An Fe-Cu binary oxide was fabricated through a simple co-precipitation process, and was used to remove Sb(III) from aqueous solution. X-ray diffraction, scanning electron microscopy, energy dispersive X-ray and N2 adsorption–desorption measurements demonstrated that the Fe-Cu binary oxide consisted of poorly ordered ferrihydrite and CuO, and its specific surface area was higher than both iron oxide and copper oxide. A comparative test indicated that Fe/Cu molar ratio of prepared binary oxide greatly influenced Sb(III) removal and the optimum Fe/Cu molar ratio was about 3/1. Moreover, a maximum adsorption capacity of 209.23 mg Sb(III)/g Fe-Cu binary oxide at pH 5.0 was obtained. The removal of Sb(III) by Fe-Cu binary oxide followed the Freundlich adsorption isotherm and the pseudo-second-order kinetics in the batch study. The removal of Sb(III) was not sensitive to solution pH. In addition, the release of Fe and Cu ions to water was very low when the pH was greater than 6.0. X-ray photoelectron spectroscopy analysis confirmed that the Sb(III) adsorbed on the surface was not oxidized to Sb(V).

INTRODUCTION

Antimony (Sb) is widely used in industrial products including flame retardants, additives in glassware, catalysts for plastics synthesis, and alloys for ammunition. Due to its toxicity and potential carcinogenicity, antimony and its compounds are included as priority pollutants by the European Union (Council of the European Communities 1976). To guard against health problems associated with Sb, the Standardization Administration of the People's Republic of China (2007) requires that the maximum level of antimony is less than 5 μg/L for drinking water. Antimony generally exists in two oxidation states (III and V) in the environment. The oxidation state Sb(V) is stable under aerobic conditions and Sb(III) persists in anoxic media. Moreover, Sb(III) has a toxicity 10 times higher than Sb(V) (Filella et al. 2007).

Recently, serious Sb pollution has been mostly caused by antimony mining and smelting industries. In the Stampede and Slate Creek watersheds, Sb concentration in stream waters was as high as 720 μg/L, which was associated with historic mining activities within the Kantishna Hills mining district (Ritchie et al. 2013). In particular in China, which has the most abundant antimony reserves in the world, pollution of water resources has become a major environmental quality issue. In 2007 alone, more than 9.45 × 105 tons of wastewater and 6.72 × 105 tons of processing residue have been produced. Dissolved antimony in the drainage from the Xikuangshan Sb2S3 deposit in Hunan Province can be as high as 29.4 mg/L (Zhu et al. 2009). However, it is only treated by precipitation or piped directly into surface water bodies. The high volumes of industrial discharges to water from the mining and smelting activities have resulted in extensive contamination of the local hydrological system and it is thought that significant amounts of Sb(III) persist in deeper waters. Therefore, developing cost-effective techniques for Sb(III) removal from aqueous solution is critical.

Major treatment techniques for antimony-contaminated water include membrane filtration (Kang et al. 2000), coagulation/flocculation (Guo et al. 2009), and adsorption (Ungureanu et al. 2015). Among these methods, adsorption is one of the most effective choices for the removal of metal ions from water because of its low cost, simplicity, rapid response and high efficiency. Previous work has demonstrated that iron oxides or hydroxides (Guo et al. 2014; Shan et al. 2014; Ungureanu et al. 2015) have higher affinity to Sb(III) than many other adsorbents. Recently, composites have been developed with more than one metal and have been found to be more effective sorbents. For instance, an Fe-Mn binary oxide exhibited a much higher Sb(III) removal capacity than both FeOOH and MnO2 (Xu et al. 2011). In addition, a Zn/Fe layered double hydroxide (Lu et al. 2015) adsorbent showed preferential removal for Sb(V). However, studies of the adsorption capacity for Sb(III) by bimetal composites are limited to the Fe-Mn binary oxide.

As a relatively cheap metal oxide, copper oxide is widely used in gas sensors and solar cells and as a catalyst. Recently, it has been reported that cupric oxide is an effective sorbent for arsenic removal over a wide pH range (Martinson & Reddy 2009). Furthermore, it has been found that an Fe-Cu binary oxide has promising sorption capacity for As(III) (Zhang et al. 2013). To the best of our knowledge, despite similar chemical properties of antimony compared to arsenic, copper-based oxide has been poorly studied in the removal of Sb(III) from aqueous solution.

This study represents the first attempt to assess the removal of Sb(III) from aqueous solution by iron-based bimetal composites. The main objectives were to: (i) prepare an Fe-Cu binary oxide by a co-precipitation technique and characterize its structural features; (ii) investigate the removal of Sb(III) by the Fe-Cu binary oxides under experimental batch tests and to optimize the Fe/Cu molar ratio in the solid phase; and (iii) assess the dominant Sb(III) removal mechanism of the Fe-Cu binary oxide.

EXPERIMENTAL PROCEDURES

Preparation and characterization of Fe-Cu binary oxides

Fe-Cu binary oxides with different Fe/Cu molar ratio were prepared by a modified co-precipitation method at laboratory temperature (23 °C). The preparation of the Fe-Cu binary oxide with an Fe/Cu molar ratio of 2/1 was as follows. First, a mixed solution containing 2.35 g FeCl3·6H2O and 0.6 g CuSO4 was prepared. Then 1.25 mol/L NaOH was added drop-wise into the solution mixture with magnetic stirring. Solution pH was kept at about 9.5 during the reaction process. After continuous stirring for 2 h, the resulting slurry was centrifuged and washed with deionized water and dried at 80 °C for 24 h. The dry solid material was ground into fine powder with a mortar and pestle and then used for Sb(III) removal. Fe-Cu binary oxides with an Fe/Cu molar ratio of 1/2, 1/1 and 3/1 were also prepared by altering the amount of FeCl3 or CuSO4 added in the first mixing step. Single copper oxide and iron oxide solids were synthesized using the same method without adding CuSO4 or FeCl3.

Particle morphology and crystallinity were characterized using scanning electron microscopy (SEM, JSM-6380LV, JEOL, Japan) and X-ray diffraction (XRD, D8 Advance, Bruker, Germany). Semi-quantitative composition of the surface was examined by means of energy-dispersive X-ray spectroscopy (EDX, JSM-6490, JEOL, Japan). Nitrogen adsorption/desorption isotherms were obtained with a NOVA 2200e surface area and pore size analyzer (Quantachrome Instruments, USA) after degassing the samples at 150 °C for 12 h. The specific surface area was calculated using the Brunauer–Emmett–Teller (BET) model and the pore size distribution was estimated from the Barrett–Joyner–Halenda (BJH) method. The chemical state of elements on the surfaces was assessed using X-ray photoelectron spectroscopy (XPS, Kratos Amicus, UK) using monochromatized MgKα X-ray source working at 180 W and 12 kV. The zeta potential of samples at different pH was determined using a JS94H micro-electrophoresis apparatus, which was manufactured by the Digital Technology Equipment of Zhongchen Co., Ltd, Shanghai, China.

Batch experiments

A stock solution of 1,000 mg/L Sb(III) was freshly prepared by dissolving 2.738 g of antimony potassium tartrate (K(SbO)C4H4O6·½H2O) in 100 mL deionized water. A simulated wastewater with required concentrations was obtained by diluting the Sb(III) stock solution with deionized water, and the Sb(III) solution was freshly prepared as it was used.

The experiments were carried out in 125 mL sealed polypropylene bottles at room temperature (23 ± 1 °C). All experiments were performed in duplicate. Comparison batch tests for Sb(III) removal were conducted using iron oxide, copper oxide and Fe-Cu binary oxides with different Fe/Cu mole ratio. In each test, 0.03 g of iron oxide, copper oxide or Fe-Cu binary oxide was added, respectively, to 100 mL of 40 mg/L Sb(III) solution after the initial pH was adjusted to 5.0 ± 0.1 with 1.0 mol/L HCl or 1.25 mol/L NaOH. These bottles were then placed on a rotary shaker at 100 rev/min. After a 24 h reaction time, 3 mL aliquots were taken from the suspension, filtered through 0.45 μm membrane filters and analyzed for the Sb remaining in water.

Kinetic experiments were carried out by taking 0.01, 0.02, 0.03 and 0.04 g of Fe-Cu binary oxide with 100 mL of 40 mg/L Sb(III) solution. The initial pH of Sb(III) solution was 5.0 ± 0.1. Samples were taken at the following intervals: 0.5, 1, 2, 4, 6, 10, 12, and 24 h after start of the reaction, and then filtered to determine the residual Sb concentration in solution.

Sorption isotherm experiments were conducted by taking 0.03 g of Fe-Cu binary oxide with 100 mL of Sb(III) solution varied from 5 to 90 mg/L, respectively. The initial pH of Sb(III) solution was 5.0 ± 0.1. After shaking at 100 rev/min for 48 h, the residual Sb concentration in water was analyzed as mentioned before.

To test the effect of solution pH on the removal ability of Fe-Cu binary oxide, the initial pH of Sb(III) solution was pre-adjusted to a desired level from 2.0 ± 0.1 to 12 ± 0.1. Then, 0.03 g of Fe-Cu binary oxide was added into a vessel containing 100 mL of 40 mg/L Sb(III) solution, agitated at 100 rev/min for 24 h. Then the residual Sb concentration in the solution was determined.

Analytical methods

The pH values of solution were measured using a basic PB-10 meter (Sartorius, Germany). The total Sb and Cu concentrations in solutions were detected by a flame atomic absorption spectrophotometer (AA-7001, East & West Analytical Instruments Inc., Beijing). The detection limit of this method was 0.2 mg/L, and the analytical regression coefficient R2 was greater than 0.997. Total Fe was determined using the 1, 10 phenanthroline spectrometric method (APHA-AWWA-WEF 1998).

RESULTS AND DISCUSSION

Characterization of iron oxide and copper oxide

The XRD patterns of the synthesized iron oxide and copper oxide are shown in Figure S1(a) and S1(b) (available with the online version of this paper). The characteristic peaks of iron oxide approximately at 35°, 55° and 62.5° were features of poorly ordered and amorphous ferrihydrite (Fe2O3·2FeOOH·2.6H2O) (Jambor & Dutrizac 1998; Zhang et al. 2013). The crystalline peaks in copper oxide which appeared at 36.5°, 39°, 49.5° and 62° were in good agreement with those of the standard patterns of CuO (Liu et al. 2015).

N2 adsorption–desorption isotherms and BJH pore size distribution (inset) of the iron oxide and copper oxide are shown in Figure S2(a) and S2(b) (availabe online). The BET surface area and average pore diameter of iron oxide were calculated to be 82.39 m2/g and 3.893 nm, respectively. The BET surface area and average pore diameter of copper oxide were determined to be 28.13 m2/g and 3.796 nm, respectively.

Characterization of Fe-Cu binary oxide

Figure 1 shows XRD patterns of the prepared binary oxide with theoretical Fe/Cu molar ratios of 1/2, 1/1, 2/1 and 3/1. It can been seen that the Fe-Cu binary oxide components were complex; both ferrihydrite characteristic peaks at 35°, 55° and 62.5°, and CuO characteristic peaks at 36.5°, 39°, 49.5° and 62° occurred. Moreover, Fe-Cu binary oxides existed mainly in amorphous form.
Figure 1

XRD patterns of prepared binary oxides with Fe/Cu molar ratio of (a) 1/2, (b) 1/1, (c) 2/1 and (d) 3/1.

Figure 1

XRD patterns of prepared binary oxides with Fe/Cu molar ratio of (a) 1/2, (b) 1/1, (c) 2/1 and (d) 3/1.

Figure 2 illustrates the SEM images of Fe-Cu binary oxide with different Fe/Cu molar ratio. It was shown that most of the binary oxides were not of uniform size: certain particles reached the nanometre grade, and some were greater than 10 μm. Moreover, they were composed of many aggregated small particles, which resulted in a rough surface. In addition, the structure of binary oxide with Fe/Cu molar ratio of 1/2 (Figure 2(a)) was relatively flaky and loose; however, with the increase of iron oxide constituents, the binary oxide with Fe/Cu molar ratio of 3/1 (Figure 2(d)) was typically massive and compact.
Figure 2

SEM images of Fe-Cu binary oxide with Fe/Cu molar ratio of (a) 1/2, (b) 1/1, (c) 2/1 and (d) 3/1.

Figure 2

SEM images of Fe-Cu binary oxide with Fe/Cu molar ratio of (a) 1/2, (b) 1/1, (c) 2/1 and (d) 3/1.

The EDX spectrum of samples was collected and the result is shown in Figure S3 (available online). It confirmed that Fe, Cu, and O were the main elemental components of the products, agreeing with the XRD results. As there is oxygen in the atmosphere, the obtained value for O was inexact. Therefore, the average atomic ratios for Fe and Cu are summarized in Table S1 (available with the online version of this paper). It can be seen that experimental values of Fe/Cu molar ratio agreed well with the theoretical value, indicating that the preparation method for the binary oxides was reliable.

Figure 3 illustrates N2 adsorption–desorption isotherms and pore size distribution analysis of the binary oxides with Fe/Cu molar ratio of 1/1 and 3/1. The isotherms of the two samples exhibited the characteristics of type IV isotherms with an H3 type hysteresis loop, which was typically observed for materials with slit-shaped mesopores of packing particles (Passe-Coutrin et al. 2008). The binary oxide with Fe/Cu molar ratio of 1/1 had a BET surface area of 119.17 m2/g, and the average pore diameter was about 3.397 nm. The BET surface area and pore diameter of binary oxide with the Fe/Cu molar ratio of 3/1 was 238.79 m2/g and 2.165 nm. Obviously, the surface area of Fe-Cu binary oxide with Fe/Cu molar ratio of 3/1 increased significantly compared to binary oxide with the Fe/Cu molar ratio of 1/1 iron oxide and copper oxide.
Figure 3

N2 adsorption–desorption isotherm and pore size distribution curve of (a) Fe-Cu binary oxide with the Fe/Cu molar ratio of 1/1 and (b) Fe-Cu binary oxide with the Fe/Cu molar ratio of 3/1.

Figure 3

N2 adsorption–desorption isotherm and pore size distribution curve of (a) Fe-Cu binary oxide with the Fe/Cu molar ratio of 1/1 and (b) Fe-Cu binary oxide with the Fe/Cu molar ratio of 3/1.

In addition, the analysis of zeta potential in Figure S4 (available online) indicated that the point of zero charge of Fe-Cu binary oxide in the presence of 0.01 M NaCl was about pH 7.7.

Comparison of Sb(III) removal by metal and bimetal oxides

After a 24 h reaction time, 56.29% and 22.5% of a 40 mg/L Sb(III) solution was removed by 0.3 g/L of iron oxide and pure copper oxide. Under the same reaction conditions, the Sb(III) removal rate for binary oxides with an Fe/Cu molar ratio of 3/1, 2/1, 1/1, 1/2 was 81.30%, 76.42%, 67.32% and 64.85% correspondingly. It was obvious that with the incorporation of copper into iron oxide, the removal of Sb(III) by Fe-Cu binary oxide increased. Moreover, the adsorption of Sb(III) on binary oxide with the Fe/Cu molar ratio of 3/1 reached a maximum and then dropped as the copper oxide constituent continued to increase. The BET surface area of the iron oxide, copper oxide, and binary oxide with Fe/Cu of 1/1 and 3/1 was 82.39, 28.13, 119.17, and 238.79 m2/g, respectively. So, the high surface area for Fe-Cu binary oxide might be an important factor in the improved Sb(III) sorption capacity. As the adsorption ability of pure copper oxide was lower than for iron oxide, excess copper oxide in the composites is not ideal. The optimum Fe/Cu molar ratio was about 3/1, which was then used in the subsequent experiments.

Kinetics of Sb(III) adsorption on Fe-Cu binary oxide surface

The kinetic experiments were conducted by equilibrating a 100 mL aliquot of 40 mg/L Sb(III) solution with a series of binary oxide samples (Fe/Cu molar ratio of 3/1) over different time intervals. Figure 4 shows the amount of Sb(III) adsorbed onto Fe-Cu binary oxide with contact time. It was observed that the removal capacity of Sb(III) increased rapidly before 4 h contact time, and then slowed down as equilibrium was approached at about 24 h. The high initial removal rate may be attributed to the existence of a large number of adsorption sites on the surface of fine particles. As the sites filled up gradually, the intra-particle diffusion process dominated in the adsorption of Sb, and the removal became slow.
Figure 4

Effect of the contact time on adsorption of Sb(III) onto the Fe-Cu binary oxide. Initial Sb(III) concentration: 40 mg/L, pH: 5.0 ± 0.1, 23 ± 1 °C.

Figure 4

Effect of the contact time on adsorption of Sb(III) onto the Fe-Cu binary oxide. Initial Sb(III) concentration: 40 mg/L, pH: 5.0 ± 0.1, 23 ± 1 °C.

Pseudo-first-order kinetic and pseudo-second-order kinetic model equations have long been widely applied in the sorption system. The linear forms of the pseudo-first-order and pseudo-second-order equations are expressed as Equations (1) and (2): 
formula
1
 
formula
2
where qe (mg/g) and qt (mg/g) were the amount of adsorbed Sb(III) at equilibrium and at any time t (h), and k1 (1/h) and k2 (g/(mg·h)) are the rate constant for pseudo-first-order and pseudo-second-order sorption, respectively. The fitting of the pseudo-first-order and pseudo-second-order kinetic model for the adsorption of Sb(III) by Fe-Cu binary oxide is shown in Figure 4. The corresponding parameters are included in Table S2 (available online). The R2 for the pseudo-first-order model indicated that the model was a poor fit to the kinetic data. However, the R2 values for pseudo-second-order kinetic model were in the range of 0.9729–0.9934, and the calculated qe values were in excellent agreement with the experimental qe values. So it can be concluded that the pseudo-second-order kinetic model was able to describe the adsorption of Sb(III) onto Fe-Cu binary oxide, and the adsorption process was mainly controlled by the chemisorption process (Ho & McKay 2000). It is known that there are substantial numbers of chemically active Fe-OOH sites on the ferrihydrite surface (Jambor & Dutrizac 1998), and Xu et al. (2011) had also reported that Fe-OOH groups were able to interact with Sb.

Adsorption isotherms

To describe the equilibrium adsorption behavior of Sb(III), Langmuir and Freundlich isotherm models were used to analyze the equilibrium data. The Langmuir sorption isotherm is valid for monolayer adsorption onto a surface containing a finite number of identical sites. The Freundlich isotherm model allows for several kinds of sorption sites on the solid and properly represents the sorption data at low and intermediate concentrations on heterogeneous surfaces. The equations of Langmuir and Freundlich isotherm models can be expressed by Equations (3) and (4): 
formula
3
 
formula
4
where qe is Sb amount adsorbed onto Fe-Cu binary oxide at equilibrium (mg/g); Ce is equilibrium concentration of Sb in the aqueous solution (mg/L); qm is the maximum sorption capacity (mg/g); KL is the Langmuir sorption constant (L/mg) which is related to the energy of adsorption. Kf (mg(1–1/n)L1/n/g) and n are the Freundlich constants, which are indicators of adsorption capacity and adsorption intensity.
Figure 5 highlights the experimental data as well as the isotherm models for the adsorption of Sb(III) on Fe-Cu binary oxide. The calculated parameters of Langmuir and Freundlich isotherm models are presented in Table S3 (available online). The higher correlation coefficients (R2) indicated that the Freundlich model gave a better fit to the adsorption data than the Langmuir model over the entire concentration range studied. The fractional value of 1/n (0 < 1/n < 1) obtained clearly corresponded to a heterogeneous surface. It was because the simultaneous presence of ferrihydrite and CuO in the binary oxide led to a heterogeneous surface of the Fe-Cu binary oxide with different sorption energies.
Figure 5

Variation of adsorbed Sb on the Fe-Cu binary oxide at equilibrium. Initial Sb(III) concentration: 5–90 mg/L, sorbent dose: 0.3 g/L, pH: 5.0 ± 0.1, 23 ± 1 °C.

Figure 5

Variation of adsorbed Sb on the Fe-Cu binary oxide at equilibrium. Initial Sb(III) concentration: 5–90 mg/L, sorbent dose: 0.3 g/L, pH: 5.0 ± 0.1, 23 ± 1 °C.

As indicated by the direct experimental data in Figure 5, the maximum removal capacity of Fe-Cu binary oxide for Sb(III) was 209.23 mg/g at the initial Sb(III) concentration of 90 mg/L. A comparison of the maximum adsorption capacities of various adsorbents for Sb(III) is shown in Table 1. It was found that Fe-Cu binary oxide in this study exhibited higher adsorption capacity than many other adsorbents, highlighting its potential as an efficient adsorbent for Sb(III) removal from aqueous solutions.

Table 1

Comparison of adsorption capacity for Sb(III) from aqueous solution reported in the literature

AdsorbentConcentration range (mg/L)pHQmax (mg/g)T (K)References
Goethite (α-FeOOH) 1.13–33.83 9.0 53.45 293 Guo et al. (2014)  
Green bean husk 2.5–100 4.0 20.14 298 Iqbal et al. (2013)  
Nano-zerovalent iron stabilized by polyvinyl alcohol 0–20 7.0 6.99 298 Zhao et al. (2014)  
Carbon nanofibers decorated with ZrO2 10–500 7.0 79.83 298 Luo et al. (2015)  
Fe-Mn binary oxide 23–225 3.0 214.28 293 Xu et al. (2011)  
Fe(III)-loaded saponified orange – 10.5 136 303 Biswas et al. (2009)  
α-Fe2O3-coated Fe3O4 1–20 4.1 36.7 298 Shan et al. (2014)  
Mercapto-functionalized hybrid sorbent 100–300 5.0 108.8 298 Fan et al. (2016)  
Fe-Cu binary oxide 5–85 5.0 209.23 296 This study 
AdsorbentConcentration range (mg/L)pHQmax (mg/g)T (K)References
Goethite (α-FeOOH) 1.13–33.83 9.0 53.45 293 Guo et al. (2014)  
Green bean husk 2.5–100 4.0 20.14 298 Iqbal et al. (2013)  
Nano-zerovalent iron stabilized by polyvinyl alcohol 0–20 7.0 6.99 298 Zhao et al. (2014)  
Carbon nanofibers decorated with ZrO2 10–500 7.0 79.83 298 Luo et al. (2015)  
Fe-Mn binary oxide 23–225 3.0 214.28 293 Xu et al. (2011)  
Fe(III)-loaded saponified orange – 10.5 136 303 Biswas et al. (2009)  
α-Fe2O3-coated Fe3O4 1–20 4.1 36.7 298 Shan et al. (2014)  
Mercapto-functionalized hybrid sorbent 100–300 5.0 108.8 298 Fan et al. (2016)  
Fe-Cu binary oxide 5–85 5.0 209.23 296 This study 

Effect of the solution pH on Sb(III) adsorption and metal leaching of Fe-Cu binary oxide

The effect of pH on Sb(III) adsorption on the Fe-Cu binary oxide is shown in Figure 6(a). It can be seen that 87.96% and 81.11% of Sb(III) was removed by 0.03 g Fe-Cu binary oxide when the initial solution pH was 2.0 and 4.0, separately. The Sb(III) removal rate was kept to about 85% in the pH range 6.0–10.0. However, Sb(III) adsorption suddenly decreased by 25.6% when pH increased from 10.0 to 12.0. Moreover, in all experiments solution pH rose by 0.2–0.5 after 24 h reaction. It was known that solution pH was one of the more important factors affecting the adsorption process. In the present study, Sb(III) existed as (SbO)C4H4O6 coordination anions. The positive charge of the Fe-Cu binary oxide surface gradually dropped with increase of pH, as illustrated in Figure S4, and the electrostatic attraction between the negatively charged antimony tartrate ion and the more negatively charged surface of Fe-Cu binary oxide became weaker, which would result in the decrease of Sb(III) adsorption. However, this was not reflected by the drop in Sb(III) removal until solution pH increased to 12.0. This phenomenon suggested that the pH-dependent electrostatic attraction did not control the adsorption process onto Fe-Cu binary oxide. Adsorption was likely to be based on chemical adsorption, which was also supported by the kinetics analysis.
Figure 6

(a) Effect of solution pH on Sb(III) removal, (b) release of Fe and Cu to solution after reaction. Initial Sb(III) concentration: 40 mg/L, sorbent dose: 0.3 g/L, pH: 5.0 ± 0.1, 23 ± 1 °C.

Figure 6

(a) Effect of solution pH on Sb(III) removal, (b) release of Fe and Cu to solution after reaction. Initial Sb(III) concentration: 40 mg/L, sorbent dose: 0.3 g/L, pH: 5.0 ± 0.1, 23 ± 1 °C.

Figure 6(b) demonstrates the concentrations of dissolved Fe and Cu in water after reaction at different pH. It was shown that the leaching of Fe and Cu was significant under acid conditions. When the pH was greater than 6.0, the Fe and Cu concentration was below 0.3 mg/L and 1 mg/L, respectively, which is the limit of the drinking water standard of China (Standardization Administration of the People's Republic of China 2007). Therefore, the prepared Fe-Cu binary oxide can be used to safely treat real antimony mine drainage waters, most of which are neutral or weakly alkaline in China (Zhu et al. 2009).

Further discussion of removal mechanism

After reaction, the Fe-Cu binary oxide was recovered from the reaction solution; then SEM and EDX were employed to characterize the morphological properties and elemental composition. The results are shown in Figure S5 (available online). Compared with the freshly prepared Fe-Cu binary oxide, which was sharp-edged block (Figure 2(d)), the morphology of the Fe-Cu binary oxide surface was markedly changed after reaction, and some flocs appeared on the surface of the material (Figure S5(a)). The EDX spectra (Figure S5(b)) revealed that Fe and Cu were the primary elements, but the Fe/Cu molar ratio was changed to 3.26, resulting from the leaching of binary oxide. In addition, 15.85 weight% Sb was observed on the Fe-Cu binary oxide surface after reaction.

To further obtain the oxidation state of elements on the surface, XPS spectra of the Fe-Cu binary oxide after reaction were collected and analyzed. As shown in Figure 7(a), Sb 3d peaks as well as Fe 2p, Cu 2p and O 1d peaks appeared, which proved once again that the reaction product of Sb precipitated on the surface of Fe-Cu binary oxide. Detailed XPS surveys on the region of Fe 2p and Cu 2p are presented in Figure 7(b) and 7(c). The particles exhibited Fe 2p3/2 and 2p1/2 binding energies of about 710.95 and 724.5 eV, which were close to the binding energies of Fe(III) species (Xu et al. 2011). The photo-electron peaks for Cu 2p1/2 and Cu 2p3/2 were present at binding energy 954 eV and 933.74 eV, indicating that Cu(II) was the predominant copper species on the surface (Biesinger et al. 2010). Due to the overlapping of O 1s and Sb 3d photo-electron peaks regularly, deconvolution of the XPS spectra of Sb 3d + O 1s was performed and plotted in Figure 7(d). Although the binding energy of Sb(III) and Sb(V) in XPS spectra was reported to be too close to allow individual quantification, the Sb 3d3/2 binding energy of Sb(III)-adsorbed sample was lower than the 540.1 binding energy for Sb(V) oxide (Izquierdo et al. 1989). The peaks of Sb 3d5/2 and Sb 3d3/2 at 529.7 eV and 539.79 eV in this study can be assigned to Sb(III). Therefore, XPS measurement demonstrated that the chemical valence state of Fe(III) and Cu(II) during the sorption process did not change, and the oxidation of Sb(III) did not occur. Consequently, it is proposed that the ferrihydrite and CuO groups were the main adsorption sites available for Sb. These results are similar to those found by Xu et al. (2011) in the case of FeOOH. However, Guo et al. (2014) found that the majority of Sb(III) adsorbed onto hydrous ferric oxide was oxidized into Sb(V), probably due to the involvement of O2 in the long duration of sample preservation.
Figure 7

(a) XPS wide survey for Fe-Cu binary oxide after reaction with Sb(III). High-resolution XPS survey for (b) Fe 2p, (c) Cu 2p and (d) O 1s + Sb 3d of Fe-Cu binary oxide after reaction.

Figure 7

(a) XPS wide survey for Fe-Cu binary oxide after reaction with Sb(III). High-resolution XPS survey for (b) Fe 2p, (c) Cu 2p and (d) O 1s + Sb 3d of Fe-Cu binary oxide after reaction.

CONCLUSIONS

This study demonstrated for the first time that Sb(III) can be effectively adsorbed by Fe-Cu binary oxide from aqueous solution. Batch adsorption experiments were carried out under various conditions of initial Sb(III) concentration, material dose, contact time, and solution pH. The result indicated that the most effective removal was for binary oxide with an Fe/Cu molar ratio of 3/1. The pseudo-second-order kinetics provided the best correlation for the adsorption process, indicating the process was controlled by the chemisorption. The adsorption isotherms of Sb(III) onto Fe-Cu binary oxide could be clearly described by the Freundlich isotherm, much better than for the Langmuir isotherm. In addition the adsorption reaction for Sb(III) was not sensitive to the solution pH. The release of Fe and Cu into the aqueous phase was below the limit of the drinking water standard of China when the pH was greater than 6.0. In addition, the XPS spectra confirmed that any Sb(III) adsorbed onto the Fe-Cu binary oxide was not oxidized into Sb(V). In conclusion, the Fe-Cu binary oxide was a promising adsorbent for Sb(III) removal from contaminated water as a result of its excellent removal performance, simple synthesis and absence of secondary pollution. Continuous treatment of practical antimony mine drainage using an Fe-Cu binary oxides packed fixed bed reactor will be investigated further.

ACKNOWLEDGEMENTS

This work was supported by the National Natural Science Foundation of China (No. 51504094).

REFERENCES

REFERENCES
APHA-AWWA-WEF
1998
Standard Methods for the Examination of Water and Wastewater
, 20th edn.
American Public Health Association/American Water Works Association/Water Environment Federation
,
Washington, DC
.
Biswas
B. K.
Inoue
J. I.
Kawakita
H.
Ohto
K.
Inoue
K.
2009
Effective removal and recovery of antimony using metal-loaded saponified orange waste
.
J. Hazard. Mater
.
172
(
2–3
),
721
728
.
Council of the European Communities
1976
Council Directive 76/464/EEC of 4 May 1976 on pollution caused by certain dangerous substances discharged into the aquatic environment of the community
.
OJ L
129
,
23
29
.
Fan
H. T.
Sun
W.
Jing
B.
Wang
Q. J.
Li
D. W.
Huang
C. C.
Wang
K. J.
Zhang
Z. G.
Li
W. X.
2016
Adsorption of antimony(III) from aqueous solution by mercapto-functionalized silica-supported organic-inorganic hybrid sorbent: mechanism insights
.
Chem. Eng. J
.
286
,
128
138
.
Guo
X. J.
Wu
Z. J.
He
M. C.
Meng
X. G.
Jin
X.
Qiu
N.
Zhang
J.
2014
Adsorption of antimony onto iron oxyhydroxides: adsorption behavior and surface structure
.
J. Hazard. Mater.
276
(
9
),
339
345
.
Izquierdo
R.
Sacher
E.
Yelon
A.
1989
X-ray photoelectron spectra of antimony oxides
.
Appl. Surf. Sci.
40
(
1–2
),
175
177
.
Kang
M.
Kawasaki
M.
Tamada
S.
Kamei
T.
Magara
Y.
2000
Effect of pH on the removal of arsenic and antimony using reverse osmosis membranes
.
Desalination
131
(
1–3
),
293
298
.
Martinson
C. A.
Reddy
K. J.
2009
Adsorption of arsenic(III) and arsenic(V) by cupric oxide nanoparticles
.
J. Colloid Interf. Sci.
336
(
2
),
406
411
.
Standardization Administration of the People's Republic of China (SAC)
2007
Standards for drinking water quality (GB5749–2006)
,
Beijing
.
Xu
W.
Wang
H. J.
Liu
R. P.
Zhao
X.
Qu
J. H.
2011
The mechanism of antimony(III) removal and its reactions on the surfaces of Fe-Mn binary oxide
.
J. Colloid Interf. Sci.
363
(
1
),
320
326
.
Zhu
J.
Wu
F. C.
Deng
Q. J.
Shao
S. X.
Mo
C. L.
Pan
X. L.
Li
W.
Zhang
R. Y.
2009
Environmental characteristics of water near the Xikuangshan antimony mine, Hunan Province
.
Acta Scientiae Circumstantiae
29
(
3
),
655
661
.