Investigating Physicochemical Methods to Recover Rare-Earth Elements from Appalachian Coals

Abstract

The demand for rare-earth elements is expected to grow due to their use in critical technologies, including those used for clean energy generation. There is growing interest in developing unconventional rare-earth element resources, such as coal and coal byproducts, to help secure domestic supplies of these elements. Within the U.S., Appalachian Basin coals are particularly enriched in rare-earth elements, but recovery of the elements is often impeded by a resistant aluminosilicate matrix. This study explores the use of calcination and sodium carbonate roasting pre-treatments combined with dilute acid leaching to recover rare-earth elements from Appalachian Basin coals and underclay. The results suggest that rare-earth element recovery after calcination is dependent on the original mineralogy of samples and that light rare-earth minerals may be more easily decomposed than heavy rare-earth minerals. Sodium carbonate roasting can enhance the recovery of both light and heavy rare-earth elements. Maximum recovery in this study, ranging from 70% to 84% of total rare-earth elements, was achieved using a combination of calcination and sodium carbonate roasting, followed by 0.25 M citric acid leaching.

1. Introduction

Rare-earth elements, including the 15 lanthanide elements plus scandium and yttrium (REY), are a group of metallic elements that serve an important role in the economy, national security, and energy generation. The elements are commonly divided into two groups based on their atomic weights and ionic radii: light rare-earth elements (LREYs—Sc and La through Eu) and heavy rare-earth elements (HREYs—Y and Gd through Lu). The global demand for REYs is expected to increase over the next several decades due to their use in the clean energy sector, where the elements are used to manufacture powerful magnets for wind turbines and electric vehicle motors [1]. Identifying new sources of REYs and securing supply chains has become a global priority [1].

Utilizing coal and coal byproducts, such as refuse and fly ash, as unconventional sources of REYs has garnered significant attention. The REY content of some coal seams and associated clays can rival the enrichment seen in conventional ores [2]. Within the United States, Appalachian Basin coals and coal byproducts are particularly enriched in REYs [3,4,5]. Coal-producing states within the basin can utilize existing infrastructure and legacy coal-mining wastes to reduce the costs associated with producing REYs from these unconventional resources. However, the REYs in Appalachian coals and associated rocks are often present in resistant phosphate minerals encased in an aluminosilicate matrix that can inhibit their recovery [6,7,8]. Several studies on Appalachian coal underclays have identified monazite, xenotime, apatite, and crandallite embedded in the aluminosilicate matrix, or as pore-filling minerals [6,7,9]. Similar REY mineral occurrences were found in Appalachian Basin coals, roof rock, and coal refuse [8,10,11]. To utilize these coals and coal byproducts as unconventional REY resources, extraction methods must be developed to efficiently decompose aluminosilicate and phosphate minerals to liberate the REYs.

Studies have been conducted to evaluate the performance of pre-treatment methods that can potentially decompose clays and phosphates to enhance REY recovery from coal, coal-adjacent rocks, and coal refuse. Several researchers have investigated the use of calcination as a pre-treatment method, followed by acid leaching of the calcined product. Zhang and Honaker [12] recovered nearly 80% of the total REYs (TREYs) from Appalachian Basin coal refuse and middlings after calcining samples at 600 °C and then leaching them with 1.2 M HCl for 5 h. However, LREYs accounted for most of the recovery, with approximately 40–60% of HREYs remaining in the residual matter [12]. In a follow-up study using similar calcination and leaching conditions, Zhang and Honaker [8] recovered close to 90% of LREYs from Illinois Basin coal refuse, but less than 50% of the HREYs were recovered. Similarly, Yang et al. [13] recovered close to 80% of LREYs from an Illinois Basin middlings sample by calcining it at 750 °C and leaching it with 1.2 M H₂SO₄, but the recovery of most individual HREYs were below 50%. Gupta et al. [14] found that the optimum calcination temperature to maximize REY recovery varied between about 600 and 800 °C for different coal-refuse samples, but in most cases, LREY recovery still far exceeded HREY recovery. While all of these studies showed that calcination can enhance overall REY recovery, a significant proportion of HREYs were left behind in the leached material.

Alkaline leaching or dry roasting is a method that has been investigated for enhancing REY recovery from coal-related rocks and byproducts such as fly ash. Alkaline additives can react with mineral phases such as silicates and phosphates, increasing the leachability of REY phases in acid [15,16,17]. Yang et al. [13] found that pre-treating a fine refuse sample with an 8 M NaOH solution before leaching with sulfuric acid improved TREY recovery from 22% to 75%. King et al. [15] compared aqueous acid and alkaline leaching methods to recover REYs from coal fly ashes from several basins. The study showed that aqueous NaOH leaching followed by 2 M HCl leaching significantly improved TREY recovery compared to acid-only leaching with 12 M HCl. Total REY recovery using the combined alkaline–acid method ranged from 44.7 to 64.5% in Illinois Basin fly ashes and 48.8 to 85.9% in Appalachian Basin ashes [15]. Pan et al. [16] tested several alkaline roasting additives to evaluate their effects on REY recovery from coal fly ash. The roasting was followed by water leaching to remove soluble phases and unreacted additives, and subsequent acid leaching using 6 M HCl to recover REYs. They found that high-temperature roasting with NaOH and Na₂CO₃ increased REY recovery from 20% (baseline leaching) to 79% and 89%, respectively [16]. Tang et al. [17] investigated the effects of temperature, stirring speed, acid molarity, leaching time, and reactant ratios on REY extraction from fly ash. The authors concluded that roasting with a 1:1 mass ratio of fly ash to Na₂CO₃, followed by leaching with 3 M HCl stirred at 400 rpm for 2 h using a solid–liquid ratio of 1:20, provided optimal results. Under these conditions, REY recovery neared 73%, but only the elements La, Ce, Pr, Nd, and Y were included in the analysis [17].

Since pre-treatment techniques, including calcination and alkaline roasting, can enhance the recovery of REYs, they may facilitate the use of less concentrated acids, including organic acids such as citric acid or EDTA, during leaching [18,19]. This study examines the effects of leaching coal and coal-related rocks with either dilute HCl or dilute citric acid after calcination and Na₂CO₃ roasting. The experimental procedure is based on Pan et al.’s study [16], which incorporates a water-leaching step to remove water-soluble species and residual reactant (in the case of alkaline roasting) before acid leaching. The use of more dilute acid solutions for REY recovery may provide economic and environmental benefits, including reducing the costs and risks associated with purchasing reagents, handling large quantities of solution, and treating and disposing of aqueous waste.

2. Materials and Methods

Coal and coal-adjacent rock samples provided by the West Virginia Geologic and Economic Survey were subjected to a series of acid leaching, calcination, and alkaline roasting experiments. Coal samples were collected from the Fire Clay and Sewell coal seams. The Fire Clay coal sample was a low-ash bituminous coal, while the Sewell sample was a mixture of high-ash coal and interbedded shale/clay. The underclay sample was also associated with the Fire Clay coal bed, but did not directly underlie the Fire Clay coal sample used in this study. Sample characteristics including formation names, geologic ages, and ash yields are shown in

Table 1. Sample characteristics.

Sample Name	Lithology	Ash Yield	Formation Name	Geologic Age
Sewell Coal	Coal with interbedded shale/clay	56.1%	New River	Lower Pennsylvanian
Fire Clay Coal	Bituminous coal	6.6%	Kanawha	Middle Pennsylvanian
Underclay	Clay	57.6%	Kanawha	Middle Pennsylvanian

.

2.1. Bulk-Sample Characterization

Sulfur and Carbon: Total sulfur and total carbon analyses were completed by Pittsburgh Mineral & Environmental Technology (PMET) in New Brighton, PA. Representative samples were ground and sieved to achieve grain sizes of ≤ 45 µm before analysis. The samples were scanned with an ELTRA CS800 Carbon–Sulfur Determinator following calibration with Alpha Resources standards. Results were reported as the average of 2–3 aliquots per sample.

Scanning electron microscopy–energy-dispersive X-ray spectroscopy: Scanning electron microscopy and energy-dispersive X-ray spectroscopy (SEM-EDS) were used to identify REY-bearing minerals in bulk rock sample chips. The analysis was completed at the National Energy Technology Laboratory in Albany, OR, using the methods detailed in Montross et al. [20] and summarized in Yesenchak et al. [10]. Briefly, portable X-ray fluorescence (pXRF) was used to screen samples for REY content, and those with the highest La, Ce, and Y concentrations were selected for SEM-EDS. Elemental maps, point spectra, and line spectra were used for the putative identification of REY-bearing mineral phases.

Bulk-Sample Digestion: The whole-rock concentrations of select minor elements and REYs were measured by digesting representative bulk samples with a modified sodium peroxide fusion method [21] at the WVU Institute for Sustainability and Energy Research (WISER) analytical lab, Morgantown, PA. Digested samples were analyzed using inductively coupled plasma optical emission spectroscopy (ICP-OES) following U.S. EPA Method 200.7 (minor elements) [22] and inductively coupled plasma mass spectrometry (ICP-MS) following U.S. EPA Method 200.8 (REY) [23]. Rare-earth element detection limits are reported in Yesenchak et al. [10].

2.2. Calcination and Alkaline Roasting Experiments

Experiments were conducted in four stages: baseline leaching of raw samples using 0.05 M acids, leaching of calcined (ashed) samples using 0.05 M acids, leaching of Na₂CO₃-roasted samples using 0.05 M acids, and leaching of Na₂CO₃-roasted samples using 0.25 M acids. Each stage of the experiment consisted of six samples: 1 g of each bulk sample (Sewell coal, Fire Clay coal, and underclay) leached with HCl and 1 g of each bulk sample leached with citric acid. The bulk coal and underclay samples were ground and sieved to a grain size of ≤150 µm before beginning experiments. The percentage of REY recovered during each experimental stage was calculated relative to the bulk-sample digestion results.

Baseline leaching of raw samples: One gram of each raw coal and underclay sample was added to a borosilicate beaker followed by 100 mL of Milli-Q deionized water. Samples were placed on a hotplate set to 90 °C and agitated using magnetic stir bars at 400 rpm for 2 h. After water leaching, samples were filtered using 0.4 µm Isopore PC membrane filters, and leachate was preserved using 1% (v/v) nitric acid. The residual samples were placed back into their original borosilicate beakers and 100 mL of 0.05 M HCl or 100 mL of 0.05 M citric acid was added. Samples were placed on a hotplate set to 60 °C and agitated using magnetic stir bars at 400 rpm for 2 h. Leachate was filtered using 0.4 µm Isopore PC membrane filters and preserved using 1% (v/v) nitric acid.

Leaching of calcined samples: Samples were calcined in a Thermolyne Type 1300 muffle furnace at 650 °C for approximately 2.5 h. One gram of each ashed sample was placed in a borosilicate beaker and subjected to water leaching and acid leaching in the same manner as described in the baseline leaching section. An additional ashed sample was saved for Na₂CO₃ roasting experiments.

Leaching of Na₂CO₃-roasted samples: One gram of each ashed sample was mixed with one gram of anhydrous Na₂CO₃ in a corundum crucible and roasted at approximately 875 °C in the Thermolyne Type 1300 muffle furnace for 2 h. Samples were then removed from the furnace and allowed to cool before transferring them into borosilicate beakers. A small amount of Milli-Q deionized water was used to fully transfer the roasted samples, and they were subsequently placed in a low-temperature oven for approximately 36 h to dry. The dried samples were comminuted using a mortar and pestle and added back to the borosilicate beakers. Samples were subjected to water leaching and acid leaching in the same manner as described in the baseline leaching section. The process was repeated for an additional set of six samples using 0.25 M HCl and 0.25 M citric acid in place of the 0.05 M acids.

3. Results

3.1. Sulfur and Carbon

Total sulfur and total carbon results are reported in

Table 2. Bulk-sample total carbon and sulfur (wt. %).

Sample	Carbon	Sulfur
Sewell coal	36.17	0.36
Fire Clay coal	78.61	0.44
Underclay	29.5	0.32

. Sulfur concentrations in all bulk samples are low, ranging from 0.32% to 0.44%. Total carbon concentrations are more variable, ranging from 29.50% (underclay) to 78.61% (Fire Clay coal).

3.2. SEM-EDS

The rare-earth-element-bearing minerals identified during SEM-EDS analysis are shown in

Table 3. Rare-earth-element-bearing minerals identified during SEM-EDS (Sewell and Fire Clay data previously reported in Yesenchak et al. [10]).

Sample	REY-Bearing Minerals Identified
Sewell Coal	Y, Gd, and Sc in xenotime grains < 5 µm in length
	Yb and Sc with Ca-oxalate in organic matter pore space
	La, Ce, Tb, Yb, and Lu in an area of mixed sulfides (Fe, Pb, Cu, Ni, Se)
	Yb associated with an unidentified Ca-S mineral
Fire Clay Coal	La, Ce, Nd and P co-located throughout clay layers
Underclay	Ce co-located with framboidal pyrite
	La, Ce, Pr, Nd, Sm, and Gd in monazite grains < 5 µm in length
	Yb and Lu in xenotime grains < 5 µm in length

. The results for the Sewell coal sample showed REYs in association with several mineral phases, including xenotime, Ca-oxalate, mixed sulfides, and an unidentified Ca-S mineral. In the underclay sample, REYs were present in monazite and xenotime, and co-located with framboidal pyrite. No discrete REY-bearing minerals were identified in the Fire Clay coal sample, but EDS mapping revealed La, Ce, and Nd co-located with P dispersed throughout the clay layers. Examples of the REY-bearing minerals identified in the underclay and Sewell coal samples are shown in Figure 1, Figure 2 and Figure 3. Backscattered electron images and EDS maps for the Fire Clay coal sample, and additional images, maps, and EDS spectra for the Sewell sample are published in Yesenchak et al. [10].

3.3. Bulk-Sample Digestion

Select minor-element concentrations are shown in

Table 4. Bulk whole-rock minor-element concentrations (wt. %).

Bulk Rock	Al	Ca	Fe	K	Mg	Mn	P	Si
Sewell Coal	9.108	0.232	1.089	2.752	0.385	0.005	0.028	13.633
Fire Clay Coal	0.839	0.128	0.732	0.141	0.029	0.002	0.002	1.527
Underclay	9.058	0.211	0.645	1.953	0.246	0.004	0.040	16.911

. Compared to the Sewell coal and underclay samples, the Fire Clay coal sample was the least enriched in all of the measured elements, except for Fe. The underclay sample contained the lowest Fe concentration and the highest Si and P concentrations. The Sewell coal sample had the highest concentrations of Al, Ca, Fe, K, Mg, and Mn.

The bulk-sample REY concentrations are shown in

Table 5. Bulk whole-rock REY concentrations (ppm).

Bulk Rock	Sc	La	Ce	Pr	Nd	Sm	Eu	Gd	Tb	Dy	Y	Ho	Er	Tm	Yb	Lu	TREY
Sewell Coal	19.48	55.59	110.27	13.67	53.28	10.37	2.04	9.74	1.38	7.72	31.88	1.48	4.11	0.60	3.85	0.54	326.00
Fire Clay Coal	2.07	6.37	14.34	1.79	7.66	1.79	0.37	2.31	0.38	2.40	13.72	0.53	1.48	0.20	1.21	0.17	56.78
Underclay	15.84	96.69	188.65	21.57	82.02	15.40	1.60	14.11	2.14	11.88	57.48	2.25	6.46	0.86	5.60	0.77	523.33

. The underclay sample had the highest whole-rock TREY concentration (523.3 ppm), followed by the Sewell coal (326.0 ppm) and the Fire Clay coal (56.8 ppm).

Rare-earth element concentrations in coal and coal-adjacent rocks are often evaluated against the average REY concentrations in the upper continental crust (UCC). After normalization to the UCC, samples can be classified as having L-type (enriched in LREYs), M-type (enriched in MREYs), or H-type (enriched in HREYs) distributions [24]. Bulk-sample UCC-normalized REY distributions are shown in Figure 4. Average UCC concentrations for lanthanides and yttrium are defined by Taylor and McLennan [25], and the average concentration of Sc is defined by McLennan [26]. The UCC-normalized REY distributions vary by sample. The Sewell coal sample has an M-type distribution, which is indicative of coals influenced by acidic fluid circulation [24,27]. The Fire Clay coal sample has an H-type distribution, which is common among world coals [24]. The underclay sample is characterized by a hybrid LREY/MREY-type distribution and a significant negative Eu anomaly.

3.4. Calcination and Alkaline Roasting Experiments

Leaching experiments using dilute HCl and citric acid were conducted in four stages: baseline leaching of raw whole-rock samples using 0.05 M acids, leaching of ashed samples using 0.05 M acids, leaching of Na₂CO₃-roasted samples using 0.05 M acids, and leaching of Na₂CO₃-roasted samples using 0.25 M acids. The concentration of REY recovered by each acid during each experimental stage is shown in

Table 6. Concentration (whole-rock ppm) of REYs recovered by HCl and citric acid (CA) during each experimental stage. BDL indicates elements below method detection limits.

Whole Rock—0.05 M HCl	Sc	La	Ce	Pr	Nd	Sm	Eu	Gd	Tb	Dy	Y	Ho	Er	Tm	Yb	Lu
Sewell Coal	0.31	0.04	0.17	0.01	0.08	0.04	0.01	0.07	0.01	0.06	0.07	0.01	0.03	0.005	0.04	0.004
Fire Clay Coal	0.05	0.10	0.25	0.04	0.16	0.03	0.01	0.04	0.005	0.03	BDL	0.01	0.02	0.002	0.02	0.001
Underclay	0.11	2.75	6.36	0.77	3.05	0.67	0.08	0.58	0.08	0.32	0.91	0.05	0.09	0.01	0.05	0.01
Whole Rock—0.05 M CA	Sc	La	Ce	Pr	Nd	Sm	Eu	Gd	Tb	Dy	Y	Ho	Er	Tm	Yb	Lu
Sewell Coal	0.91	0.003	0.01	0.002	0.02	0.02	0.01	0.04	0.01	0.04	BDL	0.01	0.02	0.004	0.03	0.004
Fire Clay Coal	0.06	0.05	0.20	0.02	0.08	0.01	0.003	0.02	0.003	0.02	BDL	0.005	0.01	0.002	0.01	0.001
Underclay	0.42	0.27	0.92	0.15	0.69	0.19	0.02	0.16	0.03	0.12	0.34	0.02	0.05	0.01	0.04	0.01
Ashed—0.05 M HCl	Sc	La	Ce	Pr	Nd	Sm	Eu	Gd	Tb	Dy	Y	Ho	Er	Tm	Yb	Lu
Sewell Coal	1.12	25.87	54.70	6.53	26.34	4.67	0.63	2.50	0.21	0.78	2.25	0.11	0.28	0.03	0.21	0.03
Fire Clay Coal	0.20	1.56	2.90	0.45	2.11	0.52	0.12	0.68	0.10	0.68	5.07	0.13	0.40	0.05	0.31	0.04
Underclay	1.13	20.64	26.94	2.54	8.45	1.15	0.14	0.93	0.12	0.69	3.24	0.13	0.38	0.05	0.30	0.04
Ashed—0.05 M CA	Sc	La	Ce	Pr	Nd	Sm	Eu	Gd	Tb	Dy	Y	Ho	Er	Tm	Yb	Lu
Sewell Coal	1.41	8.75	21.00	2.84	12.69	2.40	0.34	1.38	0.13	0.56	1.62	0.08	0.22	0.03	0.18	0.02
Fire Clay Coal	0.22	1.35	2.79	0.45	2.18	0.57	0.13	0.75	0.12	0.77	5.80	0.15	0.46	0.06	0.35	0.05
Underclay	1.41	6.43	9.90	1.03	3.55	0.58	0.07	0.55	0.09	0.53	2.66	0.11	0.31	0.04	0.26	0.04
Na₂CO₃ Roast—0.05 M HCl	Sc	La	Ce	Pr	Nd	Sm	Eu	Gd	Tb	Dy	Y	Ho	Er	Tm	Yb	Lu
Sewell Coal	1.02	12.15	8.90	2.92	11.14	2.21	0.44	1.97	0.27	1.59	7.38	0.30	0.87	0.12	0.81	0.12
Fire Clay Coal	0.14	1.52	2.22	0.42	1.73	0.41	0.09	0.52	0.08	0.55	3.89	0.12	0.34	0.05	0.28	0.04
Underclay	0.29	4.18	2.56	0.99	3.84	0.73	0.08	0.73	0.12	0.73	4.27	0.15	0.49	0.07	0.51	0.07
Na₂CO₃ Roast—0.05 M CA	Sc	La	Ce	Pr	Nd	Sm	Eu	Gd	Tb	Dy	Y	Ho	Er	Tm	Yb	Lu
Sewell Coal	11.76	35.39	59.54	9.17	36.75	7.55	1.57	6.74	0.99	5.48	20.41	1.09	3.04	0.44	2.82	0.42
Fire Clay Coal	1.07	3.78	7.96	1.14	4.77	1.20	0.27	1.52	0.27	1.70	9.30	0.37	1.06	0.14	0.87	0.13
Underclay	9.38	54.93	86.64	12.88	51.78	9.92	1.06	9.16	1.36	8.05	33.82	1.53	4.53	0.62	3.96	0.56
Na₂CO₃ Roast—0.25 M HCl	Sc	La	Ce	Pr	Nd	Sm	Eu	Gd	Tb	Dy	Y	Ho	Er	Tm	Yb	Lu
Sewell Coal	12.79	39.23	65.59	10.10	39.75	8.03	1.66	7.17	1.04	5.74	22.40	1.15	3.18	0.46	2.90	0.44
Fire Clay Coal	1.18	4.50	9.41	1.34	5.66	1.39	0.31	1.81	0.31	1.98	10.18	0.43	1.22	0.17	1.01	0.15
Underclay	10.54	63.29	84.78	14.72	58.13	11.17	1.18	10.14	1.51	8.90	39.50	1.69	4.99	0.68	4.38	0.61
Na₂CO₃ Roast—0.25 M CA	Sc	La	Ce	Pr	Nd	Sm	Eu	Gd	Tb	Dy	Y	Ho	Er	Tm	Yb	Lu
Sewell Coal	15.98	47.51	88.71	11.72	45.36	9.19	1.86	8.15	1.11	6.41	27.28	1.22	3.53	0.49	3.19	0.47
Fire Clay Coal	1.53	4.99	10.94	1.43	5.98	1.48	0.32	1.88	0.31	2.01	11.52	0.43	1.24	0.16	1.01	0.14
Underclay	11.34	69.33	113.15	16.53	65.79	12.42	1.34	11.39	1.70	9.97	40.44	1.88	5.48	0.74	4.71	0.66

.

The percentage of TREYs recovered by each acid during each experimental stage is shown in Figure 5. The baseline leaching of raw whole-rock samples recovered only 0.3% to 3.0% of TREYs using 0.05 M HCl, and 0.3% to 0.9% of TREYs using 0.05 M citric acid. The highest recovery for all samples resulted from a combination of Na₂CO₃ roasting followed by 0.25 M citric acid leaching. Recovery ranged from 70.1% (underclay) to 83.5% (Sewell coal) of TREYs using this method. Leaching with 0.25 M HCl after Na₂CO₃ roasting provided the second-highest recovery, ranging from 60.4% (underclay) to 72.3% (Fire Clay coal) of TREYs. Using more dilute 0.05 M acids after Na₂CO₃ roasting resulted in slightly lower recovery for citric acid-leached samples (55.4%–62.6% of TREYs recovered) and significantly lower recovery for HCl-leached samples (3.8%–21.8% of TREYs recovered). Calcining (ashing) samples (with no additives) prior to leaching with 0.05 M acids increased TREY recovery compared to the baseline leaching of the whole-rock samples, but overall, recovery was low, ranging from 5.3% to 28.5% of TREYs using citric acid and 12.8% to 38.8% of TREYs using HCl. As shown in Figure 6, the proportion of individual REYs recovered varied by sample and by method. However, the recovery patterns for all samples were relatively similar after roasting with Na₂CO₃ and leaching with 0.25 M acids, though the magnitude of recovery for certain elements varied.

4. Discussion

The three samples used in this study represent coal and related rock types with varying characteristics and chemical makeup. The Fire Clay coal sample contains the largest organic fraction, as evidenced by the low ash yield, high carbon content, and relatively low concentrations of common inorganic elements, including Si, Al, K, Mg, and P. This sample is also the least enriched in TREYs, and no discrete REY-bearing mineral phases were identified during SEM-EDS analysis. The Sewell coal sample includes a mixture of coal and interbedded shale/clay. The abundance of Al, K, and Mg suggests the presence of illite, which was identified in this sample via qualitative X-ray diffraction in a previous study [10]. Rare-earth elements in the Sewell sample were found in association with a variety of minerals, including xenotime, sulfides, and Ca-bearing phases. The underclay sample is an underclay associated with the Fire Clay coal seam. It contains the highest concentrations of Si, P, and TREY. Rare-earth elements in the underclay sample were identified in REY phosphates and co-located with framboidal pyrite. When normalized to the UCC, the sample exhibited a distinctly negative Eu anomaly (Figure 4). Negative Eu anomalies are typically inherited from sediment source regions containing felsic igneous rocks or volcanic ash, but can also be related to high-temperature illitization [28]. The REY content of some sections of the Fire Clay coal seam in eastern Kentucky are known to be influenced by the presence of a volcanic ash tonstein, or by more diffuse volcanic ashfall [3,29,30]. However, the extent of volcanic influence in the WV Fire Clay samples used in this study (Fire Clay coal and underclay) is unknown.

The leaching experiments in this study compared the REY recovery performances of two different pre-treatment methods: ashing (i.e., calcination) and ashing paired with alkaline roasting using Na₂CO₃. The experiments also compared the performance of dilute HCl and citric acid for leaching REYs. Initial baseline leaching using 0.05 M acids was performed on raw whole-rock samples (i.e., no pre-treatment) that were ground and sieved to achieve a top grain size of 150 µm. The performance of both acids during baseline leaching was poor, with recovery ranging from 0.3% to 3.0% of TREYs (Figure 5). Recovery was fairly consistent over the full suite of elements for all samples (Figure 6).

4.1. Calcination

Ashing the samples at approximately 650 °C as a pre-treatment method followed by 0.05 M acid leaching improved the recovery of REYs, as compared to the baseline leaching of raw whole-rock samples (Figure 5). However, the increase in recovery varied by sample, by element, and by acid used (Figure 6). Other studies have found similar variability when using calcination between 400 and 800 °C as a pre-treatment method for coal and coal refuse. The published literature suggests that calcination enhances REY recovery through two primary mechanisms: the liberation of REY minerals from the clay matrix and the thermal decomposition of the REY-bearing phases themselves [8,12,14]. As such, the mineralogy of samples and REY modes of occurrence exert control over REY recovery after calcination. Clay minerals make up a significant proportion of the inorganic fraction of coal and coal-adjacent rocks and can shield REY-bearing minerals during calcination and subsequent acid leaching. The dehydroxylation of clays during calcination increases surface area and pore volume, enhancing the solubility of clays in acid [8,12,14]. However, the extent of the decomposition of different clay types can vary under the same conditions. The destruction of the crystalline structure of kaolinite begins at around 500 °C, while reducing illite crystallinity requires temperatures above 600 °C [8,12,31]. Additionally, illite has been shown to have a slower acid dissolution rate than kaolinite after calcination [12].

Similar to clays, the decomposition of REY-bearing phases is also dependent on temperature and the specific mineralogy present. In many coals and coal-related rocks, phosphates such as monazite and xenotime are the primary REY minerals. Although the thermal decomposition of REY phosphates typically requires extremely high temperatures (approximately 2000 °C), reactions with compounds such as CaO within the coal and rock samples themselves can reduce the temperature needed for decomposition [12,32]. However, the published literature suggests that thermal treatments are more effective for LREY phosphates than for HREY phosphates [8,12,14]. The fate of REYs in non-phosphate minerals can also play a role in TREY recovery after calcination. For example, REYs can sometimes be found in association with sulfide minerals, including in the Sewell coal and underclay samples used in this study (Table 3). At temperatures above 500 °C, pyrite transforms into hematite, and can potentially sequester REYs in this low-solubility phase [12,14,18]. In addition to inorganic mineral phases, REYs can also be associated with organic matter in coal and organic-rich rocks. Organically associated REYs may be adsorbed onto the organic material or may bond directly with organic compounds [33,34,35]. Heavy REYs are more likely to be organically associated than LREYs due to their stronger complexation ability with humic acids [34,36,37]. A larger organic fraction in a coal or rock sample translates into more opportunity for REYs, especially HREYs, to form organic complexes. During calcination, the destruction of the organic material releases organically bound REYs, increasing their leachability [12,13,14].

In the present study, calcination improved TREY recovery more significantly for the coal samples than for the underclay. Differences in the recovery of LREYs versus HREYs are also apparent, but are not consistent between samples. In the Sewell coal and underclay samples, the increase in LREY recovery was more significant than the increase in HREY recovery (Figure 7). But, in the Fire Clay coal sample, calcination improved the recovery of HREYs more than LREYs. These results are at least partially related to organic matter content. The relatively large organic fraction in the Fire Clay coal sample provides ample opportunity for HREY complexation, while LREYs are more likely to be present within inorganic mineral phases. A previous sequential leaching study using a Fire Clay coal sample found that 54% of the recoverable HREYs and 24% of recoverable LREYs were associated with organic matter, although overall TREY recovery for the full leaching procedure was low [10]. The lower organic matter content of the underclay sample likely contributed to the lower post-calcination recovery. Studies have shown that calcination improves REY recovery in coal-rich samples more than in non-coal-rich samples [12,13].

Rare-earth-element-bearing monazite and xenotime grains were both identified in the underclay sample during SEM-EDS analysis. The abundant clay minerals in this sample may have shielded the REY phosphates, meaning that the minerals may have still been encapsulated even after calcination [12]. Small xenotime grains were also identified in the Sewell coal sample. Monazite was not found, but SEM-EDS cannot give a full account of all REY-bearing minerals present. Rare-earth elements were also found in association with Ca-bearing minerals in the Sewell sample, and bulk Ca content was the highest in this sample. Calcium may have acted as a reactant in the Sewell sample, reducing the temperature needed for the thermal decomposition of LREY phosphates. Rare-earth elements were found to be co-located with an area of mixed sulfides (Fe, Pb, Cu, Ni, and Se) in the Sewell sample and with framboidal pyrite in the underclay. These associations may have influenced post-calcination REY recoverability, especially as pyrite thermally transitioned to low-solubility hematite. Although calcination can improve REY recovery by thermally decomposing the aluminosilicate matrix and REY phosphates, small REY-bearing minerals can still be partially or fully encapsulated in acid-resistant clays post-calcination (Zhang & Honaker 2019). Additionally, the thermal decomposition of REY minerals is insufficient, especially for HREY phosphates (e.g., xenotime), when calcining at temperatures between 400 and 800 °C [8,12,14]. To maximize REY recovery, the clay matrix and REY-bearing minerals must be further decomposed.

4.2. Alkaline Roasting

High-temperature roasting with Na₂CO₃ can improve the recovery of REYs by decomposing the resistant aluminosilicate matrix that often surrounds REY-bearing minerals and by dissolving REY-bearing phosphate minerals themselves. Between about 600 °C and 900 °C, Na₂CO₃ reacts with clays to form more soluble Na aluminosilicates, such as nepheline (NaAlSiO₄) [16,17,38]. Around 400 °C, Na₂CO₃ begins reacting with quartz to form Na metasilicate (Na₂SiO₃), and between 850 °C and 900 °C, it significantly decomposes monazite to form rare-earth oxides (REY₂O₃) and trisodium phosphate (Na₃PO₄) [16,32].

The results of Na₂CO₃ roasting for the samples in this study were mixed, with overall TREY recovery significantly increasing in samples leached with 0.05 M citric acid but slightly decreasing in samples leached with 0.05 M HCl, as compared to the calcination results (Figure 5). In the Sewell and underclay samples, these results were due to a decrease in LREY recovery, while HREY recovery slightly increased. In the Fire Clay sample, the recovery of both LREYs and HREYs was slightly diminished (Figure 7). The Sewell and underclay results suggest that additional HREYs were liberated during alkaline roasting, either through an enhanced decomposition of the clay matrix or reactions with HREY phosphates that were not as affected during calcination. However, overall REY recovery was suppressed upon leaching with HCl. While roasting with Na₂CO₃ transforms resistant clays and quartz into phases that can be dissolved by water (Na metasilicate) or low-concentration acids (Na metasilicate and Na aluminosilicate), the presence of Si in leaching solutions can result in the formation of a low-solubility silicic acid gel layer that can inhibit REY recovery [15,16]. As shown in Figure 8, the percentage of Si recovered during 0.05 M HCl leaching was about 7–28 times higher in the Na₂CO₃-roasted samples compared to the calcined samples. The concentration of Si in the 0.05 M HCl leachate ranged from 188 to 308 mg/L for alkaline roasted samples and 9 to 39 mg/L in the calcined samples. The increase in solubilized Si was caused by the conversion of remaining clays and heat-resistant silicate minerals (e.g., quartz) into more soluble species. The higher concentration of Si in alkaline roasted solutions may have generated a silica gel layer that shielded or sequestered REYs when leached with 0.05 M HCl. The difference in Si concentration was even greater for samples leached with 0.05 M citric acid, ranging from 10 to 38 mg/L in the calcined sample leachates and 1216 to 1489 mg/L in the Na₂CO₃-roasted sample leachates. However, unlike the HCl results, the recovery of REYs was significantly improved in the alkaline roasted samples. Leaching the Na₂CO₃-roasted samples with 0.05 M citric acid recovered between 55.4% and 62.6% of the TREYs. The discrepancy between HCl and citric acid TREY recovery in the alkaline roasted samples can be explained by the nature of the acids. The leaching performance of HCl is primarily due to acidity alone, since there is little to no complexation of REY ions with Cl^- [18]. However, the performance of citric acid is enhanced by a chelating effect. Citrate species form stable complexes with metal ions (including REY) that convert the metals into aqueous chelate species, preventing re-adsorption or re-precipitation of the elements [18,39,40]. At 0.05 M concentrations, the leaching solutions were not extremely acidic (HCl pH ~ 1.5, citric acid pH ~ 3.0), so the chelation effect would have played a significant role in REY leaching.

Increasing the acid concentrations from 0.05 M to 0.25 M increased TREY recovery by more than 50 percentage points when leaching alkaline roasted samples with HCl. The improvement in TREY recovery for citric acid was less extreme, ranging from about a 15 to 21 percentage point increase. Still, citric acid leaching recovered between 70% and 84% of TREYs compared to the 60%–72% of TREYs recovered using HCl. For both acids and both acid concentrations (0.05 M and 0.25 M), HREY recovery exceeded LREY recovery in all Na₂CO₃-roasted samples (Figure 7). This is significantly different from the calcined sample leaching results, where LREY recovery exceeded HREY recovery in two out of three samples (Sewell and underclay). As shown in Figure 6, the pattern of individual REY recovery percentages is most similar for samples that were roasted with Na₂CO₃ and leached with 0.25 M acids. The recovery patterns for alkaline roasted samples leached with 0.05 M citric acid are similar to the 0.25 M acid leached patterns, but with reduced extraction efficiency. These results reflect the ability of alkaline roasting to convert aluminosilicate matrix minerals into soluble species and to decompose heat-resistant HREY-bearing phosphate minerals.

4.3. Environmental and Economic Considerations

To utilize coal and coal byproducts as REY resources, extraction techniques should be economically viable while minimizing environmental impacts. The reagents used in this study were chosen based on environmental considerations. For example, dilute acids are safer to handle at an industrial scale and less harmful if accidentally released into the environment than the more concentrated acids often used to recover REYs. Similarly, Na₂CO₃ (pH ~12) is less corrosive and often less expensive than other commonly used alkaline additives, such as NaOH (pH ~14) [41,42]. Less extreme pH levels in the resulting leachate can reduce the risks and costs associated with processing, treating, and disposing of waste. The chelating effect of citric acid has the benefit of extracting REYs at a less acidic pH than required for HCl, but inorganic acids are generally cheaper to purchase than organic acids [18,39]. Additionally, although Na₂CO₃ is less caustic than NaOH, it has a much higher melting point (851 °C vs. 351 °C), so it must be roasted at a higher temperature to fully react with clays, phosphates, and other minerals [32]. This would increase the energy costs associated with alkaline roasting. To optimize REY extraction techniques and the associated reagents, thorough cost–benefit analyses must be performed.

Developing a method that can provide consistent results for a variety of coal-related rock types may also help constrain costs and reduce the economic uncertainty of using coal and coal wastes as REY resources. Coal-mining and processing-plant wastes can contain clays (including shale) of varying mineralogy and carbon contents, coal fragments, carbonates, and other rocks that may potentially contain REYs [43]. This study and the published literature [8,12,13,14] suggest that the effectiveness of calcination as a pretreatment for recovering REYs is highly dependent on the original mineralogy and that HREY-bearing minerals are more resistant to heat treatments than LREY-bearing minerals. Incorporating alkaline roasting into pre-treatment methods may provide more consistent results. As shown in this study, roasting with Na₂CO₃ generally reduced the variability of the REY recovery patterns as compared to the calcination results, especially for samples leached with citric acid (Figure 6). Additionally, the percentage of HREYs recovered was consistently higher than for LREYs in all alkaline roasted samples. The effects of alkaline roasting on raw, un-ashed material should be investigated to determine if similar results can be obtained without first eliminating the organic matter and other volatiles present in the samples.

A final consideration when designing commercial REY recovery methods is the selective recovery and purification of REYs from coal and rock leachates. Several methods have been proposed and investigated, including precipitation, adsorption, ion exchange, solvent extraction, electrochemical techniques, and membrane separation technologies [44,45,46]. Each of these methods has its own set of environmental and economic factors to consider. One way to help reduce the costs associated with selective recovery and the purification of REYs is to reduce the amount of gangue elements present in the REY-enriched leachate. As shown in Figure 8, roasting samples with Na₂CO₃ increases the recovery of Si and Al during the water leaching step performed prior to acid leaching. Between 1.0% and 4.5% of Al and 7.0% and 22.6% of Si were recovered during water leaching. Less than 0.5% of the TREYs were recovered during this step, suggesting that water leaching could potentially be used to reduce gangue elements in the REY leachate of alkaline roasted coals and coal byproducts. Additional studies to optimize the water leaching component could focus on leaching duration, stirring mechanisms, solid-to-liquid ratios, and temperature.

5. Conclusions

This study investigated the use of calcination and alkaline roasting as pre-treatments to enhance the extraction of REYs from coal and coal-related rocks in conjunction with dilute acid leaching. The results suggest that REY recovery after calcination is highly dependent on the original mineralogy of samples and that LREY-bearing phases may be more easily decomposed than HREY-bearing phases. Roasting samples with Na₂CO₃ can enhance the recovery of HREYs by decomposing HREY minerals and the clay matrix. The results showed that Na₂CO₃ roasting increased TREY recovery by 34–50 percentage points in 0.05 M citric acid leached samples compared to calcination. The recovery of TREYs in 0.05 M HCl leached samples was slightly reduced compared to the calcination results, possibly due to the formation of a silica gel layer that shielded or sequestered the REYs. Increasing the acid concentrations to 0.25 M resulted in a TREY recovery of 60–72% in HCl-leached samples and 70–84% in citric-acid-leached samples. In all alkaline roasted samples, the percentage of HREYs recovered exceeded the LREY recovery.