Alkali Treatment Implications for Microwave-Assisted Rare Earth Elements Extraction from Coal Mine Tailings

Abstract

Coal tailings represent a promising secondary resource of rare earth elements (REEs), yet efficient extraction is limited by their complex mineralogy. This study investigated the impact of alkali pretreatment on aluminosilicate structures in coal tailings and its implications for REE extraction via acid digestion. Precombustion coal refuse was treated with 5 M NaOH at varying solid-to-liquid (S/L) ratios (5, 50 g/L) and reaction times (5, 15 min), including a multistep (five-cycle) treatment under microwave conditions. At 180 °C, XRD, ²⁹Si NMR, and thermodynamic modeling showed kaolinite transformed to hydrosodalite at high S/L ratios, while kaolinite completely dissolved at low S/L ratios. Quartz maintained crystallinity but slowly transformed to amorphous silica during prolonged alkaline treatment. Compared to untreated tailings, light REE extraction was enhanced by a factor of ∼3 when kaolinite dissolved and by ∼2 when it converted to hydrosodalite; heavy REE extraction increased by ∼2 and ∼1.5, respectively. Extending pretreatment time produced minimal additional enhancement, indicating that under microwave conditions, kaolinite concentration in alkaline solutions and hydrosodalite solubility in acidic solutions are the primary factors controlling REE release. Alkali pretreatment also promoted uranium removal prior to acid digestion, while REE extraction correlated strongly with Mg, Ca, Fe, and Ti release.

Introduction

Rare earth elements (REEs) are a group of metals that are key drivers of decarbonized energy generation and high-tech technologies including wind turbines, electric vehicles, health imaging and video displays. As demand for these technologies grows, the need for REEs continues to increase due to their unique magnetic, catalytic, and optical properties. For example, the global demand-to-supply ratio for neodymium, which is used to make strong magnets for wind turbines and electric vehicles, is projected to surpass 2 in high-demand scenarios and approach 1 under low-demand scenarios by 2030. However, a large percentage of the global REE supply is concentrated within a few countries, with China dominating approximately 69% of the market as of 2024. Although the United States domestically processed some of its mined REE ores, over 80% of the usable rare-earth compounds and metals are imported into the U.S., illustrating a potential national security risk in the supply of these critical materials.

Rare earths can be found as a mixture in several rock formations. The common traditional sources of rare-earth minerals include monazite, bastnasite, xenotime, apatite, gadolinite, loparite, fergusonite, among others. , To mitigate supply chain challenges, coal-related wastecoal tailings or refuse, power plant ash, and acid mine drainagehas been identified as a valuable nontraditional source for REE production. − In 2023, the United States produced more than 21 million tons of fly ash (postcombustion coal waste), of which approximately three-quarters was reused, mainly in construction materials, while the remainder was mostly landfilled. , Current estimates suggest that about 1.5 billion tons of landfilled fly ash could yield approximately 661 kilotons of REEs if efficiently recovered. Additionally, landfilled coal refuse (precombustion coal waste) in Pennsylvania alone is estimated at around 2 billion tons, with the potential to supply about 137 kilotons of REEs. The combination of accumulated coal waste and the annual generation of these byproducts represents a potentially substantial source for REE extraction. Beyond addressing supply chain risks, recovering REE from these waste streams could reduce the environmental impact associated with exploring new mines for REE production. However, coal waste is a complex mixture of organic and inorganic compounds, which encapsulate REE and hinder extraction.

An additional challenge in REE processing from both primary and secondary sources is the presence of uranium and thorium, typically referred to as Naturally Occurring Radioactive Materials (NORMs). These elements are frequently colocated with the rare earth minerals due to their similar chemical structure, leading to radiation concerns during extraction. , Although NORM concentration in coal waste is usually low, concentrating REEs in a leach solution can lead to simultaneous concentration of these radioactive elements, thereby increasing waste management complexity and safety concerns in the downstream separation lines.

Historically, REE extraction from ores such as monazite is sometimes preceded by alkali treatment to break the REE-Phosphate bond and produce insoluble REE–OH with trisodium phosphate as a valuable byproduct, as shown in eq . The residual material is then acid leached. This method has been explored as a potential approach for breaking down the complex matrix of coal waste to enhance the REE extraction. Most studies have focused primarily on postcombustion coal wastes, such as fly ash. − Wang et al. utilized NaOH pretreatment of fly ash followed by hydrochloric acid (HCl) leaching and reported a REE-enrichment factor of approximately 1.4 and a silica-removal efficiency of 41% relative to the untreated sample. This was achieved under conditions of 14.3 M NaOH, 150 °C, a reaction time of 2 h, and a solid-to-liquid ratio of 100 g/L, resulting in significantly improved REE extractability. Similarly, Lin et al. combined physical separation with repeated NaOH treatments, achieving a REE-enrichment factor of 2.2 after ten treatment cycles. Their process involved conditions of 5 M NaOH, 100 °C, 120 min duration, and a solid-to-liquid ratio of 0.05 g/L, with 66 wt % of the ash being removed prior to alkali treatment. Generally, REEs are reported to be predominantly associated with the aluminosilicate glass matrix present in coal fly ash. Hydrothermal alkaline treatment is employed to dissolve this glassy phase, concentrating REEs in the solid residues. Additionally, another study reported increased porosity of coal fly ash residues following alkaline treatment, thus enhancing the penetration of acid solutions during subsequent REE leaching.

$$\mathrm{REE}{{\mathrm{PO}}_4}_{(\mathrm{s})}+3\mathrm{NaOH}\to \mathrm{REE}(\mathrm{OH}{)}_{3(\mathrm{s})}+{Na}_3{\mathrm{PO}}_4$$ 1

Fewer studies have investigated alkali pretreatment of precombustion coal waste such as coal tailings or refuse, ,, which is the primary focus of this research. Kuppusamy et al. studied the simultaneous extraction of coal and rare earth elements (REE) using an alkali-acid process. Although the REE-enrichment factor was not explicitly reported after alkali treatment, the highest REE recovery occurred under conditions of 6.1 M NaOH, 190 °C, 60 min, and a solid-to-liquid ratio of 0.2 g/L. Li et al. reported REE-enrichment factors of 2.5 and 1.4 for different coal mine tailings, utilizing 5 M NaOH, a 2 h reaction time, a temperature of 90 °C, and a solid-to-liquid ratio of 40 g/L. Yang et al. documented an enrichment factor of 2 following treatment with 6 M NaOH for 2 h at 125 °C and a solid-to-liquid ratio of 0.1 g/L. The rationale for alkali treatment of precombustion waste is based on explanations provided in postcombustion waste studies; specifically, alkali treatment dissolves aluminosilicate minerals in coal waste, thereby liberating particles containing REEs.

Although hydrothermal alkali treatment dissolves aluminosilicate minerals and is hypothesized to liberate REE, the dissolution could be incongruent, i.e., lead to the formation of a new phase. , Geopolymers or zeolites could be precipitated depending on the reaction conditions. − Particularly, zeolite crystals are formed through the condensation of dissolved silicate and aluminate ions, leading to the formation of an aluminosilicate gel, which subsequently undergoes crystallization. The structural and solubility characteristics of the compound formed could significantly influence its dissolution behavior during acid digestion, potentially enhancing rare earth element (REE) release. For instance, zeolites possess a highly porous three-dimensional (3D) network with a large surface area and interconnected voids, making them more susceptible to acid attack. Due to the limitation of X-ray diffraction (XRD) to crystalline phase characterization, these amorphous aluminosilicate structures are typically studied with ²⁹Si solid-state magic angle spinning nuclear magnetic resonance (MAS NMR). The notation Q ⁿ (mAl) is widely used to describe aluminosilicate materials, including zeolites, gels, glasses, and minerals. This nomenclature denotes the connectivity of silicon atoms in tetrahedral coordination connected to aluminum and other silicon centers through oxygen bridges. The parameter “n” represents the coordination number of the silicon center (with values ranging from 0 to 4), while “m” indicates the number of neighboring aluminum atoms (where m can be any value from 0 up to n). Many studies have utilized the Q ⁿ (mAl) nomenclature to characterize aluminosilicate structures, reporting their corresponding ²⁹Si NMR chemical shift values. ,,, Therefore, this notation is adopted in this study to describe the aluminosilicate minerals found in coal tailings.

While several studies have independently investigated the enhancement of REE recovery and the formation of secondary phases in alkali-treated precombustion coal resources, they have typically examined these phenomena separately. With this background in mind, this study addresses a critical gap: the lack of understanding of how alkali-driven changes in the aluminosilicate structure influence REE release during subsequent acid digestion. To the best of our knowledge, this study is the first to simultaneously examine both phenomena in these noncombusted materials. Additionally, we investigated the separation of uranium as a function of alkali and acid treatment due to its radioactivity and the associated implications for waste management and safety in downstream processing.

In this study, microwave irradiation was used to control the temperature of the extraction process. Microwave irradiation interacts directly with materials primarily through dipole polarization and ionic conduction, leading to rapid heating as molecules and ions respond to the alternating electromagnetic field. , Unlike conventional heating, which relies on conduction (solid) or convection (liquid) from an external heat source, microwave heating enables faster reaction rates, more uniform temperature distribution, and improved control over the heating process via heating at the atomic/molecular level. Leveraging these advantages, we applied a combination of inductively coupled plasma mass spectroscopy (ICP-MS), XRD, and ²⁹Si solid-state MAS NMR to monitor the mineralogical changes in the coal tailings before and after alkali treatment, as well as the microwave-assisted extraction behavior of REEs from these tailings. This approach provides new insights into the mechanisms of REE release and the potential for optimizing alkali pretreatment of coal waste for efficient REE extraction.

Methodology

Materials and Sample Preparation

The coal tailings used in this study were sourced from a coal preparation plant in Pennsylvania and supplied by CONSOL Energy. The material was received as a sludge with a moisture content of 57% and subsequently dried in an oven at 85 °C until a constant weight was achieved, ensuring complete moisture removal. Following drying, the tailings, which exhibited a compact, cake-like texture, were subjected to ball milling using 5 mm spherical zirconia milling media to achieve uniform particle size reduction and enhance reactivity for subsequent processing. After ball milling, particle size analysis with a Gilson GA-6 autosiever showed that 65.2% of the tailings were <20 μm, 20.4% were between 20–50 μm, 9.7% were between 50–100 μm, and the remaining fraction was >100 μm. The ultimate, proximate, and mineral analyses of the tailings analyzed by Standard Laboratories, Inc., are presented in Tables S1, and , respectively, detailing their compositional characteristics.

1. Proximate Analysis of Coal Tailings on a Dry Basis with the Associated ASTM Standard Provided in Parentheses.

component	% composition
Ash (D3174)	41.62
fixed carbon (D3172)	33.85
volatile (D3175)	24.53

2. Mineral Analysis (ASTM D6721) of Coal Tailings on an Ignited Basis.

mineral component	SiO2	Al2O3	Fe2O3	K2O	CaO	SO3	TiO2	MgO
% composition	58.47	24.86	6.65	2.70	2.65	2.47	1.09	1.07

The chemicals utilized in this study included nitric acid (TraceMetal grade, 67–70% w/w), sodium hydroxide (ACS Certified, ≥ 97%), purchased from Fisher Scientific and kaolinite (03584), purchased from Sigma-Aldrich.

Microwave Heating

Alkali Pretreatment

The dried coal tailings were pretreated with 5 M NaOH in a microwave digester under two distinct solid–liquid ratios: 5 g/L and 50 g/L, while maintaining a constant solvent volume of 10 mL. For each experiment, the mass of the sample corresponding to the solid-to-liquid (S/L) ratio was measured and transferred to a Teflon liner vessel. A magnetic stir bar was added to each vessel to ensure homogeneous mixing during the reaction. Subsequently, for each vessel, the NaOH solution was added, gently agitated to achieve complete wetting of the solid material, and the vessel content was allowed to prereact for 15 min in a fume hood before sealing. The capped vessels were then loaded into a microwave digestion system (CEM BLADE) and subjected to a controlled temperature program, reaching 180 °C within 4 min, followed by a 5 min isothermal hold at the target temperature. While microwave heating was employed in this study to accelerate alkali–mineral interactions, similar transformations at longer reaction times could potentially be achieved through conventional thermal methods, which may be more practical for large-scale applications.

After pretreatment, the vessels were allowed to cool to ambient temperature before being placed back to the fume hood . The vessel contents were transferred to a 50 mL volumetric flask, and the vessels were rinsed with deionized (DI) water to recover the sample and dilute the solution up to 50 mL. The solution was centrifuged at 4000 rpm for 5 min to separate the solid residue from the supernatant. The supernatant was retained for liquid-state analysis to determine the concentration of rare earth elements extracted. The solid residue underwent two sequential washing cycles, each using 25 mL of deionized water, followed by thorough mixing and centrifugation at 4000 rpm for 5 min to remove residual alkali. The washed solid material was then dried in a desiccator for 24 h and stored for characterization and subsequent acid digestion.

For the aluminosilicate structure study, additional samples were prepared under varying experimental conditions to assess the effects of solid-to-liquid ratio (5 and 50 g/L), reaction time (5 and 15 min), and multicycle treatments (1 and 5 cycles). Structural characterization of these treated samples was compared to the raw coal tailings to evaluate the impact of alkaline digestion on the aluminosilicate phases. The treated samples were then subjected to acid digestion to investigate correlations between structural modifications and enhanced REE extraction efficiency.

Acid Digestion

Both raw coal tailings and residual solids from the alkali pretreatment phase underwent acid digestion using 68 w/v% nitric acid while maintaining a constant solvent volume of 10 mL. Nitric acid was selected over sulfuric acid to avoid the potential formation of insoluble sodium–lanthanide double sulfate phases that could arise from the residual sodium introduced during the NaOH pretreatment. − Temperature effects were evaluated by conducting digestion at both 180 and 220 °C. For samples specifically prepared for aluminosilicate structural analysis, digestion parameters were standardized at 180 °C, 5 g/L solid-to-liquid ratio, and 5 min hold time to isolate the effects of alkali pretreatment from variations in acid digestion conditions. While the acid treatment conditions are considered as digestion, complete dissolution of the feedstock was not achieved due to the presence of silicate minerals in the tailings. Consequently, REEs bound within silicate structures of the raw tailings were not readily released. This study, therefore, evaluates how modification of the silicate phases by alkaline pretreatment influences REE extractability under harsh acidic conditions, with the broader aim of translating these insights into a leaching environment.

The sample preparation protocols, microwave digestion procedures, and postdigestion processing followed the same methodology as previously described for the alkali pretreatment section, with the exception that the residual solid after acid digestion was neither washed with deionized water nor retained for further analysis.

Materials Characterization

Liquid Phase Characterization

Liquid samples were analyzed at Virginia Tech for rare earth elements and uranium concentrations using a Thermo Electron iCAP-RQ inductively coupled plasma mass spectrometer (ICP-MS) per Standard Method 3125-B. Samples and calibration standards were prepared in a matrix of 2 w/v% nitric acid by volume.

Complementary analyses of major elemental constituents were measured using inductively coupled plasma optical emission spectrometry (Thermo Scientific, iCAP PRO XP Duo). The following elements were measured at their respective analytical wavelengths (in nm): Si (251.611), Al (396.152), P (213.618), Mn (257.610), Fe (261.187), Mg (279.079), V (292.402), Ca (315.887), Ti (334.941), Ba (455.403), and K (766.490). Radial viewing mode was used for all elements except phosphorus, which required axial mode for enhanced sensitivity.

Solid Phase Characterization

The structural transformations of coal tailings subjected to alkali treatment were investigated using ²⁹Si MAS NMR and XRD. The raw and alkali-treated tailings were analyzed with X-ray Diffraction (XRD, PANalytical X’Pert PRO) to determine the mineral phases present in the sample and phase transformation as a result of alkali treatment. Diffraction patterns were collected using Cu–Kα radiation (1.54059Å), a 2θ range of 10–70°, a step size of 0.02°, a current of 40 mA, and a voltage of 45 kV. The XRD diffraction patterns of the analyzed samples were compared with reference patterns of minerals commonly found in coal products, using standards obtained from the Inorganic Crystal Structure Database (ICSD). The diffractogram of the raw tailings was refined using the Rietveld method, via the GSAS-II program.

Additionally, the solid samples were analyzed with ²⁹Si MAS NMR to determine the changes to the aluminosilicate structure and relate that to the REE extracted in the acid digestion step. ²⁹Si MAS NMR spectra were collected using a Bruker solid-state NMR spectrometer operating at a Larmor frequency of 99.325 MHz. The samples were packed into a 4 mm zirconia rotor, and spectra were acquired under magic angle spinning (MAS) at 12 kHz. A single-pulse excitation sequence was employed with high-power proton decoupling. Tetramethylsilane (0 ppm) was used as an external chemical shift reference standard. The spectrum was acquired with a recycle delay between 1 and 8 min and accumulated over multiple scans, until a good signal-to-noise ratio was achieved. Peak deconvolution was performed using Gaussian fitting with peak assignment based on phases identified in XRD.

Lastly, thermogravimetric analysis was performed using a Mettler Toledo Thermal Analysis System TGA/DSC 3+ to determine the inorganic content of the raw and alkali-treated coal tailings samples. Approximately 15 mg of each sample was placed in an alumina crucible and heated from room temperature to 1000 °C at a constant heating rate of 10 °C/min in an air environment with a flow rate of 50 mL/min.

Thermodynamic Modeling

Equilibrium phase diagrams for aluminosilicate dissolution under acidic and basic conditions were simulated using the commercial OLI Studio: Stream Analyzer v 12.0. OLI uses the mixed solvent electrolyte framework and a combination of Helgeson-Kirkham-Flowers equations of state for ideal conditions and Pitzer and UNIFAC framework for nonideal conditions. In terms of the thermochemical solubility analyses in this paper, the theory behind the method is described by Adelman et al. Kaolinite - Al₂Si₂O₅(OH)₄ - was used as the representative aluminosilicate phase, while HNO₃ and NaOH were used as the acid and base titrants, respectively. Microwave conditions were simulated as 180 °C and 10 atm.

Results and Discussion

Effect of NaOH Pretreatment on REE Extraction

Figure a compares the total rare earth elements (TREE) extracted per gram of coal tailings under alkali-only, alkali-acid, and acid-only digestion conditions. The results show that alkaline pretreatment of coal tailings prior to acid digestion significantly influences REE extraction efficiency, with minimal extraction in the alkaline solution. In the alkali treatment step, NaOH should react with REEs in the tailings to form REE-hydroxides according to eq , which are highly insoluble in alkaline solutions with the exception of scandium, which was not analyzed in this study due to instrument limitations. This explains the minimal TREE leached in the alkali solution, a result consistent with literature studies. ,,

1.

Bar chart showing the concentration of (a) total rare earth elements, (b) uranium extracted from coal tailings at different treatment conditions. Results indicate that alkaline pretreatment significantly enhances REE extraction compared to acid-only digestion, with higher recovery observed at a low solid–liquid ratio of 5 g/L. The pretreatment effectively solubilizes uranium into the alkaline solution, leaving minimal uranium in the REE-enriched residual solids.

Acid digestion of the alkali-pretreated tailings yielded higher TREE concentration compared to the acid-only digestion across all tested conditions, with the highest values observed at a solid–liquid ratio of 5 g/L at 180 °C. Increasing the digestion temperature from 180 to 220 °C had a minimal effect on the TREE extracted from both treatment methods, except for the 5 g/L-220 °C alkali-acid digested sample, which exhibited a large error bar, suggesting variability in digestion efficiency under these conditions. The S/L ratio was found to significantly influence the efficiency of TREE extraction from coal tailings, as demonstrated in Figure a. Tailings treated with a lower S/L ratio of 5 g/L consistently yielded higher concentrations of total REEs compared to those treated at 50 g/L under both 180 and 220 °C digestion temperatures. At the lower S/L ratios, a greater volume of the solvent is available per unit mass of tailings, which increases reagent availability and enhances the dissolution of REE-bearing mineral phases. Additionally, lower S/L ratios promote mass transfer of the tailings, which is critical for accessing REEs bound within the matrices. Although the observed trends indicate that a lower S/L ratio is favorable for maximizing REE extraction, practical considerations such as reagent consumption and waste volume must be balanced when scaling up the process.

The digestion results in Figure a were compared with the ASTM D6357–21b standard, which establishes procedures for determining trace element concentration, including REEs, in coal and coal combustion residues. This standard method involves hot plate digestion of 200 mg coal ash sample using a combination of 20 mL aqua regia and 20 mL HCl at 130 to 150 °C, which corresponds to a solid–liquid ratio of 5 g/L. Therefore, the TREE concentration from the standard method was obtained by digesting ash coal tailings, but reported on a whole-sample dry basis. However, in this study, digestion was performed on the raw coal tailings (i.e., without removal of the organic content) using a microwave digester. The results show that acid-only digestion of the coal tailings performed below the ASTM standard, particularly at a solid–liquid ratio of 50 g/L. The reduced REE recovery during acid-only digestion can be attributed to the high organic content of the tailings, which is about 58% as shown in Table . Since REEs are primarily associated with the inorganic mineral fraction of coal tailings, the organic content could interfere with acid access to these mineral phases, thereby limiting REE dissolution. Conversely, higher concentrations of total REEs were leached from the alkali-acid digested tailings. This enhancement was more pronounced at a solid–liquid ratio of 5 g/L, further highlighting not only the influence of reagent availability but also the effectiveness of alkali pretreatment in liberating REEs by breaking down inorganic phases and minimizing the competing effects of organic matter for acid consumption. Following alkali-acid treatment, the resulting pregnant leach solution containing REEs and impurities would typically undergo separation and purification steps such as solvent extraction, ion exchange, or biosorption, with mixed-phase REEs recovery achieved through oxalate precipitation. −

Previous studies suggest that alkali treatment improves REE extraction by dissolving inorganic gangue minerals such as aluminosilicates in which REE could be encapsulated, thereby leading to their liberation , and increased porosity of the samples, which then enhances acid penetration during digestion. Although these mechanisms explain the improved extraction of REE observed, this study further investigates the structural transformations of aluminosilicate minerals to better understand the factors that govern REE release.

Effect of NaOH Pretreatment on Uranium

Figure b presents the concentration of uranium extracted under different coal tailings treatment conditions: alkali-only, alkali-acid, and acid-only digestion. The figure shows that a significant fraction of uranium was extracted into the alkaline solution after NaOH treatment. Consequently, acid digestion of the residual solid (alkali-treated sample) extracted only minimal uranium content. This suggests that uranium in the coal tailings is readily soluble under alkaline conditions, a behavior characteristic of hexavalent uranium (U⁶⁺) species rather than tetravalent uranium (U⁴⁺). However, further characterization, such as X-ray Absorption Near Edge Structure (XANES) analysis, is required to confirm uranium's oxidation state.

The solubility of uranium is significantly influenced by its oxidation state. Hexavalent uranium (U⁶⁺) mineral is much more soluble than the reduced tetravalent (U⁴⁺) species. U⁴⁺ reacts with NaOH to form insoluble U­(OH)₄, whereas U⁶⁺, which typically exists as UO₂ ²⁺, forms soluble uranyl hydroxide complexes. Therefore, the first step in uranium extraction is usually oxidation of insoluble U⁴⁺ to soluble U⁶⁺ using an oxidizing agent such as MnO₂, KMnO₂, H₂O₂, or KClO₃. Additionally, uranium complexation in aqueous solution enhances its solubility by chelation with oxygen-containing ligands (e.g., carbonate, phosphate, hydroxide), forming soluble uranium complexes and facilitating its release into the solution. For example, carbonate complexes increase the solubility of uranium by competing with hydroxyl complexes for uranium, facilitating U⁴⁺ oxidation, and limiting the extent of sorption in oxidized waters, thereby increasing the mobility of uranium.

Notably, alkaline pretreatment did not extensively leach the rare earth elements (REEs); instead, these remained largely in the residual solid fraction for subsequent acid extraction, as shown in Figure a. Such selectivity, though not commonly highlighted in REE extraction studies, is a critical advantage for downstream processes. Gajda et al. selectively leached uranium and a small amount of vanadium from sandstones, leaving behind other major metals by using alkaline carbonate solutions with oxidizers (KMnO₄ and H₂O₂). Similarly, Abdel-Rehim demonstrated that carbonate leaching of alkali-treated monazite effectively separated uranium and thorium from REEs. In contrast to these studies, neither carbonate nor oxidizer was used to enhance selective uranium extraction. This further supports the hypothesis that uranium in the coal tailings predominantly exists in the +6 oxidation state, making it readily mobilizable under alkaline conditions.

Alteration of Aluminosilicate Structure during Alkali Treatment of Coal Tailings

The XRD analysis of the raw coal tailings (Figure a) shows kaolinite (Al₂Si₂O₅(OH)₄), quartz (SiO₂), and illite (multiple chemical formulae e.g., K­(Al₄Si₂O₉)­(OH)₃), , all of which are commonly found in coal waste. ,, Rietveld refinement of the XRD pattern (Figure S4), achieved with a weighted-profile R-factor (R _w ) of 22%, indicates that the crystalline fraction of the sample comprises approximately 68% quartz, 23% kaolinite, and 9% illite by volume. The ²⁹Si MAS NMR spectrum of the raw sample (Figure b) exhibits a sharp resonance at −92.3 ppm (FWHM = 4.07 ppm), which corresponds to the Q³(0Al) silicon environment in kaolinite, characterized by silicon atoms bonded to three neighboring SiO₄ tetrahedra and in good agreement with literature values. Another sharp resonance centered at −108 ppm (FWHM = 2.25 ppm) was observed, which represents a fully polymerized Q⁴(0Al) silicate structure of quartz, where each silicon atom is tetrahedrally coordinated to four oxygen atoms from bonded adjacent silica tetrahedra. , Further analysis using peak deconvolution (Figure S1) revealed additional silicon environments within the tailings. A distinct resonance at −86.8 ppm (FWHM = 4.9 ppm) was attributed to the Q³(1Al) environment in illite. , Moreover, two broad resonances were observed: one centered at −92.3 ppm (FWHM = 17.25 ppm), corresponding to structurally disordered kaolinite, and another at −112.8 ppm (FWHM = 15.71 ppm), characteristic of amorphous silica. ,, The presence of crystalline and amorphous phases highlights the complexity and heterogeneity of coal tailings, which could potentially affect the extraction of rare earth elements.

2.

(a) XRD patterns and (b) ²⁹Si NMR Spectra of the raw and alkali-treated coal tailings, illustrating the transformation of kaolinite to hydrosodalite at a solid-to-liquid ratio of 50 g/L, and the complete dissolution of kaolinite at 5 g/L and under multicycle alkali treatments. Note: K: Kaolinite; IL: Illite; Q: Quartz; HS: Hydrosodalite.

Alkali treatment of coal tailings at a solid–liquid ratio of 50 g/L and a reaction time of 5 min led to significant structural transformation. The characteristic ²⁹Si NMR resonance for kaolinite at −92.3 ppm was no longer observed, and a new sharp resonance appeared at −87.3 ppm, corresponding to the Q⁴(4Al) silicon environment in hydrosodalite. This transformation is indicative of the dissolution of kaolinite and subsequent precipitation of hydrosodalite, a known reaction product when kaolinite interacts with NaOH under alkaline hydrothermal conditions. − The edge at −85.5 ppm is a characteristic of fully hydrated hydrosodalite. The formation of hydrosodalite was further confirmed by the emergence of distinct hydrosodalite XRD reflections at 2θ values of 13.81°, 24.13°, 34.2°, and 42.5° in the treated sample (Figure a). However, the ²⁹Si NMR resonance for quartz at −108 ppm is still visible, and the strongest XRD quartz peak shown in the inset figure in Figure a has almost the same intensity as the raw sample, suggesting resistance to dissolution of quartz under these treatment conditions. This phase transition (kaolinite to hydrosodalite) was further confirmed by alkali treatment of pure kaolinite under the same conditions used for the coal tailings. Comparison of the XRD and ²⁹Si NMR spectra between pure kaolinite and the raw tailings, as well as their respective alkali-treated samples (Figures S5 and S6), shows strong similarity. These results support the presence of kaolinite in the raw tailings and confirm the formation of hydrosodalite as the dominant crystalline phase following alkali treatment.

Increasing the alkali treatment time of the 50 g/L sample to 15 min slightly altered the mineralogical composition. Notably, the quartz peak at −107 ppm was no longer detected in the ²⁹Si MAS NMR spectrum, suggesting extensive dissolution of quartz under prolonged alkaline treatment. This was further supported by a noticeable reduction in the intensity of the strongest XRD peak for quartz at approximately a 2θ of 26.35° (insert figure in Figure a), and the disappearance of the peak at a 2θ of 20.57°. Meanwhile, the intensity of the hydrosodalite XRD peaks increases slightly, and the characteristic ²⁹Si NMR hydrosodalite resonance at −87.3 ppm remained prominent, suggesting that extending the reaction time has a negligible effect on hydrosodalite crystal structure while facilitating quartz dissolution.

In contrast, alkali treatment of the tailings at a lower solid-to-liquid ratio of 5 g/L for 5 min led to the disappearance of the characteristic kaolinite ²⁹Si NMR resonance at −92.3 ppm, without the emergence of the resonance associated with hydrosodalite. This suggests complete dissolution of kaolinite rather than its transformation into a stable secondary crystalline phase. This interpretation was supported by the XRD diffraction pattern of the sample, where neither residual kaolinite nor hydrosodalite reflections were detected. However, the quartz ²⁹Si NMR resonance at −108 ppm remained visible, and the intensity of the strongest XRD peak for quartz is similar to the raw and 50g/L-5 min samples (insert figure in Figure a), indicating resistance to dissolution at microwave treatment time of 5 min. Additionally, a broad resonance associated with amorphous silica at approximately −112 ppm became more distinct, suggesting an increased presence of amorphous silicate phases, attributed to the dissolution of kaolinite and subsequent reprecipitation of amorphous silicate species under this treatment condition.

For coal tailings subjected to five-cycle alkali treatment at a solid-to-liquid ratio of 50 g/L (5 min per cycle), the structural transformations observed were essentially a combination of those occurring in the single-step treatments (5 g/L–5 min and 50 g/L–15 min). Specifically, the ²⁹Si MAS NMR spectrum of the multicycle-treated sample lacked the peaks at −92.3 ppm and −87.3 ppm corresponding to kaolinite and hydrosodalite, respectively, similar to observations in the single-step 5 g/L–5 min sample. Moreover, the characteristic ²⁹Si NMR peak for quartz at −108 ppm was no longer detected, and the resonance attributed to amorphous silica at approximately −112 ppm became notably more pronounced. The XRD analysis further corroborated these findings, showing a substantial reduction in intensity of the dominant quartz peak at 26.35° 2θ and the disappearance of the peak at 20.57° 2θ, aligning closely with the observations from the 50 g/L–15 min sample. This indicates extensive quartz dissolution under repeated alkali exposure conditions. Thus, it can be concluded that a S/L ratio of 50 g/L primarily favors the transformation of kaolinite into hydrosodalite, while a lower S/L ratio (5 g/L) promotes complete dissolution of kaolinite, and an extended treatment duration (≥15 min) enhances quartz dissolution. This suggests quartz dissolution is a stronger function of time and independent of concentration at the conditions evaluated, while this phenomenon is the inverse for kaolinite conversion.

Thermodynamic simulations conducted at 180 °C confirm this transformation of kaolinite under highly alkaline conditions, as shown in Figure . This figure shows the stability diagram for kaolinite under equilibrium conditions at 180 °C, 10 atm, and different molar concentrations of kaolinite. At a certain kaolinite molarity (y-axis) and pH (x-axis), this phase diagram predicts the existence of 1 or 2 solid phases (the patterned regions) or complete dissolution (the blank region). For example in a solution that contains 10^–1.5 M kaolinite, has a pH of 7, and is at 180 °C and 10 atm, kaolinite exists as a solid mixture of AlOOH and NaAlSi₃O₈. On the other hand, if the pH were increased to 10, kaolinite would exist as NaAlSi₃O₈ and if the pH were increased to 12, kaolinite would completely dissolve. At molar concentrations on the order of 10^–1, similar to the 50 g/L alkali-treated samples, kaolinite transforms into hydrosodalite. On the other hand, at molar concentrations on the order of 10^–2, similar to the 5 g/L alkali-treated samples, kaolinite is completely dissolved. These analyses further suggest that kaolinite transformation reaches equilibrium within the 5 min time frame of the experiment, with either hydrosodalite formation or complete kaolinite dissolution.

3.

Simulated phase diagram for kaolinite transformation under equilibrium at 180 °C, 10 atm, and in different acid (HNO₃), base (NaOH), and kaolinite concentrations. At a specific pH and molarity of kaolinite, the blank region is where no stable solid phases exist (i.e., complete dissolution), while the color-coded regions show the most stable solid phase(s).

Implication of Aluminosilicate Structure for REE Extraction

Figure a presents the concentration of total rare elements (TREE), silicon (Si), and aluminum (Al) extracted from raw and alkali-treated tailings, alongside the predominant mineral phases identified in each sample. The raw tailings primarily contain kaolinite and quartz by volume, as mentioned in the section above. Kaolinite, a phyllosilicate mineral, has a layered structure consisting of alternating tetrahedral silica layers (Si–O) and octahedral alumina layers (Al­(O,H)), linked through hydrogen bonding as shown in Figure S7. , Although kaolinite exhibits a relatively small surface area and limited cation exchange capacity, it can host rare earth elements (REEs) through surface adsorption mechanisms such as inner-sphere complexation and as non ion-exchangeable cations. −

4.

(a) Si, Al, and total rare elements extracted from 5 g/L acid digestion of raw and alkali-treated coal tailings. The text on top of each bar denotes the predominant mineral in each sample based on phase identification with XRD and ²⁹Si NMR (K-Kaolinite, Q-Quartz, HS-Hydrosodalite, AS-Amorphous Silica). (b) Ratio of REE extracted from alkali-treated vs raw coal tailings illustrating the enhanced impact of alkaline pretreatment on the amount extracted, especially for the Light Rare Earth Elements (LREE).

Upon alkali treatment at a solid-to-liquid ratio of 50 g/L (for 5 and 15 min), kaolinite converts to hydrosodalite, as previously discussed. Under these conditions, minimal amount of Si and Al were released into the alkali solution, as indicated by the low concentrations measured in the alkali leach solution (Figure S2) and negligible reduction in inorganic content shown by thermogravimetric analysis (TGA) in Figure S3. This structural transformation notably enhanced REE liberation during subsequent acid digestion. Specifically, as shown in Figure a, the total REE extracted from hydrosodalite-rich samples (50 g/L–5 min and 50 g/L–15 min) increased approximately by 1.95 and 1.97-fold, respectively, compared to the raw tailings. Hydrosodalite is a microporous zeolite-type aluminosilicate composed of alternating and corner-sharing SiO₄ and AlO₄ tetrahedra arranged in a cubic symmetric framework. ,,, Hydrosodalite composition can vary depending on synthesis conditions, but the general unit cell formula is often represented as Na_6+x (SiAlO₄)₆(OH) _x ·nH₂O where x could be 0 or 2. , The composition is typically classified based on the Na⁺:OH^–:H₂O stoichiometry per unit cell: a ratio of 8:2:n corresponds to basic hydrosodalite (unit cell in Figure S8), whereas 6:0:n is defined as the nonbasic form. The porous framework of hydrosodalite increases its susceptibility to acid dissolution, which explains the enhanced extraction of Si, Al, and TREE from hydrosodalite-containing samples compared to the raw tailings, as presented in Figure a. This suggests that the conversion of kaolinite into hydrosodalite facilitates REE release by producing a more acid-soluble mineral phase. Although extending the alkali treatment time from 5 to 15 min led to quartz dissolution and enhanced hydrosodalite crystallinity, as previously discussed, the longer time resulted in minimal improvements in REE extraction. This indicates kaolinite solubility, rather than quartz solubility, was the primary factor influencing REE release from hydrosodalite-rich tailings.

At a lower solid-to-liquid ratio of 5 g/L, kaolinite underwent complete dissolution, leaving quartz and amorphous silica as the dominant mineral phases. Similarly, the multicycle alkali treatment at 50 g/L (five cycles of 5 min each) produced predominantly amorphous silica. Both conditions significantly increased Al and Si release into the alkali solution (Figure S2). These treatments led to a pronounced reduction in the inorganic content, decreasing from approximately 43% in the raw and single-step alkali-treated (50 g/L, 5 and 15 min) samples to 25% in the 5 g/L treatment, and further to 15% in the multicycle-treated samples, as confirmed by TGA analysis (Figure S3). This substantial removal of gangue minerals resulted in REE enrichment, enhancing extraction efficiency during subsequent acid digestion. As illustrated in Figure a, the total REE extracted increased approximately 2.90-fold for the 5 g/L–5 min treatment and 3.35-fold for the 50 g/L multicycle treatment, compared to raw tailings. While low solid-to-liquid ratios and multicycle treatments enhance REE extraction, these conditions also generate larger volumes of alkaline waste. Future work should therefore examine the reusability of spent alkaline solutions.

In summary, alkali-induced conversion of kaolinite to hydrosodalite improves REE extraction through enhanced mineral solubility. Complete dissolution of kaolinite, particularly under low solid-to-liquid ratios and repeated treatment cycles, substantially reduces gangue material, thereby significantly increasing REE liberation and extraction efficiency. It is important to note that coal tailings from different origins may contain other aluminosilicate phases, which could influence the efficiency of alkali–acid treatment discussed in this work. However, predictive modeling of phase changes (Figure ) and experimental evaluation of the changed mineral structure (e.g., Figure S7 transformed into Figure S8) could offer some insight into aluminosilicates that differ from those in this study.

REE Selectivity and Correlation with Major Elements in Coal Tailings

To evaluate the selectivity of the alkali–acid digestion method for individual rare earth elements (REEs), Figure b compares the ratio of each REE extracted from acid digestion of alkali-treated coal tailings to that obtained from digestion of raw tailings. A ratio of 1 indicates equivalent extraction, whereas values exceeding 1 signify enhanced REE extraction from the treated samples. The results show that all REEs exhibit enrichment ratios above 1 under all tested alkali-treatment conditions, further highlighting the effect of alkali-treatment in improving REE recovery from coal tailings. However, certain REEs exhibit high selectivity under specific conditions. For instance, cerium (Ce), thulium (Tm), and lutetium (Lu) demonstrate notably high extraction from 5 g/L alkali-treated tailings, with enrichment factors of approximately 4.0, 4.6, and 5.4, respectively. Additionally, neodymium (Nd) reaches a ratio of 5.3 for 50 g/L5 cycles alkali-treated tailings, suggesting that multiple treatment cycles can further enhance extraction for specific REEs. However, the actual concentration of Nd extracted from this sample is significantly higher than that observed in other samples, as shown in Figure S9, and may represent either an outlier or the treatment of a particle enriched in Nd.

REEs are generally categorized into light REEs (LREEs) and heavy REEs (HREEs) based on their atomic number and similarities in properties. LREEs typically include lanthanum (La) to gadolinium (Gd) elements with 0 to 7 unpaired electrons in the 4f orbital, while HREEs include terbium (Tb) to lutetium (Lu), where the 4f orbitals contain paired electrons. However, yttrium (Y) is classified as a HREE due to its similar properties to the heavier lanthanides. The insert table in Figure b summarizes the concentrations of LREEs and HREEs extracted from raw and alkali-treated tailings, along with their corresponding ratios. The raw tailings exhibit an LREE-HREE ratio of about 5.2, whereas all alkali-treated samples consistently show an increased ratio greater than 7 across the different treatment conditions. This increase indicates that alkali treatment has an enhanced effect on LREE-bearing mineral phases, with higher LREE extractability and release during subsequent digestion. Accounting for the potential outlier effect of Nd (a light rare earth element) showed no significant impact on the enhanced extraction of the LREE due to alkaline pretreatment as shown in Table S2. These results suggest that LREE-bearing minerals in the tailings could be more associated with kaolinite, which either dissolves or transforms depending on the alkali treatment condition. Specifically, under dissolution conditions (5 g/L and 50 g/L multicycle treatment), LREE extraction was enhanced by factors of 3.1 and 3.6, respectively, while HREE extraction increased by factors of 2.1 and 2.4, respectively. In contrast, under transformation conditions where kaolinite converted to hydrosodalite (50 g/L–5 min and 50 g/L–15 min), LREE extraction was enhanced by a factor of ∼2.1 and HREE by ∼1.5 (Table inset in Figure b).

Figure presents the correlations between major elements and total rare elements extracted after acid digestion, quantified using the Pearson correlation coefficient (r). This coefficient measures the strength of a linear relationship between two variables, where r = 1 indicates a perfect positive relationship, r = 0 indicates the absence of a linear relationship, and r = −1 denotes a perfect negative correlation. Given the detection limits of XRD (0.1 to 1 wt % per phase) and the relatively low concentration of REEs in the coal tailings (in μg/g), REE-bearing minerals could not be detected through the solid-state characterization done in this study. Therefore, these correlations provide insight into the possible local chemical environments of REEs and the minerals with which they may be associated. A positive linear relationship suggests that the dissolution of a specific element leads to the extraction of REEs, implying a shared mineral phase.

5.

Pearson correlation of total rare earth elements (TREEs) with major elements extracted during acid digestion, indicating strong positive associations with Mg, Fe, Ca, and Ti. These correlations suggest potential shared mineral phases hosting REEs.

Rare earth elements generally do not exist in elemental form or as compounds solely with each other. Instead, they are commonly found within minerals as mixtures or through atomic substitutions for other metals. Therefore, a strong positive correlation could point to specific phases or metallic sites where REEs are integrated. In Figure , the TREE content exhibits high positive linear correlations with Mg (r = 1), Fe (r = 0.96), Ca (r = 0.92), and Ti (r = 0.91). These correlations are statistically significant at p-value ≤ 0.05. The association of REEs with Ca can arise through known host minerals such as apatite via REE substitution. In apatite, REE³⁺ and Na⁺ could either substitute for Ca²⁺or REE³⁺ and Si⁴⁺ for P⁵⁺. Although REEs do not typically substitute directly for Ti, they can be found in Ti-bearing minerals such as loparite, where REEs substitute for Ca in the crystal structure. Other Ti-containing minerals linked with REEs include euxenite and fergusonite. Furthermore, Fe-bearing minerals like allanite, which also contains Ca, frequently accommodate REEs via substitution mechanisms. While a direct association of Mg with REEs is less documented, adsorption on Mg-rich phases remains possible. A characterization study on coal fly ash reported the colocalization of LREE and HREE with Ca-rich and Fe-rich aluminosilicate, respectively. Although these correlations alone do not confirm the presence of the specified rare-earth-bearing minerals in coal tailings, they serve as indicators of possible REE-hosting phases and could guide further investigation into the local chemical environments of REEs.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.est.5c05711.

• 29Si NMR Peak deconvolution; ultimate analysis; TGA analysis; XRD spectra; kaolinite and hydrosodalite crystal structure; elemental composition (PDF)

The authors declare no competing financial interest.

Published as part of Environmental Science & Technology special issue “Advancing a Circular Economy”.