Electrolytic separation of cobalt and tungsten from cemented carbide scrap and the electrochemical behavior of metal ions

Accepted Manuscript Electrolytic separation of cobalt and tungsten from cemented carbide scrap and the electrochemical behavior of metal ions

Xiaoli Xi, Xiangjun Xiao, Zuoren Nie, Liwen Zhang, Liwen Ma PII: DOI: Reference:

S1572-6657(17)30232-1 doi: 10.1016/j.jelechem.2017.04.001 JEAC 3213

To appear in:

Journal of Electroanalytical Chemistry

Received date: Revised date: Accepted date:

20 December 2016 30 March 2017 1 April 2017

Please cite this article as: Xiaoli Xi, Xiangjun Xiao, Zuoren Nie, Liwen Zhang, Liwen Ma , Electrolytic separation of cobalt and tungsten from cemented carbide scrap and the electrochemical behavior of metal ions. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Jeac(2017), doi: 10.1016/j.jelechem.2017.04.001

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Electrolytic separation of cobalt and tungsten from cemented carbide scrap and the electrochemical behavior of metal ions

PT

Xiaoli Xi, Xiangjun Xiao, Zuoren Nie, Liwen Zhang, Liwen Ma

RI

College of Material Science and Engineering, Beijing University of Technology, Beijing 100124,

SC

China

To whom correspondence should be addressed: Zouren Nie; E-mail: [email protected];

MA

NU

Tel/Fax:86-10-67391536

Abstract

D

Molten salt electrolysis is used to separate and recycling of elemental tungsten

PT E

and cobalt from cemented carbides. WC–6 wt% Co scrap and NaCl–KCl molten salt was used as a sacrificial anode and electrolyte, respectively. The range of preparation

CE

parameters and the electrochemical behavior of tungsten and cobalt ions were

AC

investigated through electrochemical techniques, such as cyclic voltammetry (CV) and square wave voltammetry (SWV). Results showed that the dissolution potential of cobalt and WC were 0 V and 0.6 V (vs. Ag/AgCl), and the reduction potentials of Co (II) + 2e−  Co (0) and W(II) + 2e−  W (0) were −0.2 and 0.2 V (vs. Ag/AgCl), respectively. The reduction processes of Co(II) to Co and W(II) to W were both reversible reactions controlled by ion diffusion. The average diffusion coefficients of Co(II) and W(II) were determined by CP to be 5.62 × 10−5and 3.94 × 10−5cm2 s−1, 1

ACCEPTED MANUSCRIPT respectively. Furthermore, the products at cathodes Nos. 1 and 2 were characterized by scanning electron microscopy, X-ray fluorescence and X-ray diffraction analyses. Results showed that pure cobalt and WC powder with a diameter of <100 nm was obtained at cathode No. 1 and cathode No. 2. And the result of XRF shows that the

PT

purity of the products both Co and WC were higher than 90 percent. This study

SC

controlling process parameters during electrolysis.

RI

reveals that cobalt and tungsten can be separated from WC–6 wt% Co scrap by

NU

Keywords: electrochemical behavior, separation, WC–6 wt% Co scrap, molten salt

MA

1. Introduction

D

Tungsten is an important rare and strategic metal and the amount of tungsten

PT E

spread across the planet is the twentieth of the whole element. Tungsten and its alloys are widely used in wear-resistant parts, cutting tools, mine machines, and die

CE

materials because of their excellent wearability and corrosion resistance and high

AC

hardness [1][2]. With the exploitation of tungsten ores and unceasing reduction in resource reserves, tungsten alloy scrap must be recycled. Compared with the run of mine ore, the tungsten content of scrap is higher than that in ore concentrates. According to statistics, tungsten recycling has penetrated more than 50% of the total amount by 2014 in United States [3]. However, the excellent physical and chemical stability of tungsten render the recycling process extremely challenging. Traditional recycling methods based on differences in physical and chemical 2

ACCEPTED MANUSCRIPT properties between basal material and adhesives have been studied for many years [4][9]

. Venkateswaran introduced a method called “Menstruum Process Technology”

and obtained WC powder by dissolving tungsten-bearing hard scrap in a Co–C or Fe– C melt followed by deposition into tungsten carbide [10]. Edtmaier dipped

PT

WC/Co-based cemented carbide scraps into acetic acid solutions under pure oxygen

. Lin introduced another method for selective dissolution of cobalt and recovery of

SC

[11]

RI

or air atmosphere for selective removal of the cobalt binder and recovery of the WC

WC by electrolyzing the cemented tungsten carbide scrap in acid solutions containing

NU

additives [12]. Lee developed a new hydrometallurgical route to extract cobalt and

MA

form tungstate in a single step by dissolving WC–6 wt% Co hard-metal sludge in aqua regia [13]. Jung selected the optimal craft parameter for tungsten oxidation and

D

reduction from cemented carbide sludge [14].

PT E

We proposed a novel method to recycle metallic element from tungsten carbide scrap in NaCl–KCl melt [15]. Compared with conventional approaches, molten salt

CE

electrolysis has many advantages, such as high-performing, cost-effective and

AC

environmentally friendly.

In this study, molten salt electrolysis was used to separate and recycle cobalt and tungsten from WC–6 wt% Co scrap at 1023 K. NaCl–KCl salt and WC–6 wt% Co were applied as electrolyte and sacrificial anode, respectively. WC–6 wt% Co was dissolved as cobalt and tungsten ions during electrolysis. Subsequently, the ions were discharged and cobalt and tungsten were deposited on the cathode. The extra carbon probably drifted on the melt as an atom, and then the carbon atoms were adsorbed by 3

ACCEPTED MANUSCRIPT the cathode reacting with tungsten to produce tungsten carbide. This article focused on the electrochemical behavior of cobalt and tungsten ions dropped from WC–6 wt% Co sacrificial anode; redox potential, electrochemical reversibility, electron transfer number, and diffusion coefficients of cobalt and tungsten ions were investigated by a

PT

variety of electrochemical techniques [16]. The product at the cathode was

RI

characterized by scanning electron microscopy (SEM), X-ray fluorescence (XRF) and

NU

SC

X-ray diffraction (XRD) analyses.

MA

2. Experimental

2.1 Electrochemical apparatus and electrodes

PT E

D

Fig. 1 shows the hermetically sealed three-electrode cell for electrochemical experiments. The reference electrode was a Ag/AgCl electrode, consisting of silver and equimolar KCl–NaCl molten salt and 1% AgCl in terms of mole, which was put

CE

into a very thin mullite tube. The reference electrode was calibrated by the Cl2/Cl−1

AC

electrode. The potential of Ag+/Ag (vs. Cl2/Cl−1) was -1.0 V. All the electrode potentials were measured or set relative to the potential of this reference electrode in this paper. Molybdenum plates (20 mm × 4 mm) and glassy carbon electrodes (3 mm diameter) were applied as counter electrode, WC–6 wt% Co plate (4.0 mm × 4.0 mm × 18.0 mm) and platinum wire (1 mm diameter) served as the working electrode simultaneously. All the plates were tacked on the stainless steel rod with wires.

4

ACCEPTED MANUSCRIPT 2.2 Preparation and purification of the melts

An equimolar mixture of sodium chloride (>99.5%, Beijing Chemical Works, Beijing, China) and potassium chloride (>99.5%, Beijing Chemical Works, Beijing,

PT

China) was applied as the electrolyte in this study. NaCl–KCl (120 g) was weighed and placed in an alumina crucible and then dried at 523 K in vacuum for 48 h to

RI

remove residual water thoroughly. Subsequently, the mixture salt was heated from

SC

523 K to the working temperature 1023 K under a high-purity argon atmosphere.

NU

2.3 Electrochemical test and characterization of cathode products

MA

PARSTAT 4000 electrochemistry workstation with the versa studio software package was employed to provide technical support such as cyclic voltammetry (CV)

D

and square wave voltammetry (SWV) techniques for all electrochemical experiments.

PT E

In addition, the electrochemical reaction process were carried out in a protective argon gas atmosphere to avoid reoxidation of the reduced metal particles. After electrolysis,

CE

the molten salts were naturally cooled in a high-purity argon atmosphere, and the

AC

electrolytic products obtained from the cathode were rinsed with distilled water and dried in air at 40 °C. Finally, the cathode products were analyzed using XRD, XRF and SEM respectively.

5

ACCEPTED MANUSCRIPT 3. Results and discussion

3.1 Electrochemical dissolution potential

Fig. 2 shows the cyclic voltammograms of the platinum wire, WC and WC–6 wt%

PT

Co (4.0 mm × 4.0 mm × 18.0 mm) anode in NaCl–KCl molten salt at 1023 K, scan

RI

rate is 0.1 V s−1. No evident current was noted on the platinum wire until the electric

SC

potential approached the electrochemical limit of the melt. However, a sharp increase in current was observed at 0.6 V (vs. Ag/AgCl) on the WC consumable anode; the

NU

reactions that occurred at increased current were regarded as the dissolution of

MA

tungsten metal (W → Wn+ + ne−). Similarly, the WC–6 wt% Co anode achieved similar curves to those from the WC sacrificial anode, and the current rapidly

D

increased with the augmentation in potential from 0 V (vs. Ag/AgCl). The increased

PT E

current that emerged in the voltammogram on the WC–6 wt% Co consumable anode was ascribed to the oxidation of cobalt metal (Co → Con+ + ne−). Finally, the

CE

dissolution potentials of cobalt and tungsten carbide were obtained by extrapolation of

AC

the linear portion of both the curves from the platinum wire, WC, and WC–6 wt% Co consumable anode to the zero current. Using the data, the potential of potentiostatic electrolysis was set as 0.4, 0.8, and 1.2 V (vs. Ag/AgCl), which are slightly positive to the potential of the dissolution of Co and WC and slightly negative to the potential of the decomposition voltage of the electrolyte.

6

ACCEPTED MANUSCRIPT 3.2 Electrochemical behavior of cobalt ions

To avoid the dissolution of tungsten, we maintained the potential of potentiostatic electrolysis negative to the dissolution potential of tungsten carbide. Fig.

PT

3 (a) shows a typical cyclic voltammogram of the NaCl–KCl molten salt at 1023 K. For the blank NaCl–KCl melt, no redox peak was noted in the graphic except for two

RI

steep peaks near the decomposition voltage of the molten salt. Fig. 3 (a) also shows

SC

cyclic voltammograms obtained in NaCl–KCl molten salt at the temperature 1023 K

NU

after 8 h potentiostatic electrolysis with potential of 0.4 V (vs. Ag/AgCl). The redox peak was observed in the cyclic voltammograms with the cathode peak (O) at −0.20 V

MA

and the corresponding anodic peak (R) at −0.01 V (vs. Ag/AgCl). To determine the species of the ion derived from the WC–6 wt% Co anode

PT E

D

during electrolysis at 0.4 V, the cathode product and the molten salt were characterized by XRD and inductively coupled plasma atomic emission spectrometry

CE

(ICP) (Fig. 3 (b)). All the characteristic peaks of cobalt emerged in the sample. These results indicate that the product obtained from cathode after performing the constant

AC

voltage electrolysis with potential of 0.4 V (vs. Ag/AgCl) is pure cobalt metal. Moreover, the tungsten ion concentration approached a negligible level in the melt after electrolysis. We proved that the two redox peaks present in the cyclic voltammograms correspond to the dissolution of cobalt metal and the reduction of cobalt ions, respectively. The reaction of the peak can be written as Eqs. (1) and (2): 𝐶𝑜 − 𝑛𝑒 − → 𝐶𝑜𝑛+

(1)

𝐶𝑜𝑛+ + 𝑛𝑒 − → 𝐶𝑜

(2) 7

ACCEPTED MANUSCRIPT Furthermore, square wave voltammograms were carried out to investigate the transfer electrons number of cobalt ions in the NaCl–KCl salts at 1023 K after electrolysis. Fig. 4 recorded the reduction wave of cobalt ions on a platinum wire (1 mm diameter) at a potential of −0.2 V (vs. Ag/AgCl). We found that only one

PT

reduction wave emerged in the range from −0.4 V to 0.3 V, which corresponded to the

RI

reduction of Con+ to metallic Co. And the exchange electron number (n) was

SC

calculated using Eq. (3) [17]: 𝑊1/2 = 3.52𝑅𝑇/𝑛𝐹

(3)

NU

where W1/2 is the width of the peak at half of its height obtained by Gaussian fitting

MA

(red lines in Fig. 4), R is the mole gas constant, T denotes the absolute temperature in K, F is Faraday’s constant (96.485 kC/mol), and n is the exchanged electron number.

D

According to Eq. (3), n was calculated to be 1.9 at 1023 K; hence, the reduction

PT E

of cobalt ions in NaCl–KCl molten salts is a one-step process with an exchange of two electrons.

CE

NaCl–KCl–1% CoCl2 was used to further verify the species of ions derived from

AC

the WC–6 wt% Co anode after electrolysis with a potential of 0.4 V (vs. Ag/AgCl). In Fig. 5, two redox peaks emerged in the CVs at −0.2 (reduction peak) and 0 V (anodic peak), which presented in the same potential as in a previous study. This results verified that the ions emerged both CV and SWV is Co(II). Additional the shift in the reduction potential of the cobalt ions with the increasing scan rates was insignificant. These results indicate that the reduction of Co2+ ions to Co metal in NaCl–KCl melts is a reversible electrode process. And the total redox reaction in the anode and cathode 8

ACCEPTED MANUSCRIPT during electrolysis at 0.4V can be expressed by Eqs. (4) and (5): 𝐶𝑜 − 2𝑒 − → 𝐶𝑜2+

(4)

𝐶𝑜2+ + 2𝑒 − → 𝐶𝑜

(5)

PT

3.3 Electrochemical behavior of tungsten ions

RI

To investigate the electrochemical behavior of tungsten ions in the melt, positive

SC

potential was applied between the WC–6 wt% Co consumable anode and cathode. Fig. 6 recorded the CV plots in the NaCl–KCl molten salt at 1023 K after 8 h

NU

potentiostatic electrolysis under various potentials 0.8 (red line) and 1.2 V (black line).

MA

Two couples of reduction and oxidation peaks O1, R1 and O2, R2 are observed in the plots at −0.2 and 0.2 V (vs. Ag/AgCl), respectively. Previous experiments have

D

demonstrated that the redox peak O1, R1 correspond to the dissolution of cobalt metal

PT E

and the reduction of cobalt ions. Similarly, the remaining couple of redox peak O2, R2 was ascribed to the dissolution of tungsten metal and the reduction of tungsten ions.

CE

CV measurements with a wide scan rate range were carried out to further study

AC

the reversibility of the redox reaction. Fig. 7 showed a series of CVs in NaCl–KCl melts after 8 h potentiostatic electrolysis at 0.8 V at various scan rates from 0.025 V s−1 to 1.4 V s−1. As displayed in the diagram, the shift in the reduction potential of the cobalt ions with the increasing scan rates was minor. These observations indicate that the reduction of Co2+ ions to Co metal in NaCl–KCl melts is a reversible electrochemical process. For a soluble–insoluble species and reversible electrochemical process, the ratio of oxidation peak current to reduction peak current 9

ACCEPTED MANUSCRIPT (|Ipa/Ipc|) and the half-wave potential (E1/2) can be kept constant at different sweep rates. The linear relationship between the half-wave potential (E1/2), the ratio of oxidation peak current to reduction peak current (|Ipa/Ipc|), and the square root of the scan rate (v1/2) for the reduction of Co(II)/Co are shown in Fig. 8. The value of E1/2

PT

and |Ipa/Ipc| (inset of Fig. 8) was nearly constant at - 0.11V and 4.55 for all the scan

RI

rates, respectively. These results verify that the reduction of Co2+ ions to Co metal in

SC

the NaCl–KCl melts at 1023 K is a reversible electrode process.

Similarly, the reduction potential of the tungsten ions slightly shifted toward the

NU

negative potential with increasing scan rates. This result suggests that the reduction of

MA

Wn+ ions to W metal in NaCl–KCl melts at 1023 K should not be considered as a perfectly reversible process. However, the value of E1/2 remained approximately

D

constant at 0.263 V with the increase in scan rates; meanwhile, the |Ipa/Ipc| was

PT E

maintained nearly constant at 2.0 for all scan rates (inset of Fig. 9). Moreover, the reduction peak current achieved a significantly linear correlation with the square root

CE

of the scan rate. This result indicated that the reduction of Wn+ ions to W metal in the

AC

NaCl–KCl melts at 1023 K is a reversible electrochemical process controlled by the diffusion of tungsten ions. The shift of peak potential is attributed to the change in concentration of ions. Besides, the ohmic drop of the experimental circuit can also cause the shift of the reduction peak of cobalt and tungsten ions [18]. Square wave voltammograms were carried out to further study the number of exchanged electrons in the electrochemical process. Fig. 10 shows the SWV results in the NaCl–KCl melt at 1023 K after potentiostatic electrolysis at different potentials 10

ACCEPTED MANUSCRIPT 0.4, 0.8, and 1.2 V (vs. Ag/AgCl). Only one redox species was present after potentiostatic electrolysis at 0.4 V, which corresponds to the reduction of cobalt ions to metallic cobalt. The number of exchanged electrons of cobalt was confirmed in previous studies to be 1.9. However, two redox species were observed at 0.2 and −0.2

PT

V (vs. Ag/AgCl) after potentiostatic electrolysis at potentials of 0.8 and 1.2 V. This

RI

result was found to be in line with the previous experiment, corresponding to the

SC

deposition of cobalt and tungsten metal. The average number of exchanged electrons calculated using Eq. (3) is given in Table 1. Given all acquired results, the total redox

NU

reaction in the anode and cathode during electrolysis at 0.8 and 1.2V is expressed by

𝐶𝑜 − 2𝑒 − → 𝐶𝑜2+

(4)

𝑊𝐶 − 2𝑒 − → 𝑊 2+ + 𝐶

(6)

PT E

D

Anode:

MA

following Eqs:

𝐶𝑜2+ + 2𝑒 − → 𝐶𝑜

(5)

𝑊 2+ + 2𝑒 − → 𝑊

(7)

CE

Cathode:

AC

Chronopotentiometry was carried out to study the electrochemical behavior of cobalt and tungsten ions and estimate the diffusion coefficient in NaCl–KCl melt at 1023 K with different current densities from −124 mA/cm2 to −86 mA/cm2 (Fig. 11). Two plateaus are present in the chronopotentiogram at 0.2 and −0.2 V, which correspond to the reduction of W(II) to W(0) and Co(II) to Co(0). This result further verify the mechanism of electrode reaction observed in the cyclic voltammograms. For a reversible electrode process, the product of the transition time (𝜏) and the 11

ACCEPTED MANUSCRIPT current density (i) is a constant decided by the following equation [19][20]: 1

1

𝑖𝜏 2 = −0.5𝑛𝐹𝐴𝐶(𝜋𝐷)2

(8)

where i is the current density; τ is the transition time of the plateau; n is the number of electron exchanged; F is Faraday’s constant; and C is the concentration of

PT

ions obtained from ICP to be 1.78 × 10−5 mol/cm3 (cobalt ions) and 0.93 × 10−5

RI

mol/cm3 (tungsten ions). D is the concentration and diffusion coefficient of the ions.

SC

The dependence of the iτ1/2 and current density is shown in Fig. 12. We found that the value of iτ1/2 remained almost constant with the increase in current density for the

NU

cobalt and tungsten ions [21]. We suggested that the electrochemical process of cobalt

MA

and tungsten ions is reversibly controlled by diffusion rate. The diffusion coefficients for Co(II) and W(II) are calculated using Eq. (9) for many times, the average value of

D

diffusion coefficients for Co(II) is 5.62 × 10−5cm2 s−1 (from 3.73 × 10−5 to 9.68 ×

PT E

10−5cm2 s−1). That for tungsten, the average value of diffusion coefficients is 3.94× 10−5cm2 s−1 (from 2.45 × 10−5 to 9.02 × 10−5 cm2 s−1).

CE

The electrochemical reaction resistance of WC–6 wt% Co consumable anode

AC

was investigated by electrochemical impedance spectroscopy at 1023 K in NaCl–KCl molten salt at two different potentials 0.4 and 0.8 V, which were selected on the basis of a previous study. These values correspond to the potentials at which cobalt and tungsten metals dissolve, respectively. Fig. 13 (a) shows a typical impedance spectra measured at 0.4 V in the frequency range of 1 Hz–100 kHz. A semicircle is present in the high-frequency region, followed by a straight line in the low-frequency region. At this low potential, the double-layer capacitance began charging, and the cobalt metal 12

ACCEPTED MANUSCRIPT commenced being oxidized, respectively. This process was controlled by charge transfer and corresponds to the semicircle that emerged in the high-frequency region. Subsequently, the cobalt ions diffused from the anode surface to the melt. This process was generally believed to cause the linear portion [22]. In the above analysis,

PT

an equivalent electric circuit was used to fit the data (Fig. 13(b)), which is similar to

RI

that in a Randles cell, including the electrolyte resistance (Rs), double-layer

[23]

SC

capacitance (C), charge-transfer resistance (RCT), and Warburg type impedance (W) . The values of the Rs and RCT were 0.549 Ω and 0.1597 Ω, respectively.

NU

Fig. 14 (a) displays the impedance spectra measured at 0.8 V in the frequency

MA

range 0.1 Hz–100 kHz. The radius of the semicircle that obviously emerged in the high-frequency region is larger than that in the previous study at 0.4 V. Moreover, two

D

pieces of a circular arc are observed in the high-frequency region related to the

PT E

dissolution of cobalt and tungsten metal on the anode ( Fig. 14 (c)). A straight line with an angle approaching 36.4° with the real axis is notable in the low-frequency

CE

region; this line is attributed to the diffusion of the cobalt and tungsten ions in the

AC

NaCl–KCl melts. The equivalent circuit used in this experiment is also shown in Fig. 14 (b). This finding is similar to that obtained at 0.4 V. Rs and RCT were 0.348 Ω and 1.01 Ω, respectively.

3.4 Separation and recovery of elements tungsten and cobalt

In this work, potentiostatic electrolysis was carried out to extract metallic elements from WC–6 wt% Co scrap in NaCl–KCl molten salts at 1023 K and 13

ACCEPTED MANUSCRIPT different potentials for 8 h. WC–6 wt% Co and molybdenum plates were employed as sacrificial anode and cathode, respectively. According to the results from the studies on electrochemical techniques, separating elemental cobalt from tungsten matrix is practicable if we control the electrochemical dissolution reaction occurring on the

PT

anode. Therefore, the potentials of potentiostatic electrolysis were selected as 0.4, 0.8,

RI

and 1.2 V (vs. Ag/AgCl), which are slightly positive to the potential of the dissolution

SC

of Co and WC. The products were collected from the cathode after electrolysis and then rinsed with deionized water. Fig. 15 shows the current–time curves and

NU

potential–time curves during electrolysis and the XRD spectra of the cathode products.

MA

Pure cobalt was observed in the pattern after potentiostatic electrolysis at potential of 0.4 V, and the compounds of Co and WC can be obtained at 0.8 and 1.2 V. Overall, we

D

conclude that separating cobalt metal from WC–6 wt% Co scrap can be achieved by

PT E

controlling process parameters.

To collect Co and W selectively, one cathode is replaced with the other when the

CE

cobalt was recycled completely during electrolysis. Two molybdenum plates were

AC

used as cathode Nos. 1 and 2, respectively. At the beginning of electrolysis, potentiostatic electrolysis was applied in the system for 8 h to selectively dissolve the cobalt metal, and the potential of the anode was kept at 0.4V (vs. Ag/AgCl) against reference electrode throughout electrolysis. Elemental cobalt dropped from the sacrificial anode, was discharged on cathode No. 1, and deposited as cobalt metal. Meanwhile, the tungsten element remained in the anode because the applied potential was much smaller than the standard level required for the dissolution potential of the 14

ACCEPTED MANUSCRIPT WC. Subsequently, cathode No. 1 was replaced by cathode No. 2 and the potential was increased to 0.8 and 1.2 V (vs. Ag/AgCl) for recycling elemental tungsten. WC began to dissolve and the tungsten metal deposited on cathode No. 2. Additional carbon probably drifted on the melt as an atom, and the carbon atoms were migrated

PT

in the solvent to the cathode with tungsten together and reacting with tungsten to

RI

produce tungsten carbide powder. The products obtained from cathode Nos. 1 and 2

SC

were characterized by SEM, XRF and XRD. As shown in Fig. 16, pure cobalt metal and tungsten carbide powder emerged from cathode Nos. 1 and 2, respectively. The

NU

particle sizes of the cathode products were measured by SEM to be less than 100 nm.

MA

These results proved that elemental cobalt and tungsten can be separated successfully from WC–6 wt% Co scrap by controlling process parameters during electrolysis. In

D

addition, the anode consumption and the recovered cathode product were weighed,

PT E

the recovery rate of the W and Co were calculated to be around 80%. And the result of XRF shows that the purity of the products both Co and WC were higher than 90

CE

percent.

AC

4. Conclusions

Molten salt electrolysis were employed to separate and recycle elemental cobalt and tungsten from WC–6 wt% Co scrap in NaCl–KCl melts at 1023 K. WC–6 wt% Co was used as sacrificial anode in this process. The range of preparation parameters and the electrochemical behaviors of tungsten and cobalt ions derived from the WC–6 wt% Co anode were investigated using electrochemical techniques. The dissolution 15

ACCEPTED MANUSCRIPT potential of cobalt and WC were 0 V and 0.6 V (vs. Ag/AgCl), and the electroreduction of cobalt ions was achieved with a single one-step process that involved two transfer electrons. This process showed reversible electrode behavior at various scan rates from 0.025 V s−1 to 1.4 V s−1 controlled by the diffusion of cobalt

PT

cations. Similarly, the cathode process of tungsten ions employed a single one-step

RI

process with two exchanged electrons. Furthermore, the reduction of W2+ ions to W

SC

metal also demonstrated a reversible behavior, which was controlled by the diffusion rate of tungsten cations. The average diffusion coefficients of Co(II) and W(II) were

NU

determined by CP to be 5.62 × 10−5and 3.94 × 10−5cm2 s−1, respectively. Finally,

MA

potentiostatic electrolysis was applied to separate and recycle elemental cobalt and tungsten from WC–6 wt% Co scrap; the products obtained from cathode Nos. 1 and 2

D

were then analyzed by XRD and SEM. Results showed that pure cobalt powder with a

PT E

diameter of <100 nm was obtained at cathode No. 1, and WC powder with particle diameters of <100 nm was recycled successfully at cathode No. 2. And the result of

CE

XRF shows that the purity of the products both Co and WC were higher than 90

AC

percent. In summary, elemental cobalt and tungsten was separated successfully from WC–6 wt% Co scrap by controlling process parameters during electrolysis.

Acknowledgments

This research was financially supported by National Natural Science Foundation of China (No. 51422401).

16

AC

CE

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

17

ACCEPTED MANUSCRIPT References

[1] V. K. Sarin, D. Mari, and L. Llanes, Comprehensive Hard Materials. Spain: Elsevier (2014) 127-134.

PT

[2] E. Lassner, W. D. Schubert, Tungsten: properties, chemistry, technology of the elements, alloys, and chemical compounds, Springer Science & Business Media,

RI

(1999) 124-133.

SC

[3] Survey, U.S.G.: Washington, D. C: U.S. Geological Survey (2015) 174-175.

NU

[4] K. B. Shedd. Tungsten, 2007 Minerals Yearbook. Reston: U.S. Geological Survey (2009) 103-135.

Mater. Eng. 21 (2008) 28–33.

MA

[5] D. G. Ahn, Tungsten status and cemented carbide development, Trends Met.

PT E

D

[6] S. Wongsisa, P. Srichandr, N, Development of Manufacturing Technology for Direct Recycling Cemented Carbide (WC-Co) Tool Scraps, Mater. Trans. 56

CE

(2015) 70-77.

[7] K. T. Reilly, Recovery of refractory metal values from scrap cemented carbide,

AC

US Patent 4406866 (1983). [8] L. Luo, L.K. Jun, Recovery of tungsten and vanadium from tungsten alloy scrap, Hydrometallurgy 72 (2004) 1–8. [9] C. S. Freemantle, N. Sacks, M. Topic, and C. A. Pineda-Vargas, Impurity characterization of zinc-recycled WC-6% Co cemented carbides, Int. J. Refract. Met. Hard Mater. 44 (2014) 94-102. [10] S. Venkateswaran, W. D. Schubert, B. Lux, W-Scrap Recycling by the Melt Bath 18

ACCEPTED MANUSCRIPT Technique, Int. J. Refract. Met. Hard Mater. 14 (1996) 263-270. [11] C. Edtmaiera, R. Schiessera, C. Meissla, Selective removal of the cobalt binder in WC/Co based hardmetal scraps by acetic acid leaching, Hydrometallurgy 76 (2005) 63-71.

PT

[12] J. Chie, J. Y. Lin, S. P, Selective dissolution of the cobalt binder from scraps of

RI

cemented tungsten carbide in acids containing additives, Hydrometallurgy 43

SC

(1996) 47-61.

[13] J. C. Lee, E. Young, J. H. Kim, Recycling of WC–Co hardmetal sludge by a new

NU

hydrometallurgical route, Int. J. Refract. Met. Hard Mater. 29 (3) (2011) 365-371.

MA

[14] W. Gwang, Recovery of tungsten carbide from hard material sludge by oxidation and carbothermal reduction process, J. Ind. Eng. Chem. 20 (2014) 2384-2388.

D

[15] X. L. Xi, G.H. Si, Z. R. Nie, L. W. Ma, Electrochemical behavior of tungsten ions

PT E

from WC scrap dissolution in a chloride melt, Electrochim. Acta 184 (2015) 233-238.

CE

[16] A. J. Bard, L.R, Electrochemical methods: fundamentals and applications, Wiley,

AC

New York (1980).

[17] M.S. Krause, L. Ramaley, Analytical application of square wave voltammetry, Anal. Chem. 41 (1969) 1365–1369. [18] X. B. Wanga, W. Huanga, Y. Gong, Electrochemical behavior of Th(IV) and its electrodeposition from ThF4-LiCl-KCl melt, Electrochim. Acta 196 (2016) 286– 293. [19] R.W. Laity, J. D. McIntyre, Chronopotentiometric Diffusion Coefficients in Fused 19

ACCEPTED MANUSCRIPT Salts I. Theory1a, J. Am. Chem. Soc. 87 (1965) 3806–3812. [20] L. Cassayre, J. Serp, P. Soucek, Electrochemistry of thorium in LiCl–KCl eutectic melts, Electrochim. Acta 52 (2007) 7432–7437. [21] S. Vandarkuzhali, M. Chandra, S. Ghosh, Investigation on the electrochemical

PT

behavior of neodymium chloride at W, Al and Cd electrodes in molten LiCl-KCl

RI

eutectic, Electrochim. Acta 145 (2014) 86–98.

SC

[22] J. Wang. Analytical Electrochemistry. US: Wiley-VCH, 2006, 53-54. [23] G. Y. Kim, D. Y, S. Paek, Journal of Electroanalytical Chemistry, 2012, 682,

NU

128-135. A study on the electrochemical deposition behavior of uranium ion in a

MA

LiCl–KCl molten salt on solid and liquid electrode, J. Electroanal. Chem. 682

AC

CE

PT E

D

(2012) 128-135.

20

ACCEPTED MANUSCRIPT Picture Captions:

Fig. 1 Hermetically sealed three-electrode cell for electrochemical experiments. Fig. 2 Cyclic voltammograms of the platinum wire, WC, and WC–6 wt% Co anod in

PT

NaCl–KCl molten salt at 1023 K with a scan rate of 0.1 V s−1. Fig. 3 (a) Cyclic voltammograms obtained in blank NaCl–KCl molten salt (black line)

RI

at 1023 K and after 8 h of potentiostatic electrolysis at 0.4 V (vs. Ag/AgCl) potential

SC

(red line),Working electrode area: 0.12 cm2; (b) XRD pattern of the cathode product at

NU

0.4 V (vs. Ag/AgCl) potential.

Fig. 4 Square wave voltammograms of cobalt ions in NaCl–KCl melt after

MA

potentiostatic electrolysis at potential of 0.4 V (vs. Ag/AgCl). Working electrode area: 0.12 cm2.

PT E

D

Fig. 5 Cyclic voltammograms obtained in NaCl–KCl–1％CoCl2 molten salt on platinum wire at various scan rates from 0.025 V s−1 to 1.4 V s−1. Working electrode

CE

area: 0.12 cm2.

Fig. 6 CV plots in NaCl–KCl molten salt at 1023 K at a scan rate of 100 mVs−1 after 8

AC

h potentiostatic electrolysis under potentials 0.8 (black line) and 1.2 V (vs. Ag/AgCl) (red line). Working electrode area: 0.12 cm2. Fig. 7 CVs in NaCl–KCl melts at 1023 K after 8 h potentiostatic electrolysis under 0.8 V (vs. Ag/AgCl) potential at various scan rates from 0.025 V s−1 to 1.4 V s−1. Working electrode area: 0.12 cm2. Fig. 8 Dependence of cathode peak current (red line) and the half-wave potential (blue line) on the square root of the scan rate (v1/2) for Co2+/Co in NaCl–KCl melts at 21

ACCEPTED MANUSCRIPT 1023 K. Inset: Linear relationship between the value of |Ipa/Ipc| and the square root of scan rate (v1/2) for Co2+/Co in NaCl–KCl melts at 1023 K. Fig. 9 Dependence of cathode peak current (red line) and the half-wave potential (blue line) on the square root of scan rate (v1/2) for Wn+/W in NaCl–KCl melts at 1023

PT

K. Inset: Linear relationship between the value of |Ipa/Ipc| and the square root of scan

RI

rate (v1/2) for Wn+/W in NaCl–KCl melts at 1023 K.

SC

Fig. 10 (a) Square wave voltammograms in NaCl–KCl melt after potentiostatic

fitting. Working electrode area: 0.12 cm2.

NU

electrolysis at different potentials 0.4, 0.8, and 1.2 V (vs. Ag/AgCl); (b) Gaussian

MA

Table 1 Average number of exchanged electrons of cobalt and tungsten after potentiostatic electrolysis at different potentials

D

Fig. 11 Chronopotentiograms on a GC electrode in NaCl–KCl at 1023 K with

area: 0.21 cm2.

PT E

different current densities from −85 mA/cm2 to −124 mA/cm2. Working electrode

CE

Fig. 12 (a) Dependence of the iτ1/2 and current density for cobalt ions; (b)

AC

dependence of the iτ1/2 and current density for tungsten ions. Fig. 13 (a) Impedance spectra measured at 0.4 V (vs. Ag/AgCl) in the frequency range 1 Hz–100 kHz; (b) an equivalent electric circuit. Fig. 14 (a) Impedance spectra measured at 0.8 V (vs. Ag/AgCl) in the frequency range of 0.1 Hz–100 kHz; (b) an equivalent electric circuit; (c) a magnified view of the high-frequency region. Fig. 15 (a) Current–time curves during electrolysis at different potentials 0.4, 0.8, and 22

ACCEPTED MANUSCRIPT 1.2 V (vs. Ag/AgCl); (b) XRD pattern of the product obtained from cathode after potentiostatic electrolysis at different potentials 0.4, 0.8, and 1.2 V (vs. Ag/AgCl). Fig. 16 (a) XRD pattern of the product obtained from cathode Nos. 1 and 2 after potentiostatic electrolysis at potentials 0.4 and 0.8 V (vs. Ag/AgCl); (b) XRD pattern

PT

of the product obtained from cathode Nos. 1 and 2 after potentiostatic electrolysis at

RI

potentials 0.4 and 1.2 V (vs. Ag/AgCl); (c) SEM image of the product obtained from

SC

cathode Nos. 1 in Fig. 16 (a); (d) SEM image of the product obtained from cathode

AC

CE

PT E

D

MA

NU

No. 2 in Fig. 16 (a).

23

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

Fig. 1 Hermetically sealed three-electrode cell for electrochemical experiments.

24

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

MA

Fig. 2 Cyclic voltammograms of the platinum wire, WC, and WC–6 wt% Co anod in

AC

CE

PT E

D

NaCl–KCl molten salt at 1023 K with a scan rate of 0.1 V s−1.

25

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

MA

Fig. 3 (a) Cyclic voltammograms obtained in blank NaCl–KCl molten salt (black line) at 1023 K and after 8 h of potentiostatic electrolysis at potential of 0.4 V (vs. Ag/AgCl)

PT E

D

(red line); (b) XRD pattern of the cathode product at potential of 0.4 V (vs. Ag/AgCl).

AC

CE

Working electrode area: 0.12 cm2.

26

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

MA

Fig. 4 Square wave voltammograms of cobalt ions in NaCl–KCl melt after potentiostatic electrolysis at potential of 0.4 V (vs. Ag/AgCl). Working electrode area:

AC

CE

PT E

D

0.12 cm2.

27

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

MA

Fig. 5 Cyclic voltammograms obtained in NaCl–KCl–1％CoCl2 molten salt on platinum wire at various scan rates from 0.025 V s−1 to 1.4 V s−1. Working electrode

AC

CE

PT E

D

area: 0.12 cm2.

28

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

MA

Fig. 6 CV plots in NaCl–KCl molten salt at 1023 K at a scan rate of 100 mVs−1 after 8 h potentiostatic electrolysis under potentials 0.8 (red line) and 1.2 V (vs. Ag/AgCl)

AC

CE

PT E

D

(black line). Working electrode area: 0.12 cm2.

29

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

MA

Fig. 7 CVs in NaCl–KCl melts at 1023 K after 8 h potentiostatic electrolysis under 0.8 V (vs. Ag/AgCl) potential at various scan rates from 0.025 V s−1 to 1.4 V s−1.

AC

CE

PT E

D

Working electrode area: 0.12 cm2.

30

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

MA

Fig. 8 Dependence of cathode peak current (red line) and the half-wave potential (blue line) on the square root of the scan rate (v1/2) for Co2+/Co in NaCl–KCl melts at

PT E

D

1023 K. Inset: Linear relationship between the value of |Ipa/Ipc| and the square root of

AC

CE

scan rate (v1/2) for Co2+/Co in NaCl–KCl melts at 1023 K.

31

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

MA

Fig. 9 Dependence of cathode peak current (red line) and the half-wave potential (blue line) on the square root of scan rate (v1/2) for Wn+/W in NaCl–KCl melts at 1023

PT E

D

K. Inset: Linear relationship between the value of |Ipa/Ipc| and the square root of scan

AC

CE

rate (v1/2) for Wn+/W in NaCl–KCl melts at 1023 K.

32

AC

CE

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

Fig. 10 (a) Square wave voltammograms in NaCl–KCl melt after potentiostatic electrolysis at different potentials 0.4, 0.8, and 1.2 V (vs. Ag/AgCl), Working electrode area: 0.12 cm2; (b) Gaussian fitting. 33

ACCEPTED MANUSCRIPT Table 1 Average number of exchanged electrons of cobalt and tungsten after potentiostatic electrolysis at different potentials Redox species

Potential of potentiostatic electrolysis (V) 0.8

1.2

Co

1.9

2.18

2.16

W

—

2.13

PT

0.4

AC

CE

PT E

D

MA

NU

SC

RI

2.06

34

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

MA

Fig. 11 Chronopotentiograms on a GC electrode in NaCl–KCl at 1023 K with different current densities from −124 mA/cm2 to −86 mA/cm2. Working electrode

AC

CE

PT E

D

area: 0.21 cm2.

35

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

MA

Fig. 12 (a) Dependence of the iτ1/2 and current density for cobalt ions; (b) dependence

AC

CE

PT E

D

of the iτ1/2 and current density for tungsten ions.

36

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

MA

Fig. 13 (a) Impedance spectra measured at 0.4 V (vs. Ag/AgCl) in the frequency range

AC

CE

PT E

D

1 Hz–100 kHz; (b) equivalent electric circuit.

37

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

MA

Fig. 14 (a) Impedance spectra measured at 0.8 V (vs. Ag/AgCl) in the frequency range

AC

CE

PT E

high-frequency region.

D

of 0.1 Hz–100 kHz; (b) an equivalent electric circuit; (c) a magnified view of the

38

AC

CE

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

Fig. 15 (a) Current–time curves during electrolysis at different potentials 0.4, 0.8, and 1.2 V (vs. Ag/AgCl); (b) Potential–time curves during electrolysis at different potentials 0.4, 0.8, and 1.2 V (vs. Ag/AgCl); (c) XRD pattern of the product obtained 39

ACCEPTED MANUSCRIPT from cathode after potentiostatic electrolysis at different potentials 0.4, 0.8, and 1.2 V

AC

CE

PT E

D

MA

NU

SC

RI

PT

(vs. Ag/AgCl).

40

AC

CE

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

41

CE

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

Fig. 16 (a) XRD pattern of the product obtained from cathode Nos. 1 and 2 after

AC

potentiostatic electrolysis at potentials 0.4 and 0.8 V (vs. Ag/AgCl); (b) XRD pattern of the product obtained from cathode Nos. 1 and 2 after potentiostatic electrolysis at potentials 0.4 and 1.2 V (vs. Ag/AgCl); (c) SEM image of the product obtained from cathode Nos. 1 in Fig. 16 (a); (d) SEM image of the product obtained from cathode No. 2 in Fig. 16 (a).

42