Effect on ceramic grade CaF₂ recovery quality from the etching wastewater under the optimum sulfate content

Abstract

The improvement of fluorite quality is an attractive measure to recover fluoride from high F⁻ content wastewater and thus reduces the fluorite sludge generation rate, and the removal of contaminant pollutants could be helpful for the production of ceramic grade fluorite (>85%). In this work, typical etching wastewater was collected and characterized, and SO₄²⁻ was found to be the main barrier for high quality CaF₂ production, while CaSO₄ could also contribute to CaF₂ precipitation. The optimum SO₄²⁻ content was adjusted by the introduction of barium salts, to keep a balance between the CaF₂ quality and the settlement process. The ceramic grade CaF₂ of 88% was obtained under the conditions: [Ca]/[F] of 1.29, pH 8.14 and C_sulfate of 117.4 mg L⁻¹. The residual SO₄²⁻ content should be kept in the range of 100–130 mg L⁻¹, and the formation of CaSO₄, combined with CaF₂ could accelerate the CaF₂ precipitation by the co-precipitation method, with a larger particle size from 758 nm (without SO₄²⁻) to 898 nm. The gradient settlement of CaF₂ contributes to the recovery of fluoride from etching wastewater, while CaF₂ preparation should be carried out in a dynamic experiment, such as a fluidized bed, for further research.

---

Introduction

Hydrofluoric acid (HF) has been widely used for semiconductor manufacturing, electrolytic aluminum, the aerospace industry, nuclear industry and metallurgical industry, etc. Around 3.6 million tons of fluorspar have been consumed in the fluorine chemical industry in China, which accounted for nearly 60% of national consumption.¹ As a result, large quantities of wastewaters containing fluorine were generated, with high F⁻ concentrations of 500 to 2000 mg L⁻¹, especially in the etching process of the semiconductor industry.² The fluoride-containing wastewater is usually treated by chemical precipitation,³ adsorption,⁴ ultra-filtration (UF), reverse osmosis (RO), etc. The chemical precipitation with calcium salts additive is commonly used in the practical projects, while high amounts of CaF₂ sludge are generated and some of them belongs to hazardous waste according to the National Hazardous Waste List.⁵ The low quality of CaF₂ sludge had a negative impact for the local environment.

On the other hands, fluoride is a strategic resource. CaF₂ is the primary source for the fluorine chemical industry, and only 230 million tons was proven reserves in the world, while annually consumption of fluorspar reached to 6.2 million tons.¹ The recovery CaF₂ from the existing F resource is the potential way to solve this contradiction. The typical CaF₂ sludge was only <60% content, and could not be used as the minimum CaF₂ mineral quality of 85%, that is ceramic grade.⁶ If CaF₂ could be improved through the wastewater treatment process, two-win-win results could be observed for the wastewater treatment and F resource recovery.

Typically, the fluoride-containing wastewater contains sulfate ion (SO₄²⁻), organic matters and several metal ions (Cu²⁺, Ni²⁺, Al³⁺, etc.) from the semiconductor industry. The big barrier for this process is due to the high mixture of the anions and cations, especially for the SO₄²⁻ according to the typical etching solution, which reduced the CaF₂ quality due to the co-precipitation of impurities. Some flocculants, such as PAC were also introduced to increase the settlement velocity of CaF₂, and sacrifice the CaF₂ quality.⁷ To reduce the ex-chemical compounds additives, fluidizing bed reactor were used to recover CaF₂ from a simulated fluoride-containing wastewater,⁸ with a low F concentration of 100–200 mg L⁻¹, and some extra-seed crystal were introduced to form the large particle size, while the reactor was not so stable, and recovery costs increased and fluorite content decreased by introducing extra addition of seed crystal.⁹ Therefore, the removal of the impurities substances in wastewater could benefit for the CaF₂ quality and the inner core production using the substance available might be the important way for the CaF₂ settlement by the co-precipitation process.

In this study, the ceramic grade CaF₂ was recovered from the high fluoride-containing wastewater through the chemical precipitation, and the critical factors for the purified CaF₂ production were identified. The inference of SO₄²⁻ on the CaF₂ preparation and settling properties of calcium fluoride was evaluated and the optimum conditions were found using central composite design (CCD) model experimental design in response surface methodology (RSM) from the [Ca]/[F], pH of solution, C_sulfate of solution perspective.

Materials and methods

Fluoride-containing wastewater property

The fluoride-containing wastewater was procured from the etching and cleaning processes, which consisted of the etchant and cleaning agent (hydrofluoric acid, ammonium hydroxide, and sulfuric acid, etc.), were applied as the etching solutions, in a semiconductor manufacturing, Shanghai, China. The fluoride-containing wastewater were stored in a PVC container with 20 L. Due to quite acidic nature of hydrofluoric acid, sodium hydroxide is required to adjust pH of wastewater to the range of 5–10.¹⁰ The wastewater was analyzed immediately and the fundamental property was shown in Table 1.

Table 1 Characteristics of the representative sample of etching wastewater

Parameters	Unity	Value
pH	—	3.3–3.5
Conductivity	mS cm^−1	8–10
COD	mg L^−1	180–220
Anions
F^−	mg L^−1	1100–1800
SO4^2−	mg L^−1	350–420
PO4^3−	mg L^−1	17–42
Cations
Ca^2+	mg L^−1	5–7
Cu^2+	mg L^−1	8–15
Ni^2+	mg L^−1	5–10
Al^3+	mg L^−1	1–3

---

Experimental procedures

Pre-treatment was conducted to remove the metal ions and organic matter in the high concentration fluoride-containing wastewater, with the additive of 2g L⁻¹ activated carbon^11–13 for 30 min, at pH 7–9 according to previous study. Results showed that the corresponding change of wastewater property was hardly change except to metal ions (in Fig. S1†). The CaF₂ content and characteristics were displayed in Tables S1 and S2† and XRD patterns of synthesised CaF₂ prior to pre-treatment by activated carbon was presented in Fig. S2.† The schematic diagram of the calcium fluoride recovery process was presented in Fig. 1.

Fig. 1 Schematic diagram of the calcium fluoride recovery.

450 mL fluoride-containing wastewater was fed into the 1-Lpolytetra fluoroethylene (PTFE) reactor, and barium chloride (BaCl₂) was added according to the set Ba/SO₄ molar ratio at a stirring rate of 400 rpm for 30 min. Then the filtrate was used to analysis the F⁻ content after filtration by a 0.22 μm membrane. The filter residue was collected for the further analysis of the CaF₂ property.

After pretreatment by barium salts, pH was adjusted at a given value. Then the calcium chloride (CaCl₂) was introduced into the 400 mL fluoride-containing wastewater after pH adjustment. The mixture was stirred at a stirring speed of 400 rpm for 30 min, and the suspension particles settled down for 1 h. Solid was grinded and sieved through ASTM#240 sieve after dried at 105 °C for 12 h, and the precipitation particles were collected for subsequent chemical analysis. All measurements were made at room temperature (25 ± 1 °C) and atmospheric pressure under stirring conditions.

Response surface methodology

CCD in RSM was used for design, mathematical modeling, and optimization for the preparation of CaF₂, and [Ca]/[F], pH and C_sulfate were considered based on our preliminary experiments. The five-level central composite design¹⁴ with categorical factor was employed to optimize the purity of calcium fluoride (response). The factor levels were coded as −2 (lowest), −1 (low), 0 (central point or medium), +1 (high), +2 (highest). For the purpose of statistical computations, the independent variables were denoted as λ₁, λ₂, and λ₃, respectively as illustrated in Table 2. The range and value of each variable were selected and applied to transform an actual value (λ_i) into a coded value (X_i) according to the following equation:¹⁵

Table 2 Use of response surface methodology for optimum operating conditions of preparing CaF₂ in an aqueous fluoride solution by CaCl₂

Variables	Factors	Lowest	Low	Medium	High	Highest
		−2	−1	0	1	2
[Ca]/[F]	λ1	0.8	1.0	1.2	1.4	1.6
pH	λ2	4	6	8	10	12
Csulfate	λ3	60	90	120	150	180

---

where X_i is the dimensionless coded value, λ⁰_i represents the actual value at the center point, λ_i the actual value of an independent variable, and Δλ_i is the step change value.¹⁶

The experimental design matrix of 20 runs by the CCD is tabulated in Table 3. The second-order polynomial equation developed to fit the experimental data and determined the relevant model terms, as shown in eqn (2):

where Y represents the purity of calcium fluoride (%). β the constant coefficient, β_i the coefficient of the input factor X_i, β_ii the quadratic coefficient of the input factor X_i, β_ij is the different interaction coefficient between the input factors X_i and X_j (i = 1 or 2, j = 1 or 2 and i ≠ j), ε represents the error of the model. The equation expresses the relationship between the predicted response and independent variables in coded values.^17,18

Table 3 Settlement properties analysis of CaF₂ synthesized by CaCl₂ and etching wastewater after sulfate removal

Group	Zeta potential (mv)	Average diameter (nm)	Turbidity (NTU)
			Initial	Final
Control	20.20	898.35	338	25.8
Experiment	22.67	758.45	251	52.4

---

The quality of fit in the polynomial model is expressed by the correlation coefficient (R²), and its statistical significance is verified by an F-test in the same program.^19,20 The model terms are evaluated based on the P-value corresponding to a 95% confidence level. The detailed procedure for CCD under RSM can be found elsewhere.

Analytical methods

Fluoride concentration was measured using a special fluoride ion electrode (Rex PXS-270). Calcium fluoride content was determined by EDTA-Ca titrimetric method.²¹ The metallic species concentration of the fluoride-containing wastewater were analyzed using inductively coupled plasma optical emission spectrometer (ICP-AES, iCAP6300). The concentration of sulfate (SO₄²⁻), nitrate (NO₃⁻), and phosphate (PO₄³⁻) were analyzed using ion chromatography (IC, MICI). The pH and turbidity measurements were conducted with a pH meter (Metrohm 780) and turbidity meter (settling time of 1 h) (HACH 2100Q-01). X-ray power diffraction (XRD, Bruker D8 Advance X-ray Diffractometer) was used to characterize the crystalline phases of obtained precipitation. Scanning electron microscopic (SEM, OVA Nano SEM 230) and transmission electron microscope (TEM, JEM-2100) were employed to characterize the remains precipitation. In addition, zeta potential and particle size were measured by laser particle size analyzer (BECKMAN COULTER, Delsa Nano C). All tests were performed in duplicates, and the average value was represented.

Results and discussion

Sulfate concentration vs. CaF₂ quality

To investigate the effect of sulfate concentration on CaF₂ content, the sulphate concentration was effectively controlled by barium salts. The relation of removal efficiency of SO₄²⁻ and Ba²⁺/SO₄⁴⁻ molar ratio (BaCl₂ dosage) was presented in Fig. S3.† SO₄²⁻ can be removed completely through the reaction with barium salt.

Change of sulfate and fluoride concentration in the process of CaF₂ recovery was presented in Fig. 2, and C1 represented the effluent after SO₄²⁻ precipitation, C2 the effluent of CaF₂ preparation process. When the Ba²⁺/SO₄²⁻ molar ratio reach to 1.03, the sulfate removal had decreased sharply from 398 mg L⁻¹ to 36 mg L⁻¹. The control run experimental results showed that the sulfate removal decreased sharply after addition of CaCl₂ to synthesize CaF₂. Moreover, almost all F⁻ was kept in the filtrate during the BaSO₄ precipitation. During the synthesizing CaF₂, the fluoride recovery efficiency decreased from 98.6% to 71.9% if SO₄²⁻ was precipitated with Ba²⁺.

Fig. 2 Variations of sulfate and fluoride in both of sulfate removal and CaF₂ preparation processes.

CaF₂ was synthesized under the [Ca]/[F] of 1.2, pH of 7, and CaF₂ content was increased from 78.6% to 92.4% with the pre-precipitation by Ba²⁺, while CaSO₄ content was decreased from 12.4% to 0.6% (in Table S3†). Consequently, elevating the efficiency of sulfate removal, it can increase the purity of CaF₂. The change of SO₄²⁻ and the depressed of CaF₂ content indicated the existing of SO₄²⁻ anions can effect on the CaF₂ preparation, which was resulted from the competition for calcium source between sulfate and fluoride. XRD analysis of CaF₂ products and the commercial fluoriteis compared and shown in Fig. 3. The refraction peaks of the CaF₂ in the products match the commercial grade CaF₂, and other peaks are also found due to the formation of CaSO₄ in the settlement.

Fig. 3 XRD pattern analysis of solid sample: (a) CaF₂ synthesized directly, (b) CaF₂ generated by CaCl₂ and wastewater after SO₄²⁻ removal by BaCl₂, (c) commercial ceramic acid grade fluorite.

Sulfate concentration vs. CaF₂ settlement performance

To assess the settlement performance of CaF₂, the zeta potential and particle size distribution of CaF₂ were measured, and the results are shown in Table 3 and Fig. 4a. It could be found that the main distribution range scattered from 609–1479 nm to 246–4276 nm after the removal of SO₄²⁻. The corresponding average diameter decreased from 898 nm to 758 nm. With regards to turbidity, the effluent in the control group was around 25.8 NTU finally, and higher than that of control group, which reflected that more CaF₂ remained in the solution.²²

Fig. 4 (a) Particle size distribution cures of frequency and (b and c) TEM images of solid sample: (b) CaF₂ product synthesized directly, (c) CaF₂ product generated by CaCl₂ and wastewater after SO₄²⁻ removal by BaCl₂ ([Ba]/[SO₄] = 1, [Ca]/[F] = 1.2, pH = 7).

According to TEM presented in Fig. 4b and c, pellets consisted some aggregates and central sphere during the precipitation flock formation due to the heterogeneous crystallization. It could be concluded that the SO₄²⁻ could not be removed completely, and an optimum range should be control to keep the balance between the CaF₂ quality and the settlement velocity from the solution of CaF₂.

The optimum conditions of CaF₂ preparation

Regression model equation. The CaF₂ purified was optimized from the factors using CCD method, after the SO₄²⁻ were removed in the first step, and only 30–180 mg L⁻¹ SO₄²⁻ were remained in the wastewater. Based on the equation presented in Fig. S3,† the C_sulfate (60 mg L⁻¹, 90 mg L⁻¹, 120 mg L⁻¹, 150 mg L⁻¹ and 180 mg L⁻¹) was remained by the dosage of BaCl₂ (Ba/SO₄ molar ratio is 0.94, 0.85, 0.77, 0.68 and 0.6, respectively). A total number of 20 run experiments was designed for response surface modeling. The results were summarized in Table 4. The best CaF₂ were found in case 16, with the content of 88.41%.

Table 4 Central composite experimental design matrix and experimental responses

Run no.	Location	Coded values (actual values)			Response
		A [Ca]/[F]	B pH	C Csulfate (mg L^−1)	Y CaF2 content (%)
					Observed	Predicted
1	Star	−1 (1.0)	−1 (7)	−1 (90)	83.79	82.97
2	Star	+1 (1.4)	−1 (7)	+1 (150)	85.07	85.56
3	Star	+1 (1.4)	−1 (9)	−1 (90)	85.15	84.97
4	Star	+1 (1.4)	+1 (9)	+1 (150)	84.27	84.91
5	Star	−1 (1.0)	+1 (9)	−1 (90)	86.78	86.23
6	Star	+1 (1.4)	+1 (9)	−1 (90)	85.40	85.69
7	Star	−1 (1.0)	+1 (9)	+1 (150)	85.10	85.10
8	Star	−1 (1.0)	−1 (7)	+1 (150)	83.71	83.22
9	Axial	0 (1.2)	0 (8)	2 (180)	84.83	84.42
10	Axial	−2 (0.8)	0 (8)	0 (120)	82.37	83.27
11	Axial	0 (1.2)	−2 (6)	0 (120)	80.65	81.06
12	Axial	2 (1.6)	0 (8)	0 (120)	85.81	85.10
13	Axial	0 (1.2)	2 (10)	0 (120)	83.87	83.65
14	Axial	0 (1.2)	0 (8)	−2 (60)	84.37	84.96
15	Center	0 (1.2)	0 (8)	0 (120)	88.04	88.09
16	Center	0 (1.2)	0 (8)	0 (120)	88.41	88.08
17	Center	0 (1.2)	0 (8)	0 (120)	87.99	88.08
18	Center	0 (1.2)	0 (8)	0 (120)	88.12	88.08
19	Center	0 (1.2)	0 (8)	0 (120)	88.14	88.08
20	Center	0 (1.2)	0 (8)	0 (120)	87.61	88.08

---

Based on the results, the final regression equation in terms of their coded obtained from central composite design under RSM is expressed by the following second-order polynomial equation.

CaF₂ content (%) = 88.08 + 0.46A + 0.65B − 0.13C − 0.63AB + 0.082AC − 0.34BC − 0.97A² − 1.43B² − 0.85C² (3)

where A ([Ca]/[F]), B (pH) and C (C_sulfate) in all equations are in coded units.

The ANOVA results for response parameters and response value under optimum conditions are shown in Table 5. The P-values were used to estimate whether F-values was large enough to indicate statistical significant and used to check the significance of each coefficient. The model P-values was lower than 0.05, indicated that the model and model terms were significant.²³ It could conclude that quadratic model can well navigate the design space, thus the quadratic model was significant for the optimization of preparation condition. All the factors (P-values < 0.05) were significant at the 95% confidence level except for C_sulfate (λ₃), interaction of [Ca]/[F]-C_sulfate (λ₁λ₃) and pH-C_sulfate (λ₂λ₃),^24,25 which was contributed to the counteraction of independent variables on response surface. Thus it is necessary to control the variables to offset the contradiction and improve recovery efficiency of CaF₂.

Table 5 ANOVA for analysis of variance and adequacy of the quadratic model^a

Source	Sum of squares	df	Mean square	F	P (P > F)
a SD = 0.67, PRESS = 35.35, R^2 = 0.9482, R^2adj = 0.9016, Adeq Precision = 14.785.
Model	82.53	9	9.17	20.29	<0.0001
λ1	3.31	1	3.31	7.33	0.0220
λ2	6.71	1	6.71	14.84	0.0032
λ3	0.28	1	0.28	0.63	0.4473
λ1λ2	3.16	1	3.16	6.98	0.0246
λ1λ3	0.054	1	0.054	0.12	0.7359
λ2λ3	0.94	1	0.94	2.07	0.1803
λ1^2	23.86	1	23.86	52.78	<0.0001
λ2^2	51.49	1	51.49	113.9	<0.0001
λ3^2	18.01	1	18.01	39.84	<0.0001
Residual	4.52	10	0.45
Lack of fit	4.17	5	0.83	12.08	0.0081
Pure error	0.35	5	0.069
CorTotal	87.05	19

---

The high R² coefficient of 0.9482 ensures a high level of consistency between the observed and calculated values.¹⁶ The predicted values of the model response related with the actual values was presented in Fig. 5d. The data points are distributed relatively close and have linear behavior.

Fig. 5 Combined effect of influencing variables on CaF₂ content: (a) [Ca]/[F] × pH, (b) pH × C_sulfate, (c) [Ca]/[F] × C_sulfate, and (d) predicted vs. actual values plot of CaF₂ content.

Interactive effect of variables. To estimate the CaF₂ content over independent variables ([Ca]/[F] (calcium chloride dosage), pH and C_sulfate), the relevant 3D response surface plots were presented in Fig. 5a–c. It demonstrated that high qualified CaF₂ preparation was related to all the desired factors of [Ca]/[F], pH and C_sulfate directly. Clearly, the best CaF₂ content stabilized (>88%) was under the test conditions of [Ca]/[F] of 1.25–1.35, pH of 8.0–8.6 and C_sulfate of 100–130 mg L⁻¹.

Fig. 5a and c indicated that CaF₂ content increased with the increasing of Ca/F molar ratio. Nevertheless, the CaF₂ weight was decreased if calcium additive is higher than the expressed Ca²⁺ value ([Ca]/[F] = 1.35), since most of calcium salt is dissolution, and the competition between heterogeneous crystallization and homogeneous nucleation will carry out in the solutions.^16,26,27 Thus calcium ions excess will increase the supersaturation index, which will resulted in low crystallization efficiency and higher turbidity in the effluent.²

As the Fig. 5b and c shown, according to the results mentioned earlier, it was necessary to reduce sulfate content in the high fluoride-containing wastewater. In contrast, it also demonstrated that the reasonable sulfate content in wastewater can act as the seed crystal for the CaF₂ settlement, and the co-precipitation process might occur in the solution, which will be beneficial to settlement of fine CaF₂. Calcium sulfate was slightly soluble reagent. The growth of CaSO₄–CaF₂ takes place by molecular growth and aggregation between calcium sulfate and calcium fluoride, which belongs to nucleated precipitation that compete with discrete precipitation (primary and secondary nucleation) and mineral layer abrasion in the liquid phase.^27,28 It could be concluded that forming the CaSO₄ were benefit of separation of CaF₂.

As illustrated in Fig. 5a and b, CaF₂ content (>88%) was determined by the pH in solution ranging from 8.0 to 8.6. Above all, fluorine compounds mainly existed as HF⁻, HF₂⁻ at lower pH of solution (pH < 6).²⁹ On the other hand, excessive hydroxide ion of solution (pH > 12) can form large number of calcium hydroxide precipitated with other settlements, decreasing the separated CaF₂ quality. Apparently, the calcium fluoride precipitated within a broad pH (6–12). However, optimal pH value (8.14) is caused by the other cations and anions contained in the liquid system.³⁰​ Control of suspension solution pH (8.0–8.6) can be used to prevent co-precipitation of other components and optimize the operation of CaF₂ preparation process.

In Fig. 5, the highest qualified CaF₂ recovered by pretreatment and reaction with calcium chloride was 88.41%. The highest points of CaF₂ content were observed, at [Ca]/[F] of 1.29, pH of 8.14 and the sulfate concentration of 117.4 mg L⁻¹.

Optimization and validation process. In order to confirm the feasibility of the pretreatment and reliability of the model, an additional laboratory experiment was conducted using the optimum conditions obtained from CCD-RSM. The content of CaF₂ prepared by high concentration fluoride-containing wastewater and calcium solution in the optimum condition reaches to 88.2% and CaSO₄ content was kept to 4.2%.

The bulk mineralogical composition of practical synthesized sediments consisted of fluorite, which was in agreement with the predicted content value calculated by the obtained model.³¹ The small variation in the result between the simulated values and laboratory experiment confirms that CCD-RSM in the software a useful tool to obtain the excellent conditions of precipitation and separation of CaF₂ fine particles at a suitable [Ca]/[F], pH and C_sulfate.

CaF₂ property under the optimum conditions

For the CaF₂ suspension mixed by fluoride-containing wastewater and calcium salt under the optimum conditions, the zeta potential 16.97 mV and final turbidity 82.4 NTU, and the fluorite average diameter of CaF₂ particles recovered from etching solution was kept to 907.5 nm (>758.45 nm). It revealed that the residual sulfate contribute to improve the CaF₂ separation from the effluent.

Fig. 6c shows the particle size frequency distributions. The degree of aggregation of the precipitated particles correlated with the residual sulfate concentration. As shown from the TEM images in Fig. 6b, it could be seen that the CaF₂ pellets is consisted of several central particles and many spherical aggregates.¹⁰ These findings were in agreement with the theoretical models of precipitation by remaining a suitable concentration of sulfate. However, as the XRD pattern shown in Fig. 6a, the crystal phase of CaF₂ pellets is influenced by residual sulfate impurities comparison with commercial fluorite.

Fig. 6 (a) XRD for high qualified CaF₂ synthesized at optimum condition: (i) synthesized CaF₂; (ii) commercial grade fluorite, (b) TEM images and (c) particle size distribution cures of CaF₂ synthesized on optimum conditions ([Ca]/[F] = 1.28, pH = 8.14, C_sulfate = 117.4 mg L⁻¹).

Conclusions

Effect on the sulfate concentration were tested for the CaF₂ preparation by the gradient precipitation of barium salts precipitation and CaCl₂. CaSO₄–CaF₂ system was beneficial of fine particles settlement. Then, the pH, Ca/F molar ratio, SO₄²⁻ range were optimized using the CCD methods, and the he optimum range were shown as the ratio of 1.25–1.35, pH of 8.0–8.6, and C_sulfate was kept at 100–130 mg L⁻¹. The agreement of the quadratic model with the experimental data was satisfactory, which obtained from the RSM method. Under this conditions, the CaF₂ purity could reach to 88.2%, meet the standards of ceramic grad, which could be used for as the raw materials for the production of hydrofluoric acid. This methods provide a practical methods for the F resource recovery from the fluoride-containing wastewater, and reduce the hazardous materials of CaF₂ sludge.

Conflict of interest

No potential conflicts of interest relevant to this article were reported.

Acknowledgements

This work was financially supported by National Key Technology R&D Program (No. 2014BAL02B03-4), National Natural Science Foundation of China (No. 41173108) (No. 51278350), Shanghai Rising-Star Program (14QA1402400), and Key project of Science and Technology Commission of Shanghai Municipality (No. 13DZ0511600).

Notes and references

1. USGS, Mineral commodity summaries, 2013, http://minerals.usgs.gov/minerals/pubs/mcs/2013/mcs2013.pdf.
2. K. Van den Broeck, N. Van Hoornick, J. Van Hoeymissen, R. de Boer, A. Giesen and D. Wilms, IEEE Trans. Semicond. Manuf., 2003, 16, 423–428.
3. M. F. Chang and J. C. Liu, J. Environ. Eng., 2007, 133, 419–425.
4. P. Melidis, Environ. Processes, 2015, 2(1), 205–213.
5. MEPC, National Hazardous Waste List, 2015, http://law.pharmnet.com.cn/upload/0807041600204387.pdf.
6. A. Ezzeddine, A. Bedoui, A. Hannachi and N. Bensalah, Desalin. Water Treat., 2015, 54, 2280–2292.
7. M. F. Chang and J. C. Liu, J. Environ. Eng., 2007, 133, 419–425.
8. R. Aldaco, A. Irabien and P. Luis, Chem. Eng. J., 2005, 107, 113–117.
9. R. Aldaco, A. Garea and A. Irabien, Water Res., 2007, 41, 810–818.
10. J. Chen, C. Lin, P. Lin, C. Li, Y. Liang, J. Liu and S. Chen, Sep. Purif. Technol., 2014, 137, 53–58.
11. E. Demirbas, N. Dizge, M. T. Sulak and M. Kobya, Chem. Eng. J., 2009, 148, 480–487.
12. S. A. Dastgheib and D. A. Rockstraw, Carbon, 2002, 40, 1853–1861.
13. S. M. Hamza, J. Cryst. Growth, 1991, 113, 637–642.
14. M. O. Saeed, K. Azizli, M. H. Isa and M. J. K. Bashir, J. Water Process Eng., 2015, 8, 7–16.
15. P. Kundu, V. Paul, V. Kumar and I. M. Mishra, Chem. Eng. Res. Des., 2015, 104, 773–790.
16. S. Zhou and Y. Wu, Environ. Sci. Pollut. Res., 2012, 19, 347–360.
17. S. L. C. Ferreira, R. E. Bruns, H. S. Ferreira, G. D. Matos, J. M. David, G. C. Brandão, E. G. P. Da Silva, L. A. Portugal, P. S. Dos Reis, A. S. Souza and W. N. L. Dos Santos, Anal. Chim. Acta, 2007, 597, 179–186.
18. D. A. D. A. Aljuboury, P. Palaniandy, H. B. A. Aziz and S. Feroz, J. Environ. Chem. Eng., 2015, 3, 1117–1124.
19. M. A. Bezerra, R. E. Santelli, E. P. Oliveira, L. S. Villar and L. A. Escaleira, Talanta, 2008, 76, 965–977.
20. D. Bas and I. H. Boyaci, J. Food Eng., 2007, 78, 836–845.
21. GB/T 5195.1-2006, Standard for fluorspar-determination of calcium fluoride content, State Environmental Protection Administration, China, 2006.
22. J. P. Chen and H. Yu, J. Environ. Sci. Health, Part A: Toxic/Hazard. Subst. Environ. Eng., 2000, 35, 817–835.
23. P. A. Soloman, C. Ahmed Basha, M. Velan, N. Balasubramanian and P. Marimuthu, Sep. Purif. Technol., 2008, 69, 109–117.
24. M. J. K. Bashir, H. A. Aziz, M. S. Yusoff and M. N. Adlan, Desalination, 2010, 254, 154–161.
25. M. S. Bhatti, D. Kapoor, R. K. Kalia, A. S. Reddy and A. K. Thukral, Desalination, 2011, 274, 74–80.
26. M. Yang, Y. Zhang, B. Shao, R. Qi and H. Myoga, J. Environ. Eng., 2001, 127, 902–907.
27. R. Aldaco, A. Garea and A. Irabien, Chem. Eng. J., 2007, 134, 66–71.
28. A. Garea, R. Aldaco and A. Irabien, Chem. Eng. J., 2009, 154, 231–235.
29. C. Y. Tai, P. C. Chen and T. M. Tsao, J. Cryst. Growth, 2006, 290, 576–584.
30. J. W. Morse and R. S. Arvidson, Earth-Sci. Rev., 2002, 58, 51–84.
31. N. Chong and T. Quang-Minh, Desalin. Water Treat., 2015, 54, 3422–3431.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra15540e

---