The Middle Pennsylvanian, Langsettian sub-stage Clintwood coal, and the correlative Manchester, River Gem, Lily, among other names, is one of the more important coals in eastern Kentucky. While never supporting a production level of the Pond Creek (also known as Lower Elkhorn, Blue Gem, Path Fork), Fire Clay (Hazard No. 4, Dean), and some other coals, it traditionally was a significant resource in select areas, such as in Clay County (mined as the Manchester and Horse Creek coals) and in the current study area in southeastern Pike County. Eble and Hower (1995) noted that the Manchester coal marks the last appearance of the palynomorph Schulzospora rara. Thus, the top of the Clintwood coal, also the base of the marine Betsie Shale, marks the Langsettian/Duckmantian boundary (Clayton et al. 1977) (Fig. 1).

Geologic section for the Clintwood coal to Pond Creek coal (after Outerbridge and Van Vloten 1968). The Langsettian/Duckmantian boundary lies at the base of the Betsie shale

1.1 Rare earth nomenclature

In this contribution, REE (rare earth elements) designates the lanthanide elements, REY is REE + Y, and REYSc is REY + Sc. The light REE (LREE) are La through Sm and the heavy REE (HREE) are Eu through Lu (Seredin 1996; Hower et al. 1999). Upper Continental Crust (UCC) normalization of REE abundances (indicated by the subscript suffix “N”) (after Taylor and McLennan 1985) allows the description of L-type (light type: La_N/Lu_N > 1), M-type (medium type: La_N/Sm_N < 1, Gd_N/Lu_N > 1), and H-type (heavy type: La_N/Lu_N < 1) enrichment patterns (Seredin and Dai 2012). Following the UCC corrections, Bau and Dulski (1996) and Dai et al. (2016b, 2017a, 2017b) decoupled Ce, Eu, and Gd from the other REE in the distribution patterns:

While the decoupled Ce, Eu, and Gd can be used in the interpretation of the paleoenviromental settings, they also serve to simply differentiate or re-enforce other geochemical patterns. Marsac et al. (2010) and Davranche et al. (2011, 2017) discussed the relationship between La_N/Sm_N and Gd_N/Yb_N, in particular in consideration of the indication of increasing oxidation with a decrease in La_N/Sm_N.

Tuffaceous, detrital, infiltrational/leaching, and hydrothermal modes of REE enrichment are found in coals (Seredin and Dai 2012), with organic associations being possible in low-rank coals (Eskenazy 1978, 1987a—c; 1999, 2015; Eskenazy et al. 1986; Pédrot et al. 2010; Davranche et al. 2011; Aide and Aide 2012) and perhaps to the low-rank end of bituminous coals (Given 1984; Hatcher and Clifford 1996). Low LREE/HREE’s in coals might be a remnant of the organic association with the REE elements formerly in an organic association (Hower et al. 2020a). At bituminous ranks, those REE’s could be bound to clays (Eskenazy 1995, 1999; Seredin 1996) or incorporated into carbonate or phosphate minerals. The hydrothermal mobilization of HREE-depleted phosphates and HREE- and Y-enriched organic compounds may contribute to the redistribution of LREE and HREE, leading to an enrichment in LREE (Williams-Jones et al. 2012; Dai et al. 2013a, 2013b, 2016a).

1.2 Statement of purpose

The Clintwood coal, while being an important resource in portions of eastern Kentucky, has been overshadowed by the thicker, more accessible, and more extensive Pond Creek and Fire Clay coals. In this study, the organic petrology of the Clintwood is revisited using modern nomenclature and the geochemistry is expanded to include a broader array of critical elements along with a critical assessment of the paleoenvironmental and diagenetic relationships among the elements.

2.1 Sample description

Fourteen^{Footnote 1} bench samples of the Clintwood coal collected in the mid-1980’s from three locations in Pike County, Kentucky, were selected for additional chemical analyses (Fig. 2). The site numbers shown in the map refer to the top bench in each of the profiles. The remainder of the samples are stored at the Kentucky Geological Survey’s Earth Analysis Research Laboratory (https://www.uky.edu/KGS/EARL/). The proximate, ultimate, and forms of sulfur analyses were carried over from the original analytical work. Inductively coupled plasma- mass spectroscopy, a technique not widely available in the mid-1980’s, and an X-ray fluorescence investigation of the major oxides and minor elements was done for this re-examination.

Location of samples within the Lick Creek and Jamboree 7.5’ quadrangles in Pike County, Kentucky. The site numbers on the inset map represent the sample number of the top lithotype at each of the three sites

2.2 Analytical methods

The original proximate, ultimate, sulfur forms, and heating value analyses were performed on a variety of LECO instruments using the ASTM standards in place in the mid- to later-1980’s (Eble and Hower 1995). The major oxides and the select minor elements (V, Cr, Mn, Co, Ni, Cu, Zn, Rb, Sr, Zr, Mo, Ba) were determined by X-ray fluorescence (XRF) using procedures described by Hower and Bland (1989).

Lithium, Be, B, Ga, Ge, As, Nb, Sn, Sb, Pb, Sc, Y, and the lanthanides were analyzed using an Agilent 7900 Inductively coupled plasma- mass spectroscopy (ICP-MS) at the University of Kentucky, Center for Applied Energy Research (CAER). Tellurium is used as an internal standard. The laboratory digested the samples following the ASTM D6357-21b (ASTM, 2021) digestion method which utilizes heating the sample with a combination of nitric, hydrochloric, and hydrofluoric acids. The method was modified to include an additional nitric acid step at the end to handle any potential solid residue that might be present, and a sample weight of 0.15 g was utilized. The certified reference material NIST 1633b was utilized as the primary method control sample. Other standards available include the U.S. Geological Survey Brush Creek Shale (SBC-1) (Bank et al. 2016), LGC Standards Brown Coal Ash (BF BE1; https://www.lgcstandards.com/US/en), and available round-robin certified samples. Reference standards were digested in parallel with the samples to make sure that the digestion is complete. The certified values of the reference standard were in good agreement. Instrument integrity is routinely monitored as well with random blanks and analytical standards throughout each sequence.

The maceral analysis was completed at the CAER on previously made 2.54-cm-diameter epoxy-bound particulate pellets prepared to a final 0.5-µm-alumina polish on a wetted silk pad. The microscopic examination was made on a Leitz Orthoplan microscope using a 50 × reflected-light oil-immersion objective. Both white-light and blue-light-excitation illumination was used as necessary. Maceral nomenclature followed the International Committee for Coal and Organic Petrology (ICCP) System 1994 (ICCP, 1998, 2001; Pickel et al. 2017). Vitrinite reflectances determined for the Eble and Hower (1995) study were carried over for this investigation.

3.1 Petrology

Just as the Pond Creek coal thickness, megascopic and microscopic petrology, palynology, and geochemistry could be tied to it position relative to the Belfry anticline (structure after Hunt et al. 1937, and Rice et al. 1977; petrology after Hower et al. 1991, and Hower and Eble 2022), the Clintwood coal thickens from Lick Creek quadrangle site 3501 (the site closest to the anticline; site numbers refer to the uppermost bench at each site) to Jamboree quadrangle sites 3488 and 3458 to the southeast (Table 1; Fig. 3). Considering the sample 3502/3489/3461 dull-lithotype interval to be a near equal-thickness datum, note that the underlying coal not only thickens to the southeast (site 3501 to site 3458), but the lithotype structure gets more complex. Notably, below the through-going bright clarain ± fusain interval (3504/3492’s uppermost 10.8 cm), a basal dull clarain + clarain interval is found at site 3488 and a nearly 44-cm clarain comprises the base of site 3458. Between the 3502/3489/3461 dull lithotype interval and the upper dull lithotypes (portions of 3501 and 3488 and sample 3459), the bright lithologies (plus fusain) thicken and get duller to the southeast, transitioning from a bright clarain + vitrain + fusain interval at site 3501 to a clarain + fusain interval at site 3458.

Lithologic profile with durain and dull clarain marked in gray and the vitrain, bright clarain, and clarain marked in white. The top of the 3502/3489/3461 dull lithotype in considered to be a datum. The thin durains within samples 3488 and 3501 were not sampled separately

The maceral content of the Clintwood coal is presented in Tables 2 and S1. Among the maceral assemblages in the Clintwood coal, sporinite (Fig. 4 a-b) is common. Megaspores show the transitions from a cytoplasm-filled interior (Fig. 4c) to micrinite (Fig. 4d) and to micrinite and degraded cytoplasm (Fig. 4e) interiors. Inertinite-group macerals include macrinite (Fig. 5a) and numerous varieties of fusinite and semifusinite (Fig. 5a-h).

Sporinite. Scale = 50 µm on all images. a & b Miospores (sp) in vitrinite matrix. Images 3487 01 and 3491 01, respectively; c Megaspore with exine and cytoplasm. Image 3503 05; d Megaspore with exine and micrinite (mic) replacement of remnants of cytoplasm. Image 3489 04; e Megaspore with exine and micrinite (mic) replacement of remnants of cytoplasm (cy). Image 3503 04

Inertinite. Scale = 50 µm on all images. a Fusinite (f), secretinite (sec), and macrinite (mac). Image 3502 02; b Fusinite (f) and semifusinite (sf). Image 3501 14; c Fusinite (f) and semifusinite (sf) with ovoid vitrinite (v) area with resinous (r) inclusions. Image 3502 06; d Fusinite (f) and semifusinite (sf). Image 3463 03; e Fusinite (f) and semifusinite (sf). Image 3462 01; f Fusinite (f) and semifusinite (sf). Image 3463 01; g Fusinite (f) and semifusinite (sf) bands in left 2/3.^rd’s of the composite image with resinite (r) band and durite (mixed detrital inertinite, liptinite, and vitrinite) bands in the right third of the image. Image 3489 01; h Fusinite (f) and semifusinite (sf) bands in left half of the composite image and vitrinite (v) bands with fusinite, semifusinite, and sporinite (sp) in right half of the image. Image 3490 01

The lithologies (Fig. 6) are defined by simplified petrology based on the mineral-free maceral groups (Table 2) and synopsis of the palynology (arborescent lycopods, herbaceous lycopods, small ferns, and sphenopsids) after Eble and Hower (1995). As will be seen in the discussion below, the lithologies are important in the understanding of the geochemistry.

Generalized representation of lithotypes, petrology (mineral-free vitrinite-, inertinite-, and liptinite-group summaries), and palynology (arborescent lycopods, herbaceous lycopods, small ferns, and sphenopsids) for the three seam sections in the Lick Creek and Jamboree 7.5’ quadrangles. As with Fig. 3, the top of the 3502/3489/3461 dull lithotype in considered to be a datum

3.2 Geochemistry

The sample locations, thicknesses, and chemistry are given on Tables 3, 4, 5, 6, 7, 8 and S2. The lithologies (after Figs. 3 and 6) generally have low-S and low- to moderate-ash contents. Only the basal lithotype at site 3458 exceeds 1% total S (%; as-determined basis). Four of the five dull (durain and dull clarain) lithologies have the highest ash yields (Table 5).

3.2.1 Rare earth elements, Y, and Sc

The average Clintwood REE content of more than 800 µg/g (ash basis), while not at the level of the Fire Clay (Hower et al. 2020a) and the correlative Manchester (Hower et al. 2020b) coals, is above average for Central Appalachian coals (Table 8). Two of the 14 lithotypes analyzed exceed 1000-µg/g REE and seven of them exceed 1000-µg/g REYSc, with just one of the latter seven lithotypes having less than 1000-µg/g REY: 999-µg/g REY in sample 3492. The Upper continental crust normalization (after Taylor and McLennan 1985) shows that the H-type distributions for samples 3458, 3504, and, to a lesser extent, 3492 have strikingly different patterns than the other samples (Fig. 7). Samples 3504 and 3492, along with 3463, the only coal with a definitive M-type distribution, are the basal benches in their respective sections. Overall, the L-type samples have a higher REE content than the H-type samples (899 µg/g vs. 715 µg/g), a reflection of the greater abundance of the LREE end of the element distribution irrespective of the normalized La and Lu values. Hower et al. (2023; 2024) suggested that H-type distributions in basal benches of eastern Kentucky’s Leatherwood coal, among others, might reflect plant accumulation of REE from the underlying sediments.

Upper continental crust normalization of the REE concentrations (after Taylor and McLennan 1985). The coals with H-type distributions are labeled (normal font with blue lines and symbols) and the coal with an M-type distribution is labeled (italic font with green line and symbols)

Dai et al. (2015, 2016b) noted that Eu ICP-MS values can be impacted by interferences with Ba should be treated as suspect when Ba/Eu > 1000. None of the samples in this study have Ba/Eu higher than 467. The plot of Gd_N/Gd_N* vs. Eu_N/Eu_N^* (Fig. 8) shows that, disregarding sample 3458, the low point in both parameters and the sample with the highest Ba/Eu, there is no significant trend among the data points. The coincidence of positive Gd_N/Gd_N* and Eu_N/Eu_N^*, though, hints that there may have been a hydrothermal influence in the sediment sources of the REE and/or in the diagenesis of the coal (after Dai et al. 2016b).

Gd_N/Gd_N* vs. Eu_N/Eu_N^* for the Clintwood coal samples (Upper continental crust normalizations after Taylor and McLennan 1985). The numbers in the figure key indicate the uppermost sample in the respective sample suites

Plots of La_N/Sm_N vs. Gd_N/Yb_N, after discussions by Marsac et al. (2010) and Davranche et al. (2011, 2017) regarding REE partitioning in soils, are shown for all samples (Fig. 9a) and for the site 3458 samples (Fig. 9b). The latter authors considered ± 0.9 La_N/Sm_N to be a midpoint on the continuum between oxidizing (lower La_N/Sm_N) and reducing (higher La_N/Sm_N) environments. In a discussion of the Pond Creek coal, the next higher mineable coal in the same quadrangles as the Clintwood sites, Hower et al. (2024) applied the concept of the La_N/Sm_N vs. Gd_N/Yb_N distribution to multiple benches from an ombrogenous, bright-coal zone. They noted that there was an upwards decrease in La_N/Sm_N within the lithology, a trend they attributed to the transition to drier and more oxidizing depositional environments towards the top of the lithology (a discussion of this is found in Esterle et al. 1989, and Wüst and Bustin 2001). For the Clintwood coal site 3458 (Fig. 9B), the top lithotype (sample 3458) shows the greatest tendency towards being deposited in an oxidizing environment (low La_N/Sm_N) and the lowest indication of an MREE distribution (lowest Gd_N/Yb_N, albeit with relatively high Gd_N and Yb_N). The dull lithotypes (the durain area) and the bright lithotypes (clarain + bright clarain) are segregated into two regions. Within the durains, the uppermost durain shows the greater tendency towards having been deposited in an oxidizing environment. La_N/Sm_N vs. Ce_N/Ce_N* (Fig. 10) provides another perspective on the relationship among the site 3458 lithologies. As above, the durain and the clarain + bright clarain fall into separate regions. The lower right region of the plot, occupied by both the top and basal benches of the coal seam, represents the most oxidizing depositional environments based on the assumptions concerning La_N/Sm_N (Marsac et al. 2010; and Davranche et al. 2011, 2017) and Ce_N/Ce_N* (Dai et al. 2016b).

La_N/Sm_N vs. Gd_N/Yb_N for a all samples of the Clintwood coal and b the site 3458 samples. Upper continental crust normalizations after Taylor and McLennan (1985). The numbers in the figure key indicate the uppermost sample in the respective sample suites

La_N/Sm_N vs. Ce_N/Ce_N* for the site 3458 samples. The box in the lower right of the plot indicates the parameters most indicative of an oxidizing environment. Upper continental crust normalizations after Taylor and McLennan (1985)

3.2.2 Other elements

La_N/Sm_N, an indicator of reducing and oxidizing environments, possibly of the mire but also potentially of the source environment of the sediments entering the mire, and Zr (µg/g; ash basis), an indicator, along with TiO₂, among other elements, of detrital input into the mire are plotted on Fig. 11. The main trend, with or without sample 3460, a clarain near the top of the 3458 section, might be pointing towards a parallel increase in zircon and LREE-rich monazite in the detrital sediments. Excluding sample 3460, the r² = 0.485 is significant at the 1% level of confidence. Zirconium is not correlated with TiO₂ (not shown), an indication of the difficulty in correlating a minor element with a major oxide in the face of multiple depositional factors (variations in source composition, efficiency of transport, etc.) that can complicate the comparisons. Indeed, the variations in the Zr vs. La_N/Sm_N, such as the extreme position of sample 3460, may have been a consequence of those same depositional factors.

La_N/Sm_N vs. Zr (µg/g; ash basis). Upper continental crust normalizations after Taylor and McLennan (1985). The numbers in the figure key indicate the uppermost sample in the respective sample suites

The relationship between Zr and Sr (both as µg/g; ash basis) is more complicated (Fig, 12) (sample 3501 with 4108-µg/g Sr, more than twice the concentration in any other sample, is not plotted). The durain and the durain + dull clarain samples from the 3458 and 3488 sections plot in a low-Sr part of the plot but sample 3502, a durain + dull clarain + vitrain from the 3501 section, has 1494-µg/g Sr and 358-µg/g Zr. Sample 3502 is, therefore, chemically distinct, at least with the consideration of Zr and Sr, from the other dull lithologies and falls within the middle ground of the non-dull lithology and non-3501 samples. The latter 10-sample subset has an r² = 0.485 and is significant at the 5% level of confidence. Including sample 3501 in the regression, the 11-sample subset has an r² = 0.177 and is not significant. The uncertainty and the admitted loss of statistical rigor in isolating distinct subsets of the sample set notwithstanding, the contrasts in the chemical behavior of the dull lithologies versus the coals with > 600-µg/g Sr is evident. No one chemical influence is entirely responsible for the chemistry of any of the coals, but it is hypothesized that the dull lithologies were more influenced by detrital sediments than the brighter lithologies. In addition to the influx of Zr-bearing sediments, the latter were influenced by Sr-bearing carbonate- and phosphate-rich fluids.

The chemical associations of V and Cr, as reviewed by Dai et al. (2023, chapter 6), are complicated, with both having some tendencies towards organic associations. Vanadium has clay affinities (Zubovic 1966; Rimmer 1991) and Cr has spinel associations (Ruppert et al. 1996). Because plotting inertinite/(inertinite + vitrinite), a petrographic parameter, in the case of Fig. 13, isolates the dull lithotypes better than just the V and Cr chemistry, coal sample 3502 is close to the four dull lithologies plotted on Fig. 12. The graphing of inertinite/(inertinite + vitrinite) vs. V/Cr, however, leads to the relative isolation of samples 3503 and 3504, the coals above and below, respectively, a sandstone parting in the 3501 section. Considering just the population of samples exclusive of samples 3503 and 3504, the samples have an r² = 0.648 which is significant at the 1% level of confidence. The low-V/Cr in the durains suggests that Cr-bearing spinels were among the heavy minerals in the dull lithologies.

Zr vs. Sr (both as µg/g on the ash basis). Sample 3501 with 4108-µg/g Sr is not plotted because it has more than twice the Sr concentration of any other sample. The numbers in the figure key indicate the uppermost sample in the respective sample suites

Similar to many coals (Yudovich, 2003), Ge and, to a lesser extent, Ga are enriched at the base and top of the Clintwood coal. Beryllium peaks at 5 × the world hard coal average (after Ketris and Yudovic, 2009) in the lower benches in all the sections and in the top lithotype at site 3458. With the exception of the top lithotype at site 3458, Li exceeds 2 × the world coal average in all samples and exceeds 5 × the world coal average in samples 3461, 3489, and 3490.

3.2.3 Principal components analysis

As noted above, the relationships between the geochemical and petrographic variables is complicated and certainly is not as two-dimensional as is portrayed on Figs. 8–13. To more fully investigate the trends, principal components analysis (PCA; JMP© Pro 17.0.0, JMP Statistical Discovery LLC, Cary, North Carolina, USA) was employed to examine select variables and combinations of variables. With just 14 data points (excluding the sample with no REE analyses), the number of variables is limited. Using five variables, generally as a natural log transform, the groupings of inertinite/(inertinite + vitrinite), La_N/Sm_N, V/Cr, Sr, and either Zr or Ln(100(TiO₂)/Al₂O₃)^{Footnote 2} were examined (Fig. 14 and 15, with supporting data in Tables S3A, S3B, and S3C). The parameter LREE/HREE was used instead of La_N/Sm_N (Tables S3D and S3E) with results like those shown for La_N/Sm_N, therefore, just the results for La_N/Sm_N are discussed. The alternate use of Zr or Ln(100(TiO₂)/Al₂O₃) is justified on basis of the parameters being independent and uncorrelated indicators of detrital sediments.

Inertinite/(inertinite + vitrinite) vs. V/Cr. The numbers in the figure key indicate the uppermost sample in the respective sample suites

Principal components analysis eigenvector plot for of inertinite/(inertinite + vitrinite), La_N/Sm_N, V/Cr, Sr, and Zr. The triangles represent site 3458, the dots represent site 3488, and the squares represent site 3501. The light green points in the blue oval represent the durains. Sample 3502, a durain plotting near several of the brighter lithotypes, is shown. The numbers in the figure key indicate the uppermost sample in the respective sample suites

Principal components analysis eigenvector plot for of inertinite/(inertinite + vitrinite), La_N/Sm_N, V/Cr, Sr, and TiO₂/Al₂O₃. The triangles represent site 3458, the dots represent site 3488, and the squares represent site 3501. The light green points in the blue oval represent the durains. Sample 3502, a durain plotting near several of the brighter lithotypes, is shown. The numbers in the figure key indicate the uppermost sample in the respective sample suites

The use of Zr as the indicator of the detrital sedimentation (Fig. 14 and Table S3B) results in an eigenvector, accounting for 45.25% of the variation, with similar contributions from inertinite/(inertinite + vitrinite) and V/Cr and a lesser contribution from La_N/Sm_N. The second eigenvector, accounting for 31.86% of the variation, includes Sr, Zr, and La_N/Sm_N. As with several of the two-dimensional plots (Figs. 9b, 10, 12, 13), the durains plot in a different part of the field than the brighter lithotypes. Sample 3502 from site 3501 is the only durain that clusters near to the brighter lithotypes (Fig. 14).

The use of Ln(100(TiO₂)/Al₂O₃) as the indicator of the detrital sedimentation (Fig. 15 and Table S3C) results in an eigenvector, accounting for 57.13% of the variation, with nearly identical contributions from inertinite/(inertinite + vitrinite) and V/Cr and a diametrically opposed contribution from Ln(100(TiO₂)/Al₂O₃). The second eigenvector, accounting for 23.94% of the variation, includes Sr and La_N/Sm_N. In this PCA plot, the contributions of detrital minerals (TiO₂ minerals as a proxy for zircon and other heavy minerals and the Cr denominator of the V/Cr ratio), clays (the V component of the V/Cr ratio), and the inertinite are obviously related. The Sr and La_N/Sm_N components, both important in the second eigenvector, are also minor contributors to the first eigenvector, representing more of the secondary influences on the coal lithotypes. As above, the durains plot in a distinct field (Fig. 15) with sample 3502 plotting close to some of the brighter lithotypes.

The Middle Pennsylvanian, Langsettian sub-stage Clintwood coal was mined in southeastern Pike County, Kentucky. Along with its correlatives (Manchester, River Gem, Lily, among many more names), the coalbed was one of the more important energy resources in eastern Kentucky. The correlative Manchester coal marks the last appearance of Schulzospora rara (Eble and Hower 1995) and, by extension, the top of the coal and its boundary with the overlying marine Betsie Shale, marks the Langsettian/Duckmantian boundary (Clayton et al. 1977).

As with the Pond Creek coal (higher in the section) (Hower et al 1991; Hower and Eble 2022), the Clintwood coal thickens with increasing distance to the southeast away from the NE-SW-trending Belfry anticline (informal name; structure after Hunt et al. 1937, and Rice et al. 1977). Using the sample 3502/3489/3461 dull lithotype interval as a datum, the underlying coal lithologies thicken and become more complex to the southeast. Certain of the bright lithologies transition from a bright clarain + vitrain + fusain lithology to a clarain + fusain lithology to the southeast.

Two of the 14 lithotypes analyzed exceed 1000-µg/g REE and half of them exceed 1000-µg/g REYSc (ash basis). The coals with H-type REE distributions (Upper continental crust normalization after Taylor and McLennan 1985) exhibit a markedly different UCC pattern than the L-type coals. The L-type coals, though, have a higher REE content owing to the greater abundance of light REE. The positive Gd_N/Gd_N* and Eu_N/Eu_N^* in some of the coals, though, hints that there may have been a hydrothermal influence in the sediment sources of the REE and/or in the diagenesis of the coal. Examination of the La_N/Sm_N vs. Gd_N/Yb_N trends, supported by the La_N/Sm_N vs. Ce_N/Ce_N* plot, suggests that the top lithotype may have had the most exposure to an oxidizing environment.

La_N/Sm_N, as an indicator of reducing and oxidizing environments and Zr (µg/g; ash basis), an indicator of detrital input into the mire may indicate a parallel increase in both zircon and LREE-rich monazite in the detrital sediments. Zirconium is not correlated with TiO₂, another indicator of detrital sediments. Considering the correlation between Zr and Sr, most of the dull lithotypes plot in a low-Sr band distinct from the brighter lithotypes. The dull lithotypes were more influenced by the influx of detrital sediments while the brighter lithologies were influenced by Sr-bearing carbonate- and phosphate-rich fluids. The ratio of V and Cr, both with tendencies towards organic associations along with clay associations for V and spinel associations for Cr, versus inertinite/(inertinite + vitrinite) also shows a clustering of the durains apart from the rest of the field. With respect to other elements, Ge and Ga are enriched at the base and top of the coal seam and Be is high in the lower benches. Lithium exceeds 2 × the world hard coal average in all samples and exceeds 5 × the average in three samples.

Principal components analysis, evaluating variation among five variables (inertinite/(inertinite + vitrinite), La_N/Sm_N, V/Cr, Sr, and either Zr or Ln(100(TiO₂)/Al₂O₃)), also serves to illustrate the isolation of the dull lithotypes, particularly when Ln(100(TiO₂)/Al₂O₃) is used as the descriptor of the detrital sediment input. The contributions of detrital minerals (TiO₂ minerals, zircon, and other heavy minerals, with La_N/Sm_N being an indicator of LREE-rich monazite, plus the Cr-bearing spinel contribution to the V/Cr ratio), clays (the V component of the V/Cr ratio), and the inertinite are related. The Sr and La_N/Sm_N components represent the secondary influences on the coal lithotypes.

Durains and other dull lithologies represented distinct depositional environments, certainly in the maceral composition but also in the REE and minor element chemistry. Durains plot as distinct fields in the Gd_N/Yb_N vs. La_N/Sm_N (Fig. 9b), Ce_N/Ce_N* vs. La_N/Sm_N (Fig. 10), Zr vs. Sr (Fig. 12), inertinite/(inertinite + vitrinite) vs. V/Cr (Fig. 13), and the PCA plots of Zr or Ln(100(TiO₂)/Al₂O₃)) with (inertinite/(inertinite + vitrinite), La_N/Sm_N, V/Cr, and Sr (Fig. 14 or Fig. 15, respectively). As indicated in the site 3435 section, the durains’ Gd_N/Yb_N vs. La_N/Sm_N and Ce_N/Ce_N* vs. La_N/Sm_N chemistry indicates that the lithology was deposited in a more oxidizing environment than the clarain and bright clarain lithotypes.