Subscriber access provided by American Chemical Society Article Quantitative Evaluation of Minerals in Lignites and Intraseam Sediments from the Basin, Northern Nikolaos Koukouzas, Colin R. Ward, Dimitra Papanikolaou, and Zhongsheng Li Energy Fuels, Article ASAP • DOI: 10.1021/ef8010993 • Publication Date (Web): 09 March 2009 Downloaded from http://pubs.acs.org on March 20, 2009

More About This Article

Additional resources and features associated with this article are available within the HTML version: • Supporting Information • Access to high resolution figures • Links to articles and content related to this article • Copyright permission to reproduce figures and/or text from this article

Energy & Fuels is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Energy & Fuels XXXX, xxx, A

Quantitative Evaluation of Minerals in Lignites and Intraseam Sediments from the Achlada Basin, Northern Greece

Nikolaos Koukouzas,*,† Colin R. Ward,‡ Dimitra Papanikolaou,† and Zhongsheng Li‡

Centre for Research and Technology Hellas, Institute for Solid Fuels Technology and Applications, Mesogeion AVe. 357-359, GR-15231 Halandri, Athens, Greece, and School of Biological, Earth and EnVironmental Sciences, UniVersity of New South Wales, Sydney, NSW 2052, Australia

ReceiVed December 16, 2008. ReVised Manuscript ReceiVed January 27, 2009

Seven core samples (five lignite samples and two intraseam nonlignite rock samples) from the Achlada open-cut mine in northern Greece were characterized by X-ray diffraction (XRD) and X-ray fluorescence (XRF) techniques. Quantitative evaluation of the mineral phases in each sample was made from the powder X-ray diffractograms using Siroquant commercial interpretation software, which is based on Rietveld principles. The main minerals in the low-temperature ash (LTA) ash of the lignites are kaolinite and illite, with bassanite and quartz in minor proportions. The nonlignite rock samples mainly consist of illite, mica (2M1), and kaolinite (poorly ordered), along with quartz, chlorite (ferroan), feldspar (albite), rutile, and dolomite. Oriented-aggregate XRD study further shows the presence of smectite, and interstratified illite/smectite (I/S), in the clay fractions of the lignite and rock samples, with the mineral matter of the lignites being richer in kaolinite, smectite, and I/S than in mineral matter of the nonlignite materials. The differences in mineralogy between the lignite and the rock materials probably reflect selective concentration of minerals in the original peat during deposition, combined with authigenic precipitation of minerals such as kaolinte in the peat deposit. Inferred chemical analyses derived from the XRD data show reasonably good correlations with chemical data obtained by direct ash analysis, especially if the smectite and I/S are taken into account. This provides a link between mineralogical and chemical studies that may be valuable in evaluating the behavior of the lignite under different utilization conditions.

1. Introduction Achlada, and also the associated intraseam nonlignite sediments, Lignite is the major source for the generation of electricity to correlate the data with the chemical composition of the lignite in Greece, where it is used to produce some of the least- ash. The chemical composition of the combustion products from expensive and most cost-effective electric power within the this coal characteristically shows high concentrations of calcium European Community. There are 60 lignite basins in Greece,1 and sulfur, which are responsible for the unsuitability of the fly and the annual lignite production in 2006 was estimated at 62.50 ash in concrete production. For example, the concentration of Mt.2 According to data from the Public Power Corporation, this calcium for fly ash derived from the combustion of Achlada’s lignite ranges from 22% to 34%, while the concentration of lignite was consumed to generate 31977 GWh of electricity. 5 The exploitation of lignite in Greece has a very long history. sulfur fluctuates from 4% to 8%. Significant achievements and a large amount of experience, gained during many years of mining operations, have placed 2. Geological Setting the Greek lignite-mining industry in the leading position in The Achlada region is located in the eastern part of the 2,3 Europe. Basin in northwest (Greece), extending in a The Achlada open-cut mine is the most recently opened mine NNW-SSE direction from Monastiri (Former Yugoslav Re- in Greece, having started operation in 2001. Its lignite output public of Macedonia (FYROM)) up to the hills of ∼ is 2.5 Mt per year, all of which is used by the (330 through the cities of Florina, Amynteo, and Ptolemais (see 4 MW) power station. The lignite derived from the Achlada mine Figure 1). The basin is almost 100 km wide.6 is higher in rank than the lignite from other mines in the region, The largest lignite deposits of Greece, which were formed giving it a higher heating value for equivalent ash percentages. in the Monastiri-Florina-Ptolemais-Kozani graben, are clas- The objective of the present paper is to examine the sified into two types: Ptolemais-type “earthy” lignite and mineralogical composition of representative lignite samples from Komnina-type xylitic lignite.3 * Author to whom correspondence should be addressed. Tel.: +30 210 The Achlada lignite, which is of the xylitic type, is of Lower 6501771. Fax: +30 210 6501598. E-mail: [email protected]. Neogene age, along with the other lignites of the Florina † Centre for Research and Technology Hellas, Institute for Solid Fuels sedimentary basin (Vegora, Petres, , and Lofi). The Achlada Technology and Applications. ‡ School of Biological, Earth and Environmental Sciences, University lignite is older than the lignite of the Ptolemais type, which is of New South Wales. (1) Koukouzas, N. Miner. Wealth 1998, 106, 53–68. (5) Koukouzas, N.; Tsikardani, E.; Papanikolaou, D. Fly Ash Utilisation (2) Kavouridis, K.; Koukouzas, N. Energy Policy 2008, 36, 693–703. Programme (FAUP), Technology Information, Forecasting & Assessment (3) Koukouzas, N. Int. J. Coal Geol. 2007, 71, 276–286. Council (TIFAC), Department of Science & Technology (DST), 2005. (4) Koukouzas, N.; Vassilatos, C. J. Chem. Technol. Biotechnol. 2008, (6) Karakatsanis, S.; Koukouzas, N.; Pagonas, M.; Zelilidis, A. Bull. 83, 20–26. Geol. Soc. Greece 2007, Vol. XXXX, (Part 1), 76–84.

10.1021/ef8010993 CCC: $40.75  XXXX American Chemical Society B Energy & Fuels, Vol. xxx, XXXX Koukouzas et al.

Figure 1. Geological map of the area studied, modified from the Institute of Geology and Mineral Exploration (IGME) Florina and Vevi geological map. of Upper Neogene age and is found in Ptolemais, Proastio, Perdikas, and Amynteo. The strata beneath the basin include the Pelagonian zone of Figure 2. Column section represented by the seven samples collected Palaeozoic and pre-Palaeozoic crystal schists, a Mesozoic for the study. carbonate cover, and ophiolites. The Neogene sediments that fill the basin contain lignite seams, on top of which fluviotor- Table 1. Thicknesses Represented by Lignite and Rock Samples rential or terrestrial deposits are present.7 The lignite-bearing sample code sample description interval thickness (m) sediments are Late Miocene to Pliocene in age. C1 xylite 3.5 The lowermost Neogene horizon, known as the Basal R1 marl 0.3 Conglomerate, begins the succession of sediments that fill the C2 xylite 5.0 - C3 marly lignite 1.5 basin. The middle horizon (the Vevi Achlada Formation), R2 sand 0.2 which is exposed in the Vevi and Achlada lignite mines, overlies C4 marly lignite 2.0 the Basal Conglomerate and is represented by alternations of C5 lignite 4.0 clayey, sandy, and marly sediments, as well as by lignite seams. Clayey diatomite and phosphatic nodules have also been Subsamples that were taken from the five lignite samples and identified in these sediments.8 The lignite seams and associated the two rock samples were ground to fine powder. Subsequent interseam sediments have a total thickness of ∼35 m. The analysis suggests that some of these were not fully equivalent to uppermost horizon (Lofi Formation), overlying the Vevi-Achlada the samples detailed in Tables 1 and 2, because of inhomogeneities Formation, continues the Neogene succession, with alternations in the bulk material from which they were taken. However, they indicate a similar range of quality variation within the lignite seam. of clays, marly breccias, sandy conglomerates, and lime marl The powdered lignite samples were subjected to low-temperature beds. On the top of this sandy-clay horizon is a marly limestone oxygen-plasma ashing using an IPC four-chamber asher, as outlined bed, which covers the entire basin. A thin cover of limnodeltaic, in Australian Standard 1038, Part 22; the mass percentage of low- fluviotorrential, and terrestrial Quaternary sediments completes temperature ash (LTA) was determined in each case. Each LTA the lithological column in this part of the basin. was further powdered, and then analyzed via powder XRD, using a Philips Model PW1830 diffractometer with Cu KR radiation and a graphite monochromator. Diffractograms were run in a 2θ range 3. Materials and Methods of 2°-60°, with steps of 0.04° and a counting time of 2 s. Seven samples were collected from a mine exposure in the The powdered rock samples were also analyzed in this way, Achlada region, a column section of which is shown in Figure 2. without the low-temperature ashing procedure. Quantitative analyses These included five lignite samples and two rock samples. The total of the mineral phases in each LTA or rock sample were made from thickness of the section studied is 16.5 m, representing the upper the X-ray diffractograms using Siroquant, which is commercial part of the mining succession. The coding of the lignite and rock interpretation software9 that is based on principles originally samples is presented in Table 1. The five lignite samples are C1, developed by Rietveld.10 Further details of the software interpreta- C2, C3, C4, and C5, and the two rock samples are R1 and R2. tion process have been given by Ward et al.11 Proximate and ultimate analyses and heating value determinations were conducted, according to ASTM standards (see Table 2). The 4. Results chemical composition of the lignite ashes, prepared at 815 °C, as well as that of the rocks, was determined by X-ray fluorescence 4.1. Basic Lignite Properties. The Achlada lignite has a spectrometry, using a Philips PW 2400 spectrometer and associated fixed carbon content in the range of 25.5% and 45.6%, daf SuperQ software. (9) Taylor, J. C. Powder Diffr. 1991, 6, 2–9. (7) Kotis, Th.; Koukouzas, N.; Papanicolaou, C.; Foscolos, A.; Stamatakis, (10) Rietveld, H. M. J. Appl. Crystallogr. 1969, 2, 65–71. M. Ann. Geol. Pays Hell. 2004, 40, 143–158. (11) Ward, C. R.; Matulis, C E.; Taylor, J. C.; Dale, L. S. Int. J. Coal (8) Koukouzas, N. Miner. Wealth 1992, 81, 39–52. Geol. 2001, 46, 67–82. Lignites and Sediments from Northern Greece Energy & Fuels, Vol. xxx, XXXX C

Table 2. Analysis Results and Energy Contents for the Five between 5320 kcal/kg to 6250 kcal/kg (daf) (see Table 2b). Lignite Samples C1-C5 (a) on a Dry Basis and (b) on a Dry, These calorific values are higher than those of other lignites in Ash-Free (daf) Basis the region (e.g., Ptolemais 5630-6100 kcal/kg, daf), because Value of the higher rank levels. parameter C1 C2 C3 C4 C5 Besides the classical methods of analysis (Table 2a), which (a) Analysis Results and Energy Contents, Given on a Dry Basis may be considered in the homogenization processes of lignites proximate analysis derived from different deposits, mineralogical analysis of the (wt %, dry basis) feed lignite and chemical analysis of the lignitic ash are also ash 41.9 41.4 59.6 52.4 38.0 important to predict the ash behavior at high-temperatures in volatile matter 36.7 37.2 29.6 35.5 33.6 fixed carbon 21.4 21.4 10.8 12.1 28.4 boilers and the design of electrostatic filters. The chemical ultimate analysis composition of the ash is used to calculate the slagging and (wt %, dry basis) fouling indices (SI and FI) that affect boiler performance.14 The H 3.0 2.8 1.9 2.0 2.6 slagging index (SI) is the ratio of basic oxides to acidic oxides C 36.8 38.9 24.6 28.6 39.9 determined in the ash divided by the percent of combustible CO2 0.4 0.5 0.5 0.6 0.4 N 0.6 0.9 0.9 0.8 0.9 sulfur. The fouling index (FI), which indicates the tendency of S 1.1 1.1 1.5 2.8 2.4 the ash particles to adhere to superheater tubes, is estimated by O 17.8 16.3 13.5 14.8 17.5 taking the ratio of basic oxides to acidic oxides multiplied by gross heating value 3658 3869 2358 2699 3872 the Na2O content. Taking into account the results presented later (kcal/kg, dry basis) ) - net heating value 3463 3667 2317 2533 3662 in Table 6, the slagging index is evaluated at SI 0.2 1.5, (kcal/kg, dry basis) while the fouling index is calculated at FI ) 0.00-0.03, which (b) Analysis Results and Energy Contents, Given on a Dry, Ash-Free are within the accepted values, according to the specification (daf) Basis of the Public Power Corporation of Greece (FI < 0.5 and SI < proximate analysis 2.5) for the lignite-fired power plants.7 (wt %, daf) Apart from ash composition, the minerals that remain in the volatile matter 63.1 63.6 68.7 74.6 54.2 combustion chamber, because of partial or incomplete decom- fixed carbon 36.9 36.4 31.3 25.5 45.6 gross heating value 6292 6604 5835 5668 6247 position, should also be considered. (kcal/kg, daf) 4.2. Mineral Matter in Lignite and Nonlignite Samples. net heating value 5949 6256 5732 5320 5908 Tables 3 and 4 provide details of the percentages of minerals (kcal/kg, daf) in each LTA and rock sample, from the Siroquant interpreta- tions. The tables list the estimated weight percentage of the (Table 2b), which is similar to that of Ptolemais lignite individual phases recognized in each sample, together with the - (36% 48%, daf); the volatile matter content of Achlada lignite relative error in the estimation (estimated standard deviation, - is in the range of 54.2% 74.6%, daf (see Table 2b), whereas or ESD) for each individual determination. The overall level - 12 that of Ptolemais lignite is in the range of 54% 61%, daf. of fit for the Siroquant evaluation is given by the relevant global The volatile matter percentage, in particular, is used as a basic chi-squared (2) value, at the bottom of the table in each case. parameter in the power station to select appropriate combustor The total error for each mineral percentage can be calculated operating conditions that will minimize the loss of unburned from the product of the ESD associated with that mineral and fuel and maintain good flame stability. Achlada lignite exhibits the square root of the global 2 value for the Siroquant analysis. ∼ a lower moisture content ( 40%, on an as-received basis), For some minerals (e.g., rutile), the estimated percentage may ∼ compared to Ptolemais lignite ( 55%, on an as-received basis). be close to or below the indicated error for the determination. - The carbon content is in the range of 24.55% 39.94%, the The mineral noted in such cases, if present at all, occurs in - hydrogen content is 1.86% 2.96%, the nitrogen content is proportions essentially below the limit of detection of the XRD - - 0.57% 0.93%, the oxygen content is 13.45% 17.54%, and the system. The relevant data have been included, however, to - sulfur content is 1.06% 2.78%. The carbonate carbon (CO2) ensure that the possible presence of these phases has not been - content is in the range of 0.39% 0.64%. overlooked. The lignite samples in the present study have nitrogen Table 3 indicates that kaolinite and illite are the dominant < contents of 1%, which is greater than the nitrogen content of phases in the LTA of the lignite samples, with minor proportions Ptolemais lignite (0.4%, daf). of quartz and bassanite. As described by Ward et al.,11 the The sulfur of the lignites studied varies from 1% to slightly 1 bassanite (CaSO4 · /2H2O) was probably formed primarily by less than 3% (as-analyzed), which is greater than the sulfur interaction between organic sulfur and organically associated ∼ content of Ptolemais lignite ( 1.5%). Most sulfur in Greek calcium in the lignites during the low-temperature ashing 13 lignites is found in the form of pyrite and marcasite. process. Some of the bassanite might also have been formed Worldwide, the sulfur content of lignites varies between 0.2 by the decomposition of any gypsum present in the coal before wt % and 10 wt %, although internationally traded lignites are low-temperature ashing, including precipitates produced by the typically sold at 1 wt % or less. In combustion applications, interaction of calcium and sulfate (SO4) in the pore waters with the sulfur causes corrosion problems in the boiler system. It drying of the material.15 also causes problems with atmospheric pollution, because large By contrast, illite and mica are the dominant minerals in the amounts of SOx are released with the stack gases. nonlignite rock samples, with kaolinite, quartz, chlorite (ferroan), - The gross heating value is in the range of 5835 6600 kcal/ and feldspar (albite) as minor (but significant) components. The kg (daf) (see Table 2b), and the net heating value fluctuates mica is distinguished from the illite by a shaper (001) diffraction peak, and it is identified from the remainder of its pattern as (12) Anastopoulos, I. X.; Koukouzas, C. N. Geological and Geophysical Studies; Institute of Geological and Mineral Exploration: Athens, Greece, 1972; Vol. XVI, pp 117-121. (14) Cudmore, J. F. Coal Geology and Coal Technology; Ward, C. R., (13) Koukouzas, N.; Skounakis, S. Bull. Geol. Soc. Greece 1990, 25/2, Ed.; Blackwell Scientific Publications: Oxford, U.K., 1984; pp 113-150. 193–201. (15) Ward, C. R. Int. J. Coal Geol. 1991, 17, 69–93. D Energy & Fuels, Vol. xxx, XXXX Koukouzas et al.

Table 3. Mineralogy of Lignite Samples C1-C5 by X-ray Diffraction and Siroquant Value parameter C1 C2 C3 C4 C5 LTA (wt % of lignite, as analyzed) 67.3 52.5 81.2 40.8 49.6 mineral phase (wt %)a quartz 5.0 (0.2) 4.5 (0.1) 3.7 (0.1) 4.1 (0.2) 7.8 (0.2) kaolinite (poorly ordered) 46.8 (0.8) 38.3 (0.6) 47.1 (0.8) 41.5 (0.8) 44.6 (0.8) illite 40.8 (0.8) 45.4 (0.6) 41.7 (0.8) 37.4 (0.9) 38.0 (0.9) bassanite 7.4 (0.5) 11.8 (0.3) 7.5 (0.5) 17.1 (0.5) 9.6 (0.5) global chi-squared value, 2 7.59 6.39 6.23 9.97 7.24 a Value given in parentheses is the estimated standard deviation (ESD). Table 4. Mineralogy of Rock Samples R1 and R2 by X-ray with the other sediment, may possibly be explained by Diffraction and Siroquant breakdown of the mineral in the peat-forming environment. The Value greater abundance of kaolinite in the LTA of the lignites, relative parameter R1 R2 to the nonlignite sediments, which is also noted in other coal 17,20,21 mineral phase (wt %)a seams, probably reflects authigenic precipitation of Si and quartz 11.1 (0.3) 12.0 (0.3) Al within the pores of the peat deposit. kaolinite (poorly ordered) 17.4 (0.5) 13.2 (0.5) The clay fractions (effective diameter of <2 µm) of each rock chlorite. ferroan 9.3 (0.5) 5.5 (0.4) sample and three of the LTA samples were isolated by ultrasonic mica. 2M1 20.0 (0.7) 28.6 (0.8) illite 37.8 (0.7) 32.5 (0.8) dispersion in sodium hexametaphosphate (Calgon) and subse- feldspar (albite) 3.4 (0.6) 7.2 (0.5) quent settling. Oriented aggregates of each clay fraction were rutile 0.3 (0.2) 0.9 (0.2) prepared using the pipet-on-glass-slide technique,22 and the clay dolomite 0.6 (0.3) fractions were further investigated by XRD of the oriented 2 global chi-squared value,  9.11 8.83 aggregates, using glycol and heat treatment.23,24 The relative a Value given in parentheses is the estimated standard deviation proportions of the different clay minerals in this fraction for (ESD). each sample were determined by the method of Griffin.25 < mainly representing the 2M1 polymorph. Traces of rutile and Table 5 provides data on the clay minerals in the 2-µm dolomite are also indicated in some of the rock samples. fractions. Although it is not clear from the powder diffractogram Although there is a significant variation in the LTA (and ash) data (Tables 3 and 4), the samples also contain a significant percentage, as well as in lignite type (lignite, marly lignite, and proportion of expandable clay minerals, identified from their xylite), the LTA isolated from each of the lignites contains XRD patterns as smectite and irregular illite/smectite (I/S). similar proportions of the various minerals concerned. Bassanite These are more abundant, relative to illite, in the LTA of the is less abundant in the lignites with the highest LTA percentages lignite samples than in the nonlignite rock materials. Minor (C1 and C3), which probably reflects a lower abundance of proportions of chlorite are also indicated in the nonlignite rocks, organic matter with which the calcium and sulfur were originally but not in the LTA of the lignite samples. 26 associated. Kaolinite is also slightly more abundant in these Oikonomopoulos et al. have reported a similar mineralogy samples, but the difference is not as marked and is not for a series of intraseam nonlignite beds in the Achlada area. necessarily significant. The percentage of quartz is low (4%-5%) As discussed by those authors, the intercalated nonlignite in most of the LTAs, but is slightly higher in the LTA of the materials may be suitable raw materials for the production of basal lignite sample in the succession. Although textural data red stoneware and related ceramic products. The greater are currently not available, based on parallels with other studies, abundance of illite and the lower proportion of kaolinite and this may reflect a small admixture of detrital sediment to the expandable clays in these materials, relative to the clays in the original peat,16,17 or it may indicate a concentration of authigenic lignites, indicated in the present study, would also be associated quartz near the bottom of the original peat bed.18,19 with lower fusion temperatures and lesser degrees of shrinkage Despite differences in lithology (one is marl and one is sand), than similar products derived from the lignite mineral matter. the mineralogies of the two rock samples are also very similar. The differences between the clay minerals in the lignite LTA, The sand sample (R2) has slightly more feldspar and mica, compared to the nonlignite bands within the seam, may reflect whereas the marl (R1) has slightly higher proportions of clay a greater degree of degradation in the lignite swamp environment minerals (illite, kaolinite, and chlorite). Carbonate minerals are of the clay supplied generally to the basin. The higher proportion virtually absent from both materials, with only a trace of of smectite and illite/smectite in the LTA of the lignites, relative dolomite indicated in the marl sample. In this instance, at least, to the noncoal materials, may also be responsible for stickiness the absence of significant carbonate suggests that the marl may be better described as a mudstone. (20) Spears, D. A. In Coal and Coal-Bearing Strata: Recent AdVances; However, the greatest contrast is between the nonlignite rocks Scott, A. C., Ed.; London, 1987; pp 171-185. (Geological Society Special and the LTA of the lignite samples. The lower proportion of Publication.) (21) Ward, C. R.; Spears, D. A.; Booth, C. A.; Staton, I.; Gurba, L. W. quartz in the LTA of the lignites and the absence of mica and Int. J. Coal Geol. 1999, 40, 281–308. feldspar are consistent with the filtering of coarser-grained (22) Gibbs, R. J. Procedures in Sedimentary Petrology; Carver, R. E., particles from incoming sediment by the vegetation of the peat Ed.; Wiley: New York, 1971; pp 531-540. (23) Moore, D. M.; Reynolds, R. C., Jr. X-ray Diffraction and the bed. The absence of chlorite from the LTA, if it was introduced Identification and Analysis of Clay Minerals; Oxford University Press: Oxford, U.K., 1997; 378 pp. (16) Davis, A.; Russell, S. J.; Rimmer, S. M.; Yeakel, J. D. Int. J. Coal (24) Ruan, C. D.; Ward, C. R. Appl. Clay Sci. 2002, 21, 227–240. Geol. 1984, 3, 293–314. (25) Griffin, G. M. Procedures in Sedimentary Petrology; Carver, R. E., (17) Ward, C. R. Int. J. Coal Geol. 1989, 13, 455–479. Ed.; Wiley: New York, 1971; pp 541-569. (18) Sykes, R.; Lindqvist, J. K. Org. Geochem. 1993, 20 (6), 855–866. (26) Oikonomopoulos, I.; Perraki, Th.; Kaouras, G.; Antoniadis, P. Bull. (19) Susilawati, R.; Ward, C. R. Int. J. Coal Geol. 2006, 68, 171–195. Geol. Soc. Greece 2007, 40, 906–917. Lignites and Sediments from Northern Greece Energy & Fuels, Vol. xxx, XXXX E

Table 5. Mineralogy of the <2-µm Fraction of LTA and Rock Samples by Oriented-Aggregate XRD Techniques Value Coal LTA Samples Rock Samples parameter C1 C2 C3 R1 R2 composition (%) kaolinite 26 21 17 22a 23a illite 22 39 28 52 54 expandable clays 52 40 55 26 22 nature of expandable smectite + smectite + smectite + smectite + smectite + clay minerals irregular I/S irregular I/S irregular I/S irregular I/S irregular I/S a Plus minor chlorite. Table 6. Chemical Analysis of Coal 815 °C Ash Samples lignites, as indicated in Table 2a. A systematic relationship (C1-C5) and Rock Samples (R1 and R2) by X-ray Fluorescence between the 815 °C ash percentages and the proportion of LTA Spectrometry (XFS) for the lignite samples is indicated in Figure 3A, with correlation Content (wt %) studies suggesting that the proportion of LTA (mineral matter) Rock Samples is ∼1.25 times greater than the 815 °C ash yield. The difference Coal 815 °C Ash With LOI LOI-Free is due, in part, to the water of hydration associated with the component C1 C2 C3 C4 C5 R1 R2 R1 R2 clay minerals, and in part to the formation of bassanite, rather than anhydrite or lime (which would be formed at higher % ash 50.4 38.0 61.3 28.5 37.0 SiO2 53.1 51.2 53.8 47.2 51.7 45.2 51.8 55.7 57.4 temperatures), from the nonmineral calcium associated with the Al2O3 29.1 28.3 29.9 26.5 28.6 21.9 23.5 27.0 26.0 organic matter. Fe2O3 5.2 5.7 5.8 6.5 5.8 5.4 4.8 6.6 5.3 Figure 3B shows that the proportion of sulfur retained in the MnO 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 815 °C ash, expressed in Table 6 as SO3, is approximately TiO2 1.0 1.0 0.9 0.9 1.0 0.8 0.7 1.0 0.8 CaO 3.2 4.3 2.3 7.3 3.5 1.3 1.1 1.6 1.2 proportional to the CaO content of the same ash samples. This MgO 2.7 2.7 3.0 3.1 2.9 2.8 2.3 3.4 2.5 suggests that the extent of sulfur retention in the ash depends Na2O 0.1 0.0 0.1 0.1 0.0 0.2 0.7 0.2 0.8 not only on the sulfur content of the coal but also on the K2O 3.1 3.1 3.5 3.1 3.3 3.2 4.9 3.9 5.4 availability of Ca to fix at least part of the material in nonvolatile P O 0.2 0.2 0.1 0.3 0.3 0.2 0.3 0.2 0.3 2 5 form. Therefore, sulfur retention in the 815 °C ash is similar, SO3 1.9 2.8 1.2 5.7 2.5 0.2 0.2 0.2 0.2 LOI 19.6 10.8 in principle, to the formation of bassanite in LTA. The individual data points in Figure 3B plot to the right of a total 99.5 99.4 100.7 100.5 99.6 100.7 101.1 100.0 100.0 line drawn to represent the stoichiometric composition of and other handling problems, which are sometimes encountered anhydrite (CaSO4), for which the CaO/SO3 ratio is 0.7:1. The in the feed to the power station. separation of the data points from this line indicates that only 4.3. Ash and Rock Chemistry. Table 6 lists the major part of the calcium in these ashes has been used in the sulfur element oxide percentages present in the high-temperature (815 fixation process. The remainder of the calcium in the high- °C) ash of each lignite sample, and also of each nonlignite rock, temperature ashes presumably occurs as calcium-bearing phases with the latter being analyzed only after drying at 100 °C. The without a sulfur component. In the absence of other calcium- loss on ignition (LOI) for the nonlignite rocks was obtained by bearing minerals, such as calcite or feldspars (e.g., anorthite), heating the dried samples in air at 1050 °C. LOI was not the excess calcium may represent material such as poorly determined separately for the lignite ashes, as they had already crystallized lime (CaO) or portlandite (Ca(OH)2), which are been subjected to high temperatures for ash preparation. To derived from decomposition of the remaining calcium in the facilitate comparison with the lignite ashes, however, the results organic matter. for the rock samples were also normalized to remove the LOI The inferred chemical composition of each LTA or rock (i.e., expressed to an LOI-free basis). sample was calculated from the mineral percentages determined The samples have high concentrations of SiO2 and Al2O3, by the Siroquant evaluations (see Tables 3 and 4) and the with lesser (but, in some cases, significant) proportions of Fe2O3, expected chemical compositions of the individual minerals, MnO, TiO2, CaO, K2O, SO3,P2O5, MgO, and Na2O. The LOI which have been corrected to allow for dehydration and other mainly represents the loss of hydroxyl or lattice water from the changes associated with the high-temperature ashing process.20 clay minerals, which comprise most of the two rock samples. The proportions of each major element oxide inferred from the When allowance is made for the LOI, the rock samples have XRD data were then compared to the actual percentage of each slightly higher proportions of SiO2 and slightly lower propor- oxide determined directly by XRF analysis (Table 6), to evaluate ° tions of Al2O3 than the 815 C ashes of the lignite samples. more fully the links between the XRD quantifications and the The proportions of Fe2O3 are comparable in both sets of samples, ash analysis data. but the rocks have higher percentages of K2O and lower Previous studies of coal LTA by Siroquant, using an added- percentages of CaO and SO3 than the coal ash materials. These in proportion of crystalline ZnO to measure the percentage of differences are consistent with the higher proportions of quartz, amorphous material,27 has shown that the LTA is essentially illite, and mica in the rock samples, and the higher proportions completely crystalline, with minimal if any amorphous material, of kaolinite and bassanite in the LTA of the coal samples. suggesting that most of the material in the LTA in the present is represented by the crystalline phases identified from the XRD 5. Comparison of Mineralogy and Ash Chemistry data and confirming the validity of this approach in the present The proportion of mineral matter in the lignites, as indicated study. Typical published analyses were used to represent the by the LTA percentages (see Table 3), is somewhat greater than the conventional (815 °C) ash percentages of the respective (27) Ward, C. R.; Taylor, J. C. Int. J. Coal Geol. 1996, 31, 211–229. F Energy & Fuels, Vol. xxx, XXXX Koukouzas et al.

Figure 3. (A) Correlation between percentages of mineral matter (LTA) and high-temperature (815 °C) ash for the lignite samples and (B) correlation between CaO and SO3 in the high-temperature (815 °C) ash samples.

Figure 4. Correlation of percentages of selected oxides in coal ash or rock samples inferred from XRD data with actual oxide percentages (LOI- free basis), as indicated by XRF analysis. chemical composition of phases that may have variable chem- parable to the lignite ashes and the values inferred from the istry,21 such as illite and illite/smectite (I/S), in this evaluation XRD data. The diagonal line on each plot indicates a perfect process. correlation. The relationships between the inferred and observed percent- Figure 4 shows a relatively good correlation between the ages of the major oxides are plotted in Figure 4. For the lignite inferred and observed data for SiO2 and CaO, but an overesti- ashes, the observed values represent the percentages indicated mation of Al2O3 from the XRD data, relative to the observed in Table 3; for the rock samples, the data were normalized to Al2O3 percentages. This is, in part, because the expandable clay remove the LOI percentages, making the material more com- minerals (smectite and illite/smectite (I/S)) were included with Lignites and Sediments from Northern Greece Energy & Fuels, Vol. xxx, XXXX G the illite in the powder patterns from which the inferred chemical original peat during deposition, combined with authigenic data were derived. The illite composition used, as discussed by precipitation of minerals such as kaolinite in the peat deposit. Ward et al.,21 was based on a dioctahedral illite with no Because the nonlignite bands comprise only a very minor allowance for partial replacement of Al by Fe or Mg in the proportion of the total lignite seam, and because the lignite itself lattice structure. has a relatively high mineral matter content, mined products If there was such a substitution in the illite structure, and from the area would be expected to be dominated by the mineral also if the smectite contained some iron in its lattice instead of assemblages in the lignite beds, rather than in the intraseam aluminum, lower percentages of Al2O3 and higher percentages nonlignite sediments. of Fe O would have been included in the inferred chemical 2 3 A reasonably good correlation was obtained between the data. A plot showing the relationship between the total Al O 2 3 inferred ash chemistry from the XRD data and the actual ash and Fe O percentages, also shown in Figure 4, shows a much 2 3 chemistry determined by X-ray fluorescence (XRF) analysis, better correlation, possibly because these substitutions have taken place. especially for SiO2, CaO, and K2O. The proportions of Al2O3 and Fe2O3, when considered together, also provide a good The proportion of K2O inferred from the XRD data is similar to, but slightly higher than, that determined by direct chemical correlation between inferred and observed values, which is analysis. However, as indicated by Ward et al.,21 the inferred consistent with the presence of smectite and I/S in the clay fractions of the lignite and nonlignite samples. K2O percentage is based on complete saturation of the interlayer positions in the illite with K+ ions. The presence of smectite The study confirms the value of quantitative XRD analysis and I/S suggests that other ions, such as Na, may be substituted as a basis for understanding the nature of low-rank lignites, and for K in the actual clays of the sample suite, making the actual also indicates the types of minerals that might be expected in K2O percentages slightly lower than the inferred values. other Greek lignite deposits. In conjunction with the low proportions of carbonate carbon identified in conventional 6. Conclusions analysis, it also indicates that the bulk of the calcium (and the magnesium) in the lignites occurs in an organic association, Powder X-ray diffractometry (XRD) of oxygen-plasma ash indicates that the Achlada lignites contain a mineral assemblage forming bassanite and other artifacts during low-temperature that consists of kaolinite and illite, with minor proportions of ashing, and is not represented by crystalline carbonate minerals. quartz. The low-temperature ash (LTA) also contains bassanite, The study also provides a basis for better understanding the which probably is derived from the interaction of organically behavior of lignites from the Achlada mine during lignite associated calcium in the lignites with organic sulfur during utilization processes. the plasma ashing process. Detailed analysis of the clay fraction by oriented-aggregate XRD further indicates the presence of Acknowledgment. The authors gratefully acknowledge financial smectite and interstratified illite/smectite (I/S), which were not support from the General Secretariat for Research and Technology readily seen in the powder diffraction patterns. (GSRT) of the Ministry of Development. Thanks are also expressed The nonlignite bands occurring within the seam contain a to David Jacyna (Commonwealth Scientific and Industrial Research wider range of minerals, including mica, feldspar, and chlorite, Organisation (CSIRO) Energy Technology), for high-temperature as well as higher quartz and lower kaolinite percentages, ashing of the lignite samples, and to Irene Wainwright (University compared to the lignite. Traces of dolomite are also indicated of New South Wales (UNSW) Analytical Centre), for provision of in one rock sample. The clay fractions of the nonlignite bands the XRF data. The authors thank Dr. Jim Hower (University of have lower proportions of expandable clay minerals, relative Kentucky), and an anonymous referee, for the careful review of the manuscript. to illite, than the LTAs of the lignite samples. These differences probably reflect selective concentration of minerals in the EF8010993