Electric conversion treatment of cobalt-containing wastewater

Long-term accumulation of cobalt-containing wastewater may also pollute groundwater and cause a large amount of loss of valuable metals. Therefore, the comprehensive utilization of cobaltcontaining wastewater must be realized, especially as cobalt itself is a very important strategic resource. This paper proposes a membrane electroconversion method to separate cobalt ions from cobalt-containing wastewater and prepare cobalt hydroxide. In addition, the electrolysis process was optimized, and single-factor experiments such as the initial concentration, cobalt ions, current density, temperature etc., and economic calculations such as current efficiency were explored. The electrolysis product was calcined as the precursor to obtain the oxide Co3O4, and the calcination experiment was also optimized. In this concentration range, more than 90% of cobalt can be recovered within 2 h.


INTRODUCTION
Cobalt production industries produce a large amount of Co 2þ -containing wastewater during the production process, which not only results in serious pollution to the environment, but also causes a large amount of loss of valuable metals. Compounds containing cobalt are difficult to degrade naturally into harmless substances, and cobalt is extremely harmful to aquatic animals. The main domestic production process of basic cobalt carbonate is to obtain basic cobalt carbonate slurry through the substitution reaction of soluble cobalt salt (cobalt chloride or cobalt sulfate) and sodium carbonate solution. After multiple washings to remove impurity ions, and drying to obtain a basic cobalt carbonate product, the washing wastewater contains a large amount of Co 2þ . Lithium battery factories mainly use lithium cobalt oxide and graphite as the positive and negative electrodes, and use the electrolyte as the medium and auxiliary materials to produce lithium batteries (Apiratikul & Pavasant ; Karate & Marathe ). The wastewater from the battery factory contains a large amount of heavy metal cobalt. Therefore, seeking an economical and efficient treatment process to increase the recovery rate of valuable metals and reduce environmental pollution is an urgent scientific research area for cobalt production industries in the production process.
At present, the treatment methods of cobalt-containing wastewater mainly include chemical precipitation, physical chemistry, and biological methods. The chemical precipitation method uses hydroxides, sulfides, etc. to precipitate cobalt to separate cobalt from water (Cavaco et al. ; Lee et al. ; Foroutan et al. ). The treatment effect of this method is limited, and the cobalt content of the effluent often exceeds standards. Biological methods are suitable for low-concentration cobalt-containing wastewater, and rely more on the screening and culture of high-efficiency bacterial flora. Flora are difficult to treat, and are very sensitive to water quality changes, and the application of this technology is rare (Joss et al. ). Physical and chemical methods, including adsorption, membrane separation and ion exchange resin methods, can not only meet the processing requirements, but also recover the valuable metal cobalt, this approach is a hot research topic in China and elsewhere. Membrane separation is a physical separation process in which water molecules and ions under pressure pass through membranes with different pore sizes; the sieving effect of the membrane pore size achieves separation (Chiu & Chen ). The membrane separation method has the characteristics of no phase change, no pollution, and operation at room temperature (Metz et al. ). The ion-exchange method is to selectively remove certain ions through ion exchange during contact between ion-exchange resin and wastewater, and is suitable for low-concentration wastewater. The ion-exchange method has the characteristics of simple operation, good treatment effect, etc., and it is easy to realize automation, and therefore a promising technology (Xie et al. ). Due to its large specific surface area, mesoporous cobalt tetroxide has good optical, electrical and chemical properties, and is widely used in supercapacitors (Meher &  In this paper, the cobalt-containing wastewater from a certain company was used as the treatment object. First, a self-made two-chamber electrolyzer was used to conduct membrane electrolysis experiments, and cobalt hydroxide was successfully recovered from the wastewater and prepared. Secondly, the influence of different single factors on the electrolysis products in the process of membrane electrolysis was investigated. Finally, cobalt hydroxide was used as a precursor to calcination to obtain high-purity cobalt tetroxide. At the same time, the effects of calcination temperature and time on Co 3 O 4 were investigated.

EXPERIMENTAL SECTION Experimental materials and homemade electrolyzer
The cobalt-containing wastewater produced by a domestic company in the production process was used. The approximate value of the wastewater composition is shown in Table 1. Cationic film (CJMC-3) purchased from Hefei Chemical Joy Polymer Materials Co., Ltd.

Membrane electrolysis process
This study used industrial cobalt-containing wastewater as the anode solution for the electrolysis experiment (Figure 1), and investigated the influence of electrolyte concentration, temperature, current density, different surfactants, external stirring, and ultrasonic vibration on the electrolysis process. Then, the optimized concentration of magnesium chloride was used as the anolyte, and the cobalt hydroxide precipitated at the cathode was calcined, and the influence of the calcination temperature and time of the oxide was investigated. The whole electrolysis reaction process consisted of three steps. First, the bipolar reaction. The anode generates chlorine gas and the cathode generates hydrogen; then Co 2þ migrates from the anode to the cathode under the Anode reaction: Cl À À 2e ! Cl 2 ↑ Cathodic reaction: H þ þ 2e ! H 2 ↑ Precipitation process: Co 2þ þ 2OH À ! Co(OH) 2 ↓ Total reaction equation: CoCl 2 þ 2H 2 O ! Co(OH) 2 ↓ þ Cl 2 ↑ þH 2 ↑ After the electrolysis experiment had completed, the pink precipitate was collected and filtered at the cathode, then washed with deionized water and alcohol three times each, and then dried at 60 C for 12 hours.

Calculation method
In this article, the economic standards for measuring the ion-exchange membrane electrolysis process mainly include current efficiency, DC power consumption and productivity: Among them, η is the cathode current efficiency, %; m (actual) and m (theoretical) are the mass of the actual cathode product and the mass of the product calculated according to Faraday's law, g; M is the molar mass of the cathode product, g·mol À1 ; I is the current, A; Δt is the electrolysis time, s; F is Faraday's constant (about 96,500 C·mol À1 ); and n is the number of electrons passing through the electrode reaction: W is the DC power consumption, kW·h·kg À1 ; η is the cathode current efficiency, which can be calculated by Equation (2), %; and V is the cell voltage of the electrolysis process. Since this experiment uses constant current electrolysis conditions, the cell voltage in the electrolysis process was not constant, so the expected value of the cell voltage during the electrolysis process was calculated according to the arithmetic average of the cell voltage, V; k is the theoretical power consumption of the electrolysis process, kA·h·kg À1 .
Theoretical power consumption (k) is the theoretical power consumption required to produce a unit mass product. The calculation equation is shown in (3): k is the theoretical power consumption, kA·h·kg À1 ; Q is the amount of charge passed through the electrolysis process, C; m (theoretical) is the mass of the product that should be generated theoretically when the amount of charge Q passes through the electrolysis process, g; n A is the number of moles of charge passed through the electrolysis process, mol; and F is Faraday's constant (about 96,500 C·mol À1 ): The productivity of the electrolysis process is the quality of the product produced per unit time. It is an important indicator to measure the production capacity. The specific calculation method is shown in Equation (4): The unit of productivity is g·h À1 ; m (actual) is the actual product quality, g; and t is the electrolysis time, h.

Material characterization
The structure and chemical analysis of the samples was performed by powder X-ray diffraction (XRD) (CuKα radiation, Bruker D8 Advance, Germany). The surface morphology of the sample was observed using a scanning electron microscope (SEM) (Zeiss Sigma 300). The weight loss of the precursor was analyzed by differential scanning calorimetry-thermogravimetric analysis (DSC-TGA) (SDT Q600 V20.9 Build 20, 10 C/min).

Effect of initial concentration of cobalt ion on product
The X-ray diffraction pattern of the product of different Co 2þ concentration conditions is shown in Figure 2(a). It can be roughly seen in Figure 2(a) that as the concentration was increased, the crystallinity gradually increases, and the crystallinity is optimal when the cobalt ion concentration is 0.2 mol·L À1 . (001), (100), (101) and other diffraction peaks are consistent with the standard diffraction peaks of cobalt hydroxide (JCPDS card no. 30-0443), indicating that the prepared sample is hexagonal β-Co(OH) 2 (Tong et al. ; Youssry et al. ). All diffraction peaks from XRD can be retrieved for β-Co(OH) 2 hexagonal phase diffraction data (according to standard spectrum JCPDS card no. 30-0443). It can be seen from the different curve images that the degree of crystallinity gradually increases, and the characteristic peaks become more and more obvious. The high intensity of the (101) diffraction peak indicates that the crystallites of β-Co(OH) 2 have a preferred orientation along the (101) axis during their deposition (Kitahata et al. ). In particular, no diffraction peaks related to the α phase of cobalt hydroxide were observed. It can be concluded that under the applied cathodic electrodeposition conditions, pure β-phase cobalt hydroxide can be obtained (Aghazadeh et al. ). The characteristic peaks are not significantly broadened and refined, and we can infer that the grain size has not changed much (this inference is also confirmed by the scanning electron microscope later).
According to the equation, crystallinity ¼ 1 À all amorphous peak intensities Scattering intensity of 100% fully amorphous standard × 100% (5) crystallinity ¼ Sample crystal diffraction peak intensity Diffraction intensity of 100% complete crystalline standard × 100% (6) The calculation formula is simplified, no standard sample is needed, and crystallinity can be calculated from one sample according to Equation (7). The crystallinity of each sample was calculated to be 34.0%, 85.04%, 94.06%, and 99.25%. When the cobalt ion concentration was 0.05 mol/L, the crystallinity of this sample was roughly calculated to be 34.0%, and the crystallinity of this sample was poor. The curves of the latter three concentrations overlapped, indicating that the sample has good crystallinity and can produce stable crystals. As the concentration was increased, the crystallinity also increases. For samples with small to large concentrations, and crystallinity from low to high, it was concluded that the optimal solution concentration for electrodeposition preparation Co(OH) 2 was 0.2 mol·L À1 : Crystallinity ¼ Diffraction peak intensity total intensity × 100% (7) Figure 3 shows the surface morphology of β-Co(OH) 2 powder under different initial concentrations of cobalt ions. The SEM image shows that the prepared deposit has a uniform and clear sheet-like nanoscale structure. As the concentration of cobalt ions was increased, the dispersibility of cobalt hydroxide gradually increases, and the hexagonal flake shape becomes more obvious. According to rough estimation, the particle diameter ranges for small to large cobalt ion concentrations are: (a) 40-107 nm; (b) 50-110 nm; (c) 10-70 nm; (d) 30-85 nm; the thickness is only a few nanometers. The electrodeposited Co(OH) 2 has a unique spatially staggered nanosheet network morphology, and the thickness of the nanosheets is uniform with a relatively large specific surface area. Photographs in (a), (b), (c) are not as clear as (d), because the sample had been stored for too long, but the hexagonal sheet structure can still be seen. The interlaced nanoflake network morphology makes it produce rapid ion migration in the electrochemical oxidation-reduction reaction, so that the electroactive material can be fully utilized to speed up the reaction. At the same time, it is also conducive to the formation of electric double layers.

Effect of electrolysis temperature of cobalt ion on product
As shown in Figure 4(a), compared with the room temperature reaction, when the reaction temperature was lower, the viscosity of the electrolyte solution is relatively large, the conductivity is low, and the diffusion is slow, which blocks the conduction of electrons in the solution and the cobalt in the anode chamber to the cathode diffusion process. As the reaction temperature was increased, the viscosity of the electrolyte decreases and the conductivity increases, which can accelerate the diffusion of cobalt ions into the cathode, and at the same time promote the generation of chlorine and hydrogen at the cathode. Conversely, a higher temperature is conducive to the improvement of ion migration rate in the electrolyte solution, resulting in an increase in current efficiency and a decrease in DC power consumption. When the electrolysis temperature is 60 C, the electrolysis experiment is the best. As shown in Figure 4(b), as the temperature was increased, the cell voltage first decreases and then increases. The increase in temperature reduces the activation energy of the reaction. From a theoretical analysis, this phenomenon conforms to the Arrhenius formula: k ¼ Ae ÀEa=RT . As the temperature was increased, the activation energy E a decreases, which can increase the reaction rate constant k. The increase in temperature makes the electrode reaction rate of H 2 , Cl 2 , Co(OH) 2 faster, thus lowering the cell voltage, while the cell voltage at 80 C increases, indicating that the electrolysis experiment requires a lower temperature increase to meet the market conditions. It can be roughly seen from the XRD curve that the characteristic peak intensity is the largest when the temperature was 70 C, followed by 50, 80, and 60 C. The lattice constants of the samples are both a ¼ 5 Å and c ¼ 5 Å, so each Co(OH) 2 has a regular hexagonal crystal structure. According to the simplified calculation Equation (7) of crystallinity, the crystallinity of the sample under different temperature conditions was 100%. From the perspective of crystallinity, when the above different temperatures are set, the crystallinity of the sample can be as high as 100% without an amorphous peak. Figure 5 shows the surface morphology of β-Co(OH) 2 powder under different temperature conditions. Arranged in descending order of temperature, the particle diameter ranges were: (a) 10-30 nm; (b) 20-70 nm; (c) 20-65 nm; (d) 15-75 nm, but their thickness was only a few nanometers. The powder had flake-like structure that crosses horizontally and vertically. At 50 C, due to the smaller diameter, more nanoscale pores were formed with the most gaps; at 60 C, the diameter range becomes larger, and it is still an irregularly cross-layered structure; the flake structure is obvious at 70 C, and the diameter range becomes smaller than that at 60 C, and the flakes intersect and agglomerate closely to form a porous structure; at 80 C, the interlaced structure is obvious, the diameter range is relatively larger, and the crystals are denser.

Effect of current efficiency
Figure 2(c) shows that with the increase in the current density, the characteristic peaks become sharper and the intensity becomes higher and higher. By comparing with the standard XRD curve, it was found that all peak positions of the prepared Co(OH) 2 sample are consistent with the standard sample, and the XRD diffraction peak intensity is high, which proves that the Co(OH) 2 prepared by our group had good crystallinity. From the perspective of crystallinity, when the current density was 70 mA·cm À2 and 120 mA·cm À2 , the crystallinity was as high as 100%. Figure 6 shows the surface morphology of β-Co(OH) 2 powder under different current density conditions. The particle diameter ranges from small to large current density are: (a) 30-120 nm; (b) 30-80 nm; (c) 40-170 nm; and (d) 30-170 nm.
The actual crystallization mechanism of Co(OH) 2 is still an open question. At present, initial nucleation, directional aggregation and Ostwald ripening can be used to explain the process of crystal growth. The cobalt ion in the aqueous solution reacts with the OH À generated near the cathode plate to form Co(OH) 2 , and the reaction is as described in Equation (8): Subsequently, the nanoparticles began to assemble together and spontaneously aggregated into Co(OH) 2 nanosheet structures. As the high surface energy is reduced through the process of directional aggregation, layered Co(OH) 2 nanosheets are formed (Pacholski et al. ). Ostwald maturation determines the rate of growth and recrystallization (Li et al. ), and through the Ostwald maturation process, the smaller Co(OH) 2 crystal particles produced in the solution are reduced due to the curvature. Larger, higher energy particles, gradually dissolved into the surrounding electrolyte, and then re-precipitated on the surface of the larger Co(OH) 2 crystal particles, which makes the larger Co(OH) 2 crystal particles further increase.

Effect of surfactants
Figure 2(c) shows that by adding a small amount of organic additives during the electrodeposition process, it is easy to obtain a dense product deposition layer. The additives adsorbed on the electrode surface change the structure of the electric double layer, because this adsorption enhances polarization and makes the crystals finer. The adsorption of surface active substances is different for different crystal planes. The growth of crystal planes with strong adsorption of surface substances is hindered, while the development of crystal planes with less adsorption will change the growth morphology of the crystal. The external electric energy activates the ions, and the ions pass through the adsorption layer of the surfactant to discharge. Compared with the addition of complexing agents, the method of adding surfactants to control the electrode process has the advantages of low concentration, low cost, no effect on the chemical properties of metal ions in the solution, easy handling, non-toxic, etc., and is mostly used in the processing industry diaphragm electrodeposition process for wastewater. After adding the additive polyethylene glycol (PEG), the diffraction peak intensity of the sample is significantly reduced, resembling a steamed bread peak, similar to a colloidal structure, which corresponds to the surface micromorphology in Figure 7 below. In the curve comparison with and without additives, it can be seen that without additives, the crystallinity of the product is the highest, followed by the addition of the surfactant polyvinylpyrrolidone (PVP) > sodium dodecyl sulphate (SDS) > polyethylene glycol (PEG). Compared with the XRD curve with SDS, the diffraction peaks of the sample with SDS are obviously broadened and weakened. This shows that the crystallization degree of the sample with SDS added is lower. Therefore, in the experiment of preparing cobalt hydroxide by diaphragm electrolysis, the addition of surfactants reduced the crystallinity of the sample, and even changed from crystal to colloidal. The electrolytic preparation of the surfactant-free diaphragm is an obvious advantage over other preparation methods.
The addition of surfactant PVP is beneficial to improve current efficiency, reduce cell voltage, increase productivity, and reduce DC power consumption. However, the main reason why the output of the surfactant PVP is higher than that of no additives is the extra weight of the additives.
The effects of other additives, PEG and SDS, are far inferior to those without additives, and the best additive PVP produces the same effect as without additives. From an economical point of view, the experimental results without additional surfactants are the best.
Additives such as PVP, SDS and PEG will not chemically react with the electrolytic solution, and no new crystal nuclei are generated to affect the growth of the original crystal nucleus. It can be observed that the difference between the surface morphology of the product when different organic solvents is added is very small, which means that different organic solvent additives affect the morphology of Co(OH) 2 particles by affecting the way of crystal nucleus agglomeration. Some inert organic macromolecular additives are added in the initial stage of electrolytic preparation. Although they do not participate in the reaction, the additives have a steric hindrance effect on the aggregation of atoms. As far as the current research is concerned, the steric hindrance effect is helpful in preventing the anisotropic growth of the crystal grains, and can maintain the flake shape into the final shape. Although its influence on the grain growth mechanism is not fully understood, the particle shape can still be controlled. Three different surfactants all provide different force field environments for the growth of Co(OH) 2 , and finally form tiny morphological differences. In terms of morphology, it is uneconomical for the addition of surfactants to destroy the original hexagonal structure.
Co(OH) 2 as a precursor Figure 8 shows the DSC-TGA curve of the β-Co(OH) 2 sample by adding a surfactant PVP. Figure 8 shows that the TGA curve loses 3% at 30-158 C, and the weight loss in this range is attributed to the sample adsorption water desorption. The maximum weight loss occurs at 158 C, and this weight loss extends to 615 C. When the temperature reaches 242 C, the weight loss reaches 27%, which is caused by the dehydration of water and the decomposition of some anions CO 2À 3 and the oxidation of Cl À due to the dehydroxylation of hydroxide. The weight loss here corresponds to an endothermic peak at about 242 C on the DSC curve, and there may be oxidation of Co 2þ . The decrease in mass at 242-557 C is attributed to the combustion reaction of surfactant PVP, which indicates that it is not economical to prepare Co for surfactant addition if it is necessary to calcinate Co to oxide. When the temperature exceeds 557 C, the TGA curve tends to be substantially stable, indicating that the adsorbed impurities in the sample have completely decomposed.

Effect of calcination temperature
The product is single in the calcination temperature range of 300-800 C, and the X-ray diffraction peak is sharp and fine, indicating that the calcination product has good crystallinity. When the calcination temperature is 300 C, there was no need to increase the temperature, which greatly saves energy consumption. The diffraction peaks are approximately 2θ ¼ 19 , 31 , 37 , 45 , 59 , 66 , corresponding to (111), (220), (311), (222), (400), (422), (511), (440) crystal planes, corresponding to the literature value (Liao et al. ), the standard card (JCPDS card no. 42-1467) determines that the Co 3 O 4 produced after calcination is a cubic crystal system. As the calcination temperature continues to increase, the diffraction peaks become sharper and the particle size of Co 3 O 4 increases. As the temperature decreases, the diffraction peak of Co 3 O 4 broadens significantly, indicating that the Co 3 O 4 particles produced are smaller (He et al. ).
As shown in Figure 9, Co 3 O 4 has a sheet-like structure similar to Co(OH) 2 , but the surface of Co 3 O 4 is rougher and there are a large number of pore structures due to the thermal decomposition of Co(OH) 2 to produce O 2 and H 2 O (Ghosh et al. ). The particle size of Co 3 O 4 is approximately (a) 45-104 nm in diameter, 4-11 nm in thickness; (b) 34-96 nm, 5-10 nm in thickness; (c) 30-65 nm, 3-7 nm in thickness; (d) 20-110 nm, 4-10 nm thick; (e) 20-115 nm, 4-20 nm thick; (f) 20-135 nm, 4-12 nm thick. This shows that the higher the calcination temperature, the larger the particle size. The interlaced pores formed by the sintering flakes easily penetrate the electrolyte and can increase the contact area between the electrolyte ions and the electrode material. The diameter of the pores formed by cross-linking of the nanosheets is about 12 nm. These pores make it easy for the electrolyte to penetrate, which is beneficial to improve the utilization of the active sites of the electrode material, and is beneficial to the transport of electrolyte in electrochemical reactions. The thickness of the nanosheet was about 10-20 nm, and the thinner sheet structure can provide more active sites per unit mass of active material (Maile et al. ).

Effect of calcination time
The calcination time had no obvious effect on the crystallinity and surface morphology of the material (Figures 2(f) and 10). When the calcination time is 1 h, the diffraction peak of the oxide Co 3 O 4 is the sharpest. When the calcination time is increased, the intensity of the diffraction peak does not increase or even begins to weaken, indicating that a shorter calcination time can reach a certain degree of crystallinity. The samples calcined at different times showed hexagonal flake morphology similar to its precursor cobalt hydroxide, indicating that the calcined product better inherited the morphological characteristics of the precursor, but compared with the precursor Co(OH) 2 . The particle size of Co 3 O 4 is larger, and the thickness also increases, around 6-20 nm. The growth of crystal grains on the sample surface is scattered, and the surface morphology is relatively flat. The surface of the sample is smooth, the surface particles are very small, and the agglomeration of small crystal grains is obvious.

CONCLUSION
In this study, a self-made two-chamber electrolyzer was used to investigate the influence of the initial concentration, current density, reaction temperature, and additives on the electrolytic conversion process. The results show that the best electrolysis process conditions are: current density 500 A·m À2 , the initial reaction temperature is 60 C, the CoCl 2 concentration is 0.2 mol·L À1 , and there is no additive. The prepared sample is a layered β-Co(OH) 2 nanosheet with hexagonal crystal phase. In the experiment to prepare cobalt hydroxide by diaphragm electrolysis, the addition of surfactant reduced the crystallinity of the sample, and even changed it from crystal to gel. The electrolysis preparation of diaphragm without surfactant is a clear advantage over other preparation methods. The precipitate obtained by electrolytic conversion was studied by calcination experiment, and the influence of calcination temperature and time on the product was investigated. The precipitated products obtained by electrolytic conversion are all calcined Co 3 O 4 , and their microscopic morphology inherited the nanoplate structure of cobalt hydroxide. The pores formed by cross-linking between the nanosheets make it easy for electrolyte to penetrate, which is beneficial to improve the utilization of the active sites of the electrode material. The thickness of the nanosheet was about 10-20 nm, and the thinner sheet structure can provide more active sites per unit mass of active material. The best calcination temperature for electrolytic conversion precipitated product was 300 C, and the calcination time was 1 h. The products Co(OH) 2 and Co 3 O 4 can be used as catalysts for water electrolysis.

DECLARATION OF COMPETING INTEREST
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.