Application of organic environmental markers in the assessment of recent and fossil organic matter input in coal wastes and river sediments_ A case study from the Upper Silesia Coal Basin (Poland)

Available online 26 July 2018 0166-5162/ © 2018 Elsevier B.V. All rights reserved. T hydroxy fatty acids, and alkanoic, αand β-alkanedioic, and phenolic acids, along with n-alkanols and sterols, play a key role (Fiorentino et al., 2006). Humic acids, insoluble in water, do not migrate intensively in the soil profile; hence they are the most important components for the formation of soil aggregates through the process of organo-mineral stabilisation. Furthermore, they can act as useful indicators of soil development processes (Kögel-Knabner et al., 2008). General formation of soil crust can be observed one or two years after initial coal waste dumping. Under favourable conditions e.g., if the wastes contain carboniferous clays and are rich in fossil OM, colonisation by a large number of various algae, cyanobacterial groups, and mosses would be favoured. Later, these types of dumps can be more easily covered by vascular plants (Lukešová et al., 2014). These type of soil-forming processes occurred at the Wełnowiec dump despite selfheating. The aim of the research was to learn how vegetation, water washing, or self-heating affects characteristic features of geochemical markers occurring in coal wastes, thus helping to distinguish coal-related compounds in river sediments. The distribution of such vegetation markers as n-alkanes, n-alkanoic acids, n-alkanols, and sterols can constitute a clear sign of the presence of extant plants. Their occurrence can indicate initial soil-forming processes taking place on coal waste dumps, using river sediments as reference material. Furthermore, one part of a secondary process in coal wastes, called self-heating, can influence the distribution of studied compounds, e.g. polar compounds (phenols), aliphatic (n-alkanes) and aromatic (n-alkylbenzenes) hydrocarbons. The article is a continuation of previous researches, see Nádudvari and Fabiańska (2015); Nádudvari and Fabiańska (2016a), covering a range of the new compounds to find their applicability in geochemical investigations. According to our knowledge polar compounds were not applied in the research of coal wastes subjected to selfheating or river sediments containing coal dust. 2. Samples and methods The river sediments were given names according to the form AN_Mxx and they were studied with details in Nádudvari and Fabiańska (2015) and Nádudvari et al. (2018). Two of them are borehole samples: DR2_1, taken from a depth of approximately 1m from a river bed containing gravelly sediment, and DR2_2, taken from the bank of the river at a depth of 1.7 m in typically gravelly sediment (Fig. 1). SZ_F_12 and SZ_F_13 are fresh coal wastes, dumped several weeks prior to sampling, taken from the Szczygłowice dump (Fig. 1; see the earlier publication by Nádudvari and Fabiańska, 2016a). Additional samples include self-heated coal wastes AN8 and AN9 (taken from the Anna dump), WE2 and WE5 (taken from the Wełnowiec dump), CZ7 (taken from the Czerwionka-Leszczyny dump) (see Nádudvari and Fabiańska, 2016a, 2016b), and RC7, consisting of expelled pyrolytic bitumen from the Rymer dump coal wastes. Details are given in Nádudvari and Fabiańska (2016b). 2.1. Extraction and derivatisation Powdered samples (ca 18–20 g) were extracted using dichloromethane with an accelerated Dionex ASE 350 solvent extractor. The extracts of the samples were converted to trimethylsilyl (TMS) derivatives via a reaction with N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA), 1% trimethylchlorosilane, and pyridine for three hours at 70 °C. The excess reagent was then removed under blowdown with dry nitrogen and the sample mixture was dissolved in an equivalent volume of dehydrated n-hexane. 2.2. Gas chromatography-mass spectrometry Gas chromatography-mass spectrometry (GC–MS) analyses were carried out using an Agilent Technologies 7890A gas chromatograph and an Agilent 5975C Network mass spectrometer with a Triple-Axis detector (MSD) at the Faculty of Earth Sciences, Sosnowiec, Poland. Helium (grade 6.0) was used as a carrier gas at a constant flow of 2.6 ml/min. Separation was obtained on fused silica capillary column, DB-5 (60m×0.25mm i.d.; film thickness 0.25 μm) coated with a chemically bonded phase (5% phenyl, 95% methylsiloxane), for which the GC oven temperature was programmed from 45 (1min) to 100 °C at 20 °C/min, then to 300 °C (held for 60min) at 3 °C/min, with a solvent delay of 10min. The GC column outlet was connected directly to the ion source of the MSD. The GC–MS interface was set at 280 °C, while the ion source and the quadrupole analyser were set at 230 and 150 °C, respectively. Mass spectra were recorded from 45 to 550 da (0–40min), and 50 to 700 da (> 40min). The MS was operated in the electron impact mode, with an ionisation energy of 70 eV. For quality control, these samples were compared with the originals using PAH diagnostic ratios such as phenanthrene/anthracene; fluoranthene/pyrene; benz[a] anthracene/(benz[a]anthracene + chrysene); benzo[a]pyrene/benzo [ghi]perylene; and indeno[1,2,3-cd]pyrene/(indeno[1,2,3-cd]pyrene + benzo[ghi]perylene). The calculated error varied, but was consistently under 2% (Nádudvari et al., 2018). 3. Results and discussion In the studied area, the predominant sources of OM are coal and coal wastes, accompanied, in the case of river sediments, by recent vegetation and urban wastes. In the following section, the particular sources are discussed, along with an indication of the main organic compounds generated by individual sources. 3.1. Aliphatic hydrocarbons Elevated relative percentages of n-alkanes (ranging from C11 to C34) were found in the coal waste samples and in AN_M9 compared with other sediments. Typically, short-chain n-alkanes dominated over longchain n-alkanes in coal wastes (both fresh and self-heated coal wastes, except for the Wełnowiec samples) (Tables 1 and 2). The reason for the dominance of short-chain over long-chain n-alkanes differs between fresh (0.29–0.51) and self-heated (0.02–0.19 for AN8, AN9, CZ7, RC7 and 0.94–6.80 for WE2, WE5) coal wastes (Table 3). It is a common feature for short-chain n-alkanes to dominate in unaltered coal wastes where water-washing and biodegradation have not influenced the organic material (Fabiańska and Kurkiewicz, 2013; Nádudvari and Fabiańska, 2016a). However, in the case of self-heated coal wastes, the distribution reflects macromolecule cracking and the expulsion of shortchain n-alkanes (n-C11 to n-C18), a process which can be seen in the AN8, AN9, CZ7, and RC7 samples (Table 3). The distribution maximum depends on the range of heating temperatures (Misz-Kennan et al., 2007; Nádudvari and Fabiańska, 2016a, 2016b). The origin of shortchain n-alkanes (e.g. n-C17 to n-C18) in fossil OM is associated with microorganisms, including phytoplankton settled in terrestrial organic material deposited in deltaic environments (Gelpi et al., 1970; Grimalt and Albaigés, 1987; Meyers, 1997; Fabiańska et al., 2003, 2008). However, long-chain n-alkanes with carbon numbers from n-C25 to nC35 and odd carbon preponderance are related to higher terrestrial plants (Eglinton and Hamilton, 1967). Carbon Preference Index values e.g. CPI(n-C24-C34), or CPI(n-C25-C31) can distinguish coal wastes. CPI(n-C24-C34)= 1.02–1.49; CPI(n-C25-C31)= 0.94–1.60) from almost all river sediments CPI(n-C24-C34)= 2.72–9.26; CPI(n-C25C31)= 2.46–8.89; see Table 3). In river sediments, long-chain n-alkanes dominated, with a predominance of odd-over-even in the range n-C25 to n-C31, showing a bimodal distribution. Commonly n-C27 and n-C29nalkanes occurred with abundant relative percentages in relation to others, with the exception of only one sample (AN_M62), in which n-C31 was dominant (Tables 1 and 2). Typically, n-alkanes from higher terrestrial plants (trees, roots, shrubs, epicuticular waxes in plant leaves) are characterised by a strong odd-over-even carbon preference, Á. Nádudvari et al. International Journal of Coal Geology 196 (2018) 302–316


Introduction
In Upper Silesia, coal mining has had a significant impact on the original landscape, marring it with hundreds of coal waste dumps. The enormous volume of coal wastes has led to widespread pollution in the environment. Generally, in the Upper Silesian Coal Basin, types of coal range from sub-bituminous to high-volatile bituminous; sapropelic coals occur only rarely (Kotarba et al., 2002). For the presented study, river sediments (polluted by coal dust and coal ash) and coal waste samples were collected. Biomarkers preserved in both sample types appear to be useful indicators for identification of an organic-matter (OM) origin Nádudvari et al., 2018). For example, n-fatty acids can serve as useful indicators of OM origin because they are precursors of aliphatic hydrocarbons, which represent some of the main components of petroleum and coal bitumen, which in turn forms via decarboxylation and reduction reactions (Cooper and Bray, 1963;Kawamura and Ishiwatari, 1985;Dong et al., 1993). In addition, n-fatty acids, which are present in recent plant tissues, may be important components of OM involving all types of kerogen. Typical features of cutin and suberin in higher vascular plants include elevated concentrations of long-chain n-alkanoic acids (n-C 22 to n-C 32 ) and n-alkanols, with a notable predominance of even chain homologues and the occurrence of certain hydroxy, dicarboxylic, and diterpenoid acids, as well as long-chain (n-C 40 to n-C 64 ) saturated alkyl (wax) esters (Gillan and Sandstrom, 1985;Mita et al., 1998;Cooper, 1990).
In conjunction with this study, it is important to mention that the spontaneous soil-forming processes which are generally possible on coal waste dumps under the influence of vegetation and other paedogenetic factors (Wiegleb and Felinks, 2001;Frouz et al., 2008;Alday et al., 2012;Zhang et al., 2015) can be accelerated by adding various waste products, e.g. composts, lignohumate, organic waste, or salvaged materials. Such improvements are important in the successful recultivation of coal wastes (Sydnor and Redente, 2002;Mercuri et al., 2005). Fresh coal waste contains no extant organic material, but is rich in fossil OM. Over time, fossil OM, along with inorganic matter, may undergo so-called 'secondary' processes such as weathering, selfheating, water washing, chelatisation and complexation of soil mineral compounds, and migration and translocation of organo-mineral compounds (e.g. Nádudvari et al., 2018). Thus, soil formation is determined by the quality of the OM and soil biota (Kononova, 1966;Emmer and Sevnik, 1994;Bradshaw, 1997;Abakumov, 2008). In soil formation, humic acids, e.g. di-and tri-hydroxyalkanoic acids, α-, βand ω-hydroxy fatty acids, and alkanoic, αand β-alkanedioic, and phenolic acids, along with n-alkanols and sterols, play a key role (Fiorentino et al., 2006). Humic acids, insoluble in water, do not migrate intensively in the soil profile; hence they are the most important components for the formation of soil aggregates through the process of organo-mineral stabilisation. Furthermore, they can act as useful indicators of soil development processes (Kögel-Knabner et al., 2008).
General formation of soil crust can be observed one or two years after initial coal waste dumping. Under favourable conditions e.g., if the wastes contain carboniferous clays and are rich in fossil OM, colonisation by a large number of various algae, cyanobacterial groups, and mosses would be favoured. Later, these types of dumps can be more easily covered by vascular plants (Lukešová et al., 2014). These type of soil-forming processes occurred at the Wełnowiec dump despite selfheating.
The aim of the research was to learn how vegetation, water washing, or self-heating affects characteristic features of geochemical markers occurring in coal wastes, thus helping to distinguish coal-related compounds in river sediments. The distribution of such vegetation markers as n-alkanes, n-alkanoic acids, n-alkanols, and sterols can constitute a clear sign of the presence of extant plants. Their occurrence can indicate initial soil-forming processes taking place on coal waste dumps, using river sediments as reference material. Furthermore, one part of a secondary process in coal wastes, called self-heating, can influence the distribution of studied compounds, e.g. polar compounds (phenols), aliphatic (n-alkanes) and aromatic (n-alkylbenzenes) hydrocarbons. The article is a continuation of previous researches, see ; Nádudvari and Fabiańska (2016a), covering a range of the new compounds to find their applicability in geochemical investigations. According to our knowledge polar compounds were not applied in the research of coal wastes subjected to selfheating or river sediments containing coal dust.

Samples and methods
The river sediments were given names according to the form AN_Mxx and they were studied with details in  and Nádudvari et al. (2018). Two of them are borehole samples: DR2_1, taken from a depth of approximately 1 m from a river bed containing gravelly sediment, and DR2_2, taken from the bank of the river at a depth of 1.7 m in typically gravelly sediment (Fig. 1). SZ_F_12 and SZ_F_13 are fresh coal wastes, dumped several weeks prior to sampling, taken from the Szczygłowice dump ( Fig. 1; see the earlier publication by Nádudvari and Fabiańska, 2016a). Additional samples include self-heated coal wastes AN8 and AN9 (taken from the Anna dump), WE2 and WE5 (taken from the Wełnowiec dump), CZ7 (taken from the Czerwionka-Leszczyny dump) (see Fabiańska, 2016a, 2016b), and RC7, consisting of expelled pyrolytic bitumen from the Rymer dump coal wastes. Details are given in Nádudvari and Fabiańska (2016b).

Extraction and derivatisation
Powdered samples (ca 18-20 g) were extracted using dichloromethane with an accelerated Dionex ASE 350 solvent extractor. The extracts of the samples were converted to trimethylsilyl (TMS) derivatives via a reaction with N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA), 1% trimethylchlorosilane, and pyridine for three hours at 70°C. The excess reagent was then removed under blowdown with dry nitrogen and the sample mixture was dissolved in an equivalent volume of dehydrated n-hexane.

Gas chromatography-mass spectrometry
Gas chromatography-mass spectrometry (GC-MS) analyses were carried out using an Agilent Technologies 7890A gas chromatograph and an Agilent 5975C Network mass spectrometer with a Triple-Axis detector (MSD) at the Faculty of Earth Sciences, Sosnowiec, Poland. Helium (grade 6.0) was used as a carrier gas at a constant flow of 2.6 ml/min. Separation was obtained on fused silica capillary column, DB-5 (60 m × 0.25 mm i.d.; film thickness 0.25 μm) coated with a chemically bonded phase (5% phenyl, 95% methylsiloxane), for which the GC oven temperature was programmed from 45 (1 min) to 100°C at 20°C/min, then to 300°C (held for 60 min) at 3°C/min, with a solvent delay of 10 min. The GC column outlet was connected directly to the ion source of the MSD. The GC-MS interface was set at 280°C, while the ion source and the quadrupole analyser were set at 230 and 150°C, respectively. Mass spectra were recorded from 45 to 550 da (0-40 min), and 50 to 700 da (> 40 min). The MS was operated in the electron impact mode, with an ionisation energy of 70 eV. For quality control, these samples were compared with the originals using PAH diagnostic ratios such as phenanthrene/anthracene; fluoranthene/pyrene; benz[a] anthracene/(benz[a]anthracene + chrysene); benzo[a]pyrene/benzo [ghi]perylene; and indeno[1,2,3-cd]pyrene/(indeno[1,2,3-cd]pyrene + benzo[ghi]perylene). The calculated error varied, but was consistently under 2% .

Results and discussion
In the studied area, the predominant sources of OM are coal and coal wastes, accompanied, in the case of river sediments, by recent vegetation and urban wastes. In the following section, the particular sources are discussed, along with an indication of the main organic compounds generated by individual sources.

Aliphatic hydrocarbons
Elevated relative percentages of n-alkanes (ranging from C 11 to C 34 ) were found in the coal waste samples and in AN_M9 compared with other sediments. Typically, short-chain n-alkanes dominated over longchain n-alkanes in coal wastes (both fresh and self-heated coal wastes, except for the Wełnowiec samples) (Tables 1 and 2). The reason for the dominance of short-chain over long-chain n-alkanes differs between fresh (0.29-0.51) and self-heated (0.02-0.19 for AN8, AN9, CZ7, RC7 and 0.94-6.80 for WE2, WE5) coal wastes (Table 3). It is a common feature for short-chain n-alkanes to dominate in unaltered coal wastes where water-washing and biodegradation have not influenced the organic material (Fabiańska and Kurkiewicz, 2013;Nádudvari and Fabiańska, 2016a). However, in the case of self-heated coal wastes, the distribution reflects macromolecule cracking and the expulsion of shortchain n-alkanes (n-C 11 to n-C 18 ), a process which can be seen in the AN8, AN9, CZ7, and RC7 samples ( Table 3). The distribution maximum depends on the range of heating temperatures (Misz-Kennan et al., 2007;Fabiańska, 2016a, 2016b). The origin of shortchain n-alkanes (e.g. n-C 17 to n-C 18 ) in fossil OM is associated with microorganisms, including phytoplankton settled in terrestrial organic material deposited in deltaic environments (Gelpi et al., 1970;Grimalt and Albaigés, 1987;Meyers, 1997;Fabiańska et al., 2003Fabiańska et al., , 2008. However, long-chain n-alkanes with carbon numbers from n-C 25 to n-C 35 and odd carbon preponderance are related to higher terrestrial plants (Eglinton and Hamilton, 1967). Carbon Preference Index values e.g. CPI(n-C 24 -C 34 ), or CPI(n-C 25 -C 31 ) can distinguish coal wastes.
CPI(n-C 24 -C 34 ) = 1.02-1.49; CPI(n-C 25 -C 31 ) = 0.94-1.60) from almost all river sediments CPI(n-C 24 -C 34 ) = 2.72-9.26; CPI(n-C 25 -C 31 ) = 2.46-8.89; see Table 3). In river sediments, long-chain n-alkanes dominated, with a predominance of odd-over-even in the range n-C 25 to n-C 31 , showing a bimodal distribution. Commonly n-C 27 and n-C 29 nalkanes occurred with abundant relative percentages in relation to others, with the exception of only one sample (AN_M62), in which n-C 31 was dominant (Tables 1 and 2). Typically, n-alkanes from higher terrestrial plants (trees, roots, shrubs, epicuticular waxes in plant leaves) are characterised by a strong odd-over-even carbon preference, especially n-C 27 and n-C 29 alkanes, while n-alkanes from grasses and mosses are characterised by a maximum at n-C 31 (Eglinton and Hamilton, 1967;Rieley et al., 1991;Ficken et al., 2000;Pancost et al., 2002;Bi et al., 2005;Bingham et al., 2010). Fig. 2A and B distinctly separate coal wastes, river sediments mixed with coal particles with long-chain n-alkane input from recent vegetation, and samples where nalkanes related to terrestrial plants were dominant.

Polar compounds
Polar compounds were identified in both sample groups, i.e. in river sediments and coal wastes. The main groups of compounds are: sterols, stanols, phenols, n-alkanoic-and benzenedicarboxylic acid, 4-hydroxybenzaldehyde, acetophenones, and benzoic and methylbenzoic acids. However, the distributions of the listed compounds varied according to the type of samples studied.

n-Alkanoic acids
A common feature of river sediments in comparison with coal wastes was the appearance of n-alkanoic acids (ranging from n-C 6 to n-C 32 ) with elevated relative abundances (Tables 1 and 2, Fig. 3). The applied EOP index (even-over-odd predominance) points to the dominance of even long-chain n-alkanoic acids (n-C 20 to n-C 26 ) in river sediments (Table 3). Typically, high concentrations and notable predominance of even long-chain (n-C 22 to n-C 32 ) compounds are characteristic of higher-vascular-plant cutin and suberin (Mita et al., 1998). Notably, n-hexadecanoic (palmitic), n-octadecanoic (stearic), and oleic acids were dominant in several river sediment samples, e.g. DR2_1 and AN_M11_2 (see Table 1). These compounds can be generated by both bacteria and higher plants (Hedges et al., 1997;Summons et al., 2013;Walley et al., 2013;Huang et al., 2016). Application of n-alkanoic acids for different geochemical ratios calculated from the total ion chromatograms, such as (Σn-alkanoic acids + Σlong chain n-alkanes)/ Σshort chain n-alkanes, for river sediments 2.21-300.98, for coal wastes 0.18-11.25; or (Σn-alkanoic acids + Σn-alkanols)/Σn-alkanes, for river sediments 0.34-23.66, for coal wastes 0.01-0.58, closely reflected the vegetation/soil input to the sediments and Wełnowiec samples ( Table 3). The previously mentioned ratio values of the AN_M9 sample differ from those of the other sediments (Table 3; Fig. 2C) because this sample is characterised by an elevated level of inertinite (the sample contained many coal particles), as described in sample no. m29(2)). The ternary diagram (Fig. 4) shows clear separation of the coal samples from the extant organic material deposited in the river; as in the case of recent OM, n-alkanoic acids were dominant over n-alkanes. The exceptions are the WE2 and WE5 samples, where additional n-alkanoic acids derive from recent vegetation related to initial soil formation on the coal waste dump.
Á. Nádudvari et al. International Journal of Coal Geology 196 (2018) 302-316 sample contained mostly tyrosol, as well as pimaric and isopimaric acids (Table 1). Pimaric, isopimaric, and dehydroabietic acids are natural resin acid constituents, of which the last-named occurs most abundantly (Råbergh et al., 1999). Additional compounds included oxalic, succinic, and benzeneacetic acids, along with methylbenzoic acids, vanillin, and β-hydroxybutyric and levulinic acids occurring in river sediments and in those coal waste samples contaminated by initial soil formation, e.g. CZ7, WE2, and WE5. These compounds were not identified in fresh coal wastes (Table 1). Therefore, in the samples, they originated recently from vegetation. Benzeneacetic acid is widely present in vascular and non-vascular plants (Kim and Chung, 2000;Sugawara et al., 2015). Oxalic and succinic acids and vanillin have also been identified in resin components, leaves, roots, stems, fruits and seeds, prairie grassland soils, and in pyrolysis/oxidation products of  Table 1).
Á. Nádudvari et al. International Journal of Coal Geology 196 (2018) 302-316 Table 2 Calculated relative percentages of identified compounds from the TIC (total ion chromatogram) based on their peak areas. Only representative, non-overlapping compounds were selected. Dibenzofuran, phenanthrene, and retene, commonly present in coal and coal wastes, were chosen solely in order to compare their distributions with others.        '-' indicates that the compound was absent or occurred in very small amounts.
Benzoic acid was identified in most samples, with no great differences in relative percentages. This compound may be present in coal tar, but it can occur naturally in higher plants as well (Qualley et al., 2012;Pavón et al., 2016). Therefore the ratio of dimethylphenols/benzoic acid, combined with that of (Σn-alkanoic acids + Σlong chain n-alkanes)/ Σshort chain n-alkanes, clearly distinguished between origins from coal wastes and from extant vegetation in the studied samples ( Fig. 2A). One example of a human-related inorganic pollutant was phosphoric acid, identified only in river sediments. This compound is commonly released into the environment by phosphate fertilisers used in the agricultural sector (Belboom et al., 2015). The presence of 1,2-benzenedicarboxylic (phthalic acid) in the river sediments and Wełnowiec samples is related to a human-produced organic pollutant, as this compound is widely used in e.g. plasticizers, producing polyester / alkylresins and insect repellents (Bang et al., 2011).

Ketones and alcohols
Acetophenone and m-hydroxyacetophenone were identified mostly in coal waste samples and, with elevated relative percentages, in selfheated samples (Table 1). As previously indicated, these compounds can be released into the environment via petrol and residential fuel oil exhaust, coal combustion, heavy oil fractions in coal tar, waste waters from petrochemical plants, and waste incineration (Safaei-Ghomi et al., 2007;Alexieva et al., 2009;HSDB, 2010). Friedelan-3-one was identified only in river sediments as a common marker of recent vegetation input (Brassel and Eglinton, 1983;Duke, 1992;Odeh et al., 2016). In coal waste samples, cholesterol, cholestanol, stigmasterol, and sitosterol were identified with minor relative percentages as contaminants (Table 1). These compounds are usually not prominent (if present) in coals due to oxidative loss via diagenetic processes during coal deposition (Oros and Simoneit, 2000;Oros et al., 2006). Contrastingly, the river sediments were characterised by elevated percentages of sterols, such as cholesterol, β-sitosterol, campesterol, cholestanols, and cholestene, representing input from recent vegetation (Table 1; Weete, 1976;Otto and Simoneit, 2001;Killops and Killops, 2005;Oros et al., 2006). Therefore, their presence in coal waste material is secondary and indicates initial soil-formation processes or their migration with meteoric waters. Among plant triterpenoids, compounds such as αand βamyrin, lupeol, and ursolic acid were identified in river sediments, indicating recent input from vegetation, as they are commonly present in leaves, bark, wood resins, and roots (Bauer et al., 2004;Vázquez et al., 2012;Yasumoto et al., 2017). Alkanols with even-over-odd predominance in the number of carbon atoms occurred in river sediments (ranging from n-C 14 to n-C 32 ) with elevated relative percentages (Tables Fig. 4. Ternary diagram applied in order to distinguish the different sample groups; results were calculated using selected ion chromatograms (see Table 1). Σ(phenols) = phenols + cresols (o-, m-, and p-) + dimethylphenols + 2ethoxyphenol.

Table 3
The values of geochemical ratios.  1 and 2). In several coal waste samples, only n-octacosanol was identified. The predominance of n-C 26 to n-C 30 alkanols in river sediments indicates an origin from vascular-plant wax (Oros and Simoneit, 2000).

Phenols
Generally, a wide range of alkylphenols (o-, m-, and p-cresols and dimethylphenols) occurred in self-heated coal wastes with elevated percentages in comparison to river sediments (Tables 1 and 2; Fig. 3), as was preliminarily indicated by . Since phenols are relatively readily soluble in water, they can be used as tracers to identify the occurrence and extent of water-washing processes. This feature thus increases their hazardous environmental potential, as they are toxic and carcinogenic to humans (Clayton and Clayton, 1994;Michałowicz and Duda, 2007). The water-washing effect can be identified, using the Σdimethylphenols/benzoic acid ratio (Table 3), in the WE2 and WE5 samples as compared to SZ_F_12 and SZ_F_13 or the selfheated samples AN8 and AN9. The latter were taken following restructuring of the dump's shape into a trapezoidal form, which was done only a few months before sampling. As a result these samples exhibit less water washing, as they were taken from a deeper part of the dump. The origin of phenols in these samples can be explained as the result of thermal destruction of vitrinite and may represent relatively early stages of self-heating (Skręt et al., 2010). Phenol and alkylphenols are the most prominent products of the degradation of both lignin itself and lignin-derived macromolecules in coal (Saiz-Jimenez and de Leeuw, 1985;Hatcher et al., 1992;Iglesias et al., 2002). Fig. 5 indicates the distribution of phenols in two representative samples (one representing river sediment, the other self-heated coal waste). However, the compatibility of the distribution of cresols and dimethylphenols from coal waste dumps and river sediments, despite waterwashing, indicates that these compounds, found in Bierawka river sediments, originated from coal. Among phenols, tyrosol (a polyphenol and phenylethanoid) was identified only in DR2_2; its origin is associated with lignin degradation (Stefanova et al., 2004;Grasset et al., 2010). 2-Ethoxyphenol, which appears in most samples, was identified as one of the major compounds in immature OM from internal sediments in southern Poland (Rybicki et al., 2017) and may be another lignin-degradation product.

Aromatic compounds
A homologous series of n-alkylbenzenes ranging from C 14 to C 34 were found in coal waste with high relative concentrations. The predominance of short-chain n-alkylbenzenes was observed in self-heated coal wastes such as AN8 and AN9 due to influence of elevated temperatures during their formation ( Table 3). The presence of these compounds in river sediments clearly indicates input from coal material, as n-alkylbenzenes have been widely reported as common constituents of coals, coal pyrolysates, and coal smoke (Gallegos, 1981;Philp, 1985;Oros and Simoneit, 2000;Killops and Killops, 2005;Bi et al., 2008). However, in WE2, WE5, and river sediments, long-chain nalkylbenzenes were found to be dominant. According to Fabiańska et al. (2012), n-alkylbenzenes exhibit various degrees of resistance to biodegradation and water washing, similarly to e.g. n-alkanes. As a result of these processes, short-chain n-alkylbenzenes are removed from samples subjected to biodegradation and water-washing. The coal particles were previously identified in river sediments by . Therefore the identification of these compounds is well agreed with the presence of coal particles in sediments. However, the applied PAH diagnostic ratios indicate their pyrogenic origin not only due to coal wastes self-heating, but also because river sediments contained other contaminations like e.g. ash deposits Nádudvari et al., 2018).
Generally, elevated relative percentages of dibenzofuran and phenanthrene differentiate coal wastes from river sediments, except in the case of AN_M9 . The DR2_2 sample contained elevated amounts of retene, an aromatic land plant biomarker typical of terrestrial sedimentary OM (Table 2). Ramdahl (1983) also mentioned retene as wood-combustion indicator in environment.

Compounds related to an urban rubbish dump covered with coal waste and soil growth on the Wełnowiec dump
The Wełnowiec dump is a specific case, as it served as an urban rubbish dump from 1991 to 1996, when it was covered by coal wastes which began to self-heat . Phthalimide and 1Hisoindole-1,3(2H)-dione or n-(hydroxymethyl)phthalimide were found in two self-heated samples from Wełnowiec. Phthalimide, which has been detected previously in self-heated coal waste dumps, is a transformation product formed from the gaseous phase originating in the natural pyrolytic process (Rost, 1942;Jehlička et al., 2007;Lorz et al., 2007;Fabiańska et al., 2015). Compounds related to urban rubbish, such as triphenylbenzene or triphenylpyridine (plastic combustion tracers; Simoneit et al., 2005), had been previously identified at this dump . The chromone and coumarin occurred only in the Wełnowiec samples, indicating that their origin was connected with the plants which cover the dump, and thus were absent from river sediments. These compounds are widely present naturally in various plants, fruits, and vegetables (Dean, 1963;Robinson, 1963;Digiovanni, 1990). The complete series of methylbenzoic acids (o-, mand p-toluic acid), also found only in WE samples (Table 2), are common in some vascular plants, e.g. in their flowers (Kolosova et al.,  2001;Negre et al., 2003). The Wełnowiec dump features/conditions favour plant growth, e.g. an abundance of emitted CO 2 , nitrates, ammonia acting as a fertiliser, and heated warm surfaces, even in winter (Doerr and Shakesby, 2013;Komnitsas et al., 2010;González-Alcaraz et al., 2011). This enables the occurrence of special and rare species (Vanderpoorten, 2001;Ciesielczuk et al., 2015). These places are seeded by pioneer species immediately after cooling of the surface. Since the seasonal vegetation cycle is disturbed, some plants may be found seeding at the same time others are blooming or fruiting (Midgley and Bond, 2013;Ciesielczuk et al., 2015).

Conclusions
Characteristic polar compounds from the extant OM were identified as dominant in river sediments and coal wastes exposed under air for longer time periods and contaminated by soil. These compounds included αand β-tocopherol, sterols, stanols, glycerol, pimaric and isopimaric acids, oxalic, succinic, and ursolic acids, and friedelan-3-one. This input from recent vegetation was subsequently confirmed by the applied geochemical ratios which clearly separated young sediments from unaltered coal wastes (fresh coal wastes), e.g. (Σn-alkanoic acids + Σn-alkanols)/Σn-alkanes, (Σn-alkanoic acid + Σlong chain n-alkanes)/Σshort chain n-alkanes, CPI(n-C 25 -C 31 ), and CPI(n-C 24 -C 34 ) ratios. Among common features were the dominance of n-alkanes and n-alkylbenzenes in coal wastes, whereas alkanoic acid, alkanols, sterols, stanols, etc. occurred with elevated relative percentages or exclusively in river sediments. In several coal wastes, initial soil-formation processes had started, as indicated by the occurrence of cholestanol, sitosterol, oxalic acid, and methylbenzoic acids, and by a preponderance of odd-over-even long-chain n-alkanes and n-alkanoic acids.
Typical water washing and the initial stage of the self-heating effect was identified in coal waste samples using Σdimethylphenols/benzoic acid ratio. The distributions of o-, m-and p-cresols and dimethylphenols in river sediments was found to be similar to those in coal waste, indicating that these compounds connected with coal were derived from the same source. Other coal-related pollution compounds present in river sediments included n-alkylbenzenes, phenanathrene, dibenzofuran, and acetophenone. Distribution of these compounds varied due to water-washing and/or degradation processes.