*2.2. Methodology*

Leaching experiments were conducted in a triple neck round bottom flask with the middle neck connected to a total reflux condenser, which ensured that a constant volume was maintained. The reactor was placed in a water bath where the solution was heated using an immersion heater and a precise temperature control system to maintain the temperature throughout the duration of the test. Agitation was provided using a magnetic stir at a speed that could be varied up to around 1200 rpm. The leaching experiments were conducted using deionized water and trace metal grade acid (purity > 99.99%).

The investigation involved the evaluation of the following parameters on leaching recovery and kinetics: i) lixiviant type (i.e., H2SO4, HCl, and HNO3), ii) acid concentration (i.e., 0.1M, 0.5M, 1M, and 2M etc.), iii) solid-to-liquid ratio (i.e., S/L = 1/100, 1/50, 1/10, 1/5, etc.), and iv) solution temperature (i.e., 25 ◦C, 40 ◦C, 50 ◦C, 60 ◦C, and 75 ◦C). To assess leaching kinetics, samples were collected at time intervals established from the initial start of the test, i.e., 1 min, 3 min, 5 min, 10 min, 20 min, 30 min, 60 min, 90 min, and 120 min. Micro-filter (0.45 μm) plunger syringes were used to separate the leachate from the solids to immediately stop the solid–liquid reaction. The final solid residue was filtered and washed with deionized water.

The REE contents in addition to other elements of interest in the leachate and solid residue samples were determined using inductively coupled plasma optical emission spectrometry (ICP-OES). The results were used to calculate elemental and overall recovery of the REEs. Solid loss, REE content in the leach solid residue and leachate solution are presented in the supplementary materials (Table S1) that is accessible online. Leach recovery represents the amount of material in the test feed source that was solubilized into solution during the leaching process, which was quantified using the following expression:

$$\text{Leach recovery} \left( \% \right) = \frac{c\_L \* V\_L}{c\_L \* V\_L + c\_{SR} \* m\_{SR}} \times 100\% \tag{1}$$

in which *cL* is the elemental concentration in the leachate solution (μg/mL); *VL* the volume of the analyzed leachate solution (mL); *cSR* the elemental concentration in solid residue (μg/g); and *mSR* the weight of solid residue (g).

#### *2.3. Analytical Methods*

REE content was determined by digestion and analysis of the resultant solution in an ICP-OES. The solid sample preparation procedure followed the ASTM D6357 method for ashing and digestion of coal and refuse samples with modifications made to the digestion to allow for use of a digestion block apparatus. The ICP-OES unit was calibrated using a standard solution identified as VHG-SM68 multi standard, which contained 48 elements. The REE recovery of these check standards was maintained at +/− 10% relative standard deviation (RSD). A duplicate sample was chosen at random and run through the entire process to verify repeatability at the frequency of not less than one every 40 samples. A certified coal ash sample (1633b) was utilized to ensure the digestion procedure and as a reference standard for peak selection. Three standard sample were repeated with each batch of digestion and the standard deviations for the rare earth elements are: <2% for Ce, Dy, Er, Eu, Gd, Ho, La, Lu, Nd, Pr, Sc, Th, Y, and Yb; <5% for Sm and Tm; <15% for Tb.

X-ray di ffraction (XRD) analyses were conducted on feed samples using an Advance D8 instrument produced by the Bruker Company. The scanning was performed from 10◦ to 70◦ with a stepwise increase of 0.02◦ and a scanning speed of 0.5◦/min. The XRD spectra were analyzed to estimate concentrations of major mineral components using the EVA software developed by the same company.

#### **3. Results and Discussions**

#### *3.1. Particle Size E*ff*ect*

A reduction in particle size may provide two significant benefits, i.e., (1) liberation of the clay particles and other mineral matter which exposes more surface area and exchangeable REEs for lixiviants to interact and extract the RE ions and/or (2) liberation of nano-sized RE minerals and RE oxides that may be dissolved in acid solutions. Acid leaching on finer size material can also provide faster kinetic rates and higher e fficiency for REE extraction. On the negative side, by reducing the particle size to a micron level, the newly generated surface area is increased exponentially which escalates the consumption of hydrogen ions by dissolving more contaminate metal ions. As such, selectivity may be reduced, thereby elevating the cost of leaching and downstream concentration processes. Additional issues are the higher cost of energy for grinding and di fficulties associated with thickening and dewatering ultrafine solid residuals.

To assess the e ffect of particle size on leaching performance, representative samples of the Fire Clay middlings material were ground for di fferent lengths of time before the de-carbonization step to generate samples having a range of 80% passing sizes (P80). Acid leaching tests were conducted using 1.2 mol/L sulfuric acid solution with a solid concentration of 10 g/<sup>L</sup> at 75 ◦C. The REE recovery was quantified using test data generated after leaching for five hours and Equation (1).

The results shown in Table 1 indicated that reducing the particle size liberated mineral matter containing higher concentrations of REE. For example, by reducing the particle size to a P80 size of 150 μm (80 mesh top size) in the feed, the flotation tailings material contained 444 ppm of total REEs with a P80 size of 32 μm. The REE concentration of 444 ppm reflects the content of the coarser mineral matter dispersed in the middling particles. By grinding for greater lengths of time resulting in more applied energy, the P80 size was reduced to sub-micron level and the tailing material generated by flotation nearly doubled to 719 ppm of total REEs. This finding indicates that the finest mineral matter dispersed within the Fire Clay coal has the highest concentration of REEs. A previous study found that REEs in the Fire Clay coal were strongly associated with micro-dispersed kaolinite which may be liberated and released through size reduction [9].


**Table 1.** Liberation of REEs from the de-carbonized Fire Clay middlings material.

TREE: total rare earth elements.

Reducing particle size resulted in a significant increase in leach recovery from 71.2% to 84.3% over the range of P80 sizes. As such, the size reduction increased the amount of REEs reporting in the leach feed and increased the percentage of the REEs being recovered through leaching. These two positive outcomes sugges<sup>t</sup> that the REEs associated with micro-dispersed mineral matter in the Fire Clay middlings are more concentrated and more easily extractable by leaching relative to the coarser grain fractions. In addition, the finer mineral matter is, in general, more soluble as indicated by an increase in the amount of solid loss during the leaching process. As much as 20% of the solids in the finest sample tested was dissolved under the standard leaching conditions, which may reflect both the

solubility of the mineral matter and surface area exposure. A negative impact is an increase in the amount of contaminates in the leachate due to the elevated level of dissolved solids.

Based on the liberation test results, 20 min of grinding time was selected to generate the acid leach feed material used in this study. A flow sheet of the sample preparation process is shown in Figure 3 along with weight yield and content data for each feed and product stream. The decarbonization step resulted in a high-quality clean coal product containing around 7% ash-forming material while the reject material was nearly pure mineral matter as indicated by an ash content of 90.81%. The mineral flotation rougher-cleaner treatment resulted in 0.74% of the total feed reporting to the concentrate having an ash content of 75.12%. The lower ash content in the flotation concentration was an indicator of calcite flotation (CaCO3). The float product also contained 741 ppm of TREEs, which may be due to RE mineral flotation resulting from the use of octanohydroxamic acid. The flotation tailing material produced from the two stages of flotation represented 12.1% of the feed and contained 615 ppm of TREEs and 90.67% ash-forming material. The REE upgrade is 2.54:1 starting from a feed content of 242 ppm. This material was used as the acid leach feed in the subsequent leaching studies presented in this paper.

**Figure 3.** Schematic of sample preparation for the acid leach feed material using coal and mineral flotation.

Five representative samples of the acid leach feed were analyzed to assess the repeatability of the ICP-OES. The average TREE value was 607 ± 18 ppm (2.97% variation) with thorium content of 41 ± 0.6 ppm (1.46% variation). The Ce content accounted for 42% of the total REEs as shown in Figure 4a. The Fire Clay material was rich in light REEs (i.e., Ce, La, Pr, Nd, Sm, Eu, Gd, Sc) as indicated by a content of 534 ppm or 88.0% of the total. Yttrium was the most abundant heavy REE (i.e., Y, Tb, Dy, Ho, Er, Tm, Yb, and Lu) with a concentration of 47 ppm. The major minerals present in the sample were quartz, kaolinite, illite, and muscovite as shown in the XRD plot in Figure 4b.

**Figure 4.** The composition of the acid leaching feed material used in this study on the basis of (**a**) rare earth content and (**b**) mineralogy as determined by X-ray di ffraction analysis (Q—quartz, K—kaolinite, I—illite, M—muscovite).

#### *3.2. E*ff*ect of Major Variables on REE Leaching*

#### 3.2.1. E ffect of Acid Type

The lixiviant type a ffects the REE leaching characteristics by changing the solution speciation stabilities due to the existence of various anions in varying concentrations. Sulfate ions were reported to have a higher coordination ability with rare earths than chloride ions even in high monovalent concentration solutions [18]. Leaching experiments were conducted using di fferent inorganic acids at an acid concentration of 1M, solid/liquid ratio of 10 g/L, and a temperature of 75 ◦C. Sulfuric acid (H2SO4), hydrochloric acid (HCl), and nitric acid (HNO3) were used to study the e ffect on REE leaching recovery and reaction rate as shown in Figure 5. Total REE (TREEs) recovery values of 80%, 76%, and 74% were achieved after 3 h of leaching using HCl, HNO3, and H2SO4 solution, respectively. The pH of the leachate solutions at the end of the tests were 0.105, 0.113, and 0.112, respectively.

**Figure 5.** Effect of acid type on the leaching recovery of total rare earth elements contained in the Fire Clay coal middlings (75 ◦C, 530 rpm, solid-to-liquid ratio (S/L) = 10 g/L, d80 = 8.7 μm).

Hydrochloric acid provided the fastest leaching rate, which achieved 73% recovery after the first 5 min of leaching, and slowly reached equilibrium after 3 h. Nitric acid also provided fast leaching rate within the first 30 min. Sulfuric acid was the least effective under the leaching conditions and provided the slowest leaching rate. This finding was likely due to the fact that sulfate ions have a higher coordination ability with rare earths than chloride ions even in high monovalent concentration solutions [19].

The coal-based leachate contained high concentrations of trivalent ions that may coordinate with sulfate ions resulting in depression of the rare earth-sulfate coordination. In addition, sulfuric acid requires two steps of dissociation reaction to release H<sup>+</sup> into solution whereas hydrochloric acid and nitric acid dissociates more rapidly into solution. Viscosity of the sulfuric acid solution is another factor that could have resulted in the slower reaction rate as the wetting rate of the solid particle surfaces is reduced when the solution viscosity is high. Despite the negative aspects of sulfuric acid, the lixiviant is still considered a viable lixiviate due to its relatively low cost and the negative aspects of the other lixiviants including the volatility of hydrochloric acid and the decomposability of nitric acid under 75 ◦C [20].

#### 3.2.2. Stirring Speed Effect

Stirring speed affects the thickness of the film layer surrounding a solid particle suspended in the lixiviate solution. A high stirring speed creates an enhanced shear rate in solution which reduces the film layer thickness thereby increasing the mass transfer rate through the film diffusion layer [21]. The effect of stirring speed was evaluated at 300 rpm, 530 rpm, 760 rpm, and 900 rpm as shown in Figure 6. The leaching condition included 1M H2SO4 solution and a solid/liquid ratio of 10 g/<sup>L</sup> at 75 ◦C. The test results indicated that a stirring speed of 300 rpm did not provide sufficient agitation due to inadequate suspension of the slurry based on visual observations, while stirring speeds of 530 rpm to 900 rpm provided nearly equal kinetics. The recovery achieved using a 900-rpm stirring speed was slightly lower than that obtained at 760 rpm. A stirring speed of 530 rpm was established as an adequate value for the standard test conditions.

**Figure 6.** Effect of stirring speed on the leaching recovery of total rare earth elements contained in the Fire Clay coal middlings (75 ◦C, 1 M H2SO4, S/L = 10g/L, d80 = 8.7 μm).

#### 3.2.3. Solid-to-Liquid Ratio Effect

The solid-to-liquid (S/L) ratio establishes to the stochiometric ratio of reactants, which directly affects the reaction equilibrium. The effect of the S/L ratio on rare earth leaching recovery was investigated in the range of 10 g/<sup>L</sup> to 200 g/<sup>L</sup> while maintaining the other parameters constant at 75 ◦C, 1 M H2SO4, and 530 rpm. The association between reactants decreased with an increase in the solid/liquid ratio, which resulted in a decrease in the extraction rate as shown in Figure 7. Leach recovery was reduced from 74% to 40% after increasing the S/L ratio from 10 g/<sup>L</sup> to 200 g/L. The magnitude of the recovery reduction is not commonly observed in other metal leaching operations. In the metallic copper leaching process, the leaching reaction was more effective when the Cu2+ concentration in solution was higher due to Cu2+ reacting with metallic Cu to Cu+ [22]. This type of reaction mechanism does not occur in a REE solution since the REEs exist mostly as a compound. Niobium leaching from titanium oxide residues did not show any effect from the S/L ratio on leaching recovery [21]. However, Li et al. (2013) reported on a rare earth concentrate leaching study that found the S/L ratio to have a negative effect on the leaching of a rare earth concentrate when the ratio was higher than 100 g/<sup>L</sup> [23]. Therefore, the solid/liquid ratio effect varies from source-to-source in different leaching environments.

#### 3.2.4. Effect of Acid Concentration

The effect of sulfuric acid concentration on leaching recovery was studied using 0.1 M, 0.5 M, 1 M, and 2 M acid concentrations using the standard values for temperature, stirring speed, and solid-to-liquid ratio. The initial acid concentrations of 0.1 M, 0.5 M, 1 M, and2Mresulted in ending pH values of 1.04, 0.38, 0.11, and −0.25, respectively, after 3 h of leaching. As shown in Figure 8, the total REE recovery increased substantially from 40% to 74% by elevating acid concentration from 0.1 to 1 M. However, from 1 M to 2 M, the increase was marginal at around 2.5 absolute percentage points. The optimal acid concentration was selected to be 1 M since higher concentrations of acid did not provide a significant increase in recovery of REEs.

**Figure 7.** Effect of solid-to-liquid ratio on the leaching recovery of total rare earth elements contained in the Fire Clay coal middlings (75 ◦C, 1 M H2SO4, 530 rpm, D80 = 8.7 μm).

**Figure 8.** Effect of sulfuric acid solution concentration on the leaching recovery of total rare earth elements contained in the Fire Clay coal middlings (75 ◦C, 530 rpm, S/L = 10g/L, D80 = 8.7 μm).

The effect of acid concentration on individual RE leaching recovery is shown in Figure 9. Recovery of the light REEs significantly increased when acid concentration was elevated from 0.1 M to 1 M and very little improvement was realized when using a 2 M acid solution. Scandium was the least sensitive to acid concentration. The remaining un-leached portion of Sc is likely associated with a mineral structure that requires a higher level energy to break down. Yang et al. (2018) improved the scandium leaching recovery from 31% to 74% after treated by roasting with no chemical additives at 750 ◦C for 2 h from a coal-based material [24].

**Figure 9.** Effect of sulfuric acid solution concentration on individual rare earth element leaching recovery (2 h, 75 ◦C, 530 rpm, S/L = 10 g/L, D80 = 8.7 μm).

#### 3.2.5. Effect of Temperature

A leaching process that is mainly controlled by a diffusion process is more dependent on mixing conditions whereas temperature has a more significant effect on chemical reaction controlled processes [25]. The effect of temperature on REE leaching using1MH2SO4 was investigated using a stirring speed of 530 rpm and S/L ratio of 10 g/<sup>L</sup> for 2 h. Samples were taken over shorter time increments due to the relatively fast kinetics during the first 20–30 min. Figure 10 shows that REE leach recovery increased significantly with an elevation in leaching temperature. When the temperature was increased from 298 K (25 ◦C) to 348 K (75 ◦C), leaching recovery increased from 35% to 75% after 2 h of leaching. The data suggests the existence of a relatively fast leaching process during the first 20 min followed by a slow process. As such, two or more reaction mechanisms may be occurring when leaching the coal source.

**Figure 10.** Effect of leaching reaction temperature on the leaching recovery of total rare earth elements contained in the Fire Clay coal middling (1 M H2SO4, 530 rpm, S/L = 10g/L, D80 = 8.7 μm).

The effect of temperature on individual REEs is shown in Figure 11. Most of the light REEs (i.e., La, Ce, Pr, Nd, Sm) appeared to be very sensitive to temperature, which indicated that the leaching mechanism of light REEs was mostly chemical reaction controlled. The recovery of Ce, Pr, and Nd increased from 36%, 39%, and 36% to 79%, 84%, and 80%, respectively, by increasing the temperature from 25 ◦C to 75 ◦C. The heavy REEs and scandium recovery improved with higher temperature, but the increase was not as significant. Scandium recovery rose from 29% to 36%. For the recovery of elements that were relatively insensitive to temperature, the activation energy is generally low and more likely to be a result of a diffusion controlled process [26].

**Figure 11.** Effect of leaching reaction temperature on the leaching recovery of individual rare earth element (1 M H2SO4, 530 rpm, S/L = 10g/L, D80 = 8.7 μm, retention time of 120 min).
