The Bolsovian age (Westphalian C, Middle Pennsylvanian) Princess No. 3 coalbed in the Argillite 7.5′ quadrangle in Greenup County, Kentucky (Fig. 1), is approximately correlative to the lower split of the Peach Orchard or Coalburg and the Hazard No. 7 coals (among other names), which have been mined extensively elsewhere in the eastern Kentucky and southern West Virginia coalfields (based on compilations of coal names by the Kentucky Geological Survey and the Center for Applied Energy Research in the 1980s). In northeastern Kentucky, the Princess No. 3 coal, with an estimated original resource estimate of approximately 200 Mt, occurs about 60-m below the Princess No. 7 coal bed.

Stratigraphic section after Sheppard and Ferm (1962). The interval thickness is approximately 55 m

The original investigation of the coal was part of a study of methods for the enhanced recovery of coal from abandoned surface mine highwalls. Highwall mining is a hybrid of surface and underground mining, mining coal from a surface exposure, such as a contour mining operation, utilizing a continuous miner cutting head. No mining personnel are required to go underground. In the case of the study, the remaining coal was to be mined using a highwall miner (ADDCAR 2019), with a circulating fluidized bed fly ash grout pumped into the horizontal mine slots, stabilizing the roof so that the pillars could be mined on a second pass. Because of logistics, including unsafe mining conditions due to the highwall lithology, the experiment was not conducted at this site and the test was conducted elsewhere in eastern Kentucky (Robl et al. 1998).

In this study, the geochemistry, with an emphasis on the Ge and Rare Earth Element concentrations; organic and inorganic petrology; and the palynology of the Princess No. 3 coal in a small area of Greenup County, Kentucky, are investigated. With the exceptions of a regional investigation of whole-channel samples from cores (Hower and Wild 1981) and the study of the Mudseam coal in Elliot County, the correlative of the Princess No. 3 coal (Esterle et al. 1992), few detailed studies of northeastern Kentucky coals have been undertaken. As such, this investigation is an opportunity to further examine coals in a part of eastern Kentucky that is geologically more akin to the settings found in Ohio, northern West Virginia, and southwestern Pennsylvania than it is to the remainder of eastern Kentucky.

The coal was collected in lithotype intervals from two cores and one mine face site at a single mine site (Figs. 2, 3a). The underlying Princess No. 3 leader coalbed was sampled a short distance from the mine face, at a point overlying a drainage pond (Fig. 3b).

Location of Greenup County (arrow), Kentucky, with inset of sampling locations. Simplified lithologic profiles of the three sampling sites are shown on the local-area map

a 1-m-thick section of coal and clay split (samples 1144–1151). b Sampling of the 13-cm-thick leader coalbed illustrating expedient, albeit not recommended, sampling procedure

Petrology was conducted at the University of Kentucky Center for Applied Energy Research (CAER) on epoxy-bound particulate pellets prepared to final 0.05-μm alumina polish. Optical microscopy was conducted on a Leitz Orthoplan microscope with polarized reflected-light, oil-immersion optics at a final magnification of 500 ×. Vitrinite reflectance was done with the incoming light polarized at 45° and the reflected light passing through a 546-nm bandwidth filter on the path to the photomultiplier. The photomultiplier was standardized using glass standards of known reflectance. Maceral identification was based on nomenclature from the International Committee for Coal and Organic Petrology (1998, 2001) and Pickel et al. (2017).

For the basic coal analyses performed at the CAER, proximate analysis followed ASTM Standards D3173/D3173M-17a (2017), D3175-18 (2018b), and D3174-12 (2018a); total sulfur and forms of sulfur analyses followed ASTM Standards D4239-18e1 (2018c) and D2492-02 (2012), respectively; and the ultimate analysis and heating value determinations were performed based on ASTM Standards D3176-15 (2015) and D5865/D5865M-19 (2019), respectively. Ash chemistry at the CAER was analyzed by X-ray fluorescence on a Phillips PW2404 X-ray spectrometer following procedures outlined by Hower and Bland (1989). Inductively coupled plasma mass spectrometry (X series II ICP-MS), in pulse counting mode (three points per peak), was used to determine trace elements in the coal samples (samples from 1155 to 117) at China University of Mining and Technology, Beijing (CUMTB). For ICP-MS analysis, samples were digested using an UltraClave Microwave High Pressure Reactor (Milestone) (after Dai et al. 2011). Arsenic and Se were determined by ICP-MS using collision cell technology (CCT) in order to avoid disturbance of polyatomic ions (Li et al. 2014). Multi-element standards (Inorganic Ventures: CCS-1, CCS-4, CCS-5, and CCS-6; NIST 2685b and Chinese standard reference GBW 07114) were used for calibration of trace element concentrations. The method detection limit (MDL) for each of the trace elements is calculated as three times the standard deviation of the average from the blank samples (n = 10).

A field emission-SEM (FEI Quanta™ 650 FEG), in conjunction with an EDAX energy-dispersive X-ray spectrometer (Genesis Apex 4), at CUMTB was used to study the modes of occurrence of the minerals, and also to determine the occurrence of selected elements. Samples were carbon-coated using a Quorum Q150T ES sputtering coater, and were then mounted on standard aluminum SEM stubs using sticky conductive carbon tabs. The working distance of the FE-SEM–EDS was 10 mm, beam voltage 20.0 kV, aperture 6, and spot size 5.0. The images were captured via a retractable solid state back-scattered electron detector.

Palynology was investigated at the Kentucky Geological Survey. Palynomorphs were liberated by first oxidizing 2–3 g of − 20 mesh coal with Schulze’s Solution (concentrated nitric acid saturated with potassium chlorate). Following oxidation, samples were digested with 5% potassium hydroxide, repeatedly washed with distilled water, and concentrated with zinc chloride (specific gravity 1.9). Amorphous organic matter (AOM) was removed from the residues using ethylene glycol monoethyl ether (2-ethoxyethanol), ultrasonic vibration, and short centrifugation (Eble 2017). Samples were strew-mounted onto 25-mm square cover glasses with polyvinyl alcohol, and fixed to 75- × 25-mm microscope slides with acrylic resin. Spore and pollen abundances are based on a count of 250 palynomorphs for each sample. Palynomorph data are listed according to natural affinity for the following plant groups: lycopod trees, small lycopods, tree ferns, small ferns, seed ferns (pteridosperms), calamites and cordaites. Parent plant affinities of dispersed Carboniferous miospore taxa were derived from extensive summaries provided by Ravn (1986), Traverse (1988) and Balme (1995).

The Princess No. 3 at the study site and, in general, throughout the immediate vicinity of the mine (Sheppard and Ferm 1962; Ferm 1963) is overlain by a gray shale. Sheppard and Ferm (1962) and Dobrovolny et al. (1966) noted that the shale is locally channeled, but no such features were observed at the study site.

3.1 Chemistry

Samples identified as coal have a wide range of ash yields, ranging from < 5% (sample 1164; dry basis) to a clarain with > 77% ash yield (sample 1148) and a bone with nearly 52% ash yield (sample 1168) (Table 1). Megascopically, bone is nominally a high-ash, dull lithotype. Sample 1163, also identified as bone, has 6.14% ash yield, illustrating that the translation between field descriptions and actual analyses of dull lithotypes (bone, dull clarain, durain) can be difficult because much of the lithotype description is driven by the texture; the fine grain size of dull lithologies conceals the details of the composition (Hower et al. 1990). The lithologies above and, in the case of sample 1168, below the 25- to 31-cm-thick parting have the highest ash yield.

Total sulfur content ranges from 0.63% (as-determined basis) in coal sample 1168, in which case the sulfides are diluted by the clay content, to 13.40% in sample 1165, a coal with a 2.2-cm pyrite layer between the dull clarain lithologies. In general, the total sulfur is high, above 2% (as-determined basis). With the exception of sample 1165, the highest pyritic sulfur in each seam section is found in the top lithotype, a reflection of possible marine influence associated with the deposition of the overlying black to gray shale. The highest concentrations of Fe₂O₃ (expressed as a percent on the ash basis) (Table 1) and siderophile elements (such as Ni, Co, and As and, to a lesser extent, Se; ppm ash basis; Table 1)^{Footnote 1} are generally also found in the top lithotypes.

3.2 Petrology

The maceral content of the samples is given on Table 2. Since the samples were from a small area, reflectance was only done on sample 1146, which has a R_max = 0.58% and a R_random = 0.54%, indicating a high volatile C bituminous rank.

The megascopic lithology, generalized as the seam sections on Fig. 2, is simple, with dull clarain and bone being the dull lithologies and clarain being the lone bright lithology. Both in the upwards shift from clarain to a dull clarain mix and in the upwards decrease in the amount of vitrinite, the coal shows the “dulling upwards” trend found in many Appalachian coals (Esterle and Ferm 1986; Esterle et al. 1989, 1992; Hubbard et al. 2002). The total vitrinite content is dominated by telovitrinite macerals and the total inertinite content is dominated by fusinite + semifusinite. The “dulling upwards” trend also seems to be related to the absence of any obvious TiO₂-Zr enrichment in the basal lithotype. The latter chemical association is often found in basal durains (Hower and Bland 1989; Hower et al. 1994a, b; Johnston et al. 2015), suggesting a greater detrital influx into the developing mire.

Among the minerals, clays are found dispersed within macerals, as thin layers, and as cell fillings within fusinite (Fig. 4). Clay/quartz layers contain TiO₂ minerals. Pyrite and marcasite occur as euhedra, framboids, cell lumen fills, and multiple-generation overgrowths of euhedra and framboids (Figs. 5, 6). Complex two- and three-generation overgrowths of Fe-sulfides have been noted in other coals in eastern Kentucky (Hower and Pollock 1989; Ruppert et al. 2005; Diehl et al. 2012).

Kaolinite in the samples investigated in this study. a, b kaolinite in fusinite lumens. c Kaolinite and chalcopyrite in fusinite lumens. d Detrital kaolinite and quartz and La–Ce–Nd-bearing phosphate mineral in collodetrinite. a Sample 1165; b, d sample 1155; c sample 1168

SEM backscattered electron images of framboidal in coal samples. a Framboidal pyrite in vitrinite. Sample 1154. Image 301-25. Scale = 50 μm; air optics. b Framboidal pyrite in vitrinite. Sample 1154. Image 301-28. Scale = 50 μm; air optics. c Overgrown framboidal pyrite in macrinite. Sample 1148. Image 302-36. Scale = 50 μm; air optics. d Overgrown framboidal pyrite in macrinite. Sample 1145. Image 302-36. Scale = 50 μm; air optics

SEM backscattered electron images of pyrite in coal samples. a Pyrite and sulfate minerals in fusinite matrix. SEM image 1155 025. b Pyrite in fusinite lumens. SEM image 1155 027. c Pyrite in fusinite lumens. SEM image 1155 026. d Pyrite in clay and vitrinite. SEM image 1155 020. e Pyrite framboid with euhedral crystals. SEM image 1143 058. f Euhedral pyrite in fusinite lumens. SEM image 1147 051

3.3 Coal geochemistry

The percentages of major-element oxides and concentrations of trace elements in the samples investigated in this study are listed in Tables 3, 4 and 5. The highest concentrations of Fe₂O₃ (expressed as a percent on the ash basis) (Table 3) and siderophile elements [such as Ni, Co, and As and, to a lesser extent, Se; ppm ash basis; Tables 4, 5 (see footnote 1)] are generally also found in the top lithotypes.

Compared to average values for world hard coals (Ketris and Yudovich 2009) and based on the enrichment classifications by Dai et al. (2015), elements Li, Be, Zn, Ge, Se, Ta and W in HW-2 series samples are slightly enriched (2 < CC < 5; Fig. 7); elements such as Cu, In, Sn, Ba, Hf, Tl, Bi, Th, and U HW-2 series samples are depleted (CC < 0.5). Selenium HW-2 series samples is enriched (5 < CC < 10) and As, Tl and Pb are slightly enriched (2 < CC < 5) relative to average values for world hard coals (Ketris and Yudovich 2009). The elements (with an exception of Tl) that are depleted in HW-2 series samples also have a lower concentration HW-2 series samples (CC < 0.5).

Concentration coefficients (CC, the ratio of trace-element concentrations in investigated samples vs. average concentrations in world coals) for HW-2 and HW-3 series samples

As has been widely observed in other coals (Hower et al. 2002; Yudovich Ya 2003b), the Ge content is highest in the top and bottom lithotypes of the two sections. Similarly, U is highest in the uppermost lithotype, with significant concentrations in basal lithotypes, as has been observed for U concentrations in many other coals (for example, studies by Berthoud 1875; Breger et al. 1955a, b; Breger and Schopf 1955; Zubovic 1966; Szalay and Szilágyi 1969; Eskenazy 1992; Yudovich Ya 2003a). In both cases, the enrichment can be attributed, at least in part, to infiltration of mineralized fluids from the surrounding sediments (Yudovich Ya 2003a, b).

In the HW-2 and HW-3 series samples, parting samples are characterized by higher REY (rare earth elements and Y) concentrations than the coal benches, and have light-REE enrichment type. However, REY in the coal benches were weakly fractionated (Fig. 8).

REY distribution patterns of the HW-2 and HW-3 series samples (Normalized to Upper Continental Crust, UCC; Taylor and McLennan 1985)

The ratio of light to heavy rare earth elements (LREE/HREE) is lower at the top of the coalbed, coincident with the high-S/high-Fe₂O₃ zone noted above (Table 6). The transition from LREE/HREE < 3.6 at the top of the coal to > 8 (and up to 10.58) in the lower parts of the coal is primarily a function of an increase in the LREE. For example, sample 1162, the top bench, has an LREE/HREE of 3.14 while sample 1168 has a ratio of 10.58; the HREE’s are 29.6 and 28.9 ppm, respectively.

3.4 Palynology

The palynology indicates that lycopod tree-derived spores dominate the basal bench of the two columns analyzed (Table 7; Fig. 9). The correlation between abundant lycopod tree spores, particularly Lycospora micropapillata and L. granulata, and high vitrinite, seen here in the basal lithotypes, has been noted elsewhere in the region (Eble et al. 1994; Hubbard et al. 2002). The thick parting in both sequences is co-dominated by small lycopod spores (mainly Radiizonates difformis and R. rotatus) and tree fern (mainly Punctatisporites minutus) spores in the lower section of the parting. The upper section of the parting is dominated by tree fern spores (Punctatisporites minutus); tree ferns continue to be important in the upper part of the coal with small ferns (various taxa, but no dominant species), cordaites (Florinites florini), and calamites (primarily Laevigatosporites minor) present in significant amounts. Overall, the palynology of the coal lithotypes above the parting shows much greater floral diversity than is seen in the basal lithotypes.

Graphical representation of the palynology of the 1155 (top of coal) to 1161 (seat rock) section

Of all the Pennsylvanian plant types, the growth habits, reproductive biology and paleoecology of the arborescent lycopods is best known. Furthermore, much of what we know about lycopod trees has been derived from the study of coal balls, which represent permineralized peat, in the Illinois Basin (e.g., Phillips et al. 1974, 1985; DiMichele et al. 1985; DiMichele and Phillips 1994). All of the lycopod trees were heterosporous and dominated a majority of Early and Middle Pennsylvanian mire communities. Lycopod trees possessed vegetative and reproductive traits that allowed them to exploit low-nutrient swamp environments more effectively than any other contemporaneous Pennsylvanian plant group (DiMichele and Phillips 1985).

In the Princess No. 3 coal, Lycospora granulata and L. micropapillata are the dominant arborescent lycopod spores. Lycospora granulata was produced by Lepidophloios halli. Lepidophloios was best adapted to very wet environments (like paleomires). The reproductive structure of Lepidophloios was Lepidocarpon, a boat-shaped megasporanium/megasporophyll specifically adapted for water dispersal (Phillips 1979). Other vegetative features included root systems (Stigmaria) with abundant aerenchymatous tissue and outer cortex (bark) tissue that was water and rot resistant (DiMichele and Phillips 1985, 1994). Community paleoecology indicates that Lepidophloios was dominant in low-diversity ecotones. This relationship suggests high levels of abiotic stress that excluded other kinds of plants. It appears likely that a longstanding water regime was probably the major abiotic factor controlling the composition of Lepidophloios-dominant assemblages. Under such conditions, small, ground-cover and plants would have had difficulty becoming established.

Lycospora micropapillata was produced by Paralycopodites. Paralycopodites, may have been part of a pioneer type flora, based on the common occurrence of Lycospora micropapillata and L. orbicula in seat rocks (coal underclays) and basal coal layers. These palynotaxa are also prevalent within, or near, clastic layers in many coal beds. In coal ball assemblages, Paralycopodites-rich intervals are commonly associated with degraded peat and fusain, which suggest marginal conditions of peat preservation and frequent oxidation of the surficial peat through both biotic (microbial degradation) and abiotic (fire) pathways. In the Princess No. 3 coal, Lycospora micropapillata and L. orbicula are most abundant in the basal benches of both columns that were studied palynologically.

Radiizonates difformis and R. rotatus are the principle small lycopod spores in the Princess No. 3 coal bed and, biostratigraphically, serve to identify the coal as being equivalent with Bolsovian age strata of western Europe (Clayton et al. 1977), and Atokan age strata of the Eastern Interior (Illinois) Basin (Peppers 1985, 1996). Densosporites (and related crassicingulate genera, e.g. Cristatisporites) were produced by Omphalophloios, which was first described by David White (1897) from the Middle Pennsylvanian of Missouri. High percentages of Densosporites defined, in part, the “densospore” phase, which reflected the terminal part of a hydroseral series associated with peat accumulation in domed, ombrogenous mires (Smith 1957, 1962, 1964, 1968). Alternatively, high percentages of Densosporites in the terminal portions of the Lower Kittanning coal in the northern Appalachian Basin were thought to indicate a halophytic flora that developed in response to marine transgression and deposition of the overlying Columbiana Limestone (Habib 1966; Habib and Groth 1967).

Although these interpretations of Omphalophloios are highly disparate, both appear to have merit. Subsequent work by the current authors indicate that abundant Densosporites is commonly associated with (1) coal that is low in ash and sulfur, supporting formation in a domed, ombrogenous peat-forming environment, but also (2) coal with elevated amounts of ash and sulfur, which is indicative of peat formation in a planar, topogenous mire. The common element in both cases is that high percentages of Densosporites are almost always associated with increased amounts of liptinite and inertinite macerals, and relatively low percentages of vitrinite. Based on this observation, Omphalophloios was apparently a lycopod that could develop in areas where peat preservation was poor, or even minimal. Although there is no direct paleobotanical evidence that Omphalophloios was a true halophyte, some species may very well have been saline tolerant to some degree, or at least were water-stress tolerant.

Tree fern spores are both diverse and abundant in the Princess No. 3 coal. Tree ferns have their origins in the Early Pennsylvanian, rapidly expand in abundance in the late Middle Pennsylvanian (Desmoinesian, Asturian), and are a dominant element of mire floras in the Late Pennsylvanian (Stephanian). Tree ferns, and probably ferns in general, exhibited considerable ecological amplitude and adaptability. They had the least specialized method of reproduction of all the major coal-forming plant groups, exhibited a totally herbaceous growth habit, and possessed a massive root mantle for aeration and support. These features adapted them to a wide variety of swamp and swamp-like environments (Phillips and Peppers 1984).

Sphenopsid spores are a subdominant element of the Princess No. 3 coal palynoflora. Sphenopsid spores in the Pennsylvanian include Calamospora, large species of Laevigatosporites and, to a lesser extent, Vestispora, and Reticulatisporites. Sphenopsids are subdivided into two major evolutionary groups: arborescent calamites and the shrubby or vine-like Sphenophyllum (DiMichele et al. 1986). Both groups occurred in paleomires and were locally abundant in clastic substrate areas. Sphenophyllum was widespread throughout most of the Pennsylvanian, though a minor contributor to biomass because of its small size of (DiMichele et al. 1986). Calamites were also common constituents in Pennsylvanian swamps but, like Sphenophyllum, were also minor peat-biomass contributors (DiMichele et al. 1986). Although their abundance is generally associated with inorganic partings and high ash coal, exceptions exist. This type of general distribution supports the interpretation that calamites may have been centered outside of mires, with some forms persisting in both mire and clastic substrate environments (DiMichele et al. 1985).

Cordaite pollen in the Princess No. 3 coal occurs primarily as Florinites, with F. florini being the most common species. In general, Florinites is more common in the parting and coal benches that occur above the parting. Cordaites were a diverse group of plants with a broad ecological spectrum during the Pennsylvanian, ranging from dry, upland areas to waterlogged areas including mires (Raymond 1988). Coal ball studies recognize two principle stem genera, Mesoxylon and Pennsylvanioxylon (Cordaixylon). Mesoxylon was prominent in Early though mid-Middle Pennsylvanian (Langsettian—Bolsovian) mires (Phillips and Peppers 1984; Costanza 1985), and is inferred to have occupied areas of mires that were well-drained and periodically exposed. This interpretation is largely based on Mesoxylon roots lacking aerenchymatous tissues, and the poorly-preserved nature of Mesoxylon peats in coal balls. Mesoxylon produced Sullisaccites pollen, which is seldom reported in Pennsylvanian palynological literature. This is probably the result of misidentification with morphologically similar palynotaxa (e.g., Florinites, Wilsonites and Potonieisporites).

Pennsylvanioxylon was originally described as a small mangrove-like tree, with some species being salt-water tolerant (Raymond 1988). Others apparently developed in fresh water conditions. Unlike Mesoxylon, Pennsylvanioxylon roots were aerenchymatous, thus suggesting growth in water saturated, or even flooded, areas. As with Mesoxylon, Pennsylvanioxylon peats tend to be highly degraded and fusinized in coal balls. Pennsylvanioxylon produced Florinites pollen, which tends to be erratic in distribution. Samples with abundant Florinites are commonly high in ash, liptinite and inertinite, all of which are suggestive of peat accumulation under marginal conditions of preservation (DiMichele and Phillips 1994).

The high volatile C bituminous Princess No. 3 coal, correlative with the Hazard No. 7 and the lower splits of the Peach Orchard and Coalburg coals, was studied as lithotype intervals from three sites at a mine site in the Argillite 7.5′ quadrangle, Greenup County, Kentucky.

In general, the coal exhibits a “dulling upwards” trend, with decreasing vitrinite and a greater tendency towards dull clarain and bone lithotypes towards the top of the coal. The basal lithotype is marked not only by relatively high concentrations of vitrinite, but also by a dominance of lycopod tree spores. The lower portion of the middle parting is co-dominated by tree fern and small lycopod spores. Tree ferns dominate the upper bench of the coal, with contributions from small ferns, cordaites, and calamites.

The lithotypes have a wide range of ash yield contents and are generally have a moderate- to high-S content. Sulfur, Fe₂O₃, and certain siderophile elements are highest near the top of the coal. Uranium and Ge are enriched at the top and bottom margins of the coal. The rare earth chemistry at the top of the coal has a significantly lighter distribution (higher LREE/HREE) than at the base of the coal. While REE in coal attracted attention in the 2010s, as with most, if not all, coal resources, the exploitation of the coal for its REE content is likely to be driven by the need for the coal, with the REE being a valuable by-product of the coal processing and combustion. At present, coal production in eastern Kentucky is significantly less than what it was 30 years ago, decreasing from 131 Mt in 1990 to about 17 Mt in 2018 (Kentucky Energy and Environment Cabinet 2017, 2019) with much of the current production centered in central eastern Kentucky, therefore, exploitation of the coal studied here is not likely to be a near-term prospect.