ECO-EFFICIENT PROCESSING AND REFINING ROUTES FOR SECONDARY RAW MATERIALS FROM SILICON INGOT AND WAFER MANUFACTURING

ECO-EFFICIENT PROCESSING AND REFINING ROUTES FOR SECONDARY RAW MATERIALS FROM SILICON INGOT AND WAFER MANUFACTURING Martin Bellmann*1, Berhane Darsene Dimd1, Anne-Karin Søiland2, Arne Dahle3, C. Landaas4, Victorien Iwaszko5, Rene Peche6, Wolfram Palitzsch7, Philippe Lenain8, Iratxe de Meatza9, Theodora Kyratsi10, Liu Huiping11, Emanuele Milani12, Guy Chichignoud13, Stefan Fischer14, Almut Schwenke15, Eirik Nordboe16, Marco Pieterse17, Roland Riva18 1SINTEF, Trondheim Norway, 2ReSiTec AS, Kristiansand Norway, 3NorSun, Oslo Norway, 4Northern Silicon, Meråker Norway, 5ROSI Solar, Saint-Martin-d’Hères France, 6bifa Umweltinstitut GmbH, Augsburg Germany, 7LuxChemtech GmbH, Freiberg Germany, 8benkei, Lyon France, 9CIDETEC, San Sebastian Spain, 10University of Cyprus, Nicosia Cyprus, 11GRÄNGES, Finspång Sweden, 12Marelli Europe SPA, Venaria Reala Italy, 13Institut Polytechnique De Grenobl, Grenoble France, 14SGL Carbon GmbH, Meitingen Germany, 15SGL Battery Solutions, Meitingen Germany, 16Fiven Norge AS, Lillesand Norway, 17Chemconserve, Bussum The Netherlands, 18Commissariat à l’énergie atomique et aux énergies alternatives, Le Bourget-du-Lac France ABSTRACT: In the ICARUS project, 18 European partners collaborate to develop and scale innovative technologies for recovering and refining secondary raw materials from silicon photovoltaic (PV) ingot and wafer manufacturing. The production of PV modules generates significant quantities of waste, particularly silicon kerf, graphite, and silica residues. ICARUS aims to transform these waste streams into high-value secondary materials suitable for reintegration into the PV value chain and other industrial applications. Four industrial pilot-scale processes were developed, targeting the purification and reuse of these materials. Results from the pilots demonstrate both the technical feasibility and economic potential of substituting these recovered materials for virgin and critical raw materials. This work provides a viable pathway toward a more resource-efficient and circular PV manufacturing industry. Keywords: Photovoltaics, Silicon, Silicon Kerf, Silica, Recycling, Circularity 1 INTRODUCTION Global deployment of photovoltaic (PV) technology continues to accelerate, driven by declining module costs and net-zero emission targets. Alongside this growth, the silicon PV value chain generates substantial production wastes during ingot manufacturing and wafering. These include silicon kerf losses from wafering, as well as silica and graphite components from crystallization furnaces. A significant proportion of these wastes is either disposed of in landfills or diverted into low-value applications. Such practices not only result in the loss of potentially valuable materials but also conflict with sustainability and circular economy principles. In general the literature demonstrates clear progress in recovering high-purity products from silicon kerf [1], silica waste [2], and graphite [3]. However, current methods generally yield secondary materials with limited recovery efficiency and insufficient quality for high-value reuse. In addition, significant recycling techniques are not yet scalable to industrial level due to their complexity and multi-stage approach. The ICARUS project addresses this challenge by developing and scaling industrially relevant routes for the recovery, refinement, and reintegration of secondary raw materials from ingot and wafer manufacturing. The ICARUS project is a collaborative effort between 18 EU (European Union) partners. The work focuses on developing targeted technological solutions to refine and reuse silicon kerf, graphite, and silica wastes from PV manufacturing. Four industrial pilot technologies have been designed and tested at pilot scale to assess improved recycling processes, with the potential to outperform existing methods. The main goal is the production of highpurity, high-value secondary raw materials, thereby addressing the quality limitations often encountered in current recycling approaches. By 2027, the project aims to enable large-scale resource recovery, with projected capacities of 3.5 million tons of silicon kerf, 700 thousand tons of silica, and 480 thousand tons of graphite, demonstrating both scalability and economic viability. The work in ICARUS contributes directly to sustainability and circular economy goals by transforming waste into reusable raw materials, reducing environmental impacts, and improving the overall efficiency of the silicon PV value chain. 2 METHODOLOGY This section provides a brief overview of the ICARUS project and PV waste volume estimates and characterization. It also discusses the four pilot-scale recycling technologies developed for silicon kerf, graphite, and silica. 2.1 Overview of ICARUS The ICARUS project is organized into six work packages (WPs). WP1 addreses logistics, treatment, quality, quantity, and sourcing of silicon PV production waste. WP2 develops industrial routes for collecting and pre-treating silicon, silica, and graphite. WP3 reintroduces these wastes into silicon production, while WP4 focuses on controlled conditioning of purified silicon. WP5 upgrades a lab-scale reactor to semi-industrial scale for converting silicon waste into valuable materials. Finally, WP6 drives market uptake by demonstrating high-end prototypes that utilize recovered silicon, silica, and graphite. 2.2 Waste volume estimates and characterization PV ingot manufacturing generates significant waste. To assess the potential of recycling pathways, it is essential to estimate waste volumes and characterize materials. Volume estimates provide context for the scale of the challenge, while characterization supports evaluation of technical feasibility and processing needs. In this work, silicon kerf waste was estimated using factors such as annual PV installations, the ratio of production to installations, crystalline PV production volumes, and the cell-to-wafer ratio. Estimates of crucible and pot scrap waste assumed multi-batch ingot pulling with M10 and G12 wafer dimensions, while graphite waste was estimated at 100 tons per gigawatt of PV capacity, as reported by [4]. The characterization of waste materials and produced silicon was carried out using several analytical techniques. Inductively Coupled Plasma Mass Spectrometry (ICPMS) was applied to detect metallic and non-metallic impurities at trace levels, while LECO analysis was used to quantify carbon and oxygen contents. X-ray Diffraction (XRD) was employed to determine crystallographic structure and phase composition, and Scanning Electron Microscopy (SEM) was used to examine surface morphology. Together, these techniques assess chemical purity and morphology, ensuring recycled materials meet quality standards for high-performance applications. 2.3 Pilot technologies The four industrial pilot technologies which are the core of the ICARUS project are: • Pilot A: Collects and processes silicon kerf cake, graphite, and silica waste according to defined standards, delivering the material either to enduser groups for final applications or to other pilots for further refining. • Pilot B: Aims to produce silicon with higher quality and at a lower cost than conventional metallurgical silicon processes. This pilot uses secondary materials such as graphite, silica, and silicon kerf in a pilot-scale submerged arc furnace (SAF). • Pilot C: Scales up the annual controlled conditioning of silicon, based on material received from Pilot A to a capacity of 500 tons per year, aiming to improve the quality of granular silicon for PV applications. • Pilot D: Will scale up a small-scale reactor to a semi-industrial level, targeting a capacity of 50– 70 tons per year. The process is based on the reaction of free silicon in the kerf with a sodium hydroxide solution under controlled temperature and pressure conditions, producing hydrogen, sodium silicates, and reaction heat. 3 RESULT AND DISCUSSION This section reports the key findings from waste volume estimates and material characterization, along with results from the four pilot projects, with emphasis on Pilot B. 3.1 Waste volume estimates and characterization result The estimation of silicon kerf waste accounts for losses from wafering as well as cropping, squaring, chamfering, and grinding operations. Figure 1 shows the estimated global cumulative kerf generation from 2018 to 2027, indicating that annual volumes could reach about 700 kMT by 2027. Similarly, Figure 2 presents projected cumulative silica and graphite wastes from 2023 to 2027. Overall, these estimates highlight the substantial volumes of waste generated in ingot and wafer manufacturing. Although typically treated as waste, these materials represent a valuable opportunity for recycling and reintegration into the PV value chain and other applications. Figure 1: Cummulative silicon kerf waste in kMT. Figure 2: Cummulative silica and graphite waste in ton. Impurity analysis of silicon kerf samples from different sources was conducted using ICP-MS. The results show that the main dopant impurities are boron, gallium, and phosphorus, with concentrations varying across samples. The main metallic impurities include aluminum, calcium, iron, and nickel. LECO analysis indicated carbon concentrations of 1–2 wt.% and oxygen content of 4–5 wt.% in most samples. SEM and XRD analyses revealed irregular morphologies, with the kerf consisting mainly of crystalline silicon and an amorphous phase. LECO analysis of five graphite powder samples revealed significantly higher ash content compared to a virgin reference sample. ICP-MS analysis of the ash identified silicon, cobalt, and calcium as the main impurities. XRD results showed the presence of graphite 2H, SiC, and crystalline SiO2 phases, while SEM revealed multiple phases with distinct morphologies across different regions of the samples. Similar characterization was carried out on silica waste samples, which were found to consist mainly of amorphous silica and crystalline cristobalite, with calcium identified as the primary impurity. 3.2 Results from the pilots The ICARUS project aims to transform silicon kerf, graphite, and silica waste into high-value secondary raw materials by developing and scaling four pilot technologies for efficient recycling. This section presents their results, with a more focus on Pilot B, chosen as the most representative of the overall processing chain, while summarizing key outcomes from the other pilots. Pilot A: Industrial-scale processing routes for silicon kerf waste have been established at Technology Readiness Level (TRL) 7. A continuous drying system has been successfully commissioned as an alternative to the existing batch drying process for silicon kerf filter cake. This system achieves moisture levels of less than 1 wt%. Additionally, an industrial-scale post treatment process has been developed for material intended for use into lithium-ion batteries and thermoelectric modules. A pretreatment process has also been implemented to significantly reduce contaminants such as aluminum, nickel, and iron. For the silica waste, a processing route has been developed and a pilot line established, targeting raw material specification suitable for production of high purity silicon carbide. Pilot B: This pilot aimed to reintroduce silicon kerf, graphite, and silica waste into the silicon value chain through carbothermic reduction. Two experimental campaigns were carried out in a pilot-scale SAF. The process began with the collection and pretreatment of raw materials, followed by agglomeration into self-reducing briquettes. Three types of briquettes were made, Type A (pure quartz and carbon black with binder and water), Type B (pot scrap, carbon black, binder, and water), and Type C (silicon kerf with binder and water). These briquettes were then subjected to pilot-scale carbothermic reduction in the SAF to produce silicon, which was subsequently analyzed to determine its purity. The two experimental campaigns which were run over a period of three days and three nights consisted of two experiments, totaling four experiments. These are: • EXP 1: Type A briquettes and quartz lumps as charge. • EXP 2: Type A briquettes, quartz lumps, and pot scrap lumps as charge. • EXP 3: Type B briquettes and pot scrap lumps as charge. • EXP 4: Type B and Type C briquettes and pot scrap lumps as charge. Across two SAF campaigns, a total of 92 kg of silicon was produced through twelve tapping operations. In the first campaign (EXP 1 and EXP 2), 60 kg of silicon was obtained, while the second campaign (EXP 3 and EXP 4) yielded 32 kg. ICP-MS analysis of six samples from each tap (Tables I) confirmed that the produced silicon met metallurgical-grade purity requirements [5], with the second campaign achieving slightly higher purity than the first. Table I: Purity of the tapped silicon measured by ICP-MS. Purity (%) Sample Campaign 1 Campaign 2 1 97.87 98.43 2 98.27 98.16 3 98.47 98.01 4 98.39 98.47 5 98.34 98.78 6 98.79 98.05 Detailed ICP-MS analysis of Campaign 1 samples showed phosphorus as the dominant dopant, slightly above metallurgical-grade thresholds, while boron and gallium remained within limits. Aluminum exceeded the threshold in early tappings but stabilized later, iron was acceptable in half of the tappings, and titanium consistently exceeded limits; other metals were within range, confirming overall suitability for metallurgical-grade silicon. In Campaign 2, phosphorus and boron levels decreased, gallium increased, and aluminum and iron were well controlled, though calcium, titanium, chromium, and nickel exceeded thresholds. Despite these variations, it can be concluded that the produced silicon in this campaign met metallurgical-grade standards, an encouraging result given that 70% of the feedstock came from recycled pot scrap and kerf. Pilot C: A combined system for powder feeding, melting, solidification, and granulation has been developed and commissioned. The powder feeding unit has a capacity of 500 tons per year, while the melting and granulation units handle 50 tons per year. The process successfully demonstrated the production of recycled silicon at a flow rate of 1 kg/h, achieving 4N purity. Pilot D: A chemical conversion process for silicon waste has been scaled up to 87 tons per year, producing waterglass and green hydrogen for diverse market applications. End-of-life PV panels and semiconductor industry residues were also shown to be valuable alternative sources for material recovery. 4 CONCLUSIONS This study through the ICARUS project highlights the substantial volumes of silicon kerf, silica, and graphite waste generated during ingot and wafer manufacturing and demonstrates their huge potential as valuable secondary raw materials. Detailed characterization of these wastes confirmed the suitability of these wastes as substitutes for virgin materials, while the successful implementation of four pilot technologies within the ICARUS project validated the technical feasibility of their recovery and reuse. Together, these results provide a strong foundation for integrating recycling into the PV value chain, advancing both sustainability and circular economy objectives in the PV industry. 5 ACKNOWLEDGEMENT This work is part of ICARUS project funded by Horizon 2020 research and innovation programme under grant agreement No 958365. 6 REFERENCES [1] Li, J., Lin, Y., Wang, F., Shi, J., Sun, J., Ban, B., Liu, G. and Chen, J., 2021. Progress in recovery and recycling of kerf loss silicon waste in photovoltaic industry. Separation and Purification Technology, 254, p.117581. https://doi.org/10.1016/j.seppur.2020.117581. [2] Yang, S., Han, S., Chen, J., Wei, K. and Ma, W., 2024. A sustainable mineral process for silicon and quartz recovery from quartz crucible waste ash via electrical separation. Minerals Engineering, 216, p.108887. https://doi.org/10.1016/j.mineng.2024.108887. [3] Zhang, Y., Z. Chen, K. Xie, X. Chen, Y. Hu, and W. Ma. Purification of Waste Graphite from Crucibles Used in Photovoltaic Crystallization by an Alkali-Acid Method. Metals 2023, 13, 1180. 2023. https://doi.org/10.3390/met13071180. [4] Brailovsky, P., Baumann, K., Held, M., Briem, A.K., Wambach, K., Gervais, E., Herceg, S., Mertvoy, B., Nold, S. and Rentsch, J., 2023. Insights into circular material and waste flows from c-Si PV industry. EPJ Photovoltaics, 14, p.5. https://doi.org/10.1051/epjpv/2022029. [5] Metallurgical-Grade Silicon (MGS) - WikiChip. https://en.wikichip.org/wiki/metallurgicalgrade_silicon#google_vignette.