Hydrometallurgical Recovery Process of Rare Earth Elements from Waste: Main Application of Acid Leaching with Devised τ-T Diagram

2.1. Leaching

Leaching REEs from the waste is an important part of rare earth processing using hydrometallurgical route. Physically beneficiated concentrates are leached in a suitable lixiviant directly or after heat treatment to dissolve the metallic elements. The known processes range from acid leaching with H₂SO₄, HCl, HNO₃ to leaching with NaCl or (NH₄)₂SO₄ of ion adsorbed clays. Complete understanding of these processes is essential for applying them to develop more feasible methods for the recovery of REEs from the waste. Therefore, this chapter focuses primarily on the main application of the acid‐leaching process as a hydrometallurgical method needed in order to recovery REEs from waste containing rare earths.

2.2. Solvent extraction

Solvent extraction is an important technique that can usually be employed to separate and extract individual rare earths to get their mixed solutions and compounds from leached solutions after the leaching process using different cationic, anionic and solvating extractants viz. D2EHPA, Cyanex 272, PC 88A, Versatic 10, TBP, Aliquat 336, etc. For example, D2EHPA is the most widely studied extractant for rare earth separation from nitrate, sulfate, chloride and perchlorate solutions. Saponified PC 88A has been reported for their separation from chloride solutions while tributyl phosphate (TBP), a solvating extractant that extracts their nitrates from the aqueous solutions.

2.3. Ion exchange and precipitation

During the ion exchange process, different ion exchange cationic or anionic resins are employed depending on the constituent of the aqueous solution using batch or continuous mode in column to extract rare earths from leached solutions with low rare earth concentration. Cation exchange can be primarily used to obtain REEs. The affinity of the exchanged ions for the cation exchanger depends majorly on the charge, size and degree of hydration of the exchanged ions. However, the mechanism of the processes involved in anion exchangers was much more complex and could not be explained clearly. REEs show little tendency to form anionic complexes with simple inorganic ligands.

4.1. Leaching behavior of rare earth oxides (REOs) on acid leaching

Cerium oxide with in a tetravalent state is the most stable phase among the rare earth oxides, so there are many difficulties associated with dissolving it in acidic solutions. For example, the E‐pH diagram for Ce‐H₂O system shown in Figure 1a indicates that the pH required for dissolution to Ce⁴⁺ from cerium oxide is in the range of < −0.8 at atmospheric pressure [12].

Therefore, the cerium oxide can be dissolved in highly concentrated acid solutions and at elevated temperatures. For confirmation of cerium oxide dissolution in sulfuric acid solution, two chemical reactions are considered to be involved in the conversion of cerium oxide into cerium sulfate in H₂SO₄‐H₂O solution as follows. Cerium oxide reacts with sulfuric acid at atmospheric pressure according to the following dissolution reaction.

At high sulfuric acid concentrations, the concentration of mixed cerium exceeds its solubility at an early stage. Then, cerium cations dissolved under saturated conditions directly form cerium sulfate as follows:

Cerium (Ⅳ) cations can be reduced to cerium (Ⅲ) cations as follows.

However, it can be estimated that this reaction proceeds much less than the cerium sulfate formation according to the following experimental results shown in Figure 2a. This result summarized by Um and Hirato (2012) [13] showed that it took more than 48 hours to completely dissolve 0.02  mol CeO₂ powder with 2.5  µm average particle size in 100  ml of sulfuric acid (8  mol/dm³) at 125°C. Comparing the XRD patterns of the precipitate (mixture of Ce‐oxide and sulfate) provide evidence for the conversion of cerium oxide into cerium sulfate clearly (Figure 2b) and also support the results shown in Figure 2a; the intensity of the diffraction peaks originating from cerium sulfate increased with an increase in reaction time, while that from cerium oxide decreased.

Unlike cerium oxide, the dissolution of other REOs presenting as trivalent state occurred rapidly; for example, the E‐pH diagram for lanthanum‐water system shown in Figure 1b indicates that the pH required for dissolution to La³⁺ from lanthanum oxide is in the range of > at least 7.5 according to La₂O₃ + 6H⁺→La³⁺ + 3H₂O. In addition, the results summarized by Um and Hirato (2013) [14] show that the dissolution of La, Pr, and Nd oxides took less than 1 minute to dissolve in acid solution shown in Figure 3a. Under high initial amount of La₂O₃, Pr₂O₃, and Nd₂O₃ per sulfuric acid solutions, the concentration of dissolved La³⁺, Nd³⁺, and Pr³⁺ exceeds the solubility and these cations directly form sulfate. The results of the XRD patterns shown in Figure 3b provide evidence for that conversion. The chemical reactions below [Eqs. (4)–(9)] are considered to be involved in the conversion of La, Pr, and Nd oxides into precipitated sulfate in sulfuric acid solutions:

4.2. Synthesis method for purification of rare earth during acid leaching

As mentioned in Section  1, the leaching process using an acidic solution for the recovery of rare earth from the waste has its demerits with respect to separation and recovery. The materials of interest are dissolved along with undesired species in the solution, and another method is thus needed to separate these materials; if waste containing REEs is dissolved in acid solutions, difficulties in purifying REEs from non–rare earth ions inevitably arise. Therefore, the synthesis method for purification of REEs during acid leaching is one of the simpler separation methods. According to Um and Hirato (2012 and 2016) [15, 16], for purification of target rare earth from rare earth polishing powder wastes, the REEs can be precipitated in the form of rare earths and sodium double sulfate (NaRe(SO₄)₂·xH₂O) through the addition of Na₂SO₄ to H₂SO₄‐H₂O solutions during the acid leaching because NaRe(SO₄)₂·xH₂O is poorly soluble under acidic conditions.

A five‐process strategy for separation of REEs from polishing powder wastes in Figure 4 was employed in the result from Um and Hitato (2016) [16]: process 1, the high yield of synthesized NaRe(SO₄)₂·xH₂O from REOs in Na₂SO₄‐H₂SO₄‐H₂O would be controlled by the function of the reaction temperature, sulfuric acid concentration, and Na₂SO₄ concentration; process 2, NaRe(SO₄)₂·H₂O is converted into Re(OH)₃ using sodium hydroxide (NaOH) solution; process 3, the oxidation of Ce(OH)₃ to Ce(OH)₄ via the injection of air including O₂ would be advantageous for the separation of cerium, existing as the main phase in polishing powder wastes, from other rare earths; process 4, the acid leaching with hydrogen chloride (HCl) at 25°C should be limited until the pH value is between 2.5 and 3.5 because a high yield of rare earths, except cerium, is obtained after acid leaching; process 5, the residue after process 4 is added to acid solution with H₂SO₄ to separate cerium from the final impurities. As shown in Figure 4, this result suggests the possibility of target lanthanide element from the other REEs using selective synthesis method. The separation of cerium from the other REEs can be carried out by selective synthesis of the tetravalent Ce‐hydroxide in process 3, because the trivalent ceric ions, which is the major phase in polishing powder wastes, is the most likely to be oxidize to its tetravalent state by bubbling air with oxygen, unlike the other REEs presenting as trivalent cations.

5.1. Leaching kinetics

The experimental data of REOs dissolution, which was performed under various conditions of acid concentration, reaction temperature, solid‐to‐liquid ratio, particle size, etc., were made to fit the shrinking core model to REO‐dissolution vs. time curves, as follows.

The dissolution mechanism of B (solid) into A (acid‐liquid) is described as a two‐step process; the first step is B dissolution into A, and the second step is that the B cations dissolved under saturated conditions directly form B‐precipitate in an aqueous acid medium, as mentioned in Section  4.1. Because these steps occur consecutively, if one is slower than the other, the step becomes rate determining. The formation of B‐precipitate under saturated conditions is dependent on the acid concentration, the reaction temperature, etc., and its rate is much faster than the dissolution rate. Therefore, assuming that this reaction does not affect the dissolution rate, the rate of the leaching reaction between the B solid and the A liquid becomes the rate‐determining step:

According to Wadsworth and Miller (1979) [17], Liquid‐solid reaction kinetics of dense (nonporous) particles is described by the most widespread shrinking core model. In the liquid‐solid heterogeneous system, the reaction rate may be controlled as certain individual steps such as diffusion through the fluid film, product layer diffusion control, and surface chemical reaction. The experimental data of REOs dissolution, which was performed under various conditions of reaction temperature, acid concentration of the liquid, solid‐to‐liquid ratio, and particle size, were made to fit the shrinking core model with surface chemical reaction to REOs‐dissolution vs. time curves according to Um and Hirato (2012 and 2013) [13, 14]. This result shows evidence of the leaching kinetics of REOs to fit the shrinking core model, which is discussed in detail below.

If the leaching reaction between the B solid and the A acid‐liquid becomes the rate‐determining step so that the reaction is followed by the shrinking core model with surface chemical reaction as shown in Figure 5, the rate of disappearance of B (−rB; mol/m²h) can be described as follows:

St(m²) denotes the surface area of the B; dNB(mol), the amount of B disappearing; in particular, if dNB→ bdNA(disappearing acid concentration of A; mol) according to the chemical reaction and their stoichiometric coefficient in Eq. (10), it is possible to obtain the expression with k1and CA:

Here, k1(h⁻¹) is the rate constant of first‐order reaction controlled by the interaction between the A and the surface of B, and  CA(mol/m²) is the concentration of A existing on the surface of B. Also, the amount of B disappearing can be expressed using the density and the volume of B as follows:

ρ(mol/m³) presents the density of B; Vt(m³), the volume of B (solid) after dissolution time t; r(m), the radius size of Bafter dissolution time t. According to Eqs. (12) and (13), Eqs. (14) and (15) can be written as:

and

Integrating, yield the following expression for time t:

If r = 0, the time for complete dissolution reaction (τ; h) can be expressed as:

Combining these two Eqs. (16) and (17) in the form with the radius sizes of B before and after dissolution time (rR0) gives the expression for tτand k(apparent rate constant; h⁻¹):

and

For calculating k, it is necessary to confirm the change of the radius size of B before and after dissolution time. However, it is not easy to measure the size of target particle because of their heterogeneous spherical shape. Therefore, Eq. (20) can lead to advanced means of obtaining much easier measurement with the weight of B rather than that with the radius size.

Here, W0and Wtrepresent the initial and residual amount of B vs. dissolution time t, respectively. Xtrepresents the dissolved fraction vs. dissolution time t, which was then calculated as follows:

Combining these Eqs. (18)–(21), the rate equation of surface chemical reaction controlled process for heterogeneous spherical particles of B is established as follows:

The rate constant kvaries with experimental conditions and is the function of reaction temperature, acid concentration, particle size, and initial amount of B (target REEs) per A (acid‐liquid) (B/A). When B is dissolved into acid‐liquid in all experiments, with the dissolution rate constant, considered as the function of reaction temperature, sulfuric acid concentration, particle size, and B/A, can be expressed by the following equation:

where Ea is the activation energy (kJ/mol); R, the ideal gas constant, 8.314×10−3(kJ/molK); T, the reaction temperature (K); C, sulfuric acid concentration (mol/dm³); p, particle size (µm); S, B/A (mol/dm³); k0', pre‐exponential factor; and m, n, and l are constants.

It can be seen that 1−(1−Xt)13and time t are in a linear relationship and the slope of these line is kaccording to Eq. (22). In addition, the lnkvalues calculated from these kvalues were plotted against 1/T, lnC, lnp, and lnSusing equations k=k1'e−Ea/RT(relationship between rate constant and activation energy with reaction temperature), k=k2'Cm(rate constant and acid concentration constant), k=k3'pn(rate constant and particle size constant), and k=k4'Sl(rate constant and B/A constant) according to Eq. (23) where k1', k2', k3', and k4'are the pre‐exponential factors. It can also be seen that lnkand 1/T (or lnC, lnp, and lnS) are in a linear relationship and the slope of this line is −Ea/RT(or m, n, and l).

The experimental data obtained from the study of Um (2012) [18], which shows the effect of reaction temperature, sulfuric acid concentration, initial amount of cerium oxide per sulfuric acid solutions (B/A), and particle size on the conversion kinetics of cerium oxide in sulfuric acid solutions, were fitted with theoretical functions derived from Eqs. (22) and (23). Figure 6a shows the effect of reaction temperature on the dissolution rate of cerium oxide. Such results indicate that 1−(1−Xt)13and time t are in a linear relationship; the rate constants are calculated as slopes of the straight lines. The rate constants were used to determine the Arrhenius plot between lnkand 1/T (application of k=k1'e−Ea/RT) in Figure 6b. As shown in Figure 6b, the increase in dissolution rate constant with increasing temperature obeyed the Arrhenius equation with an activation energy of 123 kJ/mol. The results for the effect of sulfuric acid concentration were applied to the kinetic model and rate constants for various sulfuric acid concentrations (8, 10, and 12  mol/dm³) were determined by using the linear regressions in Figure 6c. A plot of lnkversus lnCin Figure 6d shows that the constant m was calculated to be 6.54. This large constant indicates a strong effect of acid concentration on the dissolution rate. The slops in Figure 6e determined the rate constant related to the particle sizes (2.5, 10.0, 47.5, and 112.5  µm). As seen in Figure 6e, the decrease in particle size increased the dissolution rate of cerium oxide. The plot of lnkversus lnpin Figure 6f shows that the constant n was calculated to be −0.78. In addition, the linear regressions in Figure 6g were used to calculate the reaction rate constants for various C/S (0.04, 0.12, and 0.2  mol/dm³). And the plot of lnkversus lnSin Figure 6h shows that the constant l was calculated to be −0.03, indicating that C/S has no effect on the dissolution rate of cerium oxide. Therefore, the kinetics equation on the dissolution of cerium oxide in sulfuric acid solution was k=3.26×108e−14800/TC6.54p−0.78S−0.03.

5.2. Devised τ−Tdiagram

The relationship between the time for complete dissolution reaction (τ) and the reaction temperature (T) will be the effective methods in terms of separation and recovery of REEs from the waste during the leaching process. In the case of a REOs‐H₂SO₄‐H₂O system, the relationship between τand Tcan be constructed using the kinetics equation relating to the rate constant kwith various functions of reaction temperature, sulfuric acid concentration, particle size, initial amount of REOs per acid solution, etc. If this relationship is dependent on the acid concentration, which affects the dissolution rate between the undesired species and the materials of interest (REOs), as well as the reaction temperature, the relationship can be devised using the equation of k=k5'e−Ea/RTCm(k5'; the pre‐exponential factor) obtained from Eq. (23).

Rearranging the equation of k=k5'e−Ea/RTCm, one obtains:

The equation shows that lnkand 1/Thave a linear relationship, and the slope of this line is –(EaR). Putting the activation energies of the undesired species (Eaus) and material of interest (Eaint) and the acid concentration constants of the undesired species (mus) and material of interest (mint) into the Eq. (24) respectively, the expressions between lnkand 1/Tcan be obtained as follows:

and

Converting lnkinto the time for complete dissolution reaction (τ=e−lnk) according to Eq. (19), devised τ−Tdiagrams can be obtained as follows:

and

Two types of τ−Tdiagram can be explained as shown in Figure 7.

In Type 1, the linear relation between lnkand 1/Tof undesired species [Eq. (25)] lies with that of material of interest [Eq. (26)] over the whole range of 1/T; the lnkvalues of Y‐axis are converted into τ, whereas Type 2 shows that the exponential function related to τand 1/Tof Eqs. (27) and (28) is placed over the range considered. As shown in Figure 7, the increasing difference in τbetween two linear relations (or exponential function curves) in the τ−Tdiagram can lead the material of interest to be separated from the undesired species by leaching.

From the point of view of the linear relation, the various cases of the relationship between the undesired species and the material of interest in the devised τ−Tdiagram of Type 1 can be expected as follows:

Case 1: Eaint≠ Eausand mus≐ mint

Case 2: Eaint≠ Eausand mint≠ mus

Case 2‐1: (Eaint> Eausand mint> mus) or (Eaint< Eausand mint< mus)

Case 2‐2: (Eaint< Eausand mint> mus) or (Eaint> Eausand mint< mus)

Case 3: Eaint≐ Eausand mint≠ mus

Case 4: Eaint≐ Eausand mus≐ mint

Because the difference in the activation energies as the slope of the linear relation in Eqs. (25) and (26) makes two lines angle inward toward the center, the difference in τis definitely distinctive as the reaction temperature changes. For example, in Case 1, the difference in τdecreases with increasing reaction temperature, while, in Case 2‐1, it increases. In Case 2‐2, the wide difference between two materials can be observed over the whole range of temperature (T). If there is very little difference in the activation energies (Eaint≐ Eaus), it is implied that the lines are nearly parallel with a given separation, as shown in Case 3 and Case 4; especially, in Case 4, the separation of the material of interest from the undesired species is difficult in terms of separation and recovery due to the similarity of the two lines (Figure 8).

The results from Um and Hirato (2013) [14] showing kinetics data of CeO₂ and Al₂O₃ dissolution performed under various conditions of sulfuric acid concentration and reaction temperature indicate that the τ−Tdiagram in CeO₂‐Al₂O₃‐H₂SO₄‐H₂O system could be a good example of Case 3. As for the dissolutions of CeO₂ and Al₂O₃, the following equations were obtained as follows:

CeO₂:

Al₂O₃:

Putting the calculated k5'=1.43×108(CeO₂) and k5'=1.23×1013(Al₂O₃), EaintR=14800(CeO₂) and EausR=15500(Al₂O₃), and mint=6.54(CeO₂) and mus=0.41(Al₂O₃) into Eqs. (25) and (26), the expressions between lnkand 1/Tat various acid concentrations (C) can be obtained as follows.

(1/T )×103from 2.1 to 3.1, here Cis 8:

CeO₂:

Al₂O₃:

Here C is 10:

CeO₂:

Al₂O₃:

Here C is 12:

CeO₂:

Al₂O₃:

Here C is 14:

CeO₂:

Al₂O₃:

The revised τ−Tdiagram for a CeO₂‐Al₂O₃‐H₂SO₄‐H₂O system can then be constructed as shown in Figure 9 using the expressions given previously; the boiling points of H₂SO₄‐H₂O solutions at different acid concentrations of 8, 10, 12, and 14 mol/dm³ were calculated using OLI® software. Like Case 3, the two lines of CeO₂ (material of interest) and Al₂O₃ (undesired species) are nearly parallel with a given separation at each acid concentration. Particularly, the difference in τincreases with increasing acid concentration because cerium has a higher sensitivity of acid concentration than Al₂O₃; the large acid concentration constant of 6.54 indicates a strong effect of acid concentration on the dissolution rate of CeO₂, whereas Al₂O₃ (0.41) has less effect on the dissolution rate owing to changing acid concentration.

5.3. Devised τ−Tdiagram application to a hydrometallurgical method with acid leaching process

As mentioned in Section 5.2, the devised τ−Tdiagram can be used to predict the dissolution behavior between material of interest and undesired species in the leaching process. As an example, Um and Hirato (2013) [14] reporting the recovery of cerium from rare earth polishing powder wastes presented the devised τ−Tdiagram application to a hydrometallurgical method with acid leaching process.

As shown in Figure 10, this method involved a two‐stage leaching process in series with sulfuric acid solution, which was efficient to dissolve the main elements selectively and then to recover cerium. Each dissolution experiment of stage 1 with La₂O₃, Pr₂O₃, Nd₂O₃, and CaO and stage 2 with CeO₂ with CeO₂ and Al₂O₃ by means of two‐stage leaching was studied. In particular, after Step 1, CeO₂ can be dissolved leaving Al₂O₃ in a stable solid phase under dissolution conditions calculated using the revised τ−Tdiagram for a CeO₂‐Al₂O₃‐H₂SO₄‐H₂O system shown in Figure 9; the increasing difference in τbetween cerium and aluminum in the diagram leads CeO₂ to be separated from Al₂O₃ almost completely by leaching.