The Composition and Origin of PM_1-2 Microspheres in High-Calcium Fly Ash from Pulverized Lignite Combustion

Abstract

This article presents the results of a systematic study on the composition and origin of PM_1-2 microspheres in high-calcium fly ash. The composition of individual microspheres was studied by scanning electron microscopy and energy-dispersive X-ray spectroscopy. It is shown that the compositions of the analyzed microspheres satisfy the general dependency with a high correlation coefficient: [SiO₂ + Al₂O₃] = 88.80 − 1.02 [CaO + Fe₂O₃ + MgO], r = −0.97. The formation pathway is parallel to the general trend: anorthite, gehlenite, esseneite, tricalcium aluminate, ferrigehlenite, and brownmillerite. The microspheres were classified into four groups depending on the content of major components: Group 1 (CaO > 40, SiO₂ + Al₂O₃ ≤ 35, Fe₂O₃ < 23, MgO < 16 wt %); Group 2 (30 < CaO < 40, SiO₂ + Al₂O₃ ≤ 40, Fe₂O₃ < 27, MgO < 21 wt %); Group 3 (CaO ≤ 30, 40 ≤ SiO₂ + Al₂O₃ ≤ 75, Fe₂O₃ < 10, MgO < 10 wt %); and Group 4 (14 < CaO < 40, SiO₂ + Al₂O₃ < 14, Fe₂O₃ > 30, MgO ≤ 14 wt %). A comparative analysis of the relationship between major component concentrations suggests the routes of PM_1-2 formation from feldspars and Ca–, Mg–, and Fe–humate complexes during lignite combustion.

1. Introduction

Coal-fired power plants are both the largest source of electricity generation worldwide [1] and the origin of serious environmental pollution caused by ash and slag formation, as well as of emissions of fine particles [2,3,4]. Coal combustion-generated particulate matter (PM) is of a great concern due to the serious risks to human health and significant air contamination [5,6]. PM with an aerodynamic diameter less than 10 µm (PM₁₀), and especially PM ≤ 2.5 µm (PM_2.5), can exist in the atmosphere for a long time, which contributes to its long-distance transport, deposition to remote areas, and anthropogenic pollution [7].

Comprehensive research into properties and the formation mechanisms of fine particles during coal combustion is of fundamental and practical importance in order to achieve the best solution to reduce PM emission. In recent years, extensive research has focused on determining the effect of different coal mineral forms, their occurrence and thermochemical behavior on particle size distribution (PSD), and composition and morphology of the generated PM. The physicochemical properties of PM_2.5 and PM₁₀ have been reported, and theories regarding the formation mechanisms have been discussed [4,8,9,10,11,12,13,14,15,16,17]. It is noted that PM is transformed from minerals that mainly contain Si, Al, Na, K, Ca, Fe, and Mg [4,16]. Overall, the modes of occurrence of elements in coal are classified into mineral, organic, and intimate organic associations [18]. Minerals in coal are classified as included, which are associated with the coal matrix (these are finely dispersed minerals and organically-bound elements such as Na, K, Ca, Fe, and Mg), and, conversely, as excluded discrete minerals having little or no association with carbon matter [2,4,19,20]. During the combustion of suspended coal particles, the dispersed products of thermochemical transformation of included minerals and organically bound mineral matter can interact with each other in the same coal particle; these interactions can increase the production of low-melting compounds [2,3,4].

The PSD for PM₁₀ commonly identifies three primary size modes associated with different formation mechanisms. The ultrafine mode of PM is formed via the vaporization and condensation mechanism [8,9]; the coarse mode results from char fragmentation and mineral coalescence [4,10,11,12,13], and the central mode is affected by multiple mechanisms including direct transformation and coalescence of fine included minerals [4,15,16] and char and mineral fragmentation [4,15,17]. The central mode of PM, referred to as the fine-fragmentation mode [10,21,22], is in the size range of 0.3–5.0 μm with the maximum of ~1–2 μm. Therefore, systematic research into the composition of PM_1-2 is important for addressing the emissions of fine particles in general.

Various types of mineral matter present in coals and modes of their occurrence are quite complex, differing in coal ranks and local areas. There is extensive information in the available literature on the transformation of various forms of inorganic species during the combustion of high-grade coal (e.g., sub-bituminous and bituminous coal) and its contribution to PM₁₀ emissions [12,14,15,16,23,24,25,26,27,28,29], and there are only few studies on the formation of PM from mineral matter of low-rank coal (i.e., lignite) [25,29,30,31]. It should be noted that most elements identified in low-rank coal are characterized by varying degrees of organic association, and some major elements (e.g., Ca, Mg, Fe, Al, and Ti) largely occur in nonmineral forms [18]. Understanding the essential nature of their conversion during lignite combustion is useful for reducing PM emissions.

The Kansk–Achinsk basin has significant reserves of brown coal and is one of the largest sources of lignite in Russia [32]. Despite the high efficiency of electrostatic precipitators, the Krasnoyarsk thermal power plant No. 2 (TPP-2) emits ~1000 tons of dispersed ash particles into the atmosphere per year [33]. Because of the urgent need to reduce the environmental load from the operation of coal-fired power plants and minimize pollution by dispersed components, an objective is to identify the composition and formation routes of environmentally hazardous PM. The present paper reports the results of our systematic study on the composition and origin of PM_1-2 microspheres in high-calcium fly ash generated by Irsha–Borodinsky lignite combustion at a coal-fired power plant in Russia. This information is necessary for assessing the contribution of environmentally hazardous PM_2.5 and PM₁₀ from coal combustion to anthropogenic pollution. The detailed knowledge about the chemical composition of fly ash particles is useful for evaluating the promising areas for their utilization and provides new insight into the behavior of calcium-bearing minerals during lignite combustion.

2. Materials and Methods

2.1. Sample Preparation

A narrow fraction of dispersed microspheres classified as the environmentally hazardous suspended particulate matter PM_2.5 and PM₁₀ was used as a study object. The fraction was isolated from the fly ash produced during pulverized combustion of B2-grade lignite mined at the Irsha–Borodinsky open-pit mine of the Kansk–Achinsk coal basin. The coal was burned in a BKZ-420 steam boiler at Krasnoyarsk TPP-2 (furnace temperature 1400–1500 °C, liquid slag removal, fly ash and slag content 65% and 35%, respectively. Fly ash was sampled from the fourth field of the UG-2-4-74-04 electrostatic precipitator with fine particulate collection efficiencies of 98%. This sampling point was selected because finer fly ash particles are trapped by the fourth field of the electrostatic precipitator. The content of particles sized <10 µm is as high as 80%, and their content in the first field of the electrostatic precipitator is no higher than 30%, being ~40% and 65% in the second and third fields, respectively.

According to the American Society for Testing and Materials standard ASTM C618, the fly ash used to extract the target narrow fraction is classified as class C (high-calcium fly ash). The target fraction of PM was obtained by aerodynamic classification. The physicochemical characteristics of fly ash and the technological scheme for extracting fine narrow fraction from it were reported in ref. [34].

2.2. Characterization Methods

The narrow fraction was characterized by the following parameters: bulk density, particle size distribution, average globule diameter, and chemical and phase compositions.

The bulk density of the narrow fraction was measured on an automated Autotap density analyzer (Quantachrome Instruments, Boynton Beach, FL, USA); particle size distribution was measured on a MicroTec 22 laser particle sizer (Fritsch GmbH, Idar-Oberstein, Germany).

The chemical composition of narrow fraction was determined by chemical analysis methods according to GOST 5382-91 [35]. The standard error of repeatability (S_n) and the discrepancy between the results of parallel determinations (R_max) for each component, depending on its content, did not exceed what was accepted according to GOST 5382-91.

Powder X-ray diffraction data were obtained on X’Pert PRO (PANalytical, Almelo, The Netherlands) diffractometers equipped with a solid-state detector PIXcel using Cu Kα radiation (2θ range 12–120°). The full-profile crystal structure analysis was performed using the Rietveld method with derivative difference minimization according to the procedure [36].

The composition of single microspheres sized 1–2 µm was studied by SEM-EDS using a TM-3000 scanning electron microscope (High Technologies Corporation, Hitachi, Japan) equipped with a Quantax 70 microanalysis system with a Bruker XFlash 430H energy dispersive X-ray spectrometer (EDS) at a magnification of ×$10,000$. and an accelerating voltage of $15 \mathrm{kV}$. The samples were fixed with a dusting powder on a conductive carbon adhesive tape (Ted Pella Inc., Altadena, CA, USA) secured on a flat substrate prepared from a Duopur poly(methylmethacrylate) resin (Adler, Schwaz, Austria) with a diameter of 30 mm and a thickness of 1–3 mm. The data accumulation time exceeded 10 min; the quality of the spectrum collected in this mode makes it possible to quantitatively determine the gross composition of individual globules.

Figure 1 shows an SEM image of the fine narrow fraction with individual microspheres, indicating the single globules analyzed in these areas. For each analyzed globule, the gross composition, including the elemental contents of Ca, Si, Al, Fe, Mg, S, Na, K, Ti, Mn, P, and Ba, was determined. The concentrations of elements were recalculated to the content of the corresponding oxides and normalized to 100%. Repeatability error (S_n) for each component, depending on its content, is ±0.15 to ±0.30 for CaO; ±0.15 to ±0.35 for SiO₂; ±0.10 to ±0.30 for Al₂O₃; ±0.15 to ±0.60 for Fe₂O₃; ±0.20 to ±0.40 for MgO; ±0.10 to ±0.30 for SO₃; ±0.04 to ±0.30 for Na₂O; ±0.06 to ±0.15 for K₂O; ±0.04 to ±0.20 for TiO₂; ±0.04 to ±0.10 for MnO; ±0.03 to ±0.20 for P₂O₅; and ±0.04 to ±0.30 for BaO.

3. Results and Discussion

3.1. Characterization of Fine Narrow Fraction

Bulk density, particle size distribution parameters, and chemical and phase compositions of the fine narrow fraction are summarized in

Table 1. Bulk density, particle size distribution parameters, and chemical and phase compositions of the fine narrow fraction.

Bulk Density, g/cm3		Particle Size Distribution, µm
		dav		d10		d50		d90		d99
0.89		1.6		0.5		1.3		3.1		5.4
Chemical composition, wt %
LOI	SiO2	Al2O3	Fe2O3		CaO	MgO	Na2O	K2O	SO3		TiO2
5.30	13.98	9.17	13.96		38.50	8.20	0.32	0.18	9.60		0.32
Phase composition, wt %
glass phase	Ca4Al2Fe2O10	Ca3Al2O6	CaSO4		CaCO3	CaO	Ca(OH)2	MgO	quartz		Fe spinel
41.3	14.5	8.7	14.2		0.9	1.6	8.6	7.0	1.5		1.7
Chemical composition of the glass phase, wt %
SiO2	Al2O3	Fe2O3	CaO		MgO	Na2O	K2O		SO3		TiO2
32.88	15.51	9.70	33.30		3.16	0.84	0.47		3.29		0.84

; the SEM images are shown in Figure 2. The fraction has a narrow particle size distribution and is characterized by the following parameters: d₁₀ = 0.5 μm, d₅₀ = 1.3 μm, d₉₀ = 3.1 μm, and d₉₉ = 5.4 μm; the average microsphere diameter d_av = 1.6 μm.

High-calcium fly ash particles were classified by Zhao et al. [37] into several groups according to microstructural characteristics, namely, hollowed smooth, dense, agglomerate, porous, plerosphere, and other particles with complex surface characteristics; the hollow smooth particles are often smaller than 10 μm. The SEM images of the fine narrow fraction clearly show (Figure 2) that most particles are spherical and have a smooth non-porous surface; individual globules have a contoured surface, and there are also single agglomerates and fragmented nonspherical particles.

CaO is the major component of the chemical composition of the narrow fraction; its content is the highest compared to the contents of other components (38.5 wt %). The contents of SiO₂, Fe₂O₃, Al₂O₃, SO₃, and MgO are smaller; their gross content is equal to more than half of the fraction’s composition (54.9 wt %). The contents of the remaining components (Na₂O, K₂O, and TiO₂) are less than 1 wt %.

According to the quantitative X-ray diffraction analysis (Table 1), Ca-bearing compounds (the products of the interaction between the highly reactive free calcium oxide formed during thermal destruction of humates) are the main crystalline phases of the narrow fraction [38]. These compounds are as follows: brownmillerite Ca₄Al₂Fe₂O₁₀, tricalcium aluminate Ca₃Al₂O₆, and calcium sulfate and carbonate; together with calcium hydroxide and oxide, this makes approximately 50% of the gross composition of the fraction. The other crystalline phases of the dispersed fraction are magnesium oxide, Fe spinel, and quartz (Table 1).

The content of the amorphous Ca-bearing glass phase in the fine fraction is slightly less than half. Compared to the gross chemical composition, the glass phase is usually characterized by a higher content of individual components (Table 1): mostly SiO₂, as well as Al₂O₃, Na₂O, K₂O, and TiO₂.

3.2. Single-Particle SEM-EDS Analysis

In order to identify the mineral precursors responsible for the formation of environmentally hazardous particulate matter PM_1-2, we performed a systematic study of the gross compositions of individual microspheres in the fly ash fraction characterized by a narrow particle size distribution (Table 1). A total of 178 microspheres with sizes ranging from 1 to 2 μm were analyzed, which corresponds to the distribution maximum of the fine-fragmentation mode [10,21,22] for PM₁₀. The small particle size in a narrow range of values allows one to expect that the microspheres will be as homogeneous as possible (Figure 1) and have similar compositions to the products of thermochemical conversion of lignite mineral matter.

The results of the SEM-EDS study demonstrate that dispersed microspheres refer to the complex multicomponent system comprising the following elements: Ca, Si, Al, Fe, Mg, S, Na, K, Ti, Mn, P, Ba, and O. The major components are CaO, SiO₂, Al₂O₃, Fe₂O₃, and MgO; their gross content for all the globules ranges from 75 to 95 wt %. Sulfur content varies over a broad range: SO₃ 2–23 wt %. The gross content of alkali metal oxides Na₂O and K₂O is less than 2.0 wt % for most microspheres (~80% of globules), 2–5 wt % for 16% of microspheres, and 5.5–12.5 wt % in the remaining single globules. A low barium content is also typical of most microspheres; in 1/5 of the microspheres, it is <0.01 wt %; 72% of globules contain 0.01–2.0 wt %, and individual microspheres contain up to 6.3 wt %. Titanium, phosphorus, and manganese are contained at an impurity level; their contents in individual microspheres reach the following values: TiO₂, up to 2.5 wt %; P₂O₅, up to 1.5 wt %; and MnO, up to 1.4 wt %.

The compositions of all the studied microspheres satisfy the general dependency of the contents of the major components as a sum of acidic and basic oxides with a high correlation coefficient, including the gross chemical composition of the fraction, the composition of the glass phase, and the gross composition of the crystalline phases (Figure 3):

[SiO₂ + Al₂O₃] = 88.80 − 1.02 [CaO + Fe₂O₃ + MgO], r = −0.97

(1)

This dependency runs parallel to the overall trend formed by the compositions of the following compounds: anorthite CaAlSi₂O₈; gehlenite Ca₂Al₂SiO₇; esseneite CaFeAlSiO₆; tricalcium aluminate Ca₃Al₂O₆; ferrigehlenite Ca₂FeAlSiO₇; and brownmillerite Ca₄Al₂Fe₂O₁₀, including the phases forming a low-melting (1550°C) eutectic in the quinary CaO–MgO–Fe₂O₃–Al₂O₃–SiO₂ system: SiO₂, CaAl₂Si₂O₈, CaMgSi₂O₆ (diopside), CaSiO₃, and Fe₂O₃ [39] (Figure 3). The identified dependency of the contents of major components of PM_1-2 (Figure 3) is a certain pathway of formation of fine microspheres with different compositions.

Four groups of microspheres were differentiated depending on the content of major components; each group corresponds to certain composition criteria. The minimum and maximum contents of oxide in microspheres of different groups are listed in

Table 2. The minimum and maximum values of oxide content (wt %) in PM_1-2 microspheres.

Group	CaO	SiO2	Al2O3	Fe2O3	MgO	SO3	Na2O	K2O	TiO2	SiO2/Al2O3
1	CaO > 40, SiO2 + Al2O3 ≤ 35 wt %; 64 microspheres
min	40.01	4.68	2.54	4.70	3.76	3.80	<0.01	<0.01	<0.01	0.42
max	56.99	29.29	17.98	23.17	16.11	22.58	2.72	0.46	2.52	4.01
2	30 < CaO < 40, SiO2 + Al2O3 ≤ 40 wt %; 62 microspheres
min	30.03	3.40	2.93	5.68	5.54	3.72	0.22	<0.01	<0.01	0.30
max	39.68	33.26	19.75	27.04	20.75	20.67	2.78	1.31	2.00	2.38
3	CaO ≤ 30, 40 ≤ SiO2 + Al2O3 ≤ 75, Fe2O3 < 10, MgO < 10, SO3 < 10 wt %; 45 microspheres
min	7.97	23.34	2.99	2.99	2.37	2.21	0.61	<0.01	<0.01	1.01
max	30.37	54.72	33.65	8.87	10.09	10.09	11.00	4.18	0.87	15.54
4	14 < CaO < 40, SiO2 + Al2O3 < 14, Fe2O3 > 30, MgO ≤ 14 wt %; 7 microspheres
min	14.11	4.71	4.00	30.39	5.89	4.58	0.99	<0.01	0.05	0.70
max	39.34	8.39	6.74	60.19	14.26	14.67	2.15	0.20	0.71	1.89

.

• Group 1 includes the microspheres with CaO content > 40 wt % and SiO₂ + Al₂O₃ ≤ 35 wt %; 36% of all the studied particles meet these composition criteria;
• Group 2 contains the microspheres with a smaller CaO content compared to that in Group 1 (30–40 wt %) but higher contents of MgO (up to 21 wt %) and Fe₂O₃ (up to 27 wt %); the silicon and aluminum contents are also increased. The total content of oxides of these elements increases to 40 wt %; this group contains 35% of all microspheres;
• Group 3 consists of microspheres with increased SiO₂ and Al₂O₃ contents; the total content of these oxides in globules significantly increases from 40 to 75 wt %; the content of other major components is considerably lower than that for the Group 1 and Group 2 microspheres: CaO ≤ 30, Fe₂O₃ < 10, MgO ≤ 10, and SO₃ ≤ 10 wt %; the contents of Na₂O and K₂O noticeably increase to reach 11 and 4 wt %, respectively; 25% of microspheres have this composition;
• Group 4 contains microspheres with a high Fe₂O₃ content (30–60 wt %) and with a reduced content of SiO₂ and Al₂O₃ (SiO₂ + Al₂O₃ < 14 wt %), MgO ≤ 14 wt %; it is the smallest group that includes 4% of all the studied particles.

Figure 4 shows the percentage composition of microspheres with the concentrations of the major components lying in a certain range for each group. There are typical differences in component distribution, which proves that the selected composition criteria used for classifying individual microspheres into groups are correct and will allow one to identify the general regularities in their formation.

Thus, a high content of CaO in the range from 40 to 50 wt % is characteristic only for 92% of particles of Group 1, from 30 to 40 wt % for all particles of Group 2, and for 42% of particles of Group 4. In Group 3, the CaO content from 20 to 30 wt % is found in 62% of globules.

A high content of SiO₂ > 20 wt % is observed for all globules of Group 3; in Group 1 and Group 2, these are only 6 and 21% of the particles, respectively; in Group 4, there are no such particles.

A high content of Al₂O₃ > 20 wt % is typical for 67% of the globules in Group 3; there are no such particles in other groups. Al₂O₃ < 10% is observed for all microspheres in Group 4, 67% of particles in Group 1, and 27% of particles in Group 2.

The distribution of MgO is close for particles of Group 1 and Group 4. Group 1 and Group 2 are similar in the number of particles in the specified ranges of Fe₂O₃ and SO₃ contents.

Hence, a study focusing on the compositions of individual microspheres with a size ranging from 1 to 2 μm showed that a significant portion of PM_1-2 (71% of all globules) formed during Irsha–Borodinsky lignite combustion is characterized by a high CaO content (30–57 wt %). Compared to them, one quarter of all microspheres have an increased content of aluminosilicate components and alkali metal oxides. The concentration of microspheres with a high iron content, which should be classified as ferrospheres [40,41,42,43], is 4 wt %.

3.3. Identification of Mineral Precursors

Based on the data on the composition of individual PM_1-2 microspheres, we noted that their formation during lignite combustion is controlled by the occurrence of calcium compounds in coal, which are classified as a discrete mineral matter and organically associated cation [44]. The categories of calcium-bearing minerals found in coal include carbonates, sulfates, silicates, phosphates, etc. [37,45]. Organically associated calcium is molecularly dispersed in the coal macerals and bonded to the oxygen anions in carboxyl groups [38,46]. Calcium ions are more reactive if bound to the carboxyl groups of the organic matrix than if present in the crystalline structure. Dispersed products of thermochemical transformation of organically bound calcium and aluminosilicate minerals quickly react with each other and with quartz in the coal matrix to form lower melting point species.

According to the standard of the American Society for Testing and Materials ASTM D318 “Standard Classification of Coals by Rank”, which covers the classification of coals by rank, that is, according to their degree of metamorphism, or progressive alteration, in the natural series from lignite to anthracite, the Irsha–Borodinsky coal of the Kansk–Achinsk coal basin refers to the Lignite A coal rank. The petrographic composition of Irsha–Borodinsky coal is characterized by the predominance of macerals of the vitrinite group up to 94%; the ash content is A^d~9% [32].

The Irsha–Borodinsky coals of the Kansk–Achinsk coal basin are typical humus coals; a considerable portion of calcium and magnesium, as well as noticeable amounts of iron, are present as organomineral substances in them [38]. The mineralogical composition is represented by clay minerals kaolinite, montmorillonite, and hydromica; feldspars, iron disulfides crystallizing in the pyrite (or, less commonly, marcasite) structure, quartz, and siderite and calcite impurities. Carbominerites (associations between the coal and the mineral substance) are also present, which are most frequently represented by carbargilite [47]. There are adsorption-induced associations with hydromica-, montmorillonite-, and chlorite-type clay minerals as well as their polymorphism products containing quartz and feldspar impurities.

One of the approaches for characterizing the inorganic components in coal, char, and ash is the computer-controlled scanning electron microscopy (CCSEM) technique; its detailed procedures are presented in Refs. [26,48]. Briefly, CCSEM analysis involves the classification of mineral particles into specified mineral categories based on elemental composition criteria. The classification criteria take into account the identity of the mineral chemical composition according to the detected major elements (O, Si, Al, Ti, Fe, Ca, Mg, Na, K, P, S, and Cl). Mineral categories and their elemental composition criteria were initially developed by Huggins et al. [49] and Zygarlicke et al. [50], successfully supplemented by Wen et al. [51], and are widely used by other researchers [27,28,52].

Within this approach, each analyzed microsphere was classified into a specified mineral category [50,51] according to the relative contents of elements (Figure 5). Only three mineral categories were recognized among the studied fine microspheres out of the 33 used for mineral matter analysis by the automated CCSEM technique [50,51]. The quantity of microspheres classified as mineral category Fe–Ca-aluminosilicate is 15%; gypsum/Al-silicate, 8%; and ankerite, 7%. The predominant part of the microspheres, which is 70%, according to major element contents and elemental composition criteria, cannot be assigned to any of the categories and remains unclassified (Figure 5).

In different studies performed by other researchers, the contents of mineral species unidentified by CCSEM (Unclassified category) were as high as 17% for Chinese coal of bituminous rank [26]; ~30% for Western Australia’s sub-bituminous coal [28] and for the char of lignite Chinese coal [51]; 7–44%, for fine mineral particles sized < 10 µm from Collie coal of Western Australia [52]; 48%, for Beulah–Zap lignite ash [53]; and 70% for ash of lignite Chinese coal [48].

Wu et al. [52] analyzed the data on chemical composition of unclassified mineral particles and showed that regardless of size (0.5–2.2, 2.2–4.6, or 4.6–10.0 µm), most of them have mixed compositions of components (Si + Al), (Fe + Ti), and others (Ca, K, S, and P). This observation suggested that mineral–mineral associations are present within a single fine mineral particle, leading to the formation of low-melting ash particles during coal combustion.

In our case, unclassified microspheres have a multicomponent mixed chemical composition: 14 ≤ Ca ≤ 60, 6 ≤ Fe ≤ 64, 3 ≤ Si ≤ 45, 4 ≤ Mg ≤ 31, 2 ≤ Al ≤ 25, and 2 ≤ S ≤ 15. Approximately two thirds of these globules partially meet the criterion (Fe + Ca + Al + Si) ≥ 80 for the mineral category known as Fe–Ca-aluminosilicate (Al ≥ 15, Si ≥ 20, Ca ≥ 3, Fe ≥ 3, and Fe + Ca + Al + Si ≥ 80)⁵¹, but at the same time, the content of the aluminosilicate components is insufficient (Al ≥ 2 and Si ≥ 3), and calcium and iron contents are increased (Ca ≥ 15, Fe ≥ 6).

Systematic studies of the compositions of individual microspheres and a comparative analysis of the relationships between concentrations of the major components allow one to identify the nature of mineral precursors whose thermochemical conversions give rise to ash particles. This approach has been successfully used for cenospheres [54] and ferrospheres [40,41,42,43] of different morphological types, dispersed microspheres from the non-magnetic fraction of fly ash [55], and individual microspheres localized in coal char particles with different morphology [56]. It should be noted that microspheres with a size of 1−2 μm have been identified inside coal char particles, being localized in channels and cavities; they are formed in the internal porous structure of the carbon matrix from authigenic fine included minerals of coal [56].

The important role played by silicate and aluminosilicate minerals, which act as the structural framework for many mixed species [53], in the formation of PM was assessed from the dependence SiO₂ = f(Al₂O₃) (Figure 6). The aluminosilicate compositions of individual microspheres from different groups were compared with the compositions of dehydroxylated minerals contained in the coal [57]: clay minerals (kaolinite Al₂Si₂O₅(OH)₄, K-illite K_1.5Al₄(Si_6.5Al_1.5)O₂₀(OH)₄, and montmorillonite Na_0.33(Al_1.67Mg_0.33)Si₄O₁₀(OH)₂); minerals belonging to the group of mica-like chlorites (clinochlore (Mg₅Al)(AlSi₃)O₁₀(OH)₈); and feldspars (orthoclase K(AlSi₂O₈), albite Na(AlSi₂O₈), and anorthite Ca(AlSi₂O₈)). A comparative analysis of the relationship between the major component concentrations revealed a general trend (Figure 6). Thus, approximately half of all the compositions of all microspheres satisfy the regression Equation (2) with a high correlation coefficient:

[SiO₂] = 1.18 [Al₂O₃] − 0.092, r = 0.98

(2)

The slope value (1.18) corresponds to the SiO₂/Al₂O₃ ratio in anorthite and kaolinite. The remoteness of kaolinite compositions (Figure 6) and the fact that the phase composition of the ash fraction does not contain the product of its thermochemical conversion, mullite [2,3] (Table 1), allow one to infer that this mineral was not directly involved in the formation of microspheres sized 1–2 µm, and the role of structure-forming mineral precursor was played by anorthite. This statement is true for other clay minerals, such as illite and montmorillonite. Like other members of the isomorphous feldspar series, anorthite melts during coal combustion to give rise to the glass phase; after its interaction with other components, complex aluminosilicates with different compositions are formed [2,3,38]. Gehlenite Ca₂Al₂SiO₇ [39] (Figure 6) is the conventional component of char [29] and ash [58] with a high calcium oxide content that is formed upon interaction between the products of thermochemical conversion of aluminosilicate mineral components and CaO during lignite combustion [38,58]. The ternary phase diagram of CaO – Al₂O₃ – SiO₂ proves that anorthite and gehlenite are typical high-temperature calcium-bearing compounds congruently melting and forming low-melting eutectics [39].

The compositions of esseneite CaFeAlSiO₆ and ferrigehlenite Ca₂FeAlSiO₇ [39] satisfy regression Equation (2). Calcium–-iron aluminosilicates of different composition are the products of aluminosilicate transformations upon their interaction with calcium and iron oxides [38,58]. In turn, the highly reactive CaO and Fe₂O₃ are the primary products of oxidation of Ca– and Fe–humates during lignite combustion [30,31,38], and are also formed after partial decomposition of calcite, siderite, and pyrite [2,3]. The presence of coal–mineral associations [44] with clay minerals, quartz, mica, and feldspars facilitates the interaction between the mineral matter and organically associated cation within the coal matrix during lignite combustion.

On the dependence SiO₂ = f(Al₂O₃) (Figure 6), the composition points “esseneite” and “ferrigehlenite” separate the line segments satisfying the compositions of microspheres belonging to different groups. Thus, the “anorthite–esseneite” section includes only the compositions of microspheres from Group 3 with the typical increased content of aluminosilicate components, while the ferrigehlenite section includes the compositions of Group 1 and Group 2 microspheres with increased calcium content and ferrospheres from Group 4 (Table 2).

For the remaining microspheres, chlorite-type clay minerals with an impurity of feldspars belonging to the albite–orthoclase series and quartz acted as the structure-forming aluminosilicate precursors (Figure 6).

High-iron microspheres from Group 4 and partially high-calcium microspheres from Group 1 form an individual trend on the dependence CaO = f(Fe₂O₃) with a high correlation coefficient, including the composition of ferrigehlenite (Figure 7):

[CaO] = 58.96 − 0.76 [Fe₂O₃], r = −0.95

(3)

The relationship in regression Equation (3) is similar to that revealed for the local areas of ferrospheres sized −0.04 + 0.032 mm isolated from magnetic concentrates of fly ash produced from the pulverized combustion of low-rank Berezovsky coal (Kansk–Achinsk basin, Russia) of the B2 grade (sub-C according to ASTM D388-98a) at the Berezovskaya state district power plant [43]. Further studies will be aimed at identifying the compositions of dispersed ferrospheres in high-calcium ashes, performing a comparative analysis of the relationship between macrocomponent concentrations together with the compositions of local regions of large ferrospheres [40,41,42,43] and detecting general regularities and pathways for their formation.

Therefore, our systematic study of the compositions of dispersed microspheres and the comparative analysis of the relationship between macrocomponent concentrations allows one to infer that feldspars (mainly anorthite) and Ca–, Mg–, and Fe–humate complexes are the major mineral precursors of the environmentally hazardous particulate matter PM_1-2 formed during Irsha–Borodinsky lignite combustion. During the lignite combustion, feldspars melt with the formation of a glass phase, and organically bound Ca, Mg, and Fe are easily decomposed, released, and incorporated in the aluminosilicate glass to form complex compounds of various compositions.

4. Conclusions

• A systematic study of the compositions of dispersed PM_1-2 microspheres formed in high-calcium fly ash during Irsha–Borodinsky lignite combustion was conducted for the first time. Individual microspheres (sized 1–2 μm) of the narrow fraction of the fly ash sampled from the fourth field of the electrostatic precipitator at the Krasnoyarsk TPP-2 burning coal mined at the Kansk–Achinsk basin (Russia) were analyzed by scanning electron microscopy and energy-dispersive X-ray spectroscopy.
• CaO, SiO₂, Al₂O₃, Fe₂O₃, and MgO were found to be the major components of the analyzed microspheres; their gross content ranges from 75–95 wt %. The compositions of all the microspheres under study follow the general dependence with a high correlation coefficient: [SiO₂ + Al₂O₃] = 88.80 − 1.02 [CaO + Fe₂O₃ + MgO], r = −0.97. The formation pathway for fine microspheres with different compositions is parallel to the general trend: anorthite CaAlSi₂O₈; gehlenite Ca₂Al₂SiO₇; esseneite CaFeAlSiO₆; tricalcium aluminate Ca₃Al₂O₆; ferrigehlenite Ca₂FeAlSiO₇; and brownmillerite Ca₄Al₂Fe₂O₁₀.
• The microspheres were classified into four groups depending on the content of major components: Group 1 (CaO > 40, SiO₂ + Al₂O₃ ≤ 35, Fe₂O₃ < 23, MgO < 16 wt %) and Group 2 (30 < CaO < 40, SiO₂ + Al₂O₃ ≤ 40, Fe₂O₃ < 27, MgO < 21 wt %) contain a considerable portion of PM_1-2 (71% of globules), which are characterized by a high CaO content (30–57 wt %); Group 3 (CaO ≤ 30, 40 ≤ SiO₂ + Al₂O₃ ≤ 75, Fe₂O₃ < 10, MgO < 10 wt %) contains 25% of microspheres with a typically increased content of aluminosilicate components and alkali metal oxides; and the smallest Group 4 contains 4% of microspheres with a high Fe₂O₃ content (30–60 wt %).
• A comparative analysis of the relationship between major component concentrations suggests the routes of formation of environmentally hazardous PM_1-2 from feldspars and Ca–, Mg–, and Fe–humate complexes during lignite combustion.