Sammendrag

The power distribution in submerged arc furnaces is controlled either with current or with resistance control. Current control only requires current measurements and is widely used for small furnaces. For larger furnaces, the current interaction effect is large. Alternatively, furnaces can run on resistance control, which requires unreliable voltage measurements. We propose a novel way of characterizing transformer and electrode currents that gives information about circulating current and imbalance, and outline how this characterization can be used to design current control that can neutralize the interaction effect. Our method is based on high-frequency measurements, enabling us to compute both phase and magnitude of each transformer current. The currents can then be decomposed into a circulating current that will not enter the furnace, a balanced three-phase current, which is the desired current, and an unbalanced single-phase current, which we want to minimize. This decomposition into effective balanced current and unbalanced current can be used to propose synchronized electrode movements. To correct for the imbalance, one electrode must be moved down and the other two moved up, or vice versa. Provided correctly calibrated movements, the interaction effect will be reduced or neutralized. To correct for a too high or too low total current, all electrodes are moved up or down simultaneously. This paper details the theoretical foundations. Further investigation and testing are required to develop a robust control strategy.

Keywords: symmetrical components, electrode currents, furnace control

Sammendrag

Stable and efficient production of silicon in three-phase submerged arc furnaces relies on precise heat-control of the reaction zone. As the heat is generated by the electric currents flowing through each of the three electrodes, the heat-control translates to controlling the electric power in each electrode. To determine the electrode powers, the electrode voltages must be determined through measurements. Typically, the electrode voltages are measured relative to the neutral point of the electrodes’ notional star connection using the so-called Bøckman method. The Bøckman method relies on three measurement leads outside the furnace, each connected to the furnace bottom, to cancel voltages induced by the large magnetic fields in the vicinity of the furnace. The two main sources of error associated with the Bøckman method are induction and neutral-point drift. For an idealized and symmetric furnace, the induced voltages in three symmetrically placed leads cancel, and the true neutral point of the voltage will be at the furnace centre. However, during operation the induced voltages will change with the dynamically changing current paths inside the furnace, and the neutral point will dynamically drift, leading to errors in the measured voltages. To study these effects, we have developed Finite Element Method (FEM) simulation models of industry-scale three-phase silicon furnaces. The models are implemented in COMSOL Multiphysics and are used to calculate the electromagnetic fields within the furnace (including the electrodes, linings and steel casing) as well as the induced voltages in leads placed around the furnace, mimicking the Bøckman measurement system. Through simulations we have investigated how asymmetric furnace conditions lead to neutral point drift. Furthermore, we have estimated the contribution to the measured neutral point drift from an asymmetric or poorly calibrated Bøckman system. The model results have been considered together with industrial measurements of neutral point drift.

Keywords: Bøckman system, Neutral-point drift, Submerged Arc Furnace, Finite Element Model

Sammendrag

End-of-life (EoL) photovoltaic (PV) silicon wafers are currently downcycled into lower value metallurgical silicon, losing much of its value. To address this, EoL silicon purification via vacuum refining was studied using the temperatures 1450 °C, 1600 °C, 1700 °C and a pressure of approximately 20 Pa. The sample vapor was collected by a novel approach utilizing a water-cooled plate condenser placed above the crucible holding molten silicon. The condensate and refined silicon samples were studied using glow discharge mass spectroscopy (GDMS). The results indicated that the condensation technique provided improved mass flux data. The impact of temperature indicated improved silicon yield with increased temperature. It was found that the evaporation from the melt, and hence the corresponding deposition on the condenser surface are affected by temperature and it is the highest at 1700C°, and the lowest at 1600C°. This could be explained by a surface slag layer originating from glass component and passivated layers impacted the mass transfer rate during vacuum refining by limiting the effective area of the melt due to partially surface coverage. This impact was reduced at 1700 °C temperature, where deoxidation of the melt and slag phase removal occurred. Regarding the properties of the material, a mechanism for mass transport of elements under vacuum for processing EoL PV silicon was proposed.

Keywords: Photovoltaics, Silicon, Vacuum refining, low vacuum gas flow

Sammendrag

This work addresses a critical chemical bottleneck in quantum hardware scaling by demonstrating a high-purity, high-yield conversion from isotopically enriched silicon fluoride precursors to electronic-grade monosilane-28. An analytical framework for the synthesis and chemical vapor deposition (CVD)-validation of the isotopes are imperative to bridge the gap between specialized nuclear enrichment and global semiconductor fabrication, essential for the realization and scalability of fault-tolerant quantum computing. While the chemical purification of natural silicon to electronic grade (9-11N) is a mature industrial process, the emerging field of 2nd generation quantum technologies impose new restrictions: the elimination of nuclear noise. Natural silicon contains ~4.7% of the stable isotope ²⁹Si, whose nuclear spin acts as a decoherence source for charge carrier spin states used as qubits. To enable fault-tolerant quantum computing, an industrial value chain for chemically and isotopically pure “quantum grade” ²⁸Si is required. Despite projections indicating that demand already surpasses 100 million USD annually, a scalable value chain for this critical material has yet to be developed. This work revisits a pathway for the production and analytical evaluation of high quality ²⁸Si. Isotopic enrichment is achieved via gas counter current centrifugation of silicon tetrafluoride. The subsequent transformation into viable monosilane is a critical prerequisite for further processing. This is achieved using highly potent hydrogenation processes. Gas-Mass Spectrometry (G-MS) verify chemical quality of 4 N after chemical purification. The main contaminants are traced back to carbon residues from chemical conversion. To validate the materials performance in a semiconductor environment, purified monosilane was used for the epitaxial growth of Si thin films via CVD. Secondary Ion Mass Spectrometry (SIMS) confirm the preservation of isotopic purity >99.9% throughout the deposition process and provide high-sensitivity depth profile of carbon contaminants.

Keywords: Isotopic Engineering, Silicon-28, Monosilane Synthesis, Quantum Technology, SiF₄ Conversion

Sammendrag

Si anodes are a central material for the electrification of transport in coming decades, and strategic for Europe given its lack of battery critical raw materials. They can simultaneously increase battery performance and reduce graphite consumption. Si anodes can enable electric vehicle (EV) range near 1000km and charge times below 10 minutes, while reducing pressure of natural graphite, a critical raw material, on a scale of 100s of kilo tonnes per annum (ktpa). In the processes, this could prevent accumulated CO2 emissions on mega tonne scale during battery manufacture. This presentation introduces a new method for continuous synthesis of a paper-like, freestanding material made up entirely of Silicon nanowires directly from thermocatalytic conversion of silane in a free space reactor, and its application in battery energy storage. It shows progress in the use of the Si nanowire (SiNW) paper as 100% Si anode in Li-ion batteries with high energy density and cyclability, meeting some market entry requirements.

Keywords: battery, CVD, nanowire, paper, silane, aerogel, free space

Sammendrag

Elkem is developing a two-stage silicon production process in which silicon carbide (SiC) is first produced in a dedicated reactor and subsequently used as a reducing agent in a second reactor. This approach can reduce direct emissions of dilute CO₂, enable utilisation of a concentrated CO stream, and lower energy consumption and costs. As part of this development, a pilot-scale submerged-arc furnace has been designed and constructed at Elkem’s R&D facilities in Kristiansand to demonstrate silicon production with high SiC fractions in the reduction material. To date, five pilot campaigns have been conducted with progressively increasing SiC fractions, ranging from a reference case without SiC to approximately 60 percentage points SiC in the carbon coverage. The main objective of these campaigns has been to establish operating strategies for subsequent demonstration campaigns. Material balance analyses show that silicon can be produced with a high fraction of SiC. The best campaign achieved a silicon yield of 92% based on tapped metal and 88% based on filter dust, with a specific energy consumption of 11.8 MWh per tonne Si. A solid basis has been established for the planning of forthcoming demonstration campaigns.

Keywords: two-stage silicon production, Sicalo, SiC, pilot submerged-arc furnace

Sammendrag

In the production of silicon metal, β-SiC is produced as a crucial intermediate product which occupy a significant part of the interior of the submerged arc furnace (SAF). β-SiC is produced by the reaction between carbon and SiO gas. This work investigates how the compressive hot strength of β-SiC is affected by the carbon precursors coal and charcoal. The compressive hot strength is measured indirectly by change in particle size distribution after being exposed to compressive forces at room temperature, 400 ℃, 800 ℃ and 1000 ℃ in a Spark Plasma Sintering (SPS) machine. The strength of SiC from charcoal is found to be up to 50 % lower than calcined charcoal, while there is no significant difference between SiC from coal and calcined coal. This is suggested to be caused by the deterioration of the organized cell wall structure when reacting from charcoal to SiC, while the coal has thicker cell walls and is not affected by SiC conversion in the same degree. It is also found that the strength of SiC from charcoal is more temperature dependent than SiC from coal, with an increase in strength at 800 ℃ and 1000 ℃, while SiC from coal shows approximately the same strength for all temperatures. As expected, the strength of SiC from coal is significantly higher than of SiC from charcoal.

Keywords: biocarbon, charcoal, hot strength, SiC

Sammendrag

Due to the fragility of traditional biocarbon materials, high amount of dust and fines may be generated during transport and handling when replacing fossil carbon with biocarbon in metallurgical industries as part of their decarbonization strategies. Although carbon losses and effect on furnace performance are important, the health and safety of workers are of the highest importance. Inhalable and respirable dust is a direct safety concern, and the increased risk of dust explosions are also present. As such, being able to easily predict the airborne dust formed under typical conditions is necessary. In this work, the aim is to compare one of the most common strength tests, the tumble abrasion test, to a standardized dustiness evaluation called the Heubach Dustmeter Type I analysis, which has previously not been used on these types of materials.

Keywords: Carbon materials, Dustiness, Heubach dustmeter, Strength

Sammendrag

Decarbonization of the silicon production process can be done by electrowinning from SiO2 at high temperatures using molten salts or oxides producing O2 as the anode product. Thermodynamics combined with knowledge of state-of-the-art electrowinning processes can provide an understanding of the possibilities and limitations of the approach, to best compare with today’s submerged arc furnace (SAF) process. Carbon, if used in carbothermic reduction or as a consumable anode in an electrowinning process, provides energy that is not available when the carbon is eliminated from the process. The result is that an optimized electrowinning process is likely going to require more energy than by SAF, and renewable energy must be used to eliminate the CO2 from the process. This work further details the knowledge needs related to dissolution behaviour of the SiO2, electrode materials, cell design and upscaling for both solid and liquid Si production.

Keywords: electrowinning of Si, FFC, molten oxide, molten salt, oxide dissolution

Sammendrag

Silicon carbide (SiC) is an important intermediate material in silicon production. SiC is produced from reaction between carbon and SiO gas, and if it does not react further to silicon, it may accumulate in the furnace and affect material flow, gas flow and electrical conditions. This work presents a new experimental setup for studying the reaction between SiC and SiO gas in a resistance-heated furnace. SiO gas is generated in the lower part of the chamber and passes upward to react with SiC held at a higher temperature. The reactivity of three SiC materials was compared: β-SiC from charcoal, accumulated α-SiC from a silicon furnace, and α-SiC from an Acheson furnace. Degree of reaction was evaluated from mass change and silicon content measured by SEM and XRD. The results show that the method can distinguish between SiC materials with different reactivity. Under the tested conditions, β-SiC from charcoal showed higher reactivity than accumulated α-SiC from a silicon furnace and α-SiC from an Acheson furnace. A simple diffusion-based model gave a reasonable fit to the data and BET analysis suggests that surface area may be important for the observed differences. The results indicate that SiC reactivity is in important factor when the overall efficiency of a furnace is considered.

Keywords: Silicon carbide, SiO gas, reactivity, reaction kinetics, experiments

Sammendrag

Substitution of fossil carbon by bio reductants such as charcoal is one of the prioritized directions for sustainable development of the high silicon alloy (MG-Si and FeSi) production process. The fines fraction created during the production of charcoal could potentially be effectively utilized in a traditional metallurgical submerged arc furnace (SAF) in the form of agglomerates. Compared with lumpy charcoal, agglomerates such as pellets and briquettes show several advantages, including lower moisture absorption and lower fragility which means reduced dust formation. In addition, their higher density improves logistics and utilization efficiency. Additional costs brought by the use- and processing of bio-reductants may be compensated by a decreased CO2 emission cost. Based on available laboratory-scale data, the differences in properties between agglomerates and lumpy materials are not expected to significantly affect the SiO reactivity of the carbon charge. Accordingly, potential deviations in furnace operation and energy consumption may be due to other factors such as thermal stability, charge density etc. The aim of this work was to analyze available operational data of the behavior of an industrial 23MVA submerged arc furnace producing FeSi75 at OFZ during a 20-shift trial, using biocarbon charcoal briquettes as a partial replacement for fossil reductant. The briquettes were produced by the Perechin plant, using cornstarch as binder. Results indicated similar Si-yield, energy consumption and product composition for the charcoal briquettes as with a typical carbon mix, thereby opening the path towards an increased utilization of charcoal fines and a more robust approach towards cost-efficient carbon neutrality in FeSi production.

Keywords: ferroalloys, ferrosilicon, FeSi, charcoal, biocarbon

Sammendrag

Silicon wafers are manufactured using wire sawing processes, during which a significant portion—approximately 35–40%—of the high-purity silicon is lost in the form of very fine particles known as “Kerf.” An economically meaningful use of kerf is only possible if the metallic impurities introduced during sawing can be efficiently removed. Copper poses a particular challenge in this context, as it creates deep defects in silicon, preventing it from being reused as a semiconductor material without additional processing steps. This study examines the mechanism by which copper is introduced into the kerf and outlines the challenges involved when attempting to reduce the concentration of copper in the kerf through wet chemical etching processes.

Keywords: kerf, copper, GD-OES, acidic treatment

Sammendrag

Fluorine-free routes for producing porous silicon are attractive for safer and more scalable silicon processing. In this work, urea-assisted thermochemical porosification of polycrystalline silicon microparticles was investigated at 400 °C, with emphasis on urea loading and the thermal sequence leading to pore formation. Silicon microparticles were blended with urea at urea-to-silicon (Urea:Si) mass ratios of 0.5:1, 1:1, 2:1, and 4:1, followed by a staged thermal protocol involving urea melting at 135 °C and subsequent treatment at 400 °C. Urea loading was evaluated using nitrogen sorption, BJH pore-size analysis, SEM, and EDS. The 2:1 Urea:Si blend produced a sharp increase in porosity, reaching a BET surface area of 25.7 m2 g-1 and a pore volume of 0.039 cm3 g-1, while further increasing the ratio to 4:1 gave only minor additional improvement. The 2:1 blend was selected and further examined using mercury porosimetry and TG/DSC. TG/DSC showed that pristine silicon remains mass-stable, whereas the Urea:Si mixture undergoes substantial mass loss during the 135 °C hold and displays lower-temperature DSC events than pure urea. These results suggest coupled gas-solid surface modification and stress-assisted pore formation during heating to 400 °C.

Keywords: Porous silicon, Urea etching, Mesoporosity, Surface chemistry

Sammendrag

ReSiTec has developed an industrial scale process for recycling silicon kerf-loss from the photovoltaic industry. Two high-end applications have been assessed for this material, as anode material for lithium-ion batteries (LIBs) and into silicides for thermoelectric modules (TEMs). The resulting product from the industrial process is a fine powder material of purity 3-4 N (excluding carbon and oxygen), mean particle size (d50) of 700 nm, carbon footprint of 1 kg CO2 eq/kg, and a silicon oxide layer thickness of 1-2 nm, the latter considered critical for use into LIBs. A comparison study with commercially available primary silicon for LIBs showed comparable electrochemical results between the recycled silicon kerf-loss and the primary silicon with specific capacities from 706 mAh/g to 756 mAh/g and similar cycling stability. State-of-the-art cycling stability is demonstrated for the recycled silicon kerf-loss in a Si-C composite drop-in solution, with over 400 cycles above 80% State-of-Health (SoH) in 1 Ah pouch cell testing. The recycled silicon kerf-loss was tested into manganese- and magnesium silicides for TEMs. A 16-leg prototype was made exhibiting a high-power density. Accordingly, the recycled silicon kerf-loss derived from the industrial process proves its applicability in two high-end powder applications.

Keywords: secondary silicon, recycled silicon kerf-loss, lithium-ion batteries, thermoelectric modules

Sammendrag

Replacing the fossil carbon with biocarbon is a promising and effective way to reduce dependence of the metallurgical industry on the fossil carbon and increase sustainability as well. Densification is the process of compressing loose biocarbon into regularly shaped, dense products such as pellets and briquettes. Densification technology serves as a viable and efficient strategy to refine raw biocarbon into high-value forms such as pellets and briquettes, effectively improving its mechanical performance and boosting both mass density and energy density. Despite these improvements, densified biocarbon products remain susceptible to quality degradation when exposed to rainfall or high-humidity environments for short periods during transportation and storage. Systematic study on the thermal stability of densified biocarbon is important for safe transportation and storage of biocarbon, especially in large scale under different conditions. The objectives of this study are to evaluate the mechanical stability of biocarbon pellets upon water immersion treatment and investigate self-heating and self-ignition of biocarbon pellets through different measures. Water immersion and subsequent air-drying significantly degrade the durability of biocarbon pellets, with weight fluctuations returning to baseline but causing severe structural weakening after prolonged exposure. Thermal stability evaluation reveals that pellets undergo thermal runaway and smouldering combustion at temperatures above 145–160°C, accompanied by oxygen depletion and carbon oxide emissions. Consequently, a dual-monitoring approach combining temperature sensors with CO/CO2 gas detection is recommended to manage self-heating and ignition risks during transport and storage.

Keywords: biocarbon, pellet, water immersion, self-heating, smouldering, safety

Sammendrag

Many industrial producers of metallurgical grade silicon (MG-Si) have a strategy for emission reduction involving biocarbon reductants in the near future, followed by carbon capture on the medium term. Technical and economic challenges in implementing carbon capture are related to flue gas conditions from the submerged arc furnace. This work takes a modelling approach to explore how these conditions will be affected by the upcoming transition to biocarbon. Two different scenarios are studied, one investigating mixed fossil/bio-reductant charges with increasing use of charcoal, and one investigating selected charge parameters with biocarbon as the only reductant. Six different charcoals, treated at different temperatures and from various wood sources are used in the model. A Monte Carlo approach is taken for simulations, capturing variability and uncertainty in process and furnace parameters. Results show the significance of volatiles on flue gas conditions. The charcoals contain less volatiles than coal, and thus produce flue gases with significantly lower heat content and temperatures, and with less water and CO2. There is also a drastic reduction observed on the flue gas sulfur content, although this relies on a mechanism that requires verification. In general, the results indicate that conditions could become more favourable for carbon capture implementation when biocarbon is replacing fossil carbon sources in silicon production.

Keywords: Flue gas, process modelling, charcoal, biocarbon, carbon capture

Sammendrag

Additive manufacturing (AM) is a blooming field which could allow for a shift towards more sustainable manufacturing of products while maintaining industrial scalability. Laser powder bed fusion (L-PBF) is an AM technique that allows the production of components using powders as a feedstock. Aluminium-silicide powders produced by gas atomization (GA) emerge as prospects for AM applications thanks to their low number of defects, good flowability, and casting properties. Increasing the Si content of these powders would allow to build stronger parts and tailor the materials thermal expansion coefficient for specific uses (e.g. aerospace and automotive industry). In order to optimize the printing process, deep understanding of the feedstock is necessary and, therefore, extensive characterization is required. In this work, we establish a methodology to characterize these powders, which consists of several techniques. Three batches of aluminium silicide powders were characterized by glow discharge mass spectrometry (GD-MS), scanning electron microscopy (SEM) as well as laser diffraction particle size distribution (PSD) and image analysis using ImageJ. Altogether, these techniques generate informative data on the sample and establish a characterization routine for such alloy systems which should improve the understanding of the powders by trying to correlate the physicochemical properties and the morphology of the particles to the performance in powder production for L-PBF applications.

Keywords: Aluminium-silicides, powder characterization, additive manufacturing, atomized powders

Sammendrag

The ferroalloy industry aims to reduce CO₂ emissions and achieve carbon-neutral production by 2050 by replacing fossil carbon with renewable biocarbon and developing sustainable process technologies. Although charcoal is already used as a reductant in ferroalloy production, its lower mechanical strength and higher fines generation compared with fossil carbon materials limit its use in large, closed furnaces, particularly in Norway. Therefore, improved biocarbon materials are needed to enable wider industrial adoption. This study focuses on developing sustainable, high-performance biocarbon agglomerates derived from biomass and waste/byproducts for silicon and ferroalloy production. It also focuses on improving the mechanical durability of renewable reductants through agglomeration techniques such as extrusion and briquetting, enabling the production of biomass pellets and briquettes. Strength enhancement is essential to minimize fines formation during transportation, storage, and furnace operation, thereby ensuring predictable carbon availability and stable carbon balance in industrial processes. The study systematically investigates key agglomeration parameters - including particle size distribution, binder selection, and applied compaction pressure - and their influence on pellet strength, porosity, and gas reactivity. Laboratory-scale testing evaluated the durability and gas reactivity of the developed agglomerates to identify optimal processing conditions. Overall, the results demonstrated that charcoal fines can be successfully agglomerated into mechanically robust and thermally stable materials with properties suitable for ferroalloy production processes. The outcomes supported the transition toward carbon neutral ferroalloy production by providing mechanically robust, reactive, and sustainable biocarbon-based reductants suitable for industrial furnace applications. Abstract – The ferroalloy industry aims to reduce CO₂ emissions and achieve carbon-neutral production by 2050 by replacing fossil carbon with renewable biocarbon and developing sustainable process technologies. Although charcoal is already used as a reductant in ferroalloy production, its lower mechanical strength and higher fines generation compared with fossil carbon materials limit its use in large, closed furnaces, particularly in Norway. Therefore, improved biocarbon materials are needed to enable wider industrial adoption. This study focuses on developing sustainable, high-performance biocarbon agglomerates derived from biomass and waste/byproducts for silicon and ferroalloy production. It also focuses on improving the mechanical durability of renewable reductants through agglomeration techniques such as extrusion and briquetting, enabling the production of biomass pellets and briquettes. Strength enhancement is essential to minimize fines formation during transportation, storage, and furnace operation, thereby ensuring predictable carbon availability and stable carbon balance in industrial processes. The study systematically investigates key agglomeration parameters - including particle size distribution, binder selection, and applied compaction pressure - and their influence on pellet strength, porosity, and gas reactivity. Laboratory-scale testing evaluated the durability and gas reactivity of the developed agglomerates to identify optimal processing conditions. Overall, the results demonstrated that charcoal fines can be successfully agglomerated into mechanically robust and thermally stable materials with properties suitable for ferroalloy production processes. The outcomes supported the transition toward carbon neutral ferroalloy production by providing mechanically robust, reactive, and sustainable biocarbon-based reductants suitable for industrial furnace applications. Keywords: Bio-carbon agglomerates, Pellets, Briquettes, SiO-reactivity, CO2-reactivity, Strength

Sammendrag

The transition towards lower carbon footprint silicon value chains requires increased utilisation of secondary raw materials across multiple process steps. This work addresses this challenge by developing and demonstrating industrial pathways for recycling silicon‑containing materials at Elkem Kristiansand. Two complementary activities are presented. First, ferrosilicon- and silicon‑rich skulls and slags from metallurgical ferrosilicon and silicon production are evaluated as supplementary cementitious materials, partially replacing clinker in cement applications. Material characterisation, processing routes, and performance testing demonstrate the potential for CO₂ reduction. The activity covers remelting of silicon‑rich skulls and slags to recover two distinct product streams: FeSi and a refined slag fraction. Process control focused on stable melting conditions, phase separation, and compositional consistency. Material quality was monitored through chemical analysis and particle‑size control, enabling evaluation of both FeSi and slag for downstream cement applications. Second, recycled silicon from wafer‑sawing kerf is upgraded through purification, pellet production, melting, and casting into Si99 grades. Pilot‑ and industrial‑scale trials confirm that recycled silicon can be transformed into consistent, high‑quality products suitable for demanding downstream applications. Together, these results demonstrate how integrated recycling solutions can strengthen circularity and sustainability in the silicon industry. An integrated purification–pelletisation approach was developed to stabilise kerf properties, control organic and oxide behaviour, and produce mechanically robust pellets with predictable performance during subsequent melting. Si99 is produced by remelting kerf‑based pellets under controlled furnace conditions. Pellet quality, charge composition, and ramp‑up strategy are key process parameters. Stable melting, controlled slag formation, and ladle refining enable consistent chemistry prior to casting. Systematic sampling and online analysis are applied to verify yield, impurity levels, and product stability. Elkem will in this work present operational data and results from large‑scale campaigns at the Fiskaa site in Kristiansand, demonstrating process stability, material quality, and practical feasibility under industrial conditions.

Keywords: kerf, silicon, purification, pelletisation, large scale, low-CO2, sustainability

Sammendrag

Accurate quantification of slag in silicon is essential for quality control in industrial silicon production. Conventional combustion methods provide only total oxygen content and cannot distinguish between slag-derived oxygen and that originating from surface oxides, leading to variability influenced by particle size and sample preparation. This study presents a modified analytical approach based on controlled temperature ramping to selectively differentiate oxygen contributions from surface silicon oxides (SiOx) and slag phases. The method exploits thermodynamic differences between oxygen containing species, enabling their separation through temperature resolved evolution profiles. Systematic method development was carried out on silicon materials with varying purity, including metallurgical grade (MG-Si) and electronic grade silicon (EG-Si). The results demonstrate that slag-related oxygen is released at higher temperatures than surface oxides, allowing clear signal separation. The proposed method shows strong potential for improving slag estimation accuracy and providing more robust oxygen speciation in silicon quality assessment.

Keywords: slag estimation, silicon production, quality control, oxygen measurements, inert gas fusion

Sammendrag

Superalloys, like for instance Incoloy®800H (UNS N08810), are the preferred construction material in hydrochlorination reactors due to their excellent thermal and chemical stability. These alloys typically contain a high amount of nickel and chromium. Some of the alloy components will react with the silicon compounds in the hydrochlorination process (for instance HSiCl3) to form silicides on the reactor wall. The silicides will easily oxidize if oxygen and water gets into the reactor, for instance during maintenance or turnaround of the reactor. Some of the oxygen will remain on the reactor wall after restart, and give a permanent reduction in the HSiCl3 selectivity. Silicide formation, and subsequent loss when the reactor is opened for planned and unplanned maintenance, is the main source of internal surface corrosion that ultimately determines the operational life of the very expensive reactor vessel. The formation of silicide on the superalloy reactor wall is demonstrated with small scale reactors followed by characterization with scanning electron microscope.

Keywords: Silicon, Hydrochlorination, Reactor, Corrosion, Alloy, Silicide

Sammendrag

The Si + SiO2 → 2SiO reaction was investigated under carbon free conditions in hydrogen and argon atmospheres at 1600–1700 °C to quantify the activation energy and understand how gas environment influences SiO formation. Isothermal mass loss data were evaluated using a diffusion model, which showed linearity across all temperatures (R² = 0.92–0.99), suggesting diffusion through a growing SiOx interfacial layer as the rate controlling step. Arrhenius analysis yielded activation energies (Ea) of 740 kJ·mol-1 in Ar and 582 kJ·mol-1 in H2, and the values are consistent with transport limited Si–SiO2 systems. The high Ea for both argon and hydrogen system also reveals that the reaction on the reactant surface also play an important role in the kinetics along with diffusion. The higher Ea in argon arises from the strongly suppressed SiO formation at 1600 °C, which steepens the temperature dependency. EPMA revealed coarse, sparse Si precipitates in Ar at 1600 °C, whereas hydrogen produced fine, uniformly dispersed Si domains, indicating earlier onset of defect formation and SiO transport. Condensation behaviour followed the amount of SiO formed.

Keywords: SiO formation, SiO reduction, Hydrogen based reduction, Kinetics

Sammendrag

The carbothermic reduction of quartz is a key reaction in silicon production and is closely related to quartz softening, mass-loss behaviour of carbon-containing mixtures, and phase evolution. China is a major producer of industrial silicon, and CaCl2 is sometimes added to the furnace burden in Chinese industrial silicon practice to improve furnace operation and productivity. However, its effect on quartz behaviour and carbothermic-reduction-related characteristics remains insufficiently understood. In this study, high-temperature melting experiments, TG–DTG tests, and FactSage equilibrium calculations were used to investigate the effects of CaCl2 addition on high-purity quartz, the SiO2–C system, and liquid- and gas-phase evolution. The results show that CaCl2 markedly lowers the characteristic temperatures of quartz. As the CaCl2 addition increased from 0 to 1.0 wt.%, the softening temperature (TS) decreased from 1798 °C to 1745 °C, and the hemispherical temperature (TH) decreased from 1836 °C to 1776 °C. TG–DTG results showed that the maximum mass-loss rate reached 2.90 %·min⁻¹ at 0.8 wt.% CaCl2, compared with 1.48 %·min⁻¹ for the additive-free sample. Increasing the addition to 1.0 wt.% further increased the final mass loss, but reduced the maximum mass-loss rate to 2.08 %·min⁻¹. FactSage calculations indicated that CaCl2 promotes the formation of liquid salt at approximately 770–1550 °C and shifts the rapid growth of liquid slag from about 1720 °C to 1690 °C. CaCl2 did not markedly change the equilibrium generation window of CO and SiO, whereas chlorine-containing gaseous species increased with addition. These results indicate that, under the present experimental conditions, 0.8 wt.% CaCl2 gave the highest maximum mass-loss rate, while higher additions may contribute additional mass loss through chlorine-containing volatilization.

Keywords: Quartz; Calcium chloride; High-temperature melting; Carbothermic reduction; TG–DTG; Thermodynamic analysis

Sammendrag

Latent Heat Thermal Energy Storage (LHTES) based on metallic Phase Change Materials (PCMs) represents a promising solution. Among these materials, ferrosilicon (FeSi) alloys stand out due to their exceptionally high volumetric latent heat, exceeding 1 MWh/m³, and their ability to operate at temperatures far above those of conventional LHTES systems. Such high temperature operation enables higher theoretical Carnot efficiencies and supports emerging solid state thermophotovoltaic (TPV) concepts, a framework currently being explored within the EU-funded SUNSON project. This work evaluates the thermophysical properties of available PCMs with a particular focus on the FeSi system and its ternary derivatives zfor high-temperature PCM applications. Attention is given to near‑eutectic compositions such as FeSi57 (57 wt% Si), which exhibit a narrow phase transition temperature range of less than 50 °C. Key parameters governing PCM performance, including working temperature, latent heat and density are reported. A critical challenge in the industrial application of metallic PCMs is the containment of highly reactive liquid phases. Graphite is an attractive and cost‑effective refractory option, but its interaction with molten silicon introduces both challenges and potential advantages. This review synthesizes current research on the thermodynamic stability of the resulting silicon carbide (SiC) layer and its potential to act as a self-passivating barrier. The performance of ceramic‑coated graphite systems is compared against established non‑reactive containment materials used in the silicon industry, such as Si₃N₄ and BN, along with emerging approaches such as microencapsulation. Overall, Fe–Si-based alloys combined with optimized containment strategies, such as BN-coated graphite represent a promising route toward scalable high temperature thermal energy storage systems.

Keywords: Ferrosilicon, Silicon Carbide, Wetting, Phase Change Materials (PCM), Latent Heat Energy Storage (LHTES), Refractory Corrosion

Sammendrag

The high-temperature behaviour of quartz–charcoal briquettes (QCB) and reference charcoal briquettes (CB) was investigated under conditions relevant to metallurgical silicon production. QCB briquettes contained 30 wt.% quartz, 60 wt.% calcined charcoal, and 10 wt.% corn starch binder. Briquettes were heat treated at temperatures between 1400°C and 2000°C, both with and without exposure to externally generated SiO(g), and characterised by thermogravimetric analysis (TGA), X-ray diffraction (XRD), scanning electron microscopy (SEM), image analysis, apparent and envelope density measurements, and off-gas analysis. An initial mass loss of approximately 13 wt.% was observed below 900°C, attributed to binder decomposition, devolatilisation, and moisture removal. SiC formation started at 1400°C, below the thermodynamic equilibrium onset of approximately 1500–1520°C. The main carbothermal reduction stage occurred between 1400°C and 1700°C, during which all crystalline silica phases were consumed and total SiC content reached approximately 22 wt.%, consistent with the theoretical value for complete quartz conversion. The CB briquettes showed no significant phase development in the absence of quartz. Exposure to externally generated SiO(g) resulted in additional SiC formation concentrated at the bottom and sides of both briquette types, and the QCB briquettes retained significant SiO-capture capability after extensive internal reduction. The results demonstrate that composite quartz–charcoal briquettes are reactive under silicon production conditions and show potential for improved utilisation of fine raw material fractions.

Keywords: Fines utilization, Agglomerates, Biocarbon, Silicon Production, SiO-reactivity, SiC-formation

Sammendrag

This paper describes ongoing work carried out at Wacker Chemicals Norway AS, Holla Metall, to make the silicon plant carbon neutral by 2030 through large-scale substitution of fossil reductants with biocarbon. The transition requires changes to logistics, especially related to handling and use of low-density and mechanically weak materials that generate fines and dust. Full-scale industrial test campaigns have been conducted with biocarbon shares exceeding 50% of the total carbon demand, demonstrating that stable furnace operation is possible. During testing, temporary fluctuations in silicon yield are observed, mainly associated with ramp-up phases, variability in fines generation, and controlling the effective carbon input. The work highlights differences between charcoal and agglomerated biocarbon materials in terms of reactivity, mechanical stability, and operational behavior, and highlights the need for improved control of material properties and operating strategies. A structured qualification methodology combining laboratory testing and industrial validation has been developed to support implementation. The results show that successful use of biocarbon requires looking at the full process. Major research and development projects are carried out in-house as well as in collaboration with universities, research companies and other industry including both established and new producers of biocarbon materials.

Keywords: Silicon, biocarbon, industrial testing, carbon neutrality, dust, safety

Sammendrag

The SiH2 process is a concept for production of silicon by H2 as reductant. The idea is to produce SiO gas by reaction between SiO2 and internal recycled Si. Produced SiO is reduced by H2 to Si and H2O. The two reactions take place at different temperatures. To realise the proposed concept, knowledge about the kinetics for the involved reactions is needed. Reaction between SiO2 and Si to produce SiO is discussed here. In the experiments, flow of warm SiO gas from the reaction has been measured continuously by a new method. This method shows that the reaction rate varies with time and have a peak with very high SiO generation that start as the temperature reach around 1750 °C to 1800 °C and then reduces rapidly until a lower rate is reached. Effects on reaction SiO2 + Si ⇄2SiO of various other parameters such as gas composition, gas flow rates, temperature and properties of SiO2 and Si source is also discussed.

Keywords: Silicon, hydrogen, SiO-gas, quartz

Sammendrag

Currently, silicon is a base material for more than 95% of photovoltaic (PV) modules. The deployment of PV modules is expected to increase significantly across the European Union (EU), leading to generation of millions of tons of cumulative end-of-life photovoltaic (EOL PV) waste in the coming decades. This huge expansion and waste accumulation emphasize the need for a circular economy approach. QUASAR, an EU-funded project, aims to develop a holistic recycling approach to manage the growing volume of EOL PV panels. With the help of advanced recycling technologies, QUASAR aims to enable efficient recovery of materials such as silicon, silver, polymers, glass, and other metals. As part of QUASAR, this study assesses the expected quantities of EOL PV modules and the silicon embedded within them under chosen scenarios based on historical PV installation data, EU-level installation forecasts, dynamic module mass-to-power ratios, and material composition data. In addition, the characteristics of a recovered silicon sample obtained from a preliminary refining step were evaluated using Inductively Coupled Plasma Mass Spectrometry (ICP-MS) and LECO analysis. Based on the quantity and quality assessments, the potential demand for silicon recovered from EOL PV panels as a secondary raw material in PV and other value chains has been discussed in the context of a circular economy approach.

Keywords: Circular Economy, EOL-PV Panels, Recycling , Silicon Recovery, Secondary Raw Materials

Sammendrag

Direct chlorination (DC) of metallurgical-grade silicon with anhydrous HCl is an industrial route to trichlorosilane (TCS), typically operated without intentional catalysts. In contrast, the methylchlorosilane direct synthesis (MCS) uses catalyst/promoter elements such as Cu, Zn and Sn, whose individual effects on silicon surface reactivity remain difficult to isolate because of CH3Cl cracking and coke formation. In this work, DC was used as a simplified model system to evaluate the influence of typical MCS catalyst elements on TCS selectivity, reactivity and high-boiler formation. A synthetic silicon sample containing controlled amount of Fe, Al and Ca was tested with additions of Cu, Zn or Sn in a continuous lab-scale fluidized-bed reactor at 340 °C and 1.15 barg, with continuous silicon feeding to maintain a constant reaction mass. Product gases were analysed by online GC, and reacted residues were characterized by SEM–EDS. Cu increased silicon reactivity, lowering the reaction start temperature from 366 °C to 338 °C, without significantly affecting selectivity. Zn maintained high TCS selectivity but increased HCl loss, suggesting progressive surface deactivation through chloride-rich deposits. Sn had the strongest detrimental effect, decreasing TCS selectivity and preferentially increasing Si–Si bonded high boilers. SEM–EDS showed both Sn-rich particles and dispersed Sn in reacted Si-rich regions, indicating surface modification. These results identify Sn as a critical impurity for TCS production.

Keywords: Trichlorosilanes, direct chlorination, Tin, Zinc, Copper, High boilers

Sammendrag

Charcoal continues to be a strategically important carbonaceous reductant for the metallurgical industry, including the production of metallurgical-grade silicon and ferrosilicon, due to its low ash and sulphur content and its renewable origin. However, the mechanical, thermal and water-resistance properties of conventional roller-pressed charcoal briquettes are often insufficient for the demanding handling and high-temperature conditions of submerged-arc furnaces. In the present work, cylindrical extruded charcoal briquettes (20 mm in diameter, 30–150 mm in length) were manufactured on a laboratory extruder from milled hardwood charcoal fines and a self-developed protective organic polymer binder — a water-soluble, carboxyl-functional polymer — used at a low loading, and were characterised against the requirements of EN 1860-2 and DIN 51729-10. Proximate analysis gave a fixed-carbon content of 74.0 % and an ash content of 4.5 %, with a typical alkaline-earth-rich ash composition (36.76 wt-% CaO, 16.10 wt-% SiO2, 8.59 wt-% K2O). The briquettes withstood axial compression loads of 0.5–0.9 MPa, lost only 14.4 ± 0.56 % of mass after 30 min of drum abrasion, absorbed 36.1 % water after five days of immersion without disintegration, and retained their cylindrical shape after rapid heating to 900 °C, losing 24 % of mass. The proximate and abrasion data are means ± standard deviation of three independently produced batches (n = 3); the ash analysis is a single accredited determination, the compression result is the observed range, and the water-uptake and thermal-shock values are single determinations. The combined results indicate that the obtained briquettes meet the basic mechanical, hydrothermal and thermal-shock criteria required for both barbecue applications and for use as a carbonaceous reductant in metallurgical processes.

Keywords: extruded charcoal briquettes, organic polymer binder, mechanical strength, water resistance, thermal shock, metallurgical reductant

Sammendrag

The shift to clean energy from fossil fuels will depend heavily on metals. Silicon is one of the top elements, playing a key role in the transition to green energy due to its abundance and importance in developing essential technologies. The main energy sectors are the PV industry and the new application of Si-anode elements in EV production. These industries will increase silicon output and promote new recycling methods. Economic and security concerns are present, as China holds a monopoly on production, and Europe must address the challenge of energy costs associated with smelting and refining silicon. Because of its high energy density and specific energy, silicon is also an excellent candidate as a metal fuel. Its widespread availability in silica form, commonly used in glass manufacturing, is environmentally friendly and non-toxic, requiring no specialized waste processing. Furthermore, a clean, energy-efficient method for producing hydrogen from silica has been demonstrated. Still, little is known about the combustion behaviour and explosibility parameters of pure silicon dusts or Si-H2 mixtures. This work will provide data on explosibility characterization and establish a reliable, reproducible dataset for use in risk assessment of silicon metal powder-based energy processes. The silicon powder sample tested exhibited relatively high values for overpressure and rate of pressure rise, respectively, 8.1 bar and 440 bar/sec.

Keywords: MeCRE, SSE-Chart, Safety Characteristics, Metal Explosions

Sammendrag

Silicon dust—metallurgical and higher-purity grades—exhibits ignitability and explosibility characteristics that underpin industrial applications as energetic material. Silicon has also been proposed as a possible “green fuel” for energy production through controlled combustion. However, the same energetic properties that enable these applications can, under unfavourable conditions, create hazardous situations and trigger industrial accidents. This is well recognised across the industry. Elkem has firsthand experience, including the devastating dust explosion in Bremanger in 1972. Since then, Elkem has maintained a long-standing focus on dust explosions hazards and the mechanisms that influence the explosion risk. Relevant test data for silicon dust that could be identified within the company have now been consolidated for a comprehensive review. This effort has led to an updated understanding of the behaviour and properties of silicon dust, particularly regarding the influence of dynamic variables. Among these, particle size distribution (PSD) is the most critical parameter affecting dust explosibility. Notably, PSD is not a fixed property; it is dynamic and evolves with time, mechanical forces, and degree of turbulence. Standardized dust-testing methods capture only a controlled subset of these variables, enabling comparisons across samples while simplifying the complexity of real-world conditions.

Keywords: Silicon, dust explosions, explosive properties, particle size distribution