Co-recovery of tungsten and lanthanum from photovoltaic tungsten-based busbars scrap by molten salt electrolysis

Abstract

As a rare-earth reinforced tungsten alloy, photovoltaic tungsten-based busbars have been extensively employed in silicon production in recent years. The urgent issue at hand is the need to achieve the co-recovery of tungsten and lanthanum from photovoltaic tungsten wire scrap. In this study, the complete recycling of photovoltaic tungsten scrap is achieved for the first time through a brief molten salt electrolysis process, utilizing Na₂WO₄ as the electrolyte and tungsten wire scrap as the anode. The process involves a graphite crucible as the cathode, ensuring continuous contact with the product to prevent the oxidation of tungsten powder (W) to sodium tungsten bronze (Na_xWO₃) by Na₂WO₄ electrolyte. The corundum anode basket promotes the dissolution of La₂O₃ in the electrolyte to form La₂(WO₄)₃. This compound actively participates in electrochemical reactions to facilitate the simultaneous production of W and La₂O₃. The regenerated tungsten powder possesses a uniform distribution of La₂O₃, making it a suitable raw material for the manufacture of photovoltaic tungsten-based busbars. Overall, this approach establishes a sustainable and environmentally friendly tungsten resource recycling ecosystem.

---

Green foundation

1. Photovoltaic tungsten filaments are typically processed through chemical recycling methods, which result in significant contaminant generation. These methods overlook the recovery of rare earth elements and necessitate the reintroduction of La precursors for the production of photovoltaic tungsten wires. In this study, a one-step molten salt electrolysis approach is employed to achieve the co-recovery of tungsten and La₂O₃ from photovoltaic tungsten wires, with no contaminant formation during the process.

2. One-step regeneration of photovoltaic tungsten wires by molten salt electrolysis is realized, and the prepared rare earth tungsten powder can be directly used as a raw material to produce photovoltaic tungsten wires, which shortens the recycling process and eliminates the pollution.

3. Future research should focus on optimizing current efficiency and advancing the development of continuous molten salt electrolysis equipment.

1 Introduction

Climate change, a global challenge, necessitates international cooperation to achieve the goals set by the Paris Agreement. According to “Global Market Outlook for Solar Power 2024–2028”, the world deployed 447 GW of new solar PV capacity in 2023, up from 239 GW in 2022, a 46% year-on-year increase. Total global solar power deployment is projected to exceed 2 TW in 2024 and reach 5.1 TW by 2028.¹ The burgeoning global solar photovoltaic (PV) market not only fosters the advancement of solar energy exploration but also accelerates the growth of downstream industries. The PV cutting busbar plays a crucial role in manufacturing silicon panels, where it must adhere to stringent requirements encompassing high strength (>5500 MPa) and ultra-fine dimensions (<36 μm).² Conventional high-carbon steel wire busbars struggle to meet these standards. Conversely, tungsten wires exhibit superior abrasion resistance, high strength, and low wire breakage rate, positioning them as the preferred alternative for busbar production. The substantial market potential has spurred a significant increase in production capacity, with the global annual capacity for tungsten wires now reaching 120 billion meters.^3,4 However, current production capacity still falls short of meeting the PV market's growing demand, necessitating further expansion in the future. As a result, a substantial volume of tungsten wire scrap is generated from photovoltaic applications, yet an effective recycling method remains lacking.

Photovoltaic tungsten wires are used as cutting tools and are scrapped upon breakage. To enhance their performance, tungsten is typically strengthened with rare-earth oxides.^5–7 Therefore, photovoltaic tungsten wire is actually a rare-earth-reinforced tungsten alloy. Rare-earth oxides are uniformly distributed in tungsten powder through physical or chemical methods, followed by sintering and thermal processing to form rare-earth tungsten alloys. The distribution of rare-earth oxide nanoparticles in the grains performs the role of pinning down and accumulating dislocations, enhancing the mechanical properties by the Orowan and pinning mechanisms (oxide dispersion strengthened).^8,9 Currently, the yield rate of PV tungsten wires is below 50%, meaning that every 3 tons of tungsten powder yields only 100 million meters of tungsten wire. Consequently, tungsten busbar scrap originates from two sources, production-end scrap and use-end scrap. The accumulation of waste photovoltaic tungsten wires is not only a waste of resources but also a potential risk of environmental pollution when improperly handled. The tungsten content of tungsten busbar is more than 99%, which is much higher than that of the original ore, and the recycling of photovoltaic tungsten busbar scrap can reduce the pollution caused by the original ore mining and smelting, reduce production costs, and promote the sustainable development of the photovoltaic industry. The challenge in reclaiming photovoltaic tungsten wires stems from the co-recovery of tungsten and rare earth oxides and it is difficult to directly convert PV tungsten busbar scrap into tungsten powder with uniformly dispersed rare-earth oxides. Moreover, the exceptional mechanical properties of tungsten wires hinder their regeneration through physical methods. The chemical recycling route stands out as the prevailing technology for recovering tungsten.¹⁰ Although the quality of the product prepared by this method is closest to that of the original ore, the process is cumbersome and does not comply with sustainable development. Furthermore, this recovery approach neglects rare-earth oxides, and incorporating the recovery of rare-earth elements would prolong the process and production cycle, leading to heightened environmental and economic burdens. Molten salt electrolysis, a sustainable and efficient tungsten recovery technique, has undergone extensive research and demonstrated success in recovering various tungsten products, such as tungsten (W), tungsten carbide (WC), and cemented carbide (WC-Co).^11–13 For example, Xiao et al. recovered cemented carbide scrap by double-cathode and two-stage electrolysis in NaCl–KCl molten salt and obtained cobalt and tungsten carbide, respectively.¹⁴ Xi et al. recovered WC-10Co in NaF–KF molten salt and utilized the potential difference to achieve tungsten and cobalt separation, obtaining particles less than 1 μm under 88 mA.¹⁵ Realization of the recovery of rare earth oxides in tungsten wires is another research priority. Depending on the existing tungsten wire scrap, the recyclable is La₂O₃. Conventional molten salt electrolytes use chloride or fluoride salts, and literature has been investigated on species and solubility of La₂O₃ in these electrolytes. La exists in molten FLiNaK as LaOF and La₂O₃ with a solubility of 6.81 × 10⁻⁴ wt% at 700 °C.¹⁶ The solubility of La₂O₃ could reach 8.71 wt% in molten salts of 50 wt% NaF-50 wt% (44 wt%NaCl + 56 wt%KCl).¹⁷ La indeed has solubility in the molten chlorine and fluorine, but achieving uniform distribution of La₂O₃ in tungsten powder during tungsten wire recovery is challenging. This unique distribution is expected when La³⁺ participates in electrochemical reactions and produces La₂O₃. Obviously, it is impossible in chlorine and fluorine salts, where La₂O₃ dissolves as LaOF, and the excess over the solubility becomes an anodic sludge. Hence, the primary focus of the research is to identify a suitable electrolyte and facilitate the recovery of tungsten and rare-earth oxides (La₂O₃) efficiently through streamlined processes.

In this study, we develop a recovery pathway for tungsten wires by employing Na₂WO₄ as electrolyte based on theoretical principles and experimental validation. This approach allows for the simultaneous recovery of tungsten and lanthanum from tungsten wire scrap. A cathodic protection strategy is used to avoid oxidation of tungsten powder by Na₂WO₄. The utilization of anode basket facilitates the dissolution of La₂O₃ to generate La₂(WO₄)₃. This compound can engage in the electrochemical reaction, enabling the concurrent production of tungsten powder and La₂O₃. The regenerated tungsten powder exhibits a homogeneous distribution of La₂O₃, making it suitable for direct use as a raw material in the manufacturing of photovoltaic tungsten wires. Ultimately, a sustainable ecological framework is established for the recovery of photovoltaic tungsten wires through molten salt electrolysis.

2 Experiment

2.1 Reagents and methods

The materials and reagents utilized are as follows: sodium tungsten (Na₂WO₄·2H₂O, 99.5%, Tianjin Kaida Chemical Factory, China), photovoltaic tungsten wire scrap (W, 99.75%, Xiamen Tungsten Co., Ltd, China), graphite Crucible (60 × 120 × 5 mm, 99.99%, Qingdao Baofeng Graphite Co., Ltd, China), Corundum anode basket (Al₂O₃, 20 × 300 × 3, 99.9%, Suzhou Kaifate Ceramic Technology Co., Ltd, China).

The electrolysis unit consists of photovoltaic tungsten wire anode, graphite crucible cathode and sodium tungstate electrolyte. Pre-treatment of photovoltaic tungsten wire scrap is necessary before it can be used as anode. The tungsten wire is first cleaned with pure water, after drying completely, weighed 20 g and then forged into a rectangular shape (65 × 10 × 6 mm) in a hydraulic press. The anode basket is the crucial point to achieving tungsten-lanthanum co-recovery. The corundum anode basket has a circular opening (∅ 6 mm) at 35 mm from the bottom to ensure ion transport and the photovoltaic tungsten wire scrap is placed in the basket. The graphite crucible serves as both cathode and electrolyte container, and the conductivity of the graphite crucible is achieved by a stainless-steel collector. Before configuring the electrolyte, sodium tungsten needs to be held in a vacuum oven at 200 °C for 48 h to remove water. After the electrolysis unit is assembled, it is placed in furnace (Thermcraft, Winstom-Salem, N.C. U.S.A), and the device needs to be evacuated to a vacuum before the temperature is raised. The electrolysis unit needs to be held at 573 K for 1 h to completely eliminate water interference, and then heated to 1073 K under a dry argon atmosphere. The current density used for the experiment is 15.76 mA cm⁻². Electrolysis duration is 3 h, unless otherwise specified. The product is rinsed with deionized water to remove electrolytes, then filtered through a 0.1 μm nylon microporous membrane and washed with ultrapure water until the electrolytes are completely removed. Finally, the product is completely dried in an oven at 100 °C.

2.1 Characterization

Thermodynamic calculations are performed using HSC Chemistry 6.0 software. The theoretical reduction potential is calculated using eqn (1).¹⁸

E = −ΔG/nF (1)

where n is the number of transfer electrons, F is the Faraday constant (96 485 C mol⁻¹), ΔG is Gibbs free energy, the parameter values are from HSC Chemistry.

The electrochemical window of the molten salt and the dissolution potential of tungsten are characterized by cyclic voltammetry (CV) and linear sweep voltammetry (LSV) on VersaSTAT 4 electrochemical workstation (Princeton Applied Research, Advanced Measurement Technology Inc., USA). Open circuit chronopotentiometry (OCP) and square wave voltammetry (SWV) are applied to characterize electrochemical reactions. The morphology of products is characterized by scanning electron microscopy (SEM, Sigma 300, Carl Zeiss Microscopy GmbH). The composition of products is characterized by Energy Dispersive Spectroscope (EDS), X-ray photoelectron spectroscopy (XPS, PHI5000 VersaProbe lll), X-ray diffraction (XRD, X'Pert PRO, Malvern Panalytical) and Glow Discharge Mass Spectrometry (GDMS, Element GD Plus GD-MS, Thermo Fisher Scientific Inc.). The XRD uses Co Kα radiation. Each sample is scanned from 10° to 110° and the data is compared with PDF data of MDI Jade 6.0 to determine the crystal shape. The XPS uses Al Kα radiation as the excitation source. The XPS data are processed on Avantage software and the C 1 s peak (284.8 eV) is used as a reference for the calibration of binding energies. The binding energy of elemental is based on the NIST XPS database and literature.^7,8,19,20

3 Results and discussion

3.1 Theoretical analysis and system construction

The traditional molten salt electrolysis method utilizes chloride or fluoride molten salt as electrolyte to electrochemically regenerate tungsten. The tungsten-based busbars are composed of W and La₂O₃. The regeneration of compliant tungsten powders must meet the requirement of uniform distribution of La₂O₃ in them. According to the theoretical analysis, only the presence of La₂(WO₄)₃ in the molten salt makes it possible to realize the co-production of tungsten and La₂O₃ (as reaction (2)), thereby achieving the desired.

La₂(WO₄)₃ → 3W + La₂O₃ + 4.5O₂ (2)

La₂(WO₄)₃ is not observed when using chloride/fluoride as an electrolyte. Therefore, only oxygenates such as carbonate, tungstate, sulfate, etc., can fulfill the requirements. Based on the principle of excluding impurity ions, the selection of electrolytes is constrained to tungsten oxygenates. Nonetheless, the defect of utilizing tungsten oxygenates electrolyte is that tungsten is easily oxidized to sodium tungsten bronze (Na_xWO₃) by the tungsten oxygenates, as in reactions (3) and (4).²¹

(2 − x)W + xNa₂WO₄ + (3 − 2x)O₂ → 2Na_xWO₃ (3)

xW + 3xNa₂WO₄ + (6 − 4x)WO₃ → 6Na_xWO₃ (4)

These reactions promote the anodic electrochemical dissolution of tungsten, but are detrimental to the cathodic preparation of tungsten powders. Due to the specific physicochemical traits of tungsten, the tungsten reduced cathodically tends to detach easily. Once detached, it undergoes oxidation to form Na_xWO₃ due to interaction with the electrolyte, resulting in compromised product purity and diminished current efficiency. Preventing the oxidation of the product is a critical issue that needs to be resolved when employing tungsten oxygenates as electrolytes. The reaction possibilities of La₂O₃ and tungsten oxygenates have been calculated, as shown in Fig. 1a. La₂O₃ is unable to react spontaneously with Na₂WO₄, which could lead to low solubility in Na₂WO₄. From the effect of concentration on the actual reduction potential, it can be seen that extremely low solubility prevents La₂(WO₄)₃ from participating in electrochemical processes. But La₂O₃ can react spontaneously with WO₃ or Na₂W₂O₇ to form La₂(WO₄)₃. This does not mean that Na₂W₂O₇ (Na₂WO₄–WO₃) is the ideal choice of electrolyte. The Na₂WO₄–WO₃ molten salt contains a high concentration of WO₃, which undoubtedly enhances the reaction (4), resulting in the final product that is not regenerated tungsten powder. Consequently, Na₂W₂O₇ proves unsuitable as an electrolyte.

Fig. 1 (a) Possible reactions of La₂O₃ in tungsten oxygenates, (b) the calculation of thermal stability of molten salts, (c) the theoretical potential for possible electrochemical reactions in molten salts, (d) the theoretical electrochemical reduction order at 800 °C, (e) comparison of chemical recycle route¹⁰ and molten salt electrolysis route.

Based on the above analysis, the only alternative electrolyte is Na₂WO_4. Enhancing the concentration of La₂(WO₄)₃ in the Na₂WO₄ electrolyte is pivotal for achieving co-recovery of W and La. Meanwhile, the dissolution of La₂O₃ is critically dependent on the Na₂W₂O₇ content in the electrolyte. Consequently, optimizing the concentration of Na₂W₂O₇ in the electrolyte while ensuring that the product is tungsten powder is the focus of this work. The thermal stability of the molten salt is further calculated as shown in Fig. 1b. La₂(WO₄)₃ and Na₂WO₄ have excellent thermal stability and do not decompose spontaneously. Na₂W₂O₇ spontaneously decomposes to form Na₂WO₄ and WO₃ at temperatures exceeding 800 °C. The ionic W₂O₇²⁻ reacts more readily with La₂O₃ than WO₃, thus the working temperature should be lower than the decomposition temperature of Na₂W₂O₇. The theoretical potential for possible electrochemical reactions in the electrolyte is further calculated, as in Fig. 1c. The theoretical potentials of reactions decrease with increasing temperature, which means that high temperatures are more favourable for electrochemical processes. The working temperature is chosen to be 800 °C in conjunction with calculations of the thermal decomposition of the molten salt. Fig. 1d shows the prioritization of reactions at operating temperature. At 800 °C, Na₂W₂O₇ exhibits the lowest reduction potential (−0.986 V), but Na₂W₂O₇ spontaneously binds La₂O₃ to form La₂(WO₄)₃. As a result, its impact on the anticipated sequence of electrochemical reactions is minimal. During the reduction process, La₂(WO₄)₃ preferentially produces W and La₂O₃ (−1.103 V) rather than W and La (−1.471 V). The theoretical reduction potential of Na₂WO₄ to W is −1.512 V, ensuring that electrolyte decomposition can be avoided prior to the formation of W and La₂O₃. Last but not least, the electrochemical window of Na₂WO₄ is tested using a tungsten rod as the counter electrode and a Pt wire as the reference and working electrode, and the cyclic voltammetry curve and linear voltametric curve is shown in Fig. 2a and b. Notably, sodium tungstate has an electrochemical window of −0.84 V to 1.38 V. A linear voltametric curve, as shown in Fig. 2c, signifies the dissolution potential of tungsten at −0.291 V, confirming the recovery feasibility within sodium tungstate. The aforementioned analysis provides a theoretical foundation for the recovery of photovoltaic tungsten-based busbars scraps through molten salt electrolysis. If proven feasible, this method has the potential to significantly streamline the existing chemical recycling route (Fig. 1e) and pave the way for the development of a rapid and environmentally friendly PV tungsten wire regeneration technology.

Fig. 2 Electrochemical characterization of Pt as working and reference electrodes and tungsten as counter electrode. (a) cyclic voltametric curves, (b) linear sweep voltammetry, electrochemical characterization of tungsten as working and counter electrode and Pt as reference electrode, (c) linear sweep voltammetry, (d) open circuit chronopotentiometry, (e) cyclic voltametric curves, (f) square wave voltammetry.

3.2 Validity of cathodic protection strategy

In the recovery of tungsten from Na₂WO₄ molten salts, oxidation reactions (reactions (3,4)) occur leading to product impurity and current efficiency loss. Although the theoretical basis for the recovery of tungsten and lanthanum from Na₂WO₄ has been constructed, the problem of oxidation of tungsten powder has not been solved, making it difficult to guarantee the purity of the product. Oxidation commences when tungsten detaches from the cathode, and preventing tungsten detachment is the key to avoiding oxidation. This study proposes a Cathodic Protection Strategy (CPS), involving the use of liquid metal or conductive crucibles as cathodes to prevent product flaking. This allows continuous contact between the product and cathode, minimizing oxidation risks.

In a preliminary investigation of the validity of CPS, the experiment is conducted using photovoltaic tungsten wire scrap as anode, and a comparison is made between graphite crucibles cathode and tungsten rods cathode. Photos of the molten salt and products obtained after the experiment are visualized in Fig. S1.† The XRD patterns of the product are display in Fig. 3. The product when using a tungsten rod cathode is a mixture of tungsten powder and sodium tungsten bronze (Fig. 3a), which suggests that oxidation occurs when the tungsten powder is detached from the cathode. The SEM image of the product is shown in the Fig. 3b, and its morphology is a fragment-type structure of varying sizes. The elements Na, W, and O can be clearly observed on the EDS spectra, further proving that the product is a mixture of sodium tungsten bronze and tungsten powder. To further understand the oxidation process of tungsten, the open circuit chronopotentiometry of tungsten is tested, such as Fig. 2d. Platform A is the reduction potential during electrolysis and corresponds to the WO₄²⁻ deposition process. At the end of electrolysis, the voltage is polarized from platform A to platform B and stabilized for about 20 s, which corresponds to the oxidation of tungsten to sodium tungsten bronze. La₂O₃ in the tungsten wires is not detected in the SEM image of the product, and combined with the GDMS results (Table S1†) of the product, suggests that the product does not contain lanthanum. The La₂O₃ remains dissolved in the electrolyte in the form of La₂(WO₄)₃, as a large amount of sodium tungsten bronze is present in the electrolyte. In contrast, the product is pure tungsten powder when using a graphite crucible cathode (Fig. 3c). The morphology of tungsten powder is shown in Fig. 3d, which exits in stick and sphere shape. The morphology of tungsten obtained from molten salt electrolysis is usually sphere-shaped,^22,23 whereas the stick-shaped tungsten obtained in this experiment may be caused by electrochemical dissolution of the tungsten wire during the electrolysis process. The comparative experiment is carried out using tungsten rod and wires as anode, and the products obtained were all tungsten spheres, as shown in Fig. S2a and S2b,† which proves that stick-tungsten comes from the electrochemical dissolution process of tungsten wires. Unfortunately, this tungsten of stick morphology can have a negative effect on the performance of recycled tungsten powder, as it has a relatively large size. Therefore, it is necessary to find a solution to remove this stick-shaped tungsten in subsequent experiments. The GDMS result of tungsten powder is given in Table S1,† which contains about 0.29 wt% La. The La₂O₃ observed in Fig. 3d is present as particles in the tungsten powder and does not develop a uniform distribution. Combined with theoretical analyses, it can be seen that La₂O₃ can form La₂(WO₄)₃ only if the molten salt contains WO₃ or Na₂W₂O₇. This part of the La₂O₃ is not produced by electrochemical reduction, but is formed by agglomeration of La₂O₃ during dissolution of the anode. The desired product should be a uniform distribution of La₂O₃ in the tungsten powder, thus the strategy in the theoretical analysis needs further verification. The above results show that the oxidation of tungsten powder can be inhibited by avoiding it detaching from the cathode, which is a preliminary proof of the validity of CPS. However, the graphite crucible will provide a portion of the carbon, so that the possibility of a carbothermal reduction reaction between carbon and sodium tungsten bronze cannot be excluded, and the possibility of obtaining pure tungsten powder from such a reaction also exists.²⁴

Fig. 3 (a) XRD patterns of the product by using tungsten rod cathode, (b) SEM image and EDS mapping of the product by using tungsten rod cathode, (c) XRD patterns of the product by using graphite crucible cathode, (d) SEM image and EDS mapping of the product by using graphite crucible cathode.

Therefore, the experiment is carried out using a nickel metal crucible as the cathode, as shown in Fig. S3,† which product remains pure tungsten powder. Repeated experiments are continued for 1.5, 3, 6, and 8 h. The XRD patterns of the products are all pure tungsten powder, as shown in Fig. S4,† which further excludes the effect of carbothermal reaction and proves the effectiveness of the CPS. After confirming the effectiveness of the CPS, electrochemical characterization is necessary to gain a deeper understanding of the reaction process, such as Fig. 2e and f. The number of electrons transferred is obtained by fitting the SWV and calculating viaeqn (5).²⁵

W_1/2 = 3.52RT/nF (5)

where W_1/2 is the halfwidth of the peak, R is the ideal gas constant, and F is the Faraday constant (96 485 C mol⁻¹).

It is clear from the CV that the reduction of WO₄²⁻ occurs in two steps, while the SWV fitting suggests that WO₄²⁻ is a six-electron process with a two-step reduction, with two electrons in the first step and four electrons in the second step (W⁶⁺ → W⁴⁺ → W), which is consistent with literature reports and in line with the theoretical analysis (Fig. 1d).²⁶ Finally, the current efficiency when using graphite crucible cathode is calculated as 90.02% based on the amount of product.

3.3 Overall recycling of photovoltaic tungsten wires

3.3.1 Co-recovery of W and La. The effectiveness of CPS has been demonstrated, but a new issue has emerged. Photovoltaic tungsten wires, due to their ultrafine characteristics, result in fragmentation into stick-shaped tungsten during electrochemical dissolution with larger grain sizes, which negatively impacts the performance of tungsten powder. It remains to be experimentally explored how to remove this stick-tungsten and distribute the La₂O₃ uniformly in the tungsten powder. Moreover, when tungsten rods were utilized as cathodes, the resulting product is a mixture of Na_xWO₃ and W, while lanthanum (La) remained in the molten salt. This finding suggests that the presence of sodium tungsten bronze enhances the dissolution of La₂O₃.

In the subsequent experiments, the structure of the electrolysis unit is modified, as illustrated in Fig. 4a. In this configuration, the photovoltaic tungsten wire is positioned within the anode basket, a design that facilitates the collection of sodium tungsten bronze generated through thermochemical reactions at the anode (reactions (3) and (4)), thereby promoting the dissolution of La₂O₃. After changing the structure of the electrolysis unit, the XRD spectra of the product are shown in Fig. 4b. The XRD plot reveals that the utilization of the anode basket does not alter the phase of the sample, with the resulting product still being tungsten powder. The regenerated tungsten powder morphology and elemental distribution are further characterized, as shown in Fig. 4f and Fig. S2c.† The SEM demonstrates the absence of stick-shaped tungsten in the samples, with all tungsten shapes appearing spherical, indicating the effective removal of larger grain sizes by the anode basket. This further illustrates that the stick-shaped tungsten originates from the electrochemical dissolution process of the photovoltaic tungsten wires. After employing anode basket, the stick-shaped tungsten was collected and digested in further thermochemical reaction (reaction (3)). The EDS mapping (Fig. 4f) reveals that the regenerated tungsten powder is composed of W, La, and O elements. Notably, La is uniformly distributed in the W, in alignment with previous studies on the synthesis of rare-earth tungsten alloys.^7,8,20Fig. 4c, d, and e shows XPS spectra of W 4f, La 3d, and O 1s. Each W, La, and O peaks can be clearly observed, and binding energy is corrected in 284.8 eV. Metallic W (W⁰) is depicted at 33.44 eV and 31.25 eV in Fig. 4c, and the peaks at 38.01 eV and 35.86 eV are corresponding to W⁶⁺, indicating the oxidation of a proportion of the tungsten.²⁷ This oxidation is not reflected in XRD, but areas of oxidized tungsten are clearly observed on EDS mapping (Fig. 4f). The peaks at 835.42 eV and 851.98 eV in Fig. 4d are attributed to La 3d_5/2 and La 3d_3/2, displaying the typical characteristics of La³⁺ (La₂O₃) with an interval ΔE of 16.56 eV.⁸ Additionally, the peaks at 839.22 eV and 855.76 eV are attributed to La multiplet-split peaks.^19,28 Combined with the O 1s peaks in Fig. 4e, it can be concluded that La exists as La₂O₃ in tungsten powder. The above characterization results indicate that La₂O₃ is uniformly distributed in the regenerated tungsten powder. This desirable product has been proven to be a suitable raw material for manufacturing photovoltaic tungsten wires and proves the success of the molten salt electrolysis method.

Fig. 4 (a) Schematic description of the electrolysis unit, (b) XRD pattern of product, XPS spectra at (c) W 4f, (d) La 3d, (e) O 1s, (f) SEM and EDS mapping of product.

In order to further validate the molten salt electrolysis method for the overall recovery of photovoltaic tungsten wires, the products at different electrolysis durations are studied (Fig. S5†). There is no La distribution in the tungsten at 1 h and 2 h (Fig. S5a and S5b†). This absence can be attributed to the concentration of La₂(WO₄)₃. According to the Nernst equation and Raoult's law (reactions (6) and (7)), concentration influences the actual reduction potential. Lower concentration will make the actual reduction potential rise, resulting in reduction reaction that is less competitive.

E_Ac = E° + (RT/nF)ln α₁/α₂ (6)

α_A = γ_Ax_A (7)

where E_Ac is actual reduction potential, E° is theoretical reduction potential, α_A is ionic activity, γ_A is activity coefficient and x_A is molar concentration.

Notably, the electrolyte can be recycled during industrial applications, thus portion of the lanthanum loss can be ignored. The continuous accumulation of La₂O₃ during electrolysis may result in a failure to achieve uniform distribution. The electrolysis duration is further extended from 3 h to 8 h and the products are characterized as in Fig. S5c.† La₂O₃ remains uniformly distributed in the tungsten powders. This is mainly due to the differences between La₂O₃ generation and the conventional electrochemical reduction process. The electrochemical reduction process involves the transfer of electrons, but there is no transfer of electrons in the production of La₂O₃. The reduction potential for La₂O₃ to La is −2.573 V, La₂(WO₄)₃ to La₂O₃ is −1.103 V, and La₂(WO₄)₃ to La is −1.471 V. These values indicate that the process of generating La monomers is not thermodynamically favorable or competitive. The reduction of WO₄²⁻ to W produces a large amount of O²⁻. At this point, La³⁺ combines with O²⁻ to produce La₂O₃. The ultimate manifestation is the co-deposition of tungsten and La₂O₃. The presence of tungsten powder will limit the La₂O₃ growth. Furthermore, the key to the dissolution of La₂O₃ is the presence of Na₂W₂O₇. Na₂W₂O₇ exhibits the lowest reduction potential (−0.986 V), which precludes its presence in the cathode. Therefore, the produced La₂O₃ will not revert to its ionic state. In addition, La is derived exclusively from tungsten wires, and the anodic dissolution, mass transfer, and electrochemical deposition are simultaneous, thus La₂O₃ can be consistently produced in the tungsten powder.

Ultimately, the current efficiency, estimated by the quantity of products, stands at 46.42%, demonstrating a notable decrease compared to not using anode baskets. This decline primarily arises from mass transfer constraints by anode basket.

3.3.2 La₂O₃ migration pathway. The experimental results presented above demonstrate that tungsten (W) and La₂O₃ in photovoltaic tungsten wires can be efficiently co-recovered in a single-step process via molten salt electrolysis. The regenerated tungsten powder exhibits a homogeneous distribution of rare-earth oxides and can be directly utilized as a raw material in tungsten wire production. The migration path of La₂O₃ requires further clarification.

In the absence of an anode basket, La₂O₃ detaches as anode sludge during electrolysis and eventually mixes with regenerated tungsten powder. However, when an anode basket is employed, La₂O₃ participates in the electrochemical process and is uniformly distributed in the regenerated tungsten powder. The structure of the electrolysis unit significantly influences the migration behavior of La. Specifically, the anodic dissolution of tungsten wires involves two primary mechanisms. The first is electrochemical dissolution, producing WO₄²⁻(W + 4O²⁻ → 6e⁻ + WO₄²⁻). Concurrently, O₂ is generated at the anode (O²⁻ → O₂ + 2e⁻). The second mechanism is thermochemical dissolution, where tungsten wire reacts with oxygen and sodium tungstate (W + O₂ + Na₂WO₄ → Na_xWO₃), forming Na_xWO₃. Upon modification of the electrolysis unit structure, the Na_xWO₃ is confined in the anode basket. At the end of electrolysis, the presence of Na_xWO₃ can be detected in the anode basket (Fig. 5b and Fig. S6†). Na_xWO₃, as an intermediate state oxide, can undergo further oxidized by the O₂ to form Na₂WO₄ and WO₃.²⁴ These compounds, Na₂WO₄ and WO₃, will spontaneously combine to yield Na₂W₂O₇. Analysis of the molten salt composition in the anode basket reveals a distinct characteristic peak corresponding to Na₂W₂O₇ (Fig. 5b), thereby confirming its presence. Under these conditions, La₂O₃ and Na₂W₂O₇ will spontaneously react to form La₂(WO₄)₃, which then migrate to the cathode and participates in the electrochemical reaction. This process enables the co-recovery of tungsten (W) and La₂O₃. A clear graphical representation of the La migration pathway is provided in Fig. 5a.

Fig. 5 (a) Schematic illustration and reaction equation of tungsten-lanthanum co-recovery process, (b-1) XRD patterns of molten salts in anode basket, (b-2) photo of anode basket after experiment (cross-section), (c) schematic diagram of molten salt electrolysis route for constructing tungsten resource recycling ecology.

The conventional chemical recycling route depicted in Fig. 1e involves numerous steps such as oxidation, digestion, calcination, and reduction to yield recycled tungsten powder, along with the emission and disposal of pollutants.¹⁰ In sharp contrast, the inputs to the molten salt electrolysis route are tungsten wire scrap, electricity, and heat, the molten salt electrolyte can be recycled, and the output is only the product, thus no pollutants will be generated by the process. In terms of the development and utilization of solar energy, the molten salt electrolysis route can be a perfect match with solar-thermal and solar-electricity, which can build a green and pollution-free tungsten resource recycling ecology, as shown in the Fig. 5c. Energy is a primary factor constraining the practical application of molten salt electrolysis, according to the trend of China's electric power structure, the proportion of green electricity is increasing.²⁹ As a result, the potential for industrial implementation of the molten salt electrolysis route is steadily expanding.

4 Conclusions

A comprehensive theoretical and experimental method for recycling waste photovoltaic tungsten-based busbars has been established. The study demonstrates that using a tungsten rod as the cathode produces a mixture of tungsten powder and sodium tungsten bronze due to the oxidation of tungsten powder by Na₂WO₄ after detachment from the cathode. The employment of a graphite crucible as the cathode effectively prevents this oxidation, yielding regenerated tungsten powder with an impressive current efficiency of 90.02%. Furthermore, the findings highlight the pivotal role of La₂(WO₄)₃ in the co-recovery of W and La₂O₃. La₂(WO₄)₃ actively participates in the electrochemical reaction, facilitating the simultaneous formation of W and La₂O₃. The use of an anode basket enhances the dissolution of La₂O₃ to form La₂(WO₄)₃. But it also hinders ion transport, reducing the current efficiency to 46.42%. In the absence of anode basket, the La₂O₃ falls off in the form of anode sludge. The product exhibits a uniform distribution of La₂O₃ if the electrolysis duration exceeds 3 h, with the La content closely matching that of the waste tungsten wires. Therefore, the regenerated tungsten powder can be directly utilized as raw material for producing photovoltaic tungsten wires.

Author contributions

Conceptualization: Xiaoli Xi, Zuoren Nie; data curation: Xiang Xue, Shuijie Su, Liwen Zhang; funding acquisition: Xiaoli Xi, Chunjia Liu, Qi Fang, Zuoren Nie; project administration: Liwen Zhang, Chunjia Liu; writing – original draft: Xiang Xue; writing – review & editing: Xiang Xue, Liwen Zhang; supervision and visualization: Liwen Zhang, Xiaoli Xi, Qi Fang.

Data availability

Data will be available upon request.

Conflicts of 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.

Acknowledgements

This work was supported by the National Key Research and Development Program of China (No. 2023YFB3811800), the National Natural Science Foundation of China for Distinguished Young Scholar (No. 52025042) and the Natural Science Foundation of Fujian Province (No. 2023J06041).

References

1. SolarPower Europe ( 2024): Global Market Outlook for Solar Power 2024-2028.
2. Y. Miao, Z. Wu and D. Wang, Preparation and mechanical properties of rare earth reinforced tungsten alloy for photovoltaic cutting, China Molybdenum Ind., 2023, 47, 57–60.
3. Y. Zhang and K. Wu, Photovoltaic diamond wire iteration: tungsten wire market penetration increases, China Business Journal, 2024, B15 DOI:10.38300/n.cnki.nzgjy.2024.000668.
4. Tungsten Diamond Wire Penetration Rate Speeds Up, Cost Reduction and Cost-Effectiveness Inflection Point Has Arrived, https://www.hangyan.co/reports/3328504000210273819.
5. Y. Chen, Y. C. Wu, F. W. Yu and J. L. Chen, Study on structure and property of tungsten alloy strengthened with dispersed La₂O₃, Rare Met. Mater. Eng., 2007, 36, 822–824.
6. M. Mabuchi, K. Okamoto, N. Saito, M. Nakanishi, Y. Yamada, T. Asahina and T. Igarashi, Tensile properties at elevated temperature of W-1%La₂O₃, Mater. Sci. Eng., A, 1996, 214, 174–176.
7. D. Wang, Z. Hao, P. Wang, R. Ma, Y. Shu and J. He, Preparation of spherical tungsten powder with uniform distribution of lanthania by plasma spheroidization, Int. J. Refract. Met. Hard Mater., 2023, 117, 106400.
8. E. S. Lee, Y. J. Heo, Y. I. Lee, Y. K. Jeong and S. T. Oh, Microstructural characterization and densification of W-1 wt% La₂O₃ composite powders synthesized by ultrasonic spray pyrolysis, J. Mater. Sci., 2022, 57, 17946–17956.
9. L. Huang, L. Jiang, T. D. Topping, C. Dai, X. Wang, R. Carpenter, C. Haines and J. M. Schoenung, In situ oxide dispersion strengthened tungsten alloys with high compressive strength and high strain-to-failure, Acta Mater., 2017, 122, 19–31.
10. Wolf-Dieter Schubert, Recycling of Tungsten – The technology – History, State of the Art and Peculiarities, 2019.
11. L. Zhang, X. Xi and Z. Nie, A new method for preparation of tungsten carbide powder by in situ electrochemical reduction, Electrochem. Commun., 2022, 134, 107179.
12. T. Oishi, Recycling of Tungsten by Molten Salt Process, Cham, 2022.
13. L. Zhang, Z. Nie, X. Xi, L. Ma, X. Xiao and M. Li, Electrochemical Dissolution of Tungsten Carbide in NaCl-KCl-Na₂WO₄ Molten Salt, Metall. Mater. Trans. B, 2017, 49, 334–340.
14. X. Xiao, X. Xi, Z. Nie, L. Zhang and L. Ma, Direct Electrochemical Preparation of Cobalt, Tungsten, and Tungsten Carbide from Cemented Carbide Scrap, Metall. Mater. Trans. B, 2016, 48, 692–700.
15. X. Xi, Q. Liu, Z. Nie, M. Li and L. Ma, Electrochemical preparation of tungsten and cobalt from cemented carbide scrap in NaF–KF molten salts, Int. J. Refract. Met. Hard Mater., 2018, 70, 77–83.
16. Q. Yang, J. Ge, Y. Wang and J. Zhang, Electrochemical Separation of Lanthanum Oxide in Molten FLiNaK Salt, Nucl. Technol., 2020, 206, 1769–1777.
17. P. Jang, H. Li, W. Kim, Z. Wang and F. Liu, Preparation of Al-La Master Alloy by Thermite Reaction in NaF-NaCl-KCl Molten Salt, JOM, 2015, 67, 1130–1136.
18. X. Xue, D. Gu, X. Xia, Y. Qin, H. Wu, H. Jiang, T. Jiang and B. Wang, Solar bipolar cell facilitating carbon dioxide for efficient coproduction of electricity and syngas, J. Power Sources, 2022, 521, 230985.
19. J. P. H. Li, X. Zhou, Y. Pang, L. Zhu, E. I. Vovk, L. Cong, A. P. van Bavel, S. Li and Y. Yang, Understanding of binding energy calibration in XPS of lanthanum oxide by in situ treatment, Phys. Chem. Chem. Phys., 2019, 21, 22351–22358.
20. P. Hu, X. Gong, H. Liu, W. Zhou and J. Wang, A novel dissolution-precipitation strategy to accelerate the sintering of yttrium oxide dispersion strengthened tungsten alloy with well-regulated structure, J. Mater. Sci. Technol., 2024, 184, 43–53.
21. L. Bartha, A. B. Kiss and T. Szalay, Chemistry of tungsten oxide bronzes, Int. J. Refract. Met. Hard Mater., 1995, 13, 77–91.
22. L. Zhang, Z. Nie, X. Xi and L. Ma, Electrochemical separation and extraction of cobalt and tungsten from cemented scrap, Sep. Purif. Technol., 2018, 195, 244–252.
23. G. Si, X. Xi, Z. Nie, L. Zhang and L. Ma, Preparation and characterization of tungsten nanopowders from WC scrap in molten salts, Int. J. Refract. Met. Hard Mater., 2016, 54, 422–426.
24. W. Qiu, Y. Liu, J. Ye, H. Fan and G. Wang, Molten salt synthesis and growth mechanism of WC platelet powders, Powder Technol., 2017, 310, 228–233.
25. J. Zhang, L. Zhang, X. Xi and Z. Nie, Electrochemical separation technology and mechanism of tungsten and cobalt in Na₂WO₄-WO₃-CoO molten salts, Sep. Purif. Technol., 2024, 330, 125270.
26. X. Wang, L. Zhang, L. Ma, X. Xi and Z. Nie, Multiscale modeling of Na₂WO₄–WO₃–CoO melt diffusion layer for molecular dynamics combined with finite element analysis, Mater. Chem. Phys., 2024, 311, 128589.
27. F. Y. Xie, L. Gong, X. Liu, Y. T. Tao, W. H. Zhang, S. H. Chen, H. Meng and J. Chen, XPS studies on surface reduction of tungsten oxide nanowire film by Ar+ bombardment, J. Electron Spectrosc. Relat. Phenom., 2012, 185, 112–118.
28. S. N. Ogugua, H. C. Swart and O. M. Ntwaeaborwa, The influence of post-preparation annealing atmospheres on the optical properties and energy transfer between Pr³⁺ and Dy³⁺ in mixed lanthanum-yttrium oxyorthosilicate hosts, Opt. Mater., 2018, 76, 125–140.
29. S. Jing, R. Sheng, X. Liang, D. Gu, Y. Peng, J. Xiao, Y. Shen, D. Hu and W. Xiao, Overall Carbon-neutral Electrochemical Reduction of CO₂ in Molten Salts using a Liquid Metal Sn Cathode, Angew. Chem., Int. Ed., 2023, 62, e202216315.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5gc00126a

---