Tuesday, February 11, 2020

Determining if the Char Fraction in Coal Fly & Bottom Ash Can Be a Precursor Material for Synthetic Graphite


Results from an ERA-MIN funded collaborative project (“Charphite”) between Portugal, Poland, Romania, South Africa, and Argentina are presented. The char found in coal ash is evaluated as a precursor for synthetic graphite (and subsequent application in green energy technologies). The initial coal ash samples were characterized in detail. The extraction process of char from a South African fly ash is discussed as a case study. Final characterization results on the char concentrates from four countries are presented in terms of potential graphitizability.

All initial coal ash samples were characterized using proximate, carbon forms, X-ray fluorescence, X-ray diffraction, petrography, solid-state Nuclear Magnetic Resonance, and particles size distribution analyses.

A combination of size, electrostatic, magnetic, and density extraction techniques were used to concentrate the char. It was found that the carbon grade increased from 7.04 wt. % to 82.91 wt. % LOI (loss in ignition) for the South African fly ash sample. The overall carbon recovery was 18.92 %.

All final char concentrates were characterized using proximate, ultimate, carbon forms, petrography (char form and reflectance), and Micro Raman Spectroscopy. The Romanian char concentrate showed a disordered nature and high percentages of volatile matter, organic carbon, unreacted coal, and isotropic char were encountered. Micro Raman Spectroscopy classified this material as “amorphous” with the presence of a broad G band. The char concentrates from Portugal, Poland, and South Africa were classified as transitional carbon material with the possibility of graphitization. Elemental carbon and anisotropic percentages in these three samples were high.

1. Introduction

Coal ash is described as a coal combustion by-product and globally, more than 780 million tons of ash is produced annually from coal combustion power plants (Heidrich et al., 2013). The utilization of this ash, which includes bottom ash and fly ash, is mainly in concrete and cement applications. However, this utilization depends of factors such as distance to consumer and market fluctuations, and huge amounts of ash are therefore landfilled. However, the landfilling and ponding of ash leads to numerous environmental, health and economic concerns (Carlson and Adriano, 1993; Fulekar and Dave, 1986).

This way, constant innovative and new solutions are therefore sought for recycling / re- using coal combustion ash, including: utilization in soil amelioration, waste immobilization, clay bricks, road stabilization, fillers for polymers and rubbers, treating acid mine drainage, and mine backfilling (Ahmaruzzaman, 2010; Blissett and Rowson, 2012; Kruger and Krueger, 2005; Yao et al., 2015). Certain value-added components in ash can also be extracted and utilized (Heidrich et al., 2013). One such value-added component is the carbonaceous solid residue or char.

Depending on furnace temperature, residence time, parent coal, and utility technology, unburned carbon (char and soot) percentages in ash can range from 0 – 45 %; this value is essentially an indication of the efficiency of the coal burning process (Bartoňová, 2015). Char is known to be very porous, making it ideal for mercury adsorption and activated carbon practices (Hwang and Li, 2000; Hwang et al., 2002; Li et al., 2002; Li and Maroto-Valer, 2012; Luo et al., 2004; Zhang et al., 2003). However, it may also have a high degree of structural order and a lamellar microtexture, resulting in the possibility as a synthetic alternative to natural graphite (Baltrus et al., 2001; Bartoňová, 2015; Hurt et al., 1995; Maroto-Valer et al., 2002). Cabielles et al. (2008; 2009) and Cameán and Garcia (2011) published several papers on the production of graphite materials from char. But, apart from their output, limited information on this topic can be found in literature. A definite research gap or niche thus exists.

The European Union and the United States have recently listed natural graphite as a critical raw material, classifying it as having both a high supply risk and is of great economic importance (European Commission, 2017; Fortier et al., 2018). The critical status has largely come from the increase in graphite demand; especially concerning usage in economic valuable applications such as lithium-ion batteries, fuel cells, and pebble-bed nuclear reactors. Synthetic graphite can replace natural graphite in some of these applications. Currently synthetic graphite is manufactured from petroleum coke, but anthracite has also been identified as a favorable precursor.

In 2015 an international collaborative project between Portugal, Poland, Romania, South Africa, and Argentina was funded under the 3rd ERA-MIN call of the European Union. The project “Coal char as a substituting material of natural graphite in green energy technologies – Charphite” investigates the usage of char, found in coal ash waste, to act as a possible precursor for synthetic graphite. Ultimately the formed synthetic graphite will be applied in green energy applications (electrocatalysts in fuel cells). Detailed information on the project can be obtained from the Charphite website (Charphite, 2019). The proposed collaboration methodology is given in Figure 1. The methodology is very complex with a number of smaller embedded projects. The focus of this paper (circled) will be on the characterization of coal and ash obtained from coal combustion power plants in Portugal, Poland, South Africa, and Romania; the beneficiation of char from a South African fly ash case study; and the characterization of beneficiated char samples from Portugal, Poland, South Africa, and Romania.

2. Experimental

Figure 1: Charphite proposed project methodology.

2.1 Coal and ash characterization

Fly ash and bottom ash samples were supplied by Portuguese, Polish, South African, and Romanian coal combustion power plants (Table 1). The corresponding parent / feed coal samples were also obtained; Romania was not able to supply the latter. It is assumed that all samples were obtained in a representative manner from the different coal combustion power plants following standard procedures adopted in the respective countries. A representative sub-sample from each coal and ash sample was characterized.

Table 1: Coal and ash sample identification / type, origin, and involved partners.

Proximate, ultimate, carbon forms, and X-ray fluorescence (XRF) analyses were used to obtain information on the elemental composition of the coal and ash samples. Ultimate analysis was omitted for the ash samples, seeing that their percentage of organic matter is too low for accurate speciation. The proximate and ultimate analyses were outsourced to Bureau Veritas Testing and Inspections South Africa. The standards used are given in Table 2.

Table 2: Proximate and ultimate analyses standards.

Carbon forms analysis was outsourced to Bureau Veritas Testing and Inspections Canada. Total carbon percentages were determined using the Leco method. Samples are ignited in an induction furnace where after a carrier gas sweeps up the released carbon. Released carbon is then measured by adsorption in an infrared spectrometric cell. Inorganic carbon is determined directly by measuring the released carbon when the sample is leached with perchloric acid. Elemental carbon is determined by leaching the samples with dilute HCl (to eliminate inorganic carbon), followed by ignition at 500 °C

(low temperature to prevent pyrolysis of organic carbon), and then analysing the residue by the Leco method. Organic carbon is determined by difference. Organic carbon represents material with carbon atoms bonded to other organic elements (sulphur, nitrogen, hydrogen, and oxygen), whilst elemental carbon represents material with strong carbon-carbon bonds, such as seen in graphitic material (Bjurström et al., 2014; Brown and Dykstra, 1995). Inorganic carbon represents material with carbon-inorganic bonds, such as calcite (CaCO3).

The XRF analysis was outsourced to the School of Biological, Earth and Environmental Sciences of the University of New South Wales (UNSW). Coal samples were ashed at 815 °C and the ash yield determined in each case. The ashed coal as well as the ash samples were then calcined at 1050 °C, fused with lithium metaborate, and cast into discs (Norrish and Hutton, 1969). Each disc was analysed by XRF spectrometry using a Philips PW 2400 spectrometer and SuperQ software. The results were expressed as percentages of the major element oxides in each ash sample.

X-ray diffraction (XRD) analysis was used to determine the mineralogical compositions. Samples were outsourced to the School of Biological, Earth and Environmental Sciences of the UNSW. The coals were first subjected to low-temperature oxygen-plasma ashing (LFE 4-chamber asher) (according to Australian Standard 1038.22 (2002)), and the weight percentage of low-temperature ash (LTA) determined in each case. The LTA and ash samples were then analysed by a PANalytical Empyrean diffractometer with Co K𝛼 radiation. Mineral phases were identified by the ICDD Power Diffraction File, whilst SiroquantTM software was used for quantitative analysis. The ash samples were found to contain significant proportions of amorphous material. A poorly-crystalline aluminosilicate phase (identified as “metakaolin”) was used as a substitute to represent the amorphous component (Ward and French, 2006).

A Zeiss Imager M2M reflected-light petrographic microscope with an oil immersion objective and a combined magnification of ×500 was used for morphological composition determination. The samples were pulverized to <1 mm (where necessary), mounted in epoxy resin, and polished (based on ISO 7404 part 2, 2009). Hilgers Fossil software was used for the mean random vitrinite reflectance of the coal samples (based on ISO 7404 part 5, 2009). Hilgers Diskus software was used to conduct the maceral point count for the coal samples (based on ISO 7404 part 3, 2009). Quantitative and qualitative point count analyses were conducted on the ash samples, applying the identification method presented in Table 3 (based on Hower, 2012).

Table 3: Classification scheme for the morphological composition of ash (based on Hower, 2012).

To determine the structural ordering of the coal and ash samples, X-ray powder diffraction (XRPD) and solid-state Nuclear Magnetic Resonance (ss-NMR) analyses were conducted. XRPD measurements were recorded on an Empyrean diffractometer (Panalytical, The Netherlands), with a PixCel3D area detector in a 1D scanning lineal mode. The ss-NMR results were acquired using a Bruker Avance-III HD spectrometer, equipped with a 14.1 T narrow-bore magnet operating at Larmor frequencies of 600.09 MHz for 1H, 150.91 MHz for 13C, 156.37 MHz for 27Al, and 119.21 MHz for 29Si.

Powdered samples were packed into 3.2 or 2.5 mm ZrO2 rotors and rotated at magic angle spinning (MAS) rates of 15 or 35 kHz (at room temperature) respectively. The 13C CP-MAS (cross-polarization and magic angle spinning) experiments were done in a 3.2 mm MAS probe. Glycine was used as an external reference for the 13C spectra and to set the Hartmann−Hahn matching condition in the CP-MAS experiments in 13C spectra. The contact time during CP was 2000 µs. The SPINAL64 sequence (small phase incremental alternation with 64 steps) was used for heteronuclear decoupling during acquisition. In addition, 13C direct-polarization and magic angle spinning (DP-MAS) experiments were done in order to quantify the different type of carbons, since the CP- MAS experiments are only qualitative. However, the 13C DP-MAS spectra requires an extended period of time for the acquisition of a NMR spectrum with a good signal to noise radio (>12-48 h) in comparison with the CP-MAS experiments (2-4 h). 1H, 29Si and 27Al ss-NMR spectra were done in a 2.5 mm MAS probe using single pulse excitation experiments. Chemical shifts (in ppm) are relative to (CH3)4Si for 1H and 29Si, and 1 M Al(NO3)3 aqueous solution for 27Al.

Lastly, the particle size distributions (PSD) for all ash samples were obtained using a Vibratory Sieve Shaker ANALYSETTE 3 SPARTAN and standard R20 sieves with nominal sieve openings of 150, 75, 45 and 25 µm for the fly ash samples, 100, 200, and 400 μm for the bottom ash from Poland, and 63, 90, 125, 250, 500, 1000, 2000, 4000 μm for the bottom ash from Romania. Before sieving commenced, the samples were dried in an oven for 24h at 30 °C. Each sample was sieved using portions of 100g for 10 minutes. The subsamples were weighed using a RADWAG PS 2100/C/I/CT scale, stored and sealed. The mean values per fraction were calculated and converted into a percentage.

Visual observations and literature from Hurt and Gibbins (1995), Külaots et al. (2004), and Lu et al. (2007) suggest that the char is concentrated in the larger size fractions of fly ash. For this reason, the fixed carbon percentages were also determined (following standards given in Table 2) for each size fraction of the fly ash samples.

2.2 Beneficiation of char

Each of the ash samples described in Table 1 was uniquely beneficiated at the respective host institutions to obtain a concentrated char product. The aim was to obtain a product with a high carbon grade (ideally ~90 % carbon, air-dried basis (a.d.b) fixed carbon) and a high carbon recovery. Detailed information on the country specific beneficiation techniques is available from Santos et al. (2019). In this paper, a brief description of the beneficiation of char from a South African fly ash sample is provided as a case study; the procedure developed in South Africa was in fact comparable to those developed by the collaborative institutions (e.g. size and magnetic separation).

Carbon grade and carbon recovery were used to assess the beneficiation efficiency. Carbon grade (purity of product stream) was measured by loss on ignition (LOI), with a final combustion temperature at 815 °C. Carbon recovery is similar to yield, but reported on a carbon basis. This provides an indication of the mass percentage of initial carbon ending up in the product stream. Ideally, a balance between carbon grade and carbon recovery is sought.

A flow diagram for the beneficiation process is presented in Figure 2. The starting LOI was 7.04 wt. %. A dry sieving technique was used first. Sieves with sizes 38, 53, 75, 106, 150, and 212 µm were stacked, loaded with 150 g sample, and sieved for 20 minutes in an automated shaker. The cumulative carbon grades and carbon recoveries for each sieve size were obtained, and, based on this; an ideal cut-point was obtained. The sieve product then underwent drying and subsequent double-stage electrostatic separation. The drying stage was added to ensure removal of excess moisture, as this is known to reverse particle charge (Baltrus et al., 2002; Kelly and Spottiswood, 1982; Wills and Napier-Munn, 2006; Woollacott and Eric, 1994).

A CoronaStat electrostatic separator housed at the University of Pretoria was used with an operating voltage of 20 kV and a rotor speed of 40 rpm. Although a triboelectrostatic separator is known to extract carbon material more efficiently (Bada et al., 2010), this type of equipment was unavailable in South Africa. The CoronaStat had five collection bins, namely conductive (C), middlings 1 (M1), middlings 2 (M2), middlings 3 (M3), and non-conductive (NC). The cumulative carbon grades and carbon recoveries for each bin were obtained and, based on this; an ideal cut-point was obtained.

Next, the material in the selected bins was exposed to a magnetic separator. A Frantz dry magnetic separator housed at the University of Johannesburg was used with an intensity of 0.5 A. The non-magnetic fraction was then submitted to a density (sink-float) separation technique. The density separation trials were conducted at Poland, following ISO 7936: 1992. Densities between 1.4 and 2.0 SG were tested with 0.1 SG intervals.

The procedure was repeated in South Africa on the South African fly ash. The cumulative carbon grades and carbon recoveries for each SG were obtained and, based on this, an ideal cut-point was selected.

Figure 2: Process flow diagram for the beneficiation of char from a South African fly ash.

2.3 Char characterization

The beneficiated char fractions from each country (Portugal, Poland, South Africa, and Romania) were subjected to characterization. In Table 4, the sample identification / type, origin, and carbon grades are given. Although the LOI was used to determine the carbon grade for the beneficiation of the South African fly ash sample, for uniformity, carbon grades for samples from all four countries are reported on an air-dried basis using fixed carbon (FC) (Table 2). Although LOI is a good indication of char in ash (and relative easy to apply), it is not completely accurate (Mohebbi et al., 2015). During ignition, evaporation of surface moisture, dehydration of chemically bonded water (e.g. portlandite), volatilisation of organic compounds (e.g. condensed tars), and decomposition of carbonates (e.g. calcite) will contribute, together with the oxidation of carbon, to the LOI value (Brown and Dykstra, 1995; Fan and Brown, 2001). LOI thus overestimates the char in ash.

Table 4: Sample identification / type, origin, and carbon grades subsequent to the beneficiation of char.

Proximate, ultimate, and carbon forms analyses were implemented to determine the elemental composition of the char samples. Proximate and ultimate standards were provided in Table 2. A Bureau Veritas (South Africa) in-house method was used for carbon forms analysis, to determine total carbon, elemental carbon, organic carbon, and inorganic carbon. Ideally, the char samples should have high elemental carbon percentages; indicative of a high degree of graphitization.

The char was prepared for petrographic assessment following ISO 7404-2, and the microscope and procedures as used for the ash samples were used. The morphological composition of the char samples was determined by using the classification method proposed in Table 3 (information on the anisotropy of char), the International Committee for Coal and Organic Petrology (ICCP) char classification (see Suárez-Ruiz and Valentim (2015) and Wagner et al. (2018)) (information on the porosity of char), and reflectance analysis for graphitic material (see Li et al., 2018). The latter involved the determination of mean random reflectance, the determination of maximum reflectance, the calculation of the minimum reflectance (Equation 1), the calculation of the bireflectance (Equation 2), and calculation of the anisotropy (Equation 3).

Reflectance calibration standards include a strontium-titanium 5.37, cubic-zirconia 3.24, YAG 0.900 with NG1 black glass, all supplied by Klein & Becker.

To determine the structural ordering of the char samples, Micro Raman Spectroscopy was used. Currently the char samples are also subjected to XRPD and ss-NMR analyses, but these results are not available as yet.

For Micro Raman Spectroscopy, the same blocks as per petrography were used, following re-polishing to remove the residual immersion oil. A WITec Alpha300 R confocal Raman microspectrometer (UJ Laser Raman Spectroscopy Laboratory), calibrated with a silicon standard, operating with a green line of argon (𝜆=532 nm) as excitation source, a laser strength of 3.5 mW, and a 50× objective, was used. The integration time was set at 30 seconds with a total of five cycles. At least 25 spectra were randomly obtained for each sample (different particles or different spots in individual particles). The different carbon forms were not taken into account. Information on typical Raman spectra for carbon rich particles is presented in Table 5. Currently, curve fitting is being done on the 1st order spectra using OriginPro software and results will be made available shortly. In this paper, only qualitative results will be discussed.

Table 5: Summary on 1st Order Micro Raman Spectroscopy bands for carbonaceous material.

3. Results and discussion

3.1 Coal and ash characterization

The ash samples, as well as their corresponding feed coal samples, were analyzed for their elemental (Table 6), mineralogical (Table 7), and morphological (Tables 8 and 9) compositions, structural ordering, as well as the PSDs for the ash samples (Table 11).

The ash percentage for the Portuguese coal sample is low, indicating a high quality (washed) coal being utilized in the power plant; it was imported from Colombia. The ash percentage for the South African coal is relatively high, indicative of the typical low grade coals being burned in South African power plants (South African Coal Roadmap, 2011).

For the ash samples, it can be seen that the fixed carbon percentages are all below 15 a.d.b wt. %. The char levels are thus low, which complicates the beneficiation of char and shows low volumes of available char material to be recovered.

The Al2O3 and SiO2 percentages are high for both the coal and ash samples. For the coal samples, this is a direct reflection of the large amount of clay (kaolinite and illite) and quartz minerals present in the samples. For the ash samples, the amorphous and mullite transformation products, as well as the quartz relics, are responsible for the high Al2O3 and SiO2 oxide percentages. The Poland coal sample also has a high CaO content relative to the other two samples.

All four ash samples can be classified as “F” ashes, meaning the Al2O3, Fe2O3, and SiO2 percentages are collectively higher than 70 wt. % for all samples, and the CaO percentage is lower than 10 wt. % (ASTM C618, 2012).

Table 6: Elemental composition results for the coal and ash samples (a.d.b – air-dried basis, d.a.f – dry ash free).

Quartz, kaolinite and illite are the major mineralogical phases present in the coal LTA samples. For the South African coal sample, the occurrence of illite is not as pronounced as for the other two coal samples. However, the kaolinite content for this sample is markedly higher. The Portuguese coal sample and Polish coal sample have the highest and lowest quartz contents at 33.1 wt. % and 14.0 wt. % respectively. The high calcite content for the Polish coal sample is reflected in the high CaO content.

The amorphous (glassy) phase is dominant in the ash samples. During combustion, the temperature of the coal particles greatly exceeds that of the furnace, leading to the melting of included minerals (Bryers, 1986; Flagan and Seinfeld, 1988). When expulsed from the furnace, the molten minerals will rapidly cool to form solid particles. Due to the rapid cooling, crystallization is limited and hence the amorphous nature of most of the ash (Malhotra, 2008). Quartz occurs in all four ash samples. The South African fly ash sample has considerably lower amorphous and higher quartz contents compared to the other samples. Other ash phases include: mullite, hematite, and maghemite. Romanian bottom ash also hosted some anorthite crystals (from the plagioclase group). Anorthite is commonly found in bottom ash samples and forms due to the reaction of reactive aluminium silicates (metakaolinite) with lime over an extended period of time at a temperature of approximately 1200 °C, which is characteristic of the bottom of the pulverised fuel boilers (Matjie et al., 2008; Menéndez et al., 2013).

Table 7: Mineralogical composition results for the coal and ash samples (wt. %).

The petrographic results for the coal samples are provided in Table 8, and results for the ash petrographic characterisation are presented in Table 9.

Table 8: Morphological composition results for coal samples (vol. %).

Table 9: Morphological composition results for ash samples.

Portuguese fly ash consists mainly of anisotropic char (Figure 3), whilst Romanian bottom ash consists mainly of isotropic char. Vitrinite macerals in coal will either form anisotropic or isotropic char upon combustion (Hower, 2012). The predominance of anisotropic over isotropic char depends on whether the vitrinite passed through a softening stage or not during combustion (Hower, 2012). The Portuguese coal sample has a high vitrinite content of 74.6 vol. % and is of bituminous rank origin (Table 8), which softens upon heating, explaining the higher anisotropic char content in the resultant fly ash. Although the coal from Romania was not provided or analysed, it is known that it is the Oltenia lignite (Gorj County, Romania) origin as this is the typical feed to this power plant. Lignite and other low-rank coals do not pass through a softening stage (Hower, 2012), and therefore the dominance of isotropic char in the Romanian bottom ash sample.

For the South African fly ash, the organic content consists of mainly inertinitic material. Inertinite macerals in coal will pass through the boiler with only limited amount of devolatilization and combustion (Hower et al., 2017). The South African coal has a relatively high inertinite content (48.6 vol. %) that is being reflected directly in the resultant fly ash sample.

The amorphous / glassy phase is dominant in the ash samples, in agreement with the XRD results. Solid glassy spheres, cenospheres, pseudoplerospheres, and angular glass particles were all observed in the ash samples. Portuguese fly ash showed spherical glass particles, mostly pseudoplerospheric due to the infilling of vacuoles of supermicrometric glassy spheres by micrometric glassy spheres (Figure 4A) (Valentim et al., 2016). The South African fly ash however, showed more angular glass particles of which most were of a transformation state between baked clay and glass (Figures 4B and 4C).

A fraction of the glass particles contained surface elongated and needle crystal inclusions. According to Watt and Thorne (1965), these crystals are mullite (and anorthite in the case of bottom ashes), whilst Fisher et al. (1978) suggested that it might be formed from heterogeneous nucleation on the surface of molten ash. Although mullite and anorthite were depicted as major phases with XRD analyses, it was very difficult to detect microscopically due to their interwoven nature within the glass phase.

Quartz, dense iron, and dendritic iron are the remaining inorganic phases identified. The quartz content for South African fly ash is much higher than for the other samples.

Figure 3: Example of the anisotropy of char observed in Portuguese fly ash (reflected-light, ×500, oil objective). Anisotropic char is distinguished from isotropic char by a 90° turn of the analyzer.

Figure 4: Plerosphere glass particle (A); angular glass particle showing a baked clay nature (B); and baked clay with a glassy rim (C) (reflected-light, ×500 oil objective).

For structural characterization of the coal and ash samples, XRPD and ss-NMR analyses were used. The samples were structurally divided into three different categories, considering the qualitative composition in terms of crystalline patterns observed in the XRPD profiles that were compared using the XRPD data bank (Oxford) (Figure 5).

Figure 5: Representative XRPD patterns for groups I (A), II (B), and III (C).

Group I: Coal samples from Portugal, Poland, and South Africa were categorized into Group I. This group presents an intense crystallographic signal associated with graphite content at 12.3 (2°) among other signals that can be observed in Figure 5. The other chemical substances present in the samples were mainly quartz (SiO2), mullite (Al6Si2O13) and tridymite (SiO2). The 13C DP-MAS and 13C CP-MAS ss-NMR spectra for this group is presented in Figure 6, whilst the quantitative 13C DP-MAS results are presented in Table 10. The amount of aliphatic and aromatic contents varied between 15-30% and 70-85% respectively (Table 10). From the analysis of the 13C ss-NMR spectra, it is possible to conclude that in general the coal content is related with typical carbon graphitic segments at around 125 ppm with the present of aliphatic segments between 0-40 ppm. No evidence of oxygenated carbon corresponding to the C sp3 region (like -CH2-O-) can be seen, considering the absent of chemical shift signals around 60-90 ppm. Regarding the oxygenated carbon corresponding to the C sp2 region (like phenolic groups), it might be associated with the signal at 150-160 ppm together with some other aromatic carbons bounded to the aliphatic regions.

Table 10: Aliphatic and aromatic carbon compositions obtained from the 13C DP-MAS ss-NMR experiments for Group  I samples.

Figure 6: 13C DP-MAS (A) and 13C CP-MAS (B) ss-NMR spectra for Group I samples, 13C DP-MAS (C) ss-NMR spectrum for Group II and Group III samples, and XRPD patterns for Group I (D), Group II (E), and Group III (F) samples.

Group II: Ash samples from Portugal, Poland, and South Africa were categorized into Group II. In this group, the organic composition was significantly lower than for Group I (Figure 6). For that reason, the 13C ss-NMR experiments did not show any carbon resonance signals. Moreover, the XRPD spectra do not show signals corresponding to the diffraction pattern of graphite. In fact, the components were mainly quartz (SiO2) and mullite (Al6Si2O13). Considering the high inorganic content, 1H-, 29Si- and 27Al- ss-NMR spectra were done in order to complement the crystallographic results (Figure 7). The 1H ss-NMR spectrum shows a broad and intense signal at 6.7 ppm together with a less intense signal at 2.4 ppm assigned to the amorphous graphitic segment (C-H sp2) and aliphatic hydrogen content (C-H sp3) respectively. The signals corresponding to different water populations in the solid state varied among the ash samples in terms of intensity and chemical shift, however, its presence can be clearly assigned due to their narrow resonance signals or due 2D heteronuclear correlation experiments (2D HETCOR) (Figure 8) (Travlou et al., 2018). The 29Si ss-NMR spectrum shows the typical broad signals of aluminosilicate in the chemical shift range of 80 to 110 ppm associated to “Si(nAl)” centers superimposed with the signals corresponding to quartz at 107 ppm together with some silicon material with different degree of condensation like Q2, Q3 and Q4 bands at 80 and 110 ppm (Barnes et al., 1986). The 27Al ss-NMR spectra indicate tetracoordinated aluminum at 57 ppm as the majority content (80%) with a minor content (20%) of hexacoordinated aluminum at around 5 ppm (Cecilia et al., 2017).

Group III: The ash sample from Romania was categorized into Group III. This group presents the absence of diffraction peaks associated with graphite, together with no signals regarding the range between 15-20 (2) (Figure 5). The typical signals at low angles for quartz and mullite are not observed since the high background corresponding to the amorphous silicate content masks the less intense diffraction peaks of these crystalline materials. The content of quartz, coesite and sillimanite is preponderant.

Figure 7: XRPD (A), 1H-MAS (@35 kHz) (B), 27Al-hpdec (C), and 29Si-hpdec (D) spectra corresponding to the ash samples from Group II and Group III.

Figure 8: 1H-MAS (@30 kHz) spectrum (A), and 1H-dimension extracted from the 2D 1H-13C HECTOR experiment (B) for ash samples from Group II and Group III.

In Table 11 the PSD results for the ash samples are given. The Polish bottom ash has a d50 of ~55 μm, which is relatively small compared to typical bottom ash samples as well as the Romanian bottom ash sample. The latter has a d50 of ~120 μm. The South African fly ash has a d50 of ~135 μm, which is relatively coarse compared to typical fly ash samples and the Portuguese fly ash sample (d50 ~45 μm). In fact, this sample has an even larger d50 value than the Romanian bottom ash. It is suspected that inefficient milling is the cause of this behaviour. Large quartz and char particles occurring in this sample strengthens this possibility.

Table 11: Cumulative PSD results for the ash samples.

The fractional carbon distribution for the fly ash samples are presented in Figure 9. It can be seen that the carbon content increases with an increase in particle size. This implies that possible pre-concentration of char can take place via a simple size separation scheme.

Figure 9: Fractional carbon distribution with size for fly ash samples.

The PSDs and carbon distribution results for the ash samples can also be accessed from Cruceru et al. (2017), Santos et al. (2017), and Valentim et al. (2018).

3.2 Beneficiation of char

A summary of the carbon grades and carbon recoveries obtained from the beneficiation of the South African fly ash sample is given in Table 12. The beneficiation techniques and results for the char concentrates from Portugal, Poland, and Romania can be found on the Charphite website (Santos et al., 2019).

Table 12: Carbon grades and carbon recoveries obtained from the beneficiation of the South African fly ash sample.

Overall, the carbon grade increased from 7.04 wt. % LOI to 82.91 wt. % LOI. Although the carbon recoveries for the individual steps were high, the overall carbon recovery was low at 18.92 %. The reason for the low recovery might be threefold. Firstly, all trials were conducted in batch mode (laboratory mode); once recycling streams are added higher recoveries can be expected. Secondly, small glass particles were embedded in the char matrix, resulting in an unliberated nature (Figure 10). Magnetic and density separation processes are sensitive to these types of particles and even the smallest inclusion can lead to the char particle ending up in the discard stream. Pulverization of these particles for liberation purposes will only lead to excessive fines forming and the “mini-me” effect pertaining (McMillan et al., 2015). Lastly, the fineness of the ash size distribution might lead to the agglomeration of particles. Ash particles subsequently behave as a group during beneficiation, and not as individual particles.

Figure 10: Glass embedded in massive / dense char (reflected-light, ×500 oil objective).

3.3 Char characterization

The beneficiated ash samples from Portugal, Poland, South Africa, and Romania were characterized according to their elemental (Table 13), morphological and reflectance (Tables 14 and 15) compositions, and structural ordering.

The chemical analysis shows that the Romanian sample has a different elemental composition than the samples from Portugal, Poland, and South Africa. It has a high volatile matter content, which leads to a subsequent high organic carbon percentage. The elemental carbon percentages for the other three samples are high. The carbon contents (as determined via ultimate analysis) range between 67 and 75 a.d.b wt. %.

Table 13: Elemental composition results for the beneficiated char samples.

The Portuguese char concentrate shows a higher anisotropic char content, which can lead to the conclusion that this sample, is more matured and ordered than the other three samples. The char from the Portuguese concentrate is very porous, with a high percentage of cenospheres (crassispheres) being present (Figure 11A). The Polish char can be described as mixed anisotropic and isotropic, with very large particles, mostly of networks and mixed porous nature (Figure 11B). The South African char sample contains mostly anisotropic and inertinite char, and mixed porous particles with rare glassy droplets (Figure 11C). The Romanian char is the most disordered, with high percentages of isotropic char and unreacted / partially burned coal. This explains why the volatile matter and organic carbon contents are higher for this particular sample, as it consists of a high percentage of coal particles (45.2 vol. %). The Romanian char sample also had a higher percentage of fusinoid / solid than the other three samples and was fragmented due to pulverisation during beneficiation (Figure 11D).

The bireflectance determined for mixed anisotropic and isotropic char particles ranges between 0.206 (Romania) and 2.154 (Poland). The values are very low, seeing that for graphite the bireflectance ranges between 6 and 27 (Craig and Vaughan, 1994); however, the values may increase following graphitization. The anisotropy percentages for mixed anisotropic and isotropic char particles range between 5.07 % (Romania) and 25.40 % (Poland). The low anisotropy for Romanian char is due to this sample’s high isotropic content. If only anisotropic particles are taken into account, the South African char has the highest degree of anisotropy, namely 40.28 %, and a high bireflectance.

Table 14: Morphological composition results for the beneficiated char samples (vol. %).

Figure 11: Typical char particles found in samples from Portugal (A), Poland (B), South Africa (C), and Romania (D) (Reflected-light, oil immersion, ×500).

Table 15: Reflectance results for the beneficiation of char samples.

In Figure 12A a typical Raman spectrum obtained for the char particles in the Portuguese, Polish, and South African samples is shown. All five Raman bands (G, D1, D2, D3, and D4) can be seen. For disordered material, some of the bands can merge, e.g. the G and D2 bands, as seen in Figure 12A. These two bands are located close to each other and for more disordered material the bands will be broad, ultimately merging (Kouketsu et al., 2014). Due to the strong presence of the other disordered (D) bands, it is most likely that the D2 band is dominant in this “merged” band. Following the classification from Kouketsu et al. (2014), the chars from Portugal, Poland, and South Africa were subsequently grouped as “transitional” carbonaceous material; implying they are an intermediate rank between amorphous and graphitic carbonaceous material.

A typical Raman spectrum obtained for the Romanian char is shown in Figure 12B, where there is merging of the D4 and D1 bands and the presence of a broad and intense G / D2 band. Following the classification by Kouketsu et al. (2014), the char from Romania was subsequently grouped as “amorphous” carbonaceous material.

Figure 12: Micro Raman Spectroscopy spectra of Portuguese, Polish, and South African (A) and Romanian (B) char.

4. Conclusion and recommendations

The aim of the Charphite project is to determine if the char fraction found in coal fly ash and bottom ash can be a precursor material for synthetic graphite (and subsequent usage in green energy applications). The coal and ash samples from a Portuguese, a Polish, a South African, and a Romanian power plant were characterized, techniques used to beneficiate the char from a South African fly ash were discussed, and the characterization results on the beneficiated char from Portugal, Poland, Romania, and South Africa were discussed.

The char percentages in the parent ash samples were low, indicating possible difficulties with beneficiation and low volumes of material to ultimately be recovered. All four ash samples were characterized as Class F, and major mineral phases included glass and quartz.

The South African fly ash sample was used as a case study to demonstrate the beneficiation of char steps. It was found that the carbon grade for the South African fly ash increased from 7.04 wt. % up to 82.91 wt. %, using a series of size, electrostatic, magnetic, and density beneficiation steps. The carbon recovery, however, was low, at 18.92%. The low carbon recovery might be due to embedded glass inclusions, batch mode operation, or agglomeration of fines.

From the beneficiated char characterization results, the Romanian sample is the least likely to graphitize due to its high volatile content, high organic carbon content, and isotropic nature. Micro Raman Spectroscopy classified this sample as “amorphous”, with the presence of a broad G band. The sample from Portugal is the most likely to graphitize with high percentages of elemental and anisotropic char. Micro Raman Spectroscopy classified this sample as “transitional” material.

The beneficiated char samples are currently being graphitized at the Department of Instituto Nacional del Carbón (INCAR-CSIC), Spain.

Source: Charlotte J. Badenhorst1, Nicola J. Wagner1, Bruno R.V. Valentim2, Ana C. Santos2, Alexandra Guedes2, Barbara Białecka3, Joanna Całus- Moszko3, Luminiţa G. Popescu4, Mihai Cruceru4, Georgeta Predeanu5, Karel S. Viljoen1, Juan M. Lázaro-Martínez6, Traian A. Abagiu4

1 University of Johannesburg, Johannesburg, South Africa
2 University of Porto, Institute of Earth Sciences, Porto, Portugal
3 Główny Instytut Górnictwa, Katowice, Poland
4 University “Constantin Brâncuşi” of Târgu Jiu, Târgu Jiu, România
5 University POLITEHNICA of Bucharest, Bucharest, Romania
6 University of Buenos Aires, Buenos Aires, Argentina


The 10 largest coal producers and exporters in Indonesia:

  1. Bumi Resouces
  2. Adaro Energy
  3. Indo Tambangraya Megah
  4. Bukit Asam
  5. Baramulti Sukses Sarana
  6. Harum Energy
  7. Mitrabara Adiperdana 
  8. Samindo Resources
  9. United Tractors
  10. Berau Coal

Click Here for Top World Clean Coal Technologies Contractors°

No comments:

Post a Comment

Related Posts Plugin for WordPress, Blogger...