Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

# Electrodeposition of crystalline silicon films from silicon dioxide for low-cost photovoltaic applications

## Abstract

Crystalline-silicon solar cells have dominated the photovoltaics market for the past several decades. One of the long standing challenges is the large contribution of silicon wafer cost to the overall module cost. Here, we demonstrate a simple process for making high-purity solar-grade silicon films directly from silicon dioxide via a one-step electrodeposition process in molten salt for possible photovoltaic applications. High-purity silicon films can be deposited with tunable film thickness and doping type by varying the electrodeposition conditions. These electrodeposited silicon films show about 40 to 50% of photocurrent density of a commercial silicon wafer by photoelectrochemical measurements and the highest power conversion efficiency is 3.1% as a solar cell. Compared to the conventional manufacturing process for solar grade silicon wafer production, this approach greatly reduces the capital cost and energy consumption, providing a promising strategy for low-cost silicon solar cells production.

## Introduction

Currently, global energy generation still depends strongly on fossil fuels1. Driven by the rapidly increasing energy demands and the negative environmental impact of fossil fuels, renewable energy has attracted tremendous attention in recent decades. Solar cells, utilizing sunlight to generate electricity directly, have been recognized as one of the most promising technologies for solving the energy issues1,2,3,4,5,6,7,8,9,10,11,12. Crystalline-silicon solar cells have dominated the photovoltaics market for the past several decades and are most likely to continue to be the primary technology for the photovoltaics industry in the future due to its abundant raw materials supply and non-toxicity1,6. To make silicon-based solar cells more competitive, improving the power conversion efficiency and decreasing the module costs are the most direct routes. For efficiency enhancement, innovative cell architectures involving complex processing procedures are usually required, often resulting in increased overall cost13. Reducing the silicon production cost and silicon material usage thus offers an alternative route for continued growth of crystalline-silicon photovoltaics technology in the future.

One of the largest contributions to overall module manufacturing cost still comes from silicon wafer production, involving complex processing and intensive energy consumption due to the high temperature requirement for silicon crystallization process13. To address this problem, direct production of silicon at low temperature in liquid/molten salts has been proposed and intensively investigated since 1980’s14,15,16,17,18,19,20,21,22,23,24,25,26. The major challenge for the molten salt technology, of which fluoride-based molten salt is dominant, is impurity control, due to the nature of fluoride-based molten salts and other multicomponent eutectic molten salts systems. Until now there has been no demonstration of a photovoltaic effect for a silicon film electrodeposited in fluoride-based molten salts. Therefore, chloride molten salt has been considered to be a promising alternative molten salt for silicon electrodeposition in recent years25,27,28,29.

In this work, we report the successful demonstration of a direct molten salt electrodeposition process of high purity (99.99989% (close to 6N)) crystalline silicon films in molten calcium chloride from abundant and inexpensive silicon dioxide. Calcium oxide is used as intermediate for the continuous ionization of silicon dioxide to form silicate ions in the molten salt. The doping type, either p-type or n-type, can be controlled by varying the dopants in the molten salt. Solar cell devices based on the as-prepared silicon films exhibit clear photovoltaic effects, with power conversion efficiency around 3.1%. This technique provides a promising approach for low-cost silicon solar cells production and potentially for high quality crystalline silicon film production for other applications.

## Results

### Design of electrodeposition of crystalline silicon films

Silicon dioxide is the primary source for silicon production. However, its solubility in chloride-based molten salts is generally low, inadequate for efficient electrodeposition20,21,22,25. Inspired by aluminum electrolysis in molten salt, efforts have been put into finding the right intermediate to facilitate the dissolution of silicon dioxide in chloride-based molten salts. Thanks to the considerable solubility of calcium oxide in molten calcium chloride and its reaction with silicon dioxide (Supplementary Note 1), the dissolution process from silicon dioxide to silicate ions is possible28,29, which offers access to the electrodeposition of high quality silicon films in molten calcium chloride.

As shown in Fig. 1, the only input materials for this molten salt electrodeposition of crystalline silicon films are abundant and low-cost silicon dioxide, calcium oxide and calcium chloride. Additional experimental details for the production of high-purity silicon films and reaction mechanisms can be found in Methods, Supplementary Tables 1 and 2, and Supplementary Figs. 16. Calcium oxide is added as an intermediate for continuing ionization of silicon dioxide to form silicate ions (expressed as SiOyn, including SiO32−, SiO44−, etc.), which are then electrodeposited onto substrates to form crystalline silicon films. The general reactions for the electrodeposition process can be simply expressed as follows:

$$x\mathrm{SiO}_2 + y\mathrm{CaO}\left( {\mathrm{Ca}^{2 + },\mathrm{O}^{2-}} \right) \to \mathrm{Ca}_y\mathrm{Si}_x\mathrm{O}_{(2x + y)}\left( {y\mathrm{Ca}^{2 + },\mathrm{Si}_x\mathrm{O}_{(2x + y)}^{2y-}} \right)$$
(1)
$$\mathrm{Si}_x\mathrm{O}_{(2x + y)}^{2y-} + 4xe^- \to x\mathrm{Si} + \left( {2x + y} \right)\mathrm{O}^{2-}$$
(2)

As shown in reactions (1) and (2) and Fig. 1b, the electrodeposition route is a cyclic reaction process. By periodically feeding silicon dioxide into the molten salt, crystalline silicon films can be produced continuously, which makes this method suitable for large scale production. It has been proved that various dopants can be added into molten salt, such as boric anhydride or alumina for p-type and antimony oxide or calcium phosphate for n-type, to control the doping type of deposited silicon films. In addition, a proof-of-concept demonstration of p-n junction formation all by molten salt electrodeposition is shown in Fig. 1b, c and our recent work28. Therefore, crystalline silicon films with tunable film thickness and doping type can be facilely electrodeposited. Similar to the aluminum electrolysis process, this one-step molten salt electrodeposition process offers the potential to dramatically reduce the cost of silicon products.

### Characterization of crystalline silicon films

Silicon dioxide, calcium oxide and calcium chloride mixtures were first homogenized to form a molten electrolyte at 850 °C, and then silicate ions were gradually generated during the subsequent dissolution process. The silicate ions (denoted as SiOyn, e.g. SiO32−, SiO44−, etc.) can be reduced to form silicon film on graphite substrates at approximately −1.5 V, as confirmed by CV results shown in Fig. 2a. By controlling the various dopants, either p-type and n-type silicon films can be produced. The CV results in Supplementary Fig. 7 reveal that the electrodeposition of n-type silicon film is also a simple reduction process and is similar to that of the p-type silicon films (Fig. 2a). We have demonstrated experimentally that crystalline silicon film can be deposited by using either constant potential/current density electrodeposition method or a pulse electrodeposition method. Pulse electrodeposition can yield dense and homogeneous silicon films formation, mainly due to the homogeneous silicate ion concentration at the cathode surface, which has been confirmed in our recent work28,29. The representative potential/current-time curves of the electrodeposition processes are shown in Supplementary Fig. 8. It is worth noting that calcium co-deposition would occur approximately at −2.7 V and calcium chloride would decompose at −3.2 V29. Therefore, the electrodeposition processes are all controlled at a potential less than −2.6 V. In addition, current density strongly influences the formation of dense films, with 15 to 20 mA cm−2 being the optimal current density. Higher or lower current densities will result in the formation of silicon powders and nanowires, respectively, as illustrated in Supplementary Fig. 9. By controlling current density, compact silicon films including p-n junction silicon films can be readily produced (Supplementary Figs. 1012). The film thickness can be controlled in a range of about 5 µm to more than 60 µm, on various substrates, including graphite, silicon wafers and others27,28,29,30. The X-ray diffraction patterns, as shown in Fig. 2b, confirm the good crystallinity of as-prepared films. As studied in previous work31, Ti, Cu, Ni, Cr, and Fe are impurities having the most harmful impact on crystalline silicon solar cells, thus concentrations of these impurities were analyzed by glow discharge mass spectrometry (GDMS), as shown in Fig. 2c and Supplementary Fig. 13. It is clear to see that all the impurities levels in the electrodeposited silicon are below the tolerable threshold. The overall purity based on full spectrum GDMS analysis is calculated to be 99.99989% (close to 6N, solar grade). To the best of our knowledge, this is the highest purity yet reported for electrodeposited silicon in molten salts. For the doped silicon films, the dopant concentrations of P for n-type, and Al for p-type were characterized to be 3.5 and 10 ppm, respectively. We note that the impurity control is crucial for the electrodeposition of high-purity silicon films (Supplementary Figs. 35). Periodical pre-electrolysis process (about 120 h) was used to purify the molten salt to achieve an ultra-purified system, and then metallic impurities contained in the deposited silicon films can be strictly controlled at low level (e.g., magnesium less than 0.05 ppm, tungsten less than 0.05 ppm, sodium less than 0.05 ppm, calcium is about 5 ppm, etc.). In addition, to decrease the generation of carbon dioxide gas and influence the silicon films during electrodeposition, it has been experimentally demonstrated that the atmosphere needs to be strictly controlled by high-purity argon gas (flow rate of 50 to 100 mL min−1), current density and deposition potential also need to be remained at 10 to 20 mA cm−2 and less than −2.6 V, respectively. More details about the impurity control can be found in “Methods” and Supplementary Fig. 5.

The crystallinity, thickness and morphology of the deposited silicon films generally depend on the electrodeposition time, current density, and silicate ion concentration, as shown in Supplementary Figs. 1418. In principle, p-type, n-type, and p-n junction silicon films with various thicknesses and surface morphologies can be produced by varying electrodeposition conditions, as preliminary confirmed in Fig. 2d–f and Supplementary Figs. 1518. Energy dispersive spectroscopy (EDS) analysis further confirms that dense and uniform silicon films are deposited on graphite substrates, and no obvious boundary exists in the p-n junction silicon films, as shown in Supplementary Figs. 17f and 19. However, the growth rate of silicon films during the electrodeposition process is not constant. As shown in Supplementary Fig. 20, the growth rate is high within the first 4 h and then decreases over time. In addition, small amount of silicon powders would commonly generate on the surface of silicon films, which lower the film formation efficiency, and thus the current efficiency for the deposition of silicon films is hard to be accurately calculated. However, according to our previous work28,29 and based on this experimental observation, the current efficiency for the formation of silicon film is about 60 to 80%, depending on the growth rate of silicon film, which is not constant during the electrodeposition. The loss of current efficiency is mainly attributed to the formation of silicon powder on the film’s surface, which is expected can be further optimized by programming the deposition parameters.

### Device characterization and outlook

Characterization of a liquid-junction photoelectrochemical (PEC) cell enables rapid assessment of the quality of as prepared silicon films32. The fabrication and test of real solar cell devices will be discussed later. Here, electrodeposited p-type and n-type silicon films were prepared to form silicon/liquid junctions with redox agent and then characterized photoelectrochemically, as shown in Fig. 3. For comparison, commercial p-type and n-type silicon wafers were also characterized by PEC. The light is chopped on/off during the sweep and the photocurrent can be clearly observed. Figure 3b shows the photocurrent density of the as prepared p-type silicon film and a commercial p-type silicon wafer for reduction of ethyl viologen cations (EV2+). The photocurrent density of electrodeposited p-type silicon film is approximately 50% that of the commercial p-type silicon wafer. Figure 3d shows the photocurrent density of the as prepared n-type silicon film and commercial n-type silicon wafer for the oxidation of ferrocene. The photocurrent density of the electrodeposited n-type silicon film is about 40% that of the commercial n-type silicon wafer. For comparison, the PEC result of the electrodeposited silicon film without any dopant is shown in Supplementary Fig. 21. In addition, p-type and n-type silicon films electrodeposited under different conditions exhibit different PEC performances, as shown in Supplementary Figs. 2224, suggesting the possibility to optimize the film quality by varying the electrodeposition conditions. The PEC performance of the deposited silicon film almost has no degradation after 6 months exposure in ambient (Supplementary Fig. 25).

To compare with current silicon solar cell technology, solar cell devices were fabricated on the electrodeposited p-type silicon films as example. Device current density versus voltage is shown in Fig. 4a, with 295 mV open circuit potential (Voc), 23.4 mA cm−2 short circuit current (Jsc) and 3.1% power conversion efficiency (PCE) being achieved. The cost benefit of silicon films by molten salt electrodeposition was further investigated by a detailed cost analysis33,34,35. More details of the cost analysis can be found in Supplementary Note 2 and Supplementary Table 3. Figure 4b is a brief summary of the dependence of total module cost ($Wp−1) on the module efficiency. It shows that a cell with only 6% and 10% PCE could enable 0.35 and 0.20$ Wp−1 total module cost, respectively, due to the significant reduction in the cost of the silicon wafer production, as presented in Fig. 4c. It is surprising to see that the fraction of silicon wafer cost can be reduced to 5% assuming a 10% PCE. The cell efficiencies have been enhanced steadily along with the improvement of film quality, including making uniform pin-hole free films, increasing the films thickness and reducing the impurities contents as shown in Fig. 4d and Supplementary Fig. 26.

## Discussion

In summary, we demonstrate a simple molten salt electrodeposition process for preparing crystalline silicon films for low-cost solar cells. p-type, n-type and p-n junction silicon films with tunable thicknesses can be directly produced from abundant and inexpensive silicon dioxide all in molten calcium chloride. The electrodeposited crystalline silicon films exhibit high-purity (99.99989% (close to 6N)) and clear photovoltaic effects with PCE as high as 3.1%. There is a large margin for improving the PCE with optimization of the electrodeposition process. Cost analysis further confirms that a module cost lower than 0.20 \$ Wp−1 can be achieved with PCE higher than 10%, making this technology promising for low-cost silicon solar cells.

## Methods

### Materials

Silicon dioxide (SiO2, Sigma-Aldrich, nanopowder, 10 to 20 nm, with purity of 99.5%, trace metals basis), calcium oxide (CaO, Sigma-Aldrich, with purity of 99.9%, trace metals basis) and calcium chloride (CaCl2, Sigma-Aldrich, ACS reagent, with purity of 99%, St. Louis, MO. Ba less than 0.005%, Fe less than 0.001%, K less than 0.01%, Mg less than 0.005%, NH4+ less than 0.005%, Na less than 0.02%, Sr less than 0.01%, heavy metals less than 5 ppm) were used to form a molten electrolyte for electrodeposition. Antimony oxide (Sb2O3, Sigma-Aldrich, with purity of 99.999%) and calcium phosphate (Ca3(PO4)2, Sigma-Aldrich, 4 μm, total heavy metals: less than 20 ppm) powders were used to provide antimony and phosphorus as dopants for n-type silicon film, respectively. Alumina (Al2O3, Sigma-Aldrich, with purity higher than 99.9%) and boric anhydride (B2O3, Sigma-Aldrich, with purity of 99.999%) were used to provide aluminum and boron as dopants for p-type silicon film, respectively. High-purity quartz crucible (Technical Glass Products, O.D. 40 × I.D. 37 × height 180 mm, Painesville, OH. Al: 0.5 ppm, B less than 0.2 ppm, Ca: 0.4 ppm, Cu less than 0.05 ppm, Cr less than 0.05 ppm, Fe: 0.2 ppm, K: 0.6 ppm, Li: 0.6 ppm, Mg: 0.1 ppm, Mn less than 0.05 ppm, Na: 0.7 ppm, Ni less than 0.1 ppm, P less than 0.2 ppm, Sb less than 0.003 ppm, Ti: 1.1 ppm, Zr: 0.8 ppm) was used as electrolytic cell. POCO graphite plate (AXF-5Q, Entegris POCO, Decatur, TX, US; 75 mm in length, 6–20 mm in width, and 1 mm in thickness) was used as substrate for the electrodeposition of silicon. Graphite rod (Alfa Aesar, with purity of 99.995%, diameter 6 mm, Haverhill, MA, US) or the POCO graphite plate was used as anode. Tungsten wires (Alfa Aesar, with purity of 99.9%, diameter 1 mm) were used as the electrode leads and protected by quartz tubes (Technical Glass Products, 3 mm, 6 mm and 10 mm in diameter).

### Silicon films characterization

The deposited silicon films were characterized by using scanning electron microscopy (SEM, Quanta 650 FEG, FEI Inc., Hillsboro, OR) and energy dispersive spectroscopy (EDS, XFlash Detector 5010, Bruker, Fitchburg, WI). The impurity concentration of the silicon film was analyzed by glow discharge mass spectrometry (GDMS, VG 9000, Thermo Fisher Scientific Inc., Waltham, MA, USA). X-ray diffraction spectroscopy (XRD, Philips X-ray diffractometer equipped with Cu Kα radiation) was also used to analyze the produced silicon films. The electrodeposited silicon films were tested as photoelectrodes for the PEC measurement. For the p-type silicon film, the PEC test was carried out in an argon gas-purged acetonitrile (CH3CN, 99%, Extra-dry, Acros, Fair Lawn, NJ) solution containing 0.1 M tetrabutylammonium hexafluorophosphate (TBAPF6, with purity higher than 99.9%, Fluka, Allentown, PA) as supporting electrolyte and 0.05 M ethyl viologen diperchlorate (EV(ClO4)2, Sigma-Aldrich, with purity of 98%) as the redox agent. For the n-type silicon film, the PEC test was performed in an argon gas-purged CH3CN solution containing 0.1 M TBAPF6 as supporting electrolyte and 0.05 M Ferrocene (Fe(C5H5)2, Sigma-Aldrich, St. Louis, MO) as the redox reagent. The PEC properties under UV-visible light illumination by a xenon lamp at 100 mW cm−2 were tested and compared with commercial p-type and n-type silicon wafers (University Wafers, 5 to 10 ohm-cm, (100), boron-doped (p-silicon wafer), phosphorus-doped (n-silicon wafer), Boston, MA, US).

### Device fabrications and characterization

The electrodeposited silicon films were first mechanically polished and rinsed. Then the shallow p-n junction was made by spin-on-dopant, including spin-coating and rapid thermal annealing for dopant activation (950 °C, 60 s). The top contact patterns were then made by lithography and metallization. The I-V characterization of solar cell devices was performed using a B1500A Semiconductor Device Analyzer (Agilent Technologies) and Summit 11000 AP probe station (Cascade Microtech). Solar simulator (Newport) with AM 1.5G filter, calibrated to 100 mW cm−2, was used as light source.

### Cost analysis

The cost analysis model is based on sum of costs of each fabrication steps. The cost of electrodeposition silicon films is analyzed by ownership model includes costs associated with materials, labor fees, depreciation and maintenance of equipment and facilities. Detailed data are collected from industry members, vendors and official reports. For cell fabrication processes and module assembling, costs are used based on state-of-art technologies, collected from various resources, including publicly available databases and official reports. More details can be found in Supplementary Note 2.

## Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

## References

1. Green, M. A. Commercial progress and challenges for photovoltaics. Nat. Energy 1, 15015 (2016).

2. Battaglia, C., Cuevas, A. & De Wolf, S. High-efficiency crystalline silicon solar cells: status and perspectives. Energy Environ. Sci. 9, 1552–1576 (2016).

3. Yoon, J. et al. Ultrathin silicon solar microcells for semitransparent, mechanically flexible and microconcentrator module designs. Nat. Mater. 7, 907–915 (2008).

4. Oh, J., Yuan, H.-C. & Branz, H. M. An 18.2%-efficient black-silicon solar cell achieved through control of carrier recombination in nanostructures. Nat. Nanotechnol. 7, 743–748 (2012).

5. Lewis, N. S. Research opportunities to advance solar energy utilization. Science 351, aad1920 (2016).

6. Polman, A., Knight, M., Garnett, E. C., Ehrler, B. & Sinke, W. C. Photovoltaic materials: Present efficiencies and future challenges. Science 352, aad4424 (2016).

7. Ribeyron, P.-J. Crystalline silicon solar cells: Better than ever. Nat. Energy 2, 17067 (2017).

8. Bullock, J. et al. Efficient silicon solar cells with dopant-free asymmetric heterocontacts. Nat. Energy 1, 15031 (2016).

9. Green, M. A. Silicon photovoltaic modules: A brief history of the first 50 years. Prog. Photovolt: Res. Appl. 13, 447–455 (2005).

10. Swanson, R. M. A vision for crystalline silicon photovoltaics. Prog. Photovolt: Res. Appl 14, 443–453 (2006).

11. Banerjee, A. et al. High-efficiency, multijunction nc-Si:H-based solar cells at high deposition rate. IEEE J. Photovolt. 2, 99–103 (2012).

12. Banerjee, A. et al. 12.0% efficiency on large-area, encapsulated multijunction nc-Si:H-based solar cells. IEEE J. Photovolt. 2, 104–108 (2012).

13. SEMI PV Group, “International Technology Roadmap for Photovoltaics”, (2017).

14. Rao, G. M., Elwell, D. & Feigelson, R. S. Elctrowinning of silicon from K2SiF6-molten fluoride systems. J. Electrochem. Soc. 127, 1940–1944 (1980).

15. Elwell, D. & Feigelson, R. S. Electrodeposition of solar silicon. Sol. Energy Mater. 6, 123–145 (1982).

16. Elwell, D. & Rao, G. M. Electrolytic production of silicon. J. Appl. Electrochem. 18, 15–22 (1988).

17. Haarberg, G. M., Famiyeh, L., Martinez, A. M. & Osen, K. S. Electrodeposition of silicon from fluoride melts. Electrochim. Acta 100, 226–228 (2013).

18. Nohira, T., Yasuda, K. & Ito, Y. Pinpoint and bulk electrochemical reduction of insulating silicon dioxide to silicon. Nat. Mater. 2, 397–401 (2003).

19. Yasuda, K., Nohira, T., Hagiwara, R. & Ogata, Y. H. Direct electrolytic reduction of solid SiO2 in molten CaCl2 for the production of solar grade silicon. Electrochim. Acta 53, 106–110 (2007).

20. Yasuda, K., Nohira, T., Hagiwara, R. & Ogata, Y. H. Diagrammatic representation of direct electrolytic reduction of SiO2 in molten CaCl2. J. Electrochem. Soc. 154, E95–E101 (2007).

21. Yasuda, K., Maeda, K., Nohira, T., Hagiwara, R. & Homma, T. Silicon electrodeposition in water-soluble KF-KCl molten salt: optimization of electrolysis conditions at 923 K. J. Electrochem. Soc. 163, D95–D99 (2016).

22. Jin, X., Gao, P., Wang, D., Hu, X. & Chen, G. Z. Electrochemical preparation of silicon and its alloys from solid oxides in molten calcium chloride. Angew. Chem. Int. Ed. 43, 733–736 (2004).

23. Abdelkader, A., Kilby, K. T., Cox, A. & Fray, D. J. DC voltammetry of electro-deoxidation of solid oxides. Chem. Rev. 113, 2863–2886 (2013).

24. Xiao, W. & Wang, D. The electrochemical reduction processes of solid compounds in high temperature molten salts. Chem. Soc. Rev. 43, 3215–3228 (2014).

25. Xiao, W. et al. Verification and implications of the dissolution-electrodeposition process during the electro-reduction of solid silica in molten CaCl2. RSC Adv. 2, 7588–7593 (2012).

26. Gu, J., Fahrenkrug, E. & Maldonado, S. Direct electrodeposition of crystalline silicon at low temperatures. J. Am. Chem. Soc. 135, 1684–1687 (2013).

27. Cho, S. K., Fan, F. R. F. & Bard, A. J. Electrodeposition of crystalline and photoactive silicon directly from silicon dioxide nanoparticles in molten CaCl2. Angew. Chem. Int. Ed. 51, 12740–12744 (2012).

28. Zou, X. et al. Electrochemical formation of a p-n junction on thin film silicon deposited in molten salt. J. Am. Chem. Soc. 139, 16060–16063 (2017).

29. Yang, X. et al. Toward cost-effective manufacturing of silicon solar cells: electrodeposition of high-quality Si films in a CaCl2-based molten salt. Angew. Chem. Int. Ed. 56, 15078–15082 (2017).

30. Peng, J. et al. Liquid-tin-assisted molten salt electrodeposition of photoresponsive n-type silicon films. Adv. Funct. Mater. 28, 1703551 (2018).

31. Coletti, G. Sensitivity of state-of-the-art and high efficiency crystalline silicon solar cells to metal impurities. Prog. Photovolt: Res. Appl 21, 1163–1170 (2013).

32. Hsu, H.-Y. et al. A liquid junction photoelectrochemical solar cell based on p-type MeNH3PbI3 perovskite with 1.05 V open-circuit photovoltage. J. Am. Chem. Soc. 137, 14758–14764 (2015).

33. Song, Z. et al. A technoeconomic analysis of perovskite solar module manufacturing with low-cost materials and techniques. Energy Environ. Sci. 10, 1297–1305 (2017).

34. Powell, D. M. et al. The capital intensity of photovoltaics manufacturing: barrier to scale and opportunity for innovation. Energy Environ. Sci. 8, 3395–3408 (2015).

35. Sofia, S. E. et al. Economic viability of thin-film tandem solar modules in the United States. Nat. Energy 3, 387–394 (2018).

36. Xie, H. et al. Anodic gases generated on a carbon electrode in oxide-ion containing molten CaCl2 for the electro-deoxidation process. J. Electrochem. Soc. 165, E759–E762 (2018).

## Acknowledgements

This work was financially supported by the Global Climate and Energy Project (GCEP, Agreement No. 60853646-118146), the Welch Foundation (F-0021)), and the National Science Foundation (CBET 1702944). This work was performed in part at The University of Texas Microelectronics Research Center, a member of the National Nanotechnology Coordinated Infrastructure (NNCI), which is supported by the National Science Foundation (grant ECCS-1542159). X.Z. would like to thank Shanghai Rising-Star Program (19QA1403600), the Program for Professor of Special Appointment (Eastern Scholar) at Shanghai Institutions of Higher Learning, and the CAS Interdisciplinary Innovation Team support. We sincerely appreciate Prof. Xionggang Lu (Shanghai University), Prof. Qian Xu (Shanghai University), Dr. Xiaole Chen (UT-Austin) for the kind help and Prof. Cynthiz Zoski (UT-Austin), Dr. Xiao Yang (UT-Austin), Dr. Na Gao (Xiamen University), Dr. Ji Zhao (MIT), Dr. Junjun Peng (MIT), Dr. Huayi Yin (MIT) for the valuable discussions.

## Author information

Authors

### Contributions

X.Z., L.J. and A.J.B. conceived the concept. X.Z. conducted the silicon films electrodeposition, characterization and photoelectrochemical measurements. L.J. fabricated and tested the solar cells devices as well as conducted the cost analysis. J.G. contributed to the XRD characterization. D.R.S., E.T.Y. and A.J.B. discussed all the data and assisted with the strategy design. X.Z. and L.J. prepared the manuscript. D.R.S., E.T.Y. and A.J.B. revised the manuscript. A.J.B. supervised the project.

### Corresponding author

Correspondence to Li Ji.

## Ethics declarations

### Competing interests

The authors declare no competing interests.

Peer review information Nature Communications thanks Dihua Wang and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

## Rights and permissions

Reprints and Permissions

Zou, X., Ji, L., Ge, J. et al. Electrodeposition of crystalline silicon films from silicon dioxide for low-cost photovoltaic applications. Nat Commun 10, 5772 (2019). https://doi.org/10.1038/s41467-019-13065-w

• Accepted:

• Published:

• DOI: https://doi.org/10.1038/s41467-019-13065-w

• ### Electrodeposition of Si Films from SiO2 in Molten CaCl2-CaO: The Dissolution-Electrodeposition Mechanism and Its Epitaxial Growth Behavior

• Xiang Li
• Zhongya Pang
• Xionggang Lu

Metallurgical and Materials Transactions B (2022)

• ### In-situ anodic precipitation process for highly efficient separation of aluminum alloys

• Yu-Ke Zhong
• Ya-Lan Liu
• Wei-Qun Shi

Nature Communications (2021)

• ### Facile Electrodeposition of Ti5Si3 Films from Oxide Precursors in Molten CaCl2

• Wei Tang
• Guangshi Li
• Xionggang Lu

Metallurgical and Materials Transactions B (2021)

• ### Current density enhancement for quantum dot-sensitized solar cells by modulation on the quantum dot loading amount of anatase nanowire array photoelectrodes

• Meng Wang
• Zhuoyin Peng
• Dong Huang

Journal of Solid State Electrochemistry (2021)