore and iron ore-ash mixtures normalised mass change when samples were exposed in CO and air atmospheres, respectively. The iron ore was fully reduced in the first 25–30 min of treatment (Fig. 11a). In the presence of Bca1 ash, the reduction is completed in 12 min, while the presence of Bca2 ash (OC1-Bca2 sample) does not change the reduction kinetics. Oxidation of OC1 was completed in 10 min of treatment, with the presence of Bca1 ash accelerating the oxidation (approximately 6–7 min) (Fig. 11b). However, the OC1-Bca2 ash mixture oxidises at a slightly slower rate than does iron ore.

3.4. Impact of coal ashes on oxidation and reduction kinetics To evaluate the influence of ash components on the kinetics of OC reduction and oxidation, OC1-Bca1, OC1-Bca2, OC2-Bca1 and OC2-Bca2 mixtures were studied using TGA. As-prepared mixtures were processed under reducing conditions at 900 °C, and then in an oxidising atmosphere at 950 °C. In this section, we compare the results of reduction and oxidation kinetics of the OC-Bca samples with the kinetics of iron ore (OC1) and ilmenite (OC2) without coal ashes.

Table 5 Composition (%) of phases and agglomerates in mixtures of oxygen carrier 2 (OC2) and coal ash, determined by EPMA. SiO2

Al2O3

CaO

Fe2O3

MgO

TiO2

Phase identified

High-iron and magnesia ashes, in reducing atmosphere (Fig. 8) 1 0.3 0.3 0.4 33.9 2 1.6 0 0.6 35.3 3 0.5 0 0 0.8 4 2.7 0 2.3 74.3 5 3.7 0.1 1.3 60.5

0.7 3.1 0.2 0 0.1

1.9 10.1 3.5 0.1 10.0

42.3 27.2 58.4 11.2 11.5

19.3 20.5 2.2 0.6 6.8

0.8 1.3 34.3 8.3 5.1

Silica + MgFe2O4 Former liq. phase + MgFe2O4 Magnetite + Mg-ilmenite Former liq. phase Former liq. phase + MgFe2O4

High-iron and magnesia ashes, in oxidising atmosphere (Fig. 9) 1 10.2 0.5 1.5 18.6 2 20.4 0.3 2.2 40.9 3 2.4 0.3 0.2 3.4 4 1.4 0 1.1 70.3 5 7.4 0 0 46.2

31.7 32.9 18.7 0 0.1

1.0 1.4 0.5 1.0 11.1

33.1 1.6 60.4 15.1 21.2

2.3 0.2 12.7 0.9 8.1

1.1 0.1 1.3 10.0 3.7

Silica + aluminosilicates Silica + aluminosilicates Mg-ferrite Former liq. phase Former liq. phase

High-silica ashes in reducing and oxidising atmospheres (Fig. 10) 1 2.8 0.0 1.7 56.0 2 0.6 0.0 0.0 63.2 3 1.7 0.0 1.6 51.2 4 1.3 0.0 1.3 65.9 5 5.5 0.2 1.3 53.3

6.4 0.7 1.6 4.4 18.3

6.6 0.0 0.5 0.9 2.7

12.7 18.4 19.7 13.0 8.9

4.1 0.3 0.5 0.4 3.0

4.8 11.8 14.6 8.9 2.7

Si-rich solid solution Former liq. phase Former liq. phase Former liq. phase Former liq. phase

No.

Na2O

SO3

K2O

80

A.Y. Ilyushechkin et al. / Fuel Processing Technology 147 (2016) 71–82

Fig. 10. Oxygen carrier-brown coal ash sample OC2-Bca2 (a-d) after reduction at 900 °C and (e,f) after oxidation at 950 °C, indicating phases listed in Table 5.

The kinetics of ilmenite reduction and oxidation were affected by reaction with the material of the TGA sample holder, and also have a complex character due to phase separations. Fig. 12a demonstrates the ilmenite and ilmenite–ash mixtures normalised mass change under a reducing atmosphere. Most of the ilmenite decomposition and reduction reactions occur in the first 20 min, and are almost completed after 40 min of exposure. The presence of ashes does not significantly change the ilmenite reduction kinetics. Ilmenite and ilmenite–ash mixtures oxidation profiles are shown in Fig. 12b. The oxidation of the OC2 sample was completed in 10 min. The presence of Bca2 ash does not affect the ilmenite oxidation kinetics until 90% of the mixture is oxidised; the reaction rate then slows slightly. Bca1 ash extends the time for complete oxidation to 40 min; however, the iron-containing phases from ash also uptake oxygen, and the mass gained is increased twofold. As ilmenite oxidation is also affected by interaction with the TGA sample holder material, it is difficult to estimate

the effect of ash composition on the oxidation kinetics, since the extended oxidation could be also be a result of interactions between the ash and the sample holder material.

4. Discussion When iron ore is processed with brown coal ashes, some interactions between particles are observed for both iron-rich and silica-rich ashes. In reducing atmospheres, Si, Ca, and Al-containing species from both ashes are bonded with iron oxides from iron ore and form solid solutions of alumina silicates and iron oxides by solid–solid diffusion. Agglomerates are formed between silica and iron oxide (iron ore) in the silica-rich ash mixture, and these affect iron ore reduction kinetics. In iron- and magnesium-rich ashes, Fe and Mg oxides exist as fine magnesioferrite particles, which react with iron ore and form Mg-

A.Y. Ilyushechkin et al. / Fuel Processing Technology 147 (2016) 71–82

Fig. 11. (a) Reduction and (b) oxidation kinetics of iron ore and iron ore-coal ash mixtures. OC = oxygen carrier, Bca = brown coal ash.

containing magnetite. This assists iron ore reduction and increases the reaction rate by almost twofold. In oxidation of iron ore, Na-alumina silicates from ash react with iron ore hematite by solid–solid reactions or through the formation of a liquid phase. Na-alumina silicates also react with magnesioferrites from ash. These reactions occur on the surfaces of particles and reduce the sample's surface area, mainly due to the densification of the reaction products. However, this is compensated for by the presence of microporous magnesioferrites, and the kinetics of oxidation are even improved. For a silica-rich ash mixture with iron ore, the surface of iron ore is partially blocked by bonded silica and aluminosilicate particles, which increases oxidation time. In OC2-ash mixtures, the reaction between formed ilmenite crystals and ash follows as a dissolution of ash elements in the ilmenite matrix by solid–solid reactions, or by bonding ilmenite particles by Si-rich liquid phases from molten ash. The reactions in high-iron and magnesium ashes do not decrease the porosity of the ilmenite OC. In fact, the presence of high-surface-area, fine magnesioferrite ash increases the active surface area of the bulk sample. The presence of dense silica in the silica-rich ash mixture with ilmenite OC does not significantly affect the reduction kinetics, despite some local bonding of ash particles with OC2 by the liquid phase, which partially blocks gas access to the iron-containing phases. During the oxidation of ilmenite OC with high-iron and magnesia ashes, the interaction between ash, brookite and rutile particles (former ilmenite) occurs only on the surfaces by formation of magnesioferrite rims in ash agglomerates, or by bonding silica particles with former ilmenite by silica-rich liquid phase. The last reaction may occur in the previous reduction stage of the treatment, because the composition of the former liquid phase includes similar elements (Si, Ca, Fe, Na, Mg,

81

Fig. 12. (a) Reduction and (b) oxidation kinetics of ilmenite and ilmenite-coal ash mixtures. OC = oxygen carrier, Bca = brown coal ash.

Ti-oxides) in both samples. The presence of active iron-based phases in high-iron and magnesium ashes significantly increases oxidation time. However, this could be caused by the kinetics of magnesioferrite oxidation, rather than ilmenite interactions with coal ash phases. The silica-rich ash mixture with OC2, ash components, Ca, K, Na, Mg, Al and Si-oxides formed some liquid phase, which consumed Fe and Tioxides from ilmenite. This liquid phase bonds ash silica particles with ilmenite OCs. The amount of oxidised materials decreased more than expected, due to bulk composition dilution with non-oxidised silica; however, this only slightly increased the complete oxidation. The amounts of ash used in the laboratory mixtures with iron ore and ilmenite are significant, and corresponds to the samples that would form after hundreds of CLC cycles if ash accumulated in the CLC system. If effective ash removal was undertaken in CLC processes, the negative effect of ash on the chemical activity of OCs would be minimised and ignored compared with the attrition of OCs. Since high-iron and magnesium ashes improve the reduction and oxidation kinetics of iron ore, and the ash content in the coals is significantly lower than for coals with silica-rich ashes, ash removal is not an issue for this coal-OC system. However, the TGA is a fixed-bed reactor, while CLC is conducted in fluidised bed reactors. When redox increasing rates were observed in ash-OC mixtures, it was associated with increased porosity and surface area in pelletised ash-OC mixtures that assist gas access to the active sites [15]. In practical CLC processes, oxygen carriers reacts with the gasification gases (such as CO and H2) in dynamic conditions, whereas the TGA reduction profile was obtained by processing static samples in a CO flow. Hence, the reduction kinetics in real processes could differ from the observed TGA kinetics. In the present study, the reduction in CO was used only to identify the difference in redox potential of materials with different bulk composition, and the impact of ash composition on this potential.

82

A.Y. Ilyushechkin et al. / Fuel Processing Technology 147 (2016) 71–82

Of course, inorganic materials in CLC fluidised bed reactor have shorter contact times and number of the processing cycles, which may result to different degrees of interaction between ash species and oxygen carriers. The experimental approach taken in this study was to identify the possible interactions between oxygen carriers and coal mineral matters. The degree of these interactions associated with fluidised bed conditions and the materials cycling should be investigated separately. 5. Conclusion Two types of iron-based OCs, iron ore and ilmenite, were investigated in terms of their interactions with two types of brown coal ashes: Fe and Mg-rich, and Si-rich. Local interactions between OCs and ash particles were clearly observed by microstructural analysis. The reaction products are silicarich or sodium-rich liquid phases and solid solutions, comprising different aluminosilicates, iron silicates, spinels and magnesioferrites. The main component of Fe and Mg-rich ash is fine magnesioferrite particles. Their presence in an iron ore mixture significantly increases oxidation and reduction rates, which are not affected by interactions between magnesioferrites and iron ore. In contrast, silica-rich ash, which bonds with iron ores, reduces the active surface of the OCs but only increases oxidation time. Both types of ash interact locally with calcined industrial ilmenite particles, forming liquid phases and solid solutions. This interaction slightly affects ilmenite reduction; however, it significantly increases ilmenite oxidation in the presence of iron-rich ash. We expect that ilmenite treatment cycles may be affected by increased oxidation time in the presence of high-iron and magnesium brown coal ashes, while iron ore oxidation and reduction kinetics will vary for different types of brown coal. Acknowledgements The authors acknowledge the facilities and technical assistance of the Australian Microscopy & Microanalysis Research Facility at the Centre for Microscopy and Microanalysis, The University of Queensland, and the Central Analytical Research Facility operated by the Institute for Future Environments, Queensland University of Technology. References [1] M.M. Hossain, H.I. de Lasa, Chemical-looping combustion (CLC) for inherent separations—a review, Chem. Eng. Sci. 63 (2008) 4433–4451.

[2] S. Bhavsar, M. Najera, R. Solunke, G. Veser, Chemical looping: to combustion and beyond, Catal. Today 228 (2014) 96–105. [3] J. Adanez, A. Abad, F. Garcia-Labiano, P. Gayan, L.F. de Diego, Progress in chemicallooping combustion and reforming technologies, Prog. Energy Combust. Sci. 38 (2012) 215–282. [4] A. Cuadrat, A. Abad, J. Adánez, L.F. de Diego, F. García-Labiano, P. Gayán, Behavior of ilmenite as oxygen carrier in chemical-looping combustion, Fuel Process. Technol. 94 (2012) 101–112. [5] T. Mendiara, L.F. de Diego, F. García-Labiano, P. Gayán, A. Abad, J. Adánez, On the use of a highly reactive iron ore in chemical looping combustion of different coals, Fuel 126 (2014) 239–249. [6] F. Mayer, A.R. Bidwe, A. Schopf, K. Taheri, M. Zieba, G. Scheffknecht, Comparison of a new micaceous iron oxide and ilmenite as oxygen carrier for chemical looping combustion with respect to syngas conversion, Appl. Energy 113 (2014) 1863–1868. [7] H. Leion, T. Mattisson, A. Lyngfelt, Use of ores and industrial products as oxygen carriers in chemical-looping combustion, Energy Fuel 23 (2009) 2307–2315. [8] R. Xiao, L. Chen, C. Saha, S. Zhang, S. Bhattacharya, Pressurized chemical-looping combustion of coal using an iron ore as oxygen carrier in a pilot-scale unit, Int. J. Greenhouse Gas Control 10 (2012) 363–373. [9] C. Saha, S. Zhang, K. Hein, R. Xiao, S. Bhattacharya, Chemical looping combustion (CLC) of two Victorian brown coals — part 1: assessment of interaction between CuO and minerals inherent in coals during single cycle experiment, Fuel 104 (2013) 262–274. [10] C. Saha, S. Zhang, R. Xiao, S. Bhattacharya, Chemical looping combustion (CLC) of two Victorian brown coals – part 2: assessment of interaction between CuO and minerals inherent in coals during multi cycle experiments, Fuel 96 (2012) 335–347. [11] S. Rajendran, S. Zhang, R. Xiao, S. Bhattacharya, Use of synthetic oxygen carriers for chemical looping combustion of Victorian brown coal, Proc. Combust. Inst. 35 (2015) 3619–3627. [12] C. Saha, S. Bhattacharya, Chemical looping combustion of low-ash and high-ash low rank coals using different metal oxides — a thermogravimetric analyser study, Fuel 97 (2012) 137–150. [13] C. Saha, S. Bhattacharya, Experimental investigation of fluidized bed chemical looping combustion of Victorian brown coal using hematite, J. Environ. Chem. Eng. 2 (2014) 1642–1654. [14] M. Keller, M. Arjmand, H. Leion, T. Mattisson, Interaction of mineral matter of coal with oxygen carriers in chemical-looping combustion (CLC), Chem. Eng. Res. Des. 92 (2014) 1753–1770. [15] J. Bao, Z. Li, N. Cai, Interaction between iron-based oxygen carrier and four coal ashes during chemical looping combustion, Appl. Energy 115 (2014) 549–558. [16] R. Siriwardane, H. Tian, G. Richards, T. Simonyi, J. Poston, Chemical-looping combustion of coal with metal oxide oxygen carriers, Energy Fuel 23 (2009) 3885–3892. [17] A. Rubel, Y. Zhang, K. Liu, J. Neathery, Effect of Ash on Oxygen Carriers for the Application of Chemical Looping Combustion to a High Carbon Char, Oil Gas Sci. Technol. – Rev. IFP Energies nouvelles 66 (2011) 291–300. [18] A. Rubel, Y. Zhang, J.K. Neathery, K. Liu, Comparative study of the effect of different coal fly ashes on the performance of oxygen carriers for chemical looping combustion, Energy Fuel 26 (2012) 3156–3161. [19] B. Wang, R. Yan, H. Zhao, Y. Zheng, Z. Liu, C. Zheng, Investigation of chemical looping combustion of coal with CuFe2O4 oxygen carrier, Energy Fuel 25 (2011) 3344–3354. [20] M.M. Azis, H. Leion, E. Jerndal, B.M. Steenari, T. Mattisson, A. Lyngfelt, The effect of bituminous and lignite ash on the performance of ilmenite as oxygen carrier in chemical-looping combustion, Chem. Eng. Technol. 36 (2013) 1460–1468. [21] FactSage, Ecole Polytechnique, (http://www.factsage.com/, Montréal) 2008. [22] J. Adanez, A. Cuadrat, A. Abad, P. Gayan, L.F. de Diego, F. Garcia-Labiano, Energy Fuel 24 (2010) 1402–1413.