China's rare earth element (REE) reserves that are of industrial grade are mainly classified into three categories: mixed bastnäsite and monazite (accounting for 83.7% of China's total REE reserves, located in Bayan Obo, Inner Mongolia), bastnäsite (10.6%, in Sichuan and Shandong provinces) and REE ions adsorbed on clays (2.9%, in seven provinces of southern China) (Dołęgowska & Migaszewski, Reference Dołęgowska and Migaszewski2013; Yang et al., Reference Yang, Lin, Li, Wu, Zhou and Chen2013; Weng et al., Reference Weng, Haque, Mudd and Jowitt2016). The demand for REEs and their compounds in the high-tech industry and the ongoing development of advanced technologies has increased considerably over recent years (Morais & Ciminelli, Reference Morais and Ciminelli2004; Binnemans et al., Reference Binnemans, Jones, Blanpain, Gerven, Yang, Walton and Buchert2013; Liu et al., Reference Liu, Gan and Feng2016; Schaeffer et al., Reference Schaeffer, Grimes and Cheeseman2016; Song et al., Reference Song, Cui, Wu, Yang and Zhang2016). Medium and heavy REEs (MREEs and HREEs) find more applications and are of greater value than light REEs. Bayan Obo REE deposits are rich in light REEs (accounting for 97% of the total rare earths of the deposit), whereas the deposits of REE ions adsorbed on clays are rich in MREEs and HREEs (accounting for >80% of world's total MREEs and HREEs) (Kanazawa & Kamitani, Reference Kanazawa and Kamitani2006; Chi & Tian, Reference Chi and Tian2008; Castor, Reference Castor2008). REEs in weathered rare earth ores exist in four phases (Chi et al., Reference Chi, Tian, Li, Peng, Wu, Li, Wang and Zhou2005): (1) the water-soluble phase, which refers to the REEs dissolved in water, accounting for <1 in 10,000 total REEs and should be considered negligible; (2) the ion-exchangeable phase, which refers to the REEs adsorbed on clay minerals by electrostatic interaction, accounting for >80% of total REEs. It may be released easily into leaching liquor upon ion exchange with NH4+, Mg2+, Fe2+ and Fe3+; (3) the colloidal sediment phase, which refers to the REEs adsorbed on iron manganese colloids by coordination and to the REE oxides or hydroxides bonded with or deposited on minerals, mainly CeO2/Ce(OH)4; and (4) the mineral phase, which includes the REE minerals with ionic compounds as well as diffusion of REE ions to displace mineral crystals in the ores, such as bastnäsite, apatite and so on. Utilizing physical processing methods such as flotation, magnetic, gravity and electrostatic separation may improve the grade of rare earth oxide (REO), and ion exchange is indispensable for extracting REEs from the concentrate (Kul et al., Reference Kul, Topkaya and Karakaya2008).
The leaching methods employed in the extraction of ion-adsorption clays have been greatly improved based on three different leaching processes, including the first-generation leaching process with NaCl solution, the second-generation leaching process with (NH4)2SO4 solution and the third-generation in situ leaching process with (NH4)2SO4 solution (Huang et al., Reference Huang, Long, Li, Ying, Zhang and Xue2005). Among the chemical leaching approaches, the in situ leaching technology is advantageous in terms of protection of surface vegetation because it reduces the need for soil excavation. However, it has also caused serious environmental problems, such as underground water contamination, mine collapse, landslides and plant-growth difficulties, because the ammonium salt is still the major leaching agent in this process (Huang et al., Reference Huang, Long, Li, Ying, Zhang and Xue2005; Yang et al., Reference Yang, Lin, Li, Wu, Zhou and Chen2013). Therefore, China needs to implement an integrated rare earth resource management approach to meet global demands, preserve resources for future generations and protect the environment by adopting a more environmentally friendly leaching agent.
Much effort is being expended in the research and development of hydrometallurgical leaching processes for the specially weathered crust elution-deposited rare earth ores to improve the intensification of the leaching process or reduce the consumption of leaching agents by introducing magnetochemistry, sesbania gum, compound leaching agents and impurity-inhibited leaching (Qiu et al., Reference Qiu, Luo, Fang, Hu, Cheng and Hao2002, Reference Qiu, Zhu, Fang, Zeng, Gao and Zhu2014; Tian et al., Reference Tian, Tang, Yin, Luo, Rao and Jiang2013; Yang & Zhang, Reference Yang and Zhang2015). However, the major leaching agent in these processes is still ammonium salt, which fails to tackle the ammonia–nitrogen pollution problem. Chinese scientists and engineers have been engaged in the research and development of many lixiviants without ammonia–nitrogen such as CaCl2, MgCl2, Fe2(SO4)3, Al2(SO4)3, citrates and so on, to replace the (NH4)2SO4 lixiviant. To date, all of the lixiviants introduced would not only increase processing costs, but also bring about new environmental issues with respect to Al/Fe poisoning of plants and high Ca and Na contents in soil (Huang et al., Reference Huang, Zhong, Wu, Ling and Xu2008, Reference Huang, Yu, Feng and Zhao2013). Magnesium fertilizer is lacking in the soil of southern China due to the lack of ion-exchangeable magnesium (Bai et al., Reference Bai, Jin and Yang2004). Based on China's water-quality standard, the Mg ion content in the underground and surface water is relatively low; using magnesium sulfate as the lixiviant in the in situ leaching could, therefore, act as a source of magnesium fertilizer for the soil.
In the present study, magnesium sulfate (MgSO4) was chosen as a leaching agent to remove REEs from a weathered crust elution-deposited REE ore to reduce or even eliminate ammonia–nitrogen pollution. Experimental work was conducted to investigate the influence of MgSO4 concentration, leaching time and liquid:solid ratio (L:S) on REE-leaching efficiency. The results provide a theoretical basis for and scientific approach to achieving high-efficiency and optimized leaching conditions in industrial practice and to obtaining more evidence and a better understanding of MgSO4 as a productive leaching agent.
METHODS
Rare earth ore sample and experimental reagents
The experimental weathered crust elution-deposited REE ore sample was kindly supplied by Chalco Guangxi Nonferrous Chongzuo Rare-Earth Development Ltd Co. The chemical composition of the sample was determined by inductively coupled plasma atomic emission spectrometry (ICP-AES; VARIAN, 720-ES) and X-ray fluorescence spectrometry (SHIMADZU Co., Ltd, Japan). The sample contained only 0.11% total REE (~0.072 mmol REE/100 g clay; i.e. its grade is very low) (Table 1).
The chemical compositions (%) of the ore sample.
REO = rare earth oxide.
The chemical composition of the REE ore sample was determined at the National Tungsten & Rare-Earth Product Quality Supervision Testing Center, Ganzhou, China, using ICP-AES (Table 2). The ore contained yttrium and MREEs. The precision of ICP-AES analyses varies between 1.3% and 5.5% (Ai et al., Reference Ai, Tao and Li2001). The mineralogical composition of the sample was identified by X-ray diffraction (XRD; PANalytical EMPYREAN) using Cu-Kα radiation in the range 5 to 90°2θ with a scanning speed of 2°2θ/min. Fourier-transform infrared (FTIR) spectra of the sample were collected with a Nicolet-740 FTIR spectrometer using KBr pellets.
Partitioning of the ion-exchangeable phases, in the Chongzuo rare earth ore (mass fraction %).
The leaching reagent used in this experiment was the magnesium sulfate (MgSO4) solution. All chemicals used in the analysis were of analytical reagent grade.
Apparatus and experimental procedure
A given weight (100 g) of ground REE ore was placed into a tri-neck glass flask and then the leaching solution of known concentration was added at the desired L:S. Leaching experiments were conducted under stirring at the specified conditions. After completion of the experiments, the suspensions were centrifuged and the supernatant collected. The total REE content and individual REE concentrations were determined by ICP-AES.
The fraction of the ion-exchangeable REEs was determined as follows: 100 g of the REE ores was leached by 20 g/L (NH4)2SO4 for 12 h in a beaker, with magnetic stirring at 100 rpm. Subsequently, the suspension was filtered through a quantitative neutral filter paper. The concentration of the REEs in the filtrate was determined by EDTA volumetric titration.
The rare earth leaching efficiency (η) was calculated according to equation 1:
$$ {\rm \eta} = \displaystyle{{{\rm \varepsilon} _t} \over {{\rm \varepsilon} _o}} \times 100 \percnt $$where ετ is the total amount of leached REEs at a reaction time t and εο is the ion-exchangeable REE fraction in the original ore sample.
RESULTS AND DISCUSSION
Characterization of the weathered crust elution-deposited rare earth ores
The XRD pattern of the ore sample is presented in Fig. 1. The ore consists mainly of quartz (~53%) and kaolinite (29%) and minor illite-mica (~6%) and K-feldspar (~6%). REOs or other REE minerals were not detected probably due to their low abundance. Quantitative estimation of the minerals present in the ore was obtained from the chemical composition of the ore (Table 1) and from the fact that the detection limit for illite-mica with XRD would be 5%.
XRD pattern of the REO sample.
According to the research results published by Chi et al. (Reference Chi, Tian, Li, Peng, Wu, Li, Wang and Zhou2005), the surface hydroxyls of kaolinite react with REE ions, so the trivalent rare earth ions behave as divalent or monovalent cations:
$$ {R}{E}^{3 + } + {O}{H}^{-} \to \left( {{RE} {\rm OH}} \right)^{2 + } $$
$${\rm or} \quad {R}{E}^{3 + } + 2{\rm O}{\rm H}^{-} \to \left( {{RE} {\rm OH}} \right)^ + $$In this case, kaolinite may adsorb REE ions at a ratio of 2:1 or even 1:1, rather than 3:1. Consequently, the cation exchange capacity (CEC) of kaolinite increases with the decrease in the valence state of rare earth cations. Moreover, the CEC of kaolinite is ~0.03–0.05 mmol/g (Chi et al., Reference Chi, Tian, Li, Peng, Wu, Li, Wang and Zhou2005) and the total CEC of the ore sample would be ~8.855–14.759 mmol/kg. The average molecular weight of MREEs is 152, which means that the total CEC of this sample might be ~1.346–2.243 g/kg. The detection limit of REO by ICP-AES of 1.1 g/kg is slightly lower than the average level.
The FTIR spectrum (Fig. 2) shows characteristic absorption bands for the clay minerals present. The large bands observed from 3600 to 4000 cm–1 are attributed to the OH-stretching vibration modes, and those at 1600–1660 cm–1 to bending vibrations of adsorbed water molecules. The absorption bands at 912 and 778 cm–1 are attributed to the stretching vibration of Al–OH and Si–O, respectively. The band at 467 cm–1 is assigned to the stretching vibration of Fe–O (Meng et al., Reference Meng, Jia and Wei2004). The FTIR results are in accord with the mineralogical composition (Fig. 1).
FTIR spectrum of the rare earth ore sample.
Leaching experiments
The leaching agent, MgSO4, is less expensive than Na2SO4 and K2SO4 and it has a higher leaching efficiency because the adsorption capacity of the clay is stronger for bivalent than for monovalent cations (Jiang et al., Reference Jiang, Shen and Song1992). When comparing MgSO4 with MgCl2 in usual hydrometallurgy systems, the stability constants of sulfates are an order of magnitude greater than those of chlorides, making MgSO4 a good choice for forming REE complexes (de Carvalho & Choppin, Reference de Carvalho and Choppin1967; Wood, Reference Wood1990; Millero, Reference Millero1992).
The original pH of 3% MgSO4 solution was set as the leaching acidity based on the analysis in Fig. 3. The Al3+ concentration decreases with increasing pH (Fig. 3). Excessive leaching of Al3+ might cause low rare earth yield, low product purity and an increase in difficulty in subsequent processes. The REE leaching efficiency is maximum at pH 4.4 (original 3 wt.% MgSO4), declining thereafter. In addition, the pH should be less than the hydrolysis pH of rare earth ions, because otherwise it may result in a lower leaching efficiency of REE. The pH value was kept constant by adding dilute H2SO4 or NaOH.
Effect of the acidity of the leaching agent on rare earth leaching and Al3+ concentration. MgSO4 concentration: 3%, L:S ratio = 3:1, room temperature.
The effect of temperature on the leaching efficiency of REEs was tested over the range 25–85°C (Fig. 4). The leaching efficiency of REEs was virtually constant after 30 min in the temperature range used. Based on these results, room temperature was selected in further experiments.
Effect of temperature on the leaching efficiency using MgSO4 and (NH4)2SO4 lixiviants. MgSO4 and (NH4)2SO4 concentration: 3 wt.%, L:S ratio = 3:1, pH = 4.4.
The effect of stirring speed on the leaching process is shown in Table 3. Stirring did not affect the leaching efficiency, suggesting that the leaching rate is controlled by inner diffusion, which provides the basis of analysis in studying the effect of MgSO4 concentration on leaching (Xiao et al. Reference Xiao, Chen, Feng, Huang, Huang, Long and Cui2015). A stirring speed of 500 rpm was selected as optimal for the leaching experiments.
Effect of stirring speed on rare earth leaching (temperature = 25°C, MgSO4 concentration = 3 wt.%, L:S ratio = 3:1).
Leaching chemistry in the weathered crust elution-deposited rare earth ore
REEs exist mainly in the ion-exchangeable phase adsorbed on clay minerals in the weathered crust elution-deposited ore, accounting for 75–95% of total REEs, and is the only part that can be extracted at present (Chi et al., Reference Chi, Tian, Li, Peng, Wu, Li, Wang and Zhou2005; Tian et al., Reference Tian, Yin, Tang, Chen, Luo and Rao2013). Therefore, ion-exchange leaching is the only method to extract REEs from this type of ore. The REEs are adsorbed on kaolinite ([Al2Si2O5(OH)4]·nRE 3+), halloysite ([Al2Si2O5(OH)4]·xH2O·nRE 3+) and muscovite ([KAl2[AlSi3O10](OH)2]·nRE 3+) (Yang & Zhang, Reference Yang and Zhang2015). The absorbed REEs might be easily and selectively desorbed and substituted on the substrate by the cations of the leaching agent and transferred into the solution as soluble REE sulfates (Moldoveanu & Papangelakis, Reference Moldoveanu and Papangelakis2012). The leaching reaction of the ion-exchangeable REEs in the ore with MgSO4 is as follows:
$$\eqalign{ 2&{\rm Clay}-RE({\rm s}) + 3{\rm MgS}{\rm O}_{\rm 4}({\rm aq}) \cr & \to {\rm Cla}{\rm y}_2-{\rm M}{\rm g}_{\rm 3}({\rm s}){\mkern 1mu} + {\mkern 1mu} RE_2({\rm S}{\rm O}_{\rm 4})_{\rm 3}({\rm aq})} $$where ‘s’ and ‘aq’ represent solid phase and aqueous phase, respectively, and ‘Clay’ stands for the clay mineral.
Leaching time, the concentration of leaching agent and L:S are the three main factors affecting the leaching process. These three aspects were studied in detail.
Effect of leaching time on rare earth leaching
The leaching reaction in equation 4 is a typical non-catalytic heterogeneous reaction in solid–liquid systems. To investigate the leaching process with MgSO4 solution, the effect of leaching time on the leaching efficiency was examined with 2.0% or 3.0% (w/v) concentration MgSO4 solution and 5:1 or 3:1 L:S ratio. The REE leaching efficiency increased rapidly initially before subsequently decreasing with time, indicating that the rare earth leaching behaviour was controlled by kinetics in the initial stage (Fig. 5). After 7 min, the leaching efficiency increased slowly, suggesting that the leaching process approached equilibrium.
The effect of leaching time on the leaching efficiency. MgSO4 concentration: 3 wt.%, L:S ratio = 3:1, ω = 500 rpm, pH = 4.4.
The equilibrium time above is in accordance with results reported in the literature. The adsorption of Nd onto the kaolinite surface at 25°C is relatively fast and remains essentially unchanged after a reaction time of 15 min (Aja, Reference Aja1998). Leaching of 1% H-montmorillonite for 5 min led to the same retention of REEs as leaching for 3 h (Bruque et al., Reference Bruque1980). In view of other experimental factors such as pH or temperature, it was decided that there is no need to study the above conditions because the process kinetics were very fast (Moldoveanu et al., 2013). Based on the experimental results, 30 min was considered to be sufficient for reaching maximum extraction of REEs.
Effect of concentration of the leaching agent on REE leaching
The effect of leaching-agent concentration on the leaching efficiency is illustrated in Fig. 6. The leaching efficiency increased slowly at low concentrations of MgSO4 and tended to be stable at increasing initial concentrations.
The effect of concentration of the leaching agent on the leaching efficiency (L:S ratio = 3:1, 30 min leaching time, ω = 500 rpm, pH = 4.4.
As mentioned above, the leaching process was controlled by inner diffusion. According to Fick's law, the leaching rate of inner diffusion control is described by equation 5 (Ma, Reference Ma2007):
$$ {\rm \mu} = \displaystyle{{d {\rm \varphi} } \over {dt}} = SD_0\displaystyle{{d_{\rm c}} \over {d_{\rm r}}} = \displaystyle{{4 {\rm \pi} r_0^2 D_0} \over {\rm \delta} }(C_{\rm l} - C_{\rm s})$$where S is the surface area of the ore-sample particle, D 0 (a function of temperature) is the mass transfer diffusivity coefficient of the product layer, C l and C s are the concentrations of the leaching agent in the liquid phase and the surface of the ore sample particles, respectively, r 0 is the average initial radius of the ore sample and δ is the thickness of the liquid film, which is a function of the flow rate of the leaching agent and the radius of the ore-sample particle. Clearly, the leaching process might be enhanced by raising the reaction temperature, enlarging the contact area, increasing the concentration of the leaching agent and so on.
Based on equation 5, increasing the leaching agent concentration decreases the resistance of diffusion and improves the leaching reaction rate. When the concentration of the leaching agent is <3%, the leaching efficiency is low because of the small concentration gradient. When the concentration of the leaching agent is >3%, the concentration gradient is large enough for efficient diffusion, so the leaching efficiency would remain stable with increasing leaching agent concentration. When the leaching agent concentration is adequate, the effect of the leaching agent concentration on mass transfer is insignificant. This indicates that the leaching reagent concentration should not be too high in real production. Considering the consumption, the concentration of MgSO4 solution should be 3%.
Effect of L:S ratio on rare earth leaching
The L:S affects the probability of contact of the rare earth ion with the leaching agent. These experiments were carried out in 3% MgSO4 solution for 30 min of leaching time. The mass of ore was 100 g and the volume of MgSO4 solution was such to keep an L:S of 1:1–7:1. The results are shown in Fig. 7. The leaching efficiency increases with increasing L:S, reaching a maximum value at L:S = 3:1 and becoming relatively constant at higher L:S values. This behaviour might be explained by the mass transportation equilibrium between RE 3+ and Mg2+.
The effect of L:S ratio on the leaching efficiency. MgSO4 concentration: 3 wt.%, 30 min leaching time, ω = 500 rpm, pH = 4.4.
In practice, a higher L:S is sometimes used to ensure complete slurry suspension (i.e. high solid–liquid contact interface) and efficient agitation. However, a larger electrolyte fraction is not deemed necessary because sufficient stoichiometric excess is delivered by using greater L:S and material consumption. In the present study, however, L:S = 1:1 led to a viscous slurry that was impossible to mix due to the extremely fine nature of the clays used and the general tendency of dry clays to adsorb water and expand. Therefore, L:S = 3:1 was chosen here, and the leaching efficiency exceeded 75% (Fig. 7).
Leaching efficiency
In practical applications, leaching is a continuous process. If the leaching efficiency in stage 1 is unsatisfactory, the leaching operation may be prolonged. Figure 8 shows the detailed two-stage leaching process.
Influence of two-stage leaching on overall rare earth-extraction levels using different lixiviants.
The two-stage leaching process was employed to improve the overall leaching efficiency. The ore sample was initially leached in 3% MgSO4 solution for 30 min to obtain the stage 1 leaching solution (L1). The residue from stage 1 was re-leached in fresh 3% MgSO4 solution under the same conditions to produce the stage 2 leaching solution (L2). As was expected, the overall leaching efficiency of the two-stage leaching (Ltot = L1 + L2 = 96.19%) was higher than that of one-stage leaching (L1 = 75.48%). Furthermore, one-stage leaching employing 3% (NH4)2SO4 solution achieved 80.8% extraction. After application of a second stage, the leaching efficiency improved to 93.87% (i.e. slightly less than the 96.19% achieved by the 3% MgSO4 system).
Effect of partitioning of rare earths on rare earth leaching
The individual REE leaching efficiencies using different leaching agents are shown in Fig. 9. Leaching varies from element to element, with Ce presenting the lowest leaching efficiency. This may be attributed to the fact that Ce3+ is easily oxidized to Ce4+ by atmospheric oxygen (O2) (electrode potential φCe3+/Ce4+ = –1.72 V; Bard et al., Reference Bard, Parsons and Jordan1985). The Ce4+ precipitates as CeO2, inhibiting the extraction of CeO2 in the leaching liquor.
Individual lanthanide leaching with MgSO4 lixiviants. MgSO4 concentration: 3 wt.%, 30 min leaching time, L:S ratio = 3:1, ω = 500 rpm, pH = 4.4.
Table 4 lists the partitioning of REEs in the leaching solution at different temperatures and pH values in order to clarify the leaching behaviour of individual REEs. The partition in the leaching solution under different conditions barely changes with temperature and pH, which is almost in agreement with the REE partitions of the ion-exchangeable phase in the raw ore, except for Ce.
Individual rare earth partitioning in leaching solution under various conditions (MgSO4 concentration = 3 wt.%, L:S ratio = 3:1, leaching time = 30 min).
CONCLUSIONS
This study investigates the influence of experimental conditions such as leaching time, lixiviant concentration and L:S on the leaching efficiency of REEs using MgSO4 instead of (NH4)2SO4 as a lixiviant. It was observed that:
(1) The leaching efficiency for REE increases with increasing leaching time, lixiviant concentration and L:S within a certain range of conditions (1–7 min leaching time, lixiviant concentration of 1–3 wt.%, L:S from 1:1 to 3:1).
(2) The leaching efficiency for REEs is up to 75.48% for optimal conditions of a leaching time of 30 min, a lixiviant concentration of 3% and an L:S ratio of 3:1. The leaching efficiency of REEs reaches 96.19% by extending the leaching stages.
(3) Leaching of individual lanthanides varies from element to element, with Ce presenting the lowest leaching efficiency, and with the partition in the leaching solution being in agreement with that in raw ore, except for Ce.
These results will be useful for finding a novel approach to improving REE leaching efficiency and reducing NH4+–N pollution towards green chemistry.
ACKNOWLEDGMENTS
Financial aid from the following programme is gratefully acknowledged: National Natural Science Foundation of China (51504116).
