Separation of fluorine/cerium from fluorine-bearing rare earth sulfate solution by selective adsorption using hydrous zirconium oxide

Abstract

Separation of fluorine/cerium from fluorine-bearing rare earth sulfate solution using selective adsorption was studied using hydrous zirconium oxide as an adsorbent. The relevant parameters studied for fluoride adsorption were the effects of contact time, pH, n_F/n_Ce ratio, initial fluoride concentration and coexisting ions. The material was characterized using XRD, SEM, EDS, Raman, FTIR and XPS. The Raman, FTIR and XPS spectra suggest that an ion-exchange reaction between the hydroxyl ion on hydrous zirconium oxide and fluoride is involved. Most of the fluoride adsorption takes place in the first 3 min. The adsorption capacities of fluoride and cerium both increase with an increase in solution pH, and the optimum pH is suggested to be 0.3–0.6. The loss of cerium is higher at a low n_F/n_Ce ratio. High initial fluoride concentration is unfavorable for the separation. The accompanied rare earth ions have no significant influence on the adsorption of fluoride. The adsorbed fluoride can be desorbed effectively with 0.1 mol l⁻¹ NaOH which shows utility of the adsorbent in a sustainable manner. In addition, the effectiveness of the method was also evaluated using real bastnaesite sulfuric acid leaching solution. This work presents a potential use of hydrous zirconium oxide for adsorption and separation of fluorine from cerium in fluorine-bearing rare earth sulfate solution and it is expected to eliminate the influence of fluorine on the extraction separation of rare earths.

---

1 Introduction

Bastnaesite (ReCO₃F, Re = rare earths) is one of the most important mineral resources containing about 50 wt% cerium, 0.2–0.3 wt% thorium and 8–10 wt% fluorine.^1,2 There are several hydrometallurgical methods to extract rare earths from bastnaesite. Recently, the main technology for bastnaesite treatment has been “oxidation roasting–acid leaching–solvent extraction”. According to the selectivity in the separation of rare earth elements, the type of gangue minerals in the ore and the reagents to be used in the further extraction procedures, sulfuric, hydrochloric and nitric acids can be chosen as the leaching reagents.³ The sulfuric acid leaching process can dissolve almost all the Ce⁴⁺, RE³⁺ and F⁻, which makes the resources in bastnaesite fully utilized. It is considered to be a more advanced technology for bastnaesite comprehensive utilization in the future.^4–7

During the process of bastnaesite treatment, some issues still remain. Bastnaesite theoretically contains about 8–10% fluorine. Due to its high electronegativity and small ionic size, the fluoride ion mainly exists in the form of [CeF₂]²⁺ and REF₃ complexes in the sulfuric acid system,^8–10 which makes it difficult to extract Ce⁴⁺ and RE³⁺, and brings about the formation of the third phase in the extraction process. Therefore, the key point of the sulfuric acid leaching process is to eliminate the influence of fluorine. Lots of investigations about defluorination in the hydrometallurgical process of bastnaesite have been reported.^11–14 In recent years, the most commonly used methods for fluoride removal are adsorption,^15,16 reverse osmosis,¹⁷ ion exchange,¹⁸ precipitation,¹⁹ etc. Adsorption has been found to be the most important one for easy operation, affordable cost and water quality.²⁰ A large number of materials have been studied as adsorbents, such as activated alumina,^21,22 activated carbon,²³ bone char,²⁴ clay²⁵ and zeolite.²⁶ The amorphous hydrous zirconium oxide has been known to have a remarkable selectivity for fluoride ions, and has attracted attention for defluorination due to the simplicity of preparation, the large surface area and the strong regeneration capacity.^27–30 It can be used for defluoridation of groundwater and industrial wastewater. To our knowledge, a systemic study of fluorine/cerium separation in the refining of bastnaesite through adsorption has not been seen yet.

In this paper, we chose hydrous zirconium oxide as the adsorbent due to its applicability in strong acidic solution. The hydrous zirconium oxide was prepared through precipitation, and the separation of fluorine/cerium from fluorine-bearing rare earth sulfate solution using selective adsorption onto hydrous zirconium oxide has been studied. The objective of this paper is mainly to eliminate the influence of fluorine on the extraction separation of rare earths.

2 Experimental

2.1 Materials

Sulfuric acid (98%), sodium fluoride (NaF), cerium sulfate tetrahydrate (Ce(SO₄)₂·4H₂O), zirconium oxychloride (ZrOCl₂·8H₂O), lanthanum oxide (La₂O₃), praseodymium oxide (Pr₆O₁₁), neodymium oxide (Nd₂O₃), samarium oxide (Sm₂O₃), etc. of analytical grade were purchased from Shenyang Guoyao Group Chemical Reagent Co., Ltd. All of the chemicals were diluted with deionized water.

The mineral samples of bastnaesite used in this work were obtained from the Sichuan Mianning mine in China.

2.2 Instruments

A NOVA 1200e surface area & pore size analyzer (Quantachrome) was used to analyze the BET surface area and pore size distribution using N₂ adsorption at 77 K. A BT-9300H laser scattering particle analyzer (Bettersize Instruments Ltd.) was used to detect the particle size distribution of the product. An Ultima IV X-ray diffraction system (Rigaku) was used to examine the phase structure in 2θ of 10–80°at a rate of 8° min⁻¹ with Cu Kα radiation. A S-3400N scanning electron microscope (Hitachi) with an energy dispersive spectrum (EDS) technique was used to detect the surface atomic composition of the adsorbent. A Nicolet-380 Fourier transform infrared spectrometer (Thermo Electron Corporation) was used to detect the chemical structure of the product, and the preparation of the sample consisted of dispersing and gently grinding the powder in KBr. A HR800UV laser Raman spectrometer (Horibajobin Yvon) was used to record the Raman spectra of the product in the range of 100–1000 cm⁻¹ using the excitation wavelength of 325 nm at a resolution of 2 cm⁻¹. An ESCALAB250 X-ray photoelectron spectrometer (XPS) (Thermo VG) with an Al Kα excitation source (1486.6 eV) was used to examine the surface chemical composition of the samples. A C1s calibration energy of 284.6 eV was used as a reference to correct the XPS results. The XPS data were fitted using a nonlinear least-squares curve fitting program (XPSPEAK41 software).

A PXSJ-216 ion meter connected with a PF-1 fluoride ion electrode and saturated calomel electrode (Shanghai Precision Scientific Instrument Co., Ltd.) was used to determine the electric potential of fluoride. A PHS-3C digital pH meter connected with an E-201-C pH electrode (Shanghai Precision Scientific Instrument Co., Ltd.) was used to detect the pH values of the solutions. A HY-4 speed control multi-use oscillator (Jiangsu Jintan Dadi Automation Instrument Plant) was used to carry out the adsorption experiments.

2.3 Synthesis of hydrous zirconium oxide

The hydrous zirconium oxide was prepared through dissolution of 25 g of zirconium oxychloride in deionized water. An excess of ammonium hydroxide solution (1 : 1) was slowly added dropwise to the solution while stirring magnetically at 60 °C until the pH reached 8–9, and the precipitate obtained was hydrous zirconium oxide. After precipitation, it was activated at 65–70 °C with constant stirring for 2 h, then filtered and washed several times with distilled water until free of Cl⁻. The product was dried in an oven at 100–110 °C until a constant weight was achieved, then it was ground for detection.³¹

2.4 Batch adsorption experiments with synthetic solution

The synthetic fluoride stock solution was prepared through dissolution of cerium sulfate tetrahydrate, sodium fluoride, rare earth oxides and sulfuric acid in deionized water and was further diluted to the desired concentration for practical use. All adsorption studies were carried out by mixing known weights of adsorbent material with 100 ml of fluorine-bearing rare earth sulfate solution in a plastic bottle. The mixture was stirred thoroughly at room temperature, and then the sample was filtered with a 0.45 μm filter. The filtrate was then analyzed for fluoride and other ions concentrations. The experimental variables considered were: (1) contact time between the adsorbent and solution, 0–10 min; (2) pH of the solution, 0–1; (3) n_F/n_Ce ratio, 1–2; (4) initial fluoride concentration, 0.5–8 g l⁻¹; (5) coexisting rare earth ions, La³⁺, Pr³⁺, Nd³⁺, Sm³⁺.

Cerium concentration was determined through a titration with standard (NH₄)₂Fe(SO₄)₂ using sodium diphenylamine sulfonate as the indicator. Trivalent rare earth concentrations were determined through a titration with standard EDTA using xylenol orange as the indicator. Zirconium concentration was determined through a thermal titration with standard EDTA with xylenol orange as the indicator under HNO₃ conditions, and using aluminum salt as the fluorine complexing agent, with hydroxylamine hydrochloride as the Ce⁴⁺ reducing agent. Fluoride concentration was analyzed with a fluoride-selective electrode connected to an ion meter. Before determination, total ionic strength adjusting buffer (TISAB) was added to the solution in order to maintain ionic strength and the pH, and eliminate the interference effect of complexing ions.

2.5 Adsorption experiment using real bastnaesite sulfuric acid leaching solution

The effectiveness of the method was also evaluated using real bastnaesite sulfuric acid leaching solution. The water sample was collected from the sulfuric acid leaching of roasted bastnaesite. The typical concentration ranges in sulfuric acid leaching solution are CeO₂ (15–50 g l⁻¹), CeO₂/REO (47–50%), F (0.5–8 g l⁻¹), C_H⁺ (0.2–3 mol l⁻¹) and other trace elements. The n_F/n_Ce ratio is about 1–2. The adsorption and separation experiments were conducted with varying adsorbent dose from 5 to 30 g l⁻¹ of the sample by the batch process. Hydrous zirconium oxide of a definite amount was added in 100 ml of sulfuric acid leaching solution sample, and mixed well using a mechanical stirrer. The filtrate was analyzed for fluoride and cerium concentrations.

2.6 Desorption and regeneration

The desorption and regeneration of the fluoride-loaded hydrous zirconium oxide was performed following a sequence of 1st to 5th cycle of batch operation study. The sample was treated in sodium hydroxide solution (pH = 13) at room temperature for 4 h with a solid/solution ratio of 1 : 5–10 under stirring. Then the sample was filtered, washed several times with deionized water, and dried in an oven at 100–110 °C for 1–2 h. Regenerated hydrous zirconium oxide was attained. The fluoride concentration in the filtered solutions was analyzed to calculate the desorption ratio.

2.7 Equation

The adsorption rate η (%), adsorption capacity q_t (mg g⁻¹) and distribution coefficient D (ml g⁻¹) were calculated as the following expressions:where c_i denotes the initial ion concentrations in solution, g l⁻¹; c_f denotes the final ion concentrations, g l⁻¹; V denotes the volume of solution, ml; m denotes the mass of the adsorbent material, g.

3 Results and discussion

3.1 Characterization of the adsorbent

The specific surface area and the average micropore diameter of the adsorbent as determined using a BET technique were found to be 171.53 m² g⁻¹ and 3.41 nm, respectively. The particle size distribution of the hydrous zirconium oxide is given in Fig. 1. The mean diameter of the adsorbent is about 16.82 μm.

Fig. 1 Particle size distribution of the hydrous zirconium oxide.

The X-ray diffraction pattern of the synthesized hydrous zirconium oxide is shown in Fig. 2. No crystalline peak is detected in the pattern, indicating the amorphous characteristics of the material. The surface morphology of the hydrous zirconium oxide before and after adsorption is shown in Fig. 3. The micrographs show that the hydrous zirconium oxide has a significantly rougher surface with lots of pores, indicating that the adsorbent is porous. After adsorption, the fluoride-adsorbed adsorbent has smaller particles and less pores. The EDS spectrum shows the elemental composition of the fluoride-loaded hydrous zirconium oxide. The F signal reveals that fluoride is successfully adsorbed on the hydrous zirconium oxide.

Fig. 2 X-ray diffraction pattern of the prepared hydrous zirconium oxide.

Fig. 3 SEM images of the hydrous zirconium oxide before (a) and after (b) adsorption and (c) an EDS spectrum of the fluoride-adsorbed material.

3.2 Mechanism of fluoride adsorption

The Raman spectra of hydrous zirconium oxide before and after adsorption are shown in Fig. 4. The shapes of the spectra indicate an amorphous structure which is consistent with the X-ray diffraction pattern. The broad band at 533 cm⁻¹ is due to the bridging-hydroxy perpendicular-to-plane Zr–O stretch.³² The bands at 207 cm⁻¹ and 400 cm⁻¹ are indicative of the Raman-active mode for ZrO₂.³³ Both of the characteristic bands weakened after fluoride adsorption, indicating a chemisorption process involving the hydrous zirconium oxide and fluoride. The time dependence is also shown in Fig. 4. It is observed that the fluoride adsorption is a fast process and can take place within 10 min.

Fig. 4 Raman spectra of hydrous zirconium oxide before and after adsorption with laser excitation at 325 nm.

The FTIR spectra of hydrous zirconium oxide before and after adsorption are presented in Fig. 5. The peaks at 3573 and 3469 cm⁻¹ are attributed to the symmetric and asymmetric stretching modes of molecular water coordinated to the hydrous zirconium oxide and the bending vibration of the –OH bonds of chemisorbed water.³⁴ The peak at 3415 cm⁻¹ is due to the stretching modes of the –OH bonds related to free water adsorbed on the surface of the adsorbent. The peak at 1625 cm ⁻¹ is due to the bending modes of H–O–H bonds. The peak observed at 470 cm⁻¹ is assigned to the Zr–O bonds in ZrO₂.³⁵ The peak at 1353 cm⁻¹ (ref. 36 and 37) is for the bending vibration of the Zr–OH groups which decreases significantly after fluoride sorption. This evidence indicates that fluoride has replaced a substantial fraction of the –OH groups bound to zirconium. The peak at 1125 cm⁻¹ can be assigned to the adsorbed SO₄²⁻ groups. It is also found from the spectra that the intensity of the Zr–OH peak decreases sharply in the initial 5 min, and has no evident change as the time goes on, which indicates that the adsorption of fluoride on hydrous zirconium oxide is a very fast process.

Fig. 5 FTIR spectra of the hydrous zirconium oxide before and after adsorption.

To further investigate the interactions between the fluoride and hydrous zirconium oxide, XPS studies of the hydrous zirconium oxide before and after adsorption were conducted and the results are displayed in Fig. 6–8. The mechanism between the fluoride and functional groups on the hydrous zirconium oxide can be concluded through the characteristic peak shift and intensity change in the XPS spectra. As shown in Fig. 6, a strong F1s peak at 685.5 eV is found after adsorption,²⁷ clearly indicating that surface reactions have occurred between fluoride and zirconium. Fig. 7 gives the O1s XPS spectra before and after adsorption. For the original adsorbent, the O1s peak can be fitted into two peaks at 529.8 and 531.6 eV, which are assigned to the Zr–O and Zr–OH groups, respectively.³⁸ After adsorption, the Zr–O and Zr–OH peaks shift to higher binding energies by about 0.5 eV and 0.3 eV. This must be due to the formation of a new zirconium species. The area ratio of the Zr–OH peak decreases from 63.2 to 39.4%, while the area ratio of the Zr–O peak increases from 36.8 to 60.6%, suggesting that the hydroxyl groups on the hydrous zirconium oxide surface surely participated in fluoride sorption, which is consistent with the result of the FTIR spectra.

Fig. 6 F1s core-level spectra of the hydrous zirconium oxide before and after adsorption.

Fig. 7 O1s core-level spectra of the hydrous zirconium oxide before and after adsorption.

Fig. 8 Zr3d core-level spectra of the hydrous zirconium oxide before and after adsorption.

Fig. 8 shows the Zr3d XPS spectra before and after adsorption. The prominent peaks in the spectra before adsorption at 182.1 and 184.4 eV can be attributed to Zr3d_5/2 and Zr3d_3/2, which broadened and shifted to higher binding energies by about 0.3 eV after adsorption. This is what would be expected when zirconium becomes bonded to fluoride. Bonding to fluoride causes the loss of electron density at zirconium, which in turn raises the Zr3d_5/2 and Zr3d_3/2 binding energies.²⁷ The deconvoluted Zr3d spectra of the adsorbent after adsorption is exhibited in Fig. 8(b) as two overlapped peaks for both Zr3d_3/2 and Zr3d_5/2. The new peaks at 182.9 and 184.9 eV can be attributed to Zr-oxy fluoride species. Similar results have been reported by previous studies.^27,39,40

According to the above results, the mechanism is supported with the evidence of the FTIR and XPS spectra for the involvement of the –OH group in fluoride adsorption. The most probable reaction is given as:

–Zr–OH_(s) + F⁻_(aq) ⇄ –Zr–F_(s) + OH⁻_(aq) (4)

It has been reported that F⁻ coordinates with Ce⁴⁺ to form a [CeF₂]²⁺ complex in fluorine-bearing rare earth sulfate solution at high acidity.⁹ The hydrous zirconium oxide can take the F⁻ from [CeF₂]²⁺, leading to the separation of fluorine and cerium. The separation mechanism can be described by the reaction below:

2–Zr–OH_(s) + [CeF₂]²⁺_(aq) ⇄ 2–Zr–F_(s) + Ce⁴⁺_(aq) + 2OH⁻_(aq) (5)

3.3 Batch adsorption experiments with synthetic solution

3.3.1 Effect of contact time. According to the Raman and FTIR analyses, the fluoride adsorption is a fast process. The effect of contact time on the separation of fluorine and cerium was conducted within 10 min.

Fig. 9 depicts the effect of contact time on the variation of adsorption rates of fluoride and cerium. It is observed that most of the fluoride adsorption takes place in the first 3 min. This could be attributed to the availability of a high surface area as well as the porous structure facilitating the adsorption of fluoride. The cerium adsorption is about 0 during the whole process, indicating that hydrous zirconium oxide possesses strong selective absorption capacity to fluoride in F–Ce solution. In addition, there was a possibility that some of the hydrous zirconium oxide dissolved in the acid solution, so the concentration of zirconium was measured after the adsorption, and the results are shown in Fig. 9. The dissolved amount is quite low in the initial 5 min of the reaction but increases with an increase in time, so 3 min was chosen as the period of contact time for further studies.

Fig. 9 The effect of contact time on the separation of fluoride and cerium (concentration of fluoride: 1 g l⁻¹; n_F/n_Ce: 2; pH: 0.3; adsorbent dose: 10 g l⁻¹).

3.3.2 Effect of solution pH. The solution pH is one of the most important parameters which directly influences the existing state of cerium and the characteristics of the adsorbent. It is reported that cerium only exists in the form of Ce⁴⁺ at pH 1.0 or lower, and [Ce(OH)]³⁺, [Ce(OH)]²⁺ or [CeO]²⁺ at higher pH values.^41,42 Therefore, the effect of solution pH on the separation of fluorine and cerium from fluorine-bearing rare earth sulfate solution was examined over the pH range of 0 to 1 and the results are illustrated in Table 1. The results show that the adsorption capacities of fluoride and cerium both increase with increase in solution pH. The dissolved amount of hydrous zirconium oxide increases with the rise of solution acidity as shown in Fig. 10, leading to a decrease of fluoride adsorption amount. The high D_Ce at pH around 0.8 can be attributed to the strong hydrolysis properties of cerium at lower acidity, which leads to the loss of cerium. Shifts of pH before and after adsorption were also observed. The pH of the treated solutions increased. The mechanism of fluoride adsorption using hydrous zirconium oxide was thought to be anion exchange between the available surface –OH groups and fluoride. The increased pH after adsorption can be attributed to the liberation of hydroxyl ions on the hydrous zirconium oxide. This can also be supported by the evidence of the FTIR and XPS spectra.

Table 1 The effect of initial pH on the separation of fluoride and cerium (concentration of fluoride: 1 g l⁻¹; n_F/n_Ce: 2; adsorbent dose: 10 g l⁻¹)

Initial pH	η/%		qt/(mg g^−1)		D/(ml g^−1)
	F	Ce	F	Ce	F	Ce
0.00	67.23	0.00	67.23	0.00	205.18	0.00
0.12	68.20	0.00	68.20	0.00	214.46	0.00
0.32	68.98	1.68	68.98	6.21	222.41	1.71
0.61	72.59	9.28	72.59	34.25	264.80	10.23
0.82	74.70	50.10	74.77	184.50	296.29	100.00

---

Fig. 10 The effect of initial pH on the dissolution of the adsorbent and the variation of final pH against initial pH.

To obtain a high separation efficiency with a lower dissolved amount of hydrous zirconium oxide, the optimum pH is suggested to be 0.3–0.6.

3.3.3 Effect of n_F/n_Ce. The n_F/n_Ce ratio is an important parameter of fluorine-bearing rare earth sulfate solution which affects the complex state of fluoride and cerium. The effect of the n_F/n_Ce ratio on the separation of fluorine and cerium was investigated as shown in Table 2. It is found that the adsorption capacities of fluoride and cerium decrease with increasing n_F/n_Ce from 1 to 2. Theoretically, the hydrous zirconium oxide has no adsorption ability for cerium, but there are some residual free Ce⁴⁺ at a low n_F/n_Ce ratio which will enter the skeleton and channel of the hydrous zirconium oxide by mechanical mixing,⁴³ resulting in the loss of cerium. With an increase in the n_F/n_Ce ratio, F⁻ and Ce⁴⁺ exist mainly in [CeF₂]²⁺ complexes, and a few free Ce⁴⁺ are left in solution, leading to less of a loss of cerium and a high separation efficiency of fluorine and cerium. As shown in Fig. 11, the release of zirconium is in an acceptable range and has a slight rise with the increase of n_F/n_Ce, which is probably because the uptake of cerium can reduce the dissolution of hydrous zirconium oxide.

Table 2 The effect of n_F/n_Ce ratio on separation of fluoride and cerium (concentration of fluoride: 1 g l⁻¹; pH: 0.3; adsorbent dose: 10 g l⁻¹)

nF/nCe	η/%		qt/(mg g^−1)		D/(ml g^−1)
	F	Ce	F	Ce	F	Ce
1.0	81.49	9.98	81.49	73.58	440.37	11.09
1.2	80.15	8.35	80.15	51.35	403.78	9.11
1.5	79.55	4.47	79.55	22.00	389.03	4.68
1.8	78.54	3.29	78.54	13.49	366.15	3.40
2.0	77.98	2.23	77.98	8.22	334.52	2.28

---

Fig. 11 The effect of the n_F/n_Ce ratio on dissolution of the adsorbent.

3.3.4 Effect of the initial fluoride concentration. The effect of initial fluoride concentration on the separation of fluorine and cerium is shown in Table 3. It is noticed that the adsorption rate and distribution coefficient of fluoride decrease sharply with the increase of initial fluoride concentration, while the adsorption capacity of fluoride increases. This may be because at a fixed dose of adsorbent, the adsorption on the surface is saturated faster at a higher concentration of fluoride showing a higher q_e value.⁴⁴ Additionally, the dissolved amount of zirconium gradually increases with increasing fluoride concentration as shown in Fig. 12. This may be because the high fluoride concentration promotes the formation of HF, and HF can exist in the solution destroying the Zr–O–Zr bonding of hydrous zirconium oxide.⁴⁵ This result shows that high initial fluoride concentration is unfavorable for the separation. The optimum concentration should be less than 5 g l⁻¹.

Table 3 The effect of the initial concentration of fluoride on the separation of fluoride and cerium (n_F/n_Ce: 2; pH: 0.3; adsorbent dose: 15 g l⁻¹)

Initial concentration of fluoride	η/%		qt/(mg g^−1)		D/(ml g^−1)
	F	Ce	F	Ce	F	Ce
0.5	85.86	0.00	28.62	0.00	405.01	0.00
1.0	85.15	0.00	56.77	0.00	382.25	0.00
2.0	84.16	4.87	112.21	23.92	354.15	3.41
3.0	80.33	6.74	160.67	49.72	272.33	4.82
4.0	77.11	7.69	205.61	75.69	224.52	5.56
5.0	69.98	7.78	233.28	95.71	155.43	5.63
6.0	65.72	2.73	262.87	40.24	127.80	1.87
7.0	51.92	1.48	242.31	25.43	72.00	1.01
8.0	45.48	0.00	242.55	0.00	55.61	0.00

---

Fig. 12 The effect of the initial concentration of fluoride on the dissolution of the adsorbent.

3.3.5 Effect of coexisting rare earth ions. A variety of other rare earths generally accompany cerium in fluorine-bearing rare earth sulfate solutions, such as La, Pr, Nd and Sm. The influence of coexisting rare earth ions on the adsorption of fluorine and the separation of fluorine and rare earths was investigated. The results are shown in Table 4. The RE³⁺ ions were prepared by dissolving certain amounts of their oxides in solution, and were equimolar with cerium. It is found that the distribution coefficients of fluoride and cerium are almost stable, and the hydrous zirconium oxide has no adsorption for RE³⁺, implying that the accompanying ions have no significant influence on the adsorption of fluorine, and it is possible to separate fluorine and rare earths from fluorine-bearing rare earth sulfate solution using hydrous zirconium oxide because of the high selectivity and anti-interference ability for fluoride.

Table 4 The effect of coexisting rare earth ions on the separation of fluoride and cerium (concentration of fluoride: 1 g l⁻¹; n_F/n_Ce: 2; pH: 0.3; adsorbent dose: 10 g l⁻¹)

RE^3+	η/%			qt/(mg g^−1)		D/(ml g^−1)
	F	Ce	RE	F	Ce	F	Ce
La	76.51	1.96	0.00	76.51	7.23	325.60	2.01
Pr	77.12	1.09	0.00	77.12	4.01	337.12	1.16
Nd	75.89	3.28	0.00	75.89	12.13	314.73	3.45
Sm	77.31	1.67	0.00	77.31	6.16	340.85	1.74

---

3.4 Experiment using real bastnaesite sulfuric acid leaching solution

The main components of the used bastnaesite sulfuric acid leaching solution sample estimated were CeO₂ 32.71 g l⁻¹, CeO₂/REO 48.25%, F 4.96 g l⁻¹, and C_H⁺ 2 mol l⁻¹. The pH of the water sample was adjusted to be 0.3 with ammonium hydroxide. The synthetic solution was also prepared according to the composition of the realistic solution. The effect of adsorbent dose on the separation of fluorine and cerium from the synthetic and realistic solution samples was studied as shown in Fig. 13. It is apparent that a higher adsorbent dose corresponds to a higher distribution coefficient of fluoride. This agrees well with the increase of a solid dose for a fixed solute, and surface sites heterogeneity of the adsorbent.⁴⁴ The distribution coefficient of cerium is extremely low when the dose is below 20 g l⁻¹, but obviously increases when the dose is above 20 g l⁻¹. Thus, a good separation of fluorine and cerium can be obtained at an adsorbent dose of 20 g l⁻¹ with a higher distribution coefficient of fluoride and a lower loss of cerium.

Fig. 13 The separation of fluoride and cerium in synthetic and realistic solution samples with a varying adsorbent dose.

Besides, a higher distribution coefficient is obtained from the synthetic solution in comparison to the realistic bastnaesite sulfuric acid leaching solution. The reason could be due to the fact that the real solution is always associated with a large number of cations and anions, which could interfere with the normal adsorption process.⁴⁶

3.5 Desorption and regeneration

The regeneration ability is an important factor for an advanced adsorbent, and the cost can be reduced if the adsorbent can be regenerated and reused in many cycles of operation. The desorption experiments of fluoride-adsorbed hydrous zirconium oxide were performed using 0.1 mol l⁻¹ NaOH solution as the regeneration fluid.⁴⁴ The EDS pattern of the regenerated zirconium oxide is shown in Fig. 14. The absence of F in the EDS spectrum reveals that the fluoride can be desorbed effectively with a desorption ratio reaching 91%. Fluoride adsorption experiments were also performed with the original and regenerated hydrous zirconium oxide. The results are shown in Fig. 15. After five cycles of the desorption–adsorption process, the regenerated hydrous zirconium oxide material can still show a high adsorption ability just with a little reduction.

Fig. 14 The EDS pattern of the regenerated adsorbent.

Fig. 15 Regeneration test of the hydrous zirconium oxide.

4 Conclusions

In this study, amorphous hydrous zirconium oxide prepared through precipitation was used for the adsorption and separation of fluorine from cerium in fluorine-bearing rare earth sulfate solution. The adsorbent was characterized using XRD, SEM and EDS. The batch adsorption studies show that: most of the fluoride adsorption takes place in the first 3 min; the adsorption capacities of fluoride and cerium both increase with an increase in solution pH, and the optimum pH is suggested to be 0.3–0.6; the loss of cerium is higher at a low n_F/n_Ce ratio; high initial fluoride concentration is unfavorable for the separation; the accompanying rare earth ions have no significant influence on the adsorption of fluorine. A possible mechanism of the adsorption process involving an ion-exchange reaction between the hydroxyl ion on the hydrous zirconium oxide and fluoride is proposed based on the Raman, FTIR and XPS studies. The effectiveness of the method was also evaluated using real bastnaesite sulfuric acid leaching solution. The adsorbed fluoride can be desorbed effectively using 0.1 mol l⁻¹ NaOH with the desorption ratio reaching 91%, and the adsorbent can be utilized sustainably for a number of cycles. This work represents a potential use of hydrous zirconium oxide for the adsorption and separation of fluorine from cerium in fluorine-bearing rare earth sulfate solution and hence it is expected to eliminate the influence of fluorine on the extraction separation of rare earths.

Acknowledgements

Financial aids from the following programs are gratefully acknowledged. They are the key program of the National Natural Science Foundation of China (NSFC: 50934004), National Natural Science Foundation of China (51274061), Major State Basic Research Development Program of China (973 Program: 2012CBA01205), and the Fundamental Research Supporting Project of Northeastern University (N110602006).

References

1. G. X. Xu, Rare Earths, Metallurgical Industry Press, Beijing, 2 edn, 1995, pp. 377–399.
2. M. Kul, Y. Topkaya and İ. Karakaya, Hydrometallurgy, 2008, 93, 129–135.
3. G. M. Ritcey and A. W. Ashbrook, Amsterdam, 1979, 386–420.
4. X. W. Huang, Z. Q. Long, H. W. Li, W. J. Ying, G. C. Zhang and X. X. Xue, J. Rare Earths, 2005, 23, 1–4.
5. S. L. Zhao, C. G. He, Z. J. Cao, Y. Liu and F. Zhao, Chin. Rare Earths, 2003, 24, 32–35.
6. Q. W. Zhang and F. Saito, Hydrometallurgy, 1998, 47, 231–241.
7. A. Yörükoğlu and A. Obut, İ. Girgin, Hydrometallurgy, 2003, 68, 195–202.
8. R. M. Sawant, R. K. Rastogi, M. A. Mahajan and N. K. Chaudhuri, Talanta, 1996, 43, 89–94.
9. J. Qiao, C. R. Zhang, S. G. Liu and X. K. Hao, Chin. Rare Earths, 1997, 18, 64–67.
10. Z. F. Zhang, H. F. Li, F. Q. Guo, S. L. Meng and D. Q. Li, Sep. Purif. Technol., 2008, 63, 348–352.
11. X. H. Luo, X. W. Huang, Z. W. Zhu, Z. Q. Long and Y. Liu, J. Rare Earths, 2009, 27, 119–122.
12. J. N. Li, X. W. Huang, Z. W. Zhu, Z. Q. Long, X. L. Peng and D. L. Cui, J. Rare Earths, 2007, 25, 55–58.
13. X. W. Huang, J. N. Li, Z. Q. Long, Y. Q. Zhang, X. X. Xue and Z. W. Zhu, J. Rare Earths, 2008, 26, 410–413.
14. Z. Q. Long, X. W. Huang, W. M. Huang and G. C. Zhang, J. Chin. Rare Earths Sci., 2000, 18, 18–20.
15. Y. Ku and H. M. Chiou, Water, Air, Soil Pollut., 2002, 133, 349–360.
16. X. P. Liao and B. Shi, Environ. Sci. Technol., 2005, 39, 4628–4632.
17. S. V. Joshi, S. H. Mehta, A. P. Rao and A. V. Rao, Water Treat., 1992, 10, 307–312.
18. Y. Ku, H. M. Chiou and W. Wang, Sep. Sci. Technol., 2002, 37, 89–103.
19. M. Yang, Y. Zhang, B. Shao, R. Qi and H. Myoga, J. Environ. Eng., 2001, 127, 902–907.
20. S. Meenakshi and R. C. Maheshwari, J. Hazard. Mater., 2006, 137, 456–463.
21. S. Ghorai and K. K. Pant, Sep. Purif. Technol., 2005, 42, 265–271.
22. R. S. Wang, H. M. Li and W. Feng, J. Environ. Sci., 1992, 12(3), 333–339.
23. Y. H. Li, S. G. Wang, X. F. Zhang, J. Q. Wei, C. L. Xu, Z. K. Luan and D. H. Wu, Mater. Res. Bull., 2003, 38, 469–476.
24. N. A. Medellin-Castillo, R. Leyva-Ramos, R. Ocampo-Perez, R. F. Garcia de la Cruz, A. Aragon-Piña, J. M. Martinez-Rosales, R. M. Guerrero-Coronado and L. Fuentes-Rubio, Ind. Eng. Chem. Res., 2007, 46, 9205–9212.
25. M. Mahramanlioglu, I. Kizilcikli and I. O. Bicer, J. Fluorine Chem., 2002, 115, 41–47.
26. M. S. Onyango, T. Y. Leswifi, A. Ochieng, D. Kuchar, F. O. Otieno and H. Matsuda, Ind. Eng. Chem. Res., 2008, 48, 931–937.
27. X. M. Dou, D. Mohan, C. U. Pittman Jr and S. Yang, Chem. Eng. J., 2012, 198–199, 236–245.
28. J. A. Blackwell and P. W. Carr, J. Chromatogr. A, 1991, 549, 43–57.
29. K. Biswas, D. Bandhoyapadhyay and U. Ghosh, Adsorption, 2007, 13, 83–94.
30. Y. B. Sun, Q. H. Fang, J. P. Dong, X. W. Cheng and J. Q. Xu, Desalination, 2011, 277, 121–127.
31. R. Chitrakar, S. Tezuka, A. Sonoda, K. Sakane, K. Ooi and T. Hirotsu, J. Colloid Interface Sci., 2006, 297, 426–433.
32. P. Southon, Structure Evolution During the Preparation and Heating of Nanophase Zirconia Gels, PhD Dissertation, University of Technology, Sydney, 2000.
33. X. Zhao and D. Vanderbilt, Phys. Rev. B: Condens. Matter Mater. Phys., 2002, 65, 075105.
34. J. M. Pan, X. H. Zou, X. Wang, W. Guan, Y. S. Yan and J. Han, Chem. Eng. J., 2010, 162, 910–918.
35. S. K. Swain, S. Mishra, T. Patnaik, R. K. Patel, U. Jha and R. K. Dey, Chem. Eng. J., 2012, 184, 72–81.
36. M. L. Toullec, C. J. Simmons and J. H. Simmons, J. Am. Ceram. Soc., 1988, 71, 219–224.
37. H. L. Liu, X. F. Sun, C. Q. Yin and C. Hu, J. Hazard. Mater., 2008, 151, 616–622.
38. A. Roustila, J. Chêne and C. Séverac, Int. J. Hydrogen Energy, 2007, 32, 5026–5032.
39. S. D. Wolter, J. R. Piascik and B. R. Stoner, Appl. Surf. Sci., 2011, 257, 10177–10182.
40. C. G. Pantano and R. K. Brow, J. Am. Ceram. Soc., 1988, 71, 577–581.
41. P. H. Tedesco, V. B. De Rumi and J. A. González Quintana, J. Inorg. Nucl. Chem., 1967, 29, 2813–2817.
42. Y. Liu, Z. Q. Long, W. M. Huang, G. C. Zhang and X. W. Huang, J. Rare Earths, 2001, 19, 320–323.
43. H. R. Chen, J. L. Shi, L. Wang, T. D. Chen, J. N. Yan and D. S. Yan, J. Chin. Ceram. Soc., 2001, 29, 18–20.
44. K. Biswas, K. Gupta and U. C. Ghosh, Chem. Eng. J., 2009, 149, 196–206.
45. C. F. Chang, C. Y. Chang and T. L. Hsu, Desalination, 2011, 279, 375–382.
46. S. K. Swain, R. K. Dey, M. Islam, R. K. Patel, U. Jha, T. Patnaik and C. Airoldi, Sep. Sci. Technol., 2009, 44, 2096–2116.

---