Si Feedstock Research Articles

High-Rate Epitaxial Growth of Silicon Using Electron Beam Evaporation at High Temperatures

This paper describes the high-rate (~1.5 μm/min) growth of Si films on Si supporting substrates with (100) crystallographic orientation at 600 °C, 800 °C, and 1000 °C in a vacuum environment of ~1 × 10−5 mbar using electron beam (e-beam) evaporation. The microstructure, crystallinity, and conductivity of such films were investigated. It was established that fully crystalline (Raman spectroscopy, EBSD) and stress-free epi-Si layers with a thickness of approximately 50 µm can be fabricated at 1000 °C, while at 600 °C and 800 °C, some poly-Si inclusions were observed using Raman spectroscopy. Hall effect measurements showed that epi-Si layers deposited at 1000 °C had resistivity, carrier concentration, and mobility comparable to those obtained for c-Si wafers fabricated through ingot growth and wafering using the same solar grade Si feedstock used for the e-beam depositions. The dislocation densities were determined to be ∼2 × 107 cm−2 and ∼5 × 106 cm−2 at 800 and 1000 °C, respectively, using Secco etch. The results highlight the potential of e-beam evaporation as a promising and cost-effective alternative to conventional CVD for the growth of epi-Si layers and, potentially, epi-Si wafers. Some of the remaining technical challenges of this deposition technology are briefly indicated and discussed.

Numerical analysis in Ar-H<sub>2</sub> coupled-coil inductively coupled thermal plasma with Si feedstock for stable operation

<span lang="EN-US">In nanopowder synthesis, the starting powder to be evaporated is infused in a plasma torch through the upper coil and the lower coil in the coupled model of inductively coupled thermal plasma (coupled-coil inductively coupled thermal plasma (ICTP)). Mixing these evaporated materials to form the coupled ICTP significantly influences the thermodynamic and transport properties. It is essential to understand these complex interactions between coupled ICTP and feedstock evaporation. This research investigated the thermal interactions between silicon raw material powder (Si) with ICTP in coupled 99%Ar/1%H<sub>2</sub> through the numerical model developed for the synthesis of Si nanopowder. The feed rate of the Si raw material was set at 0.05, 0.1, and 0.5 g/min. This implies that the increased Si feed is too heavy to vaporize all the injected feed.</span>

Recovery of porous silicon from waste crystalline silicon solar panels for high-performance lithium-ion battery anodes.

A low-cost and easy-available silicon (Si) feedstock is of great significance for developing high-performance lithium-ion battery (LIB) anode materials. Herein, we employ waste crystalline Si solar panels as silicon raw materials, and transform micro-sized Si (m-Si) into porous Si (p-Si) by an alloying/dealloying approach in molten salt where Li+ was first reduced and simultaneously alloyed with m-Si to generate Li-Si alloy at the cathode. Subsequently, the as-prepared Li-Si alloy served as the anode in the same molten salt to release Li+ into the molten salt, resulting in the production of p-Si by taking advantage of the volume expansion/contraction effect. In the whole process, Li+ was shuttled between the electrodes in molten LiCl-KCl, without consuming Li salt. The obtained p-Si was applied as an anode in a half-type LIBs that delivered a capacity of 2427.7 mAh g-1 at 1 A g-1 after 200 cycles with a capacity retention rate of 91.5% (1383.3 mAh g-1 after 500 cycles). Overall, this work offers a straightforward way to convent waste Si panels to high-performance Si anodes for LIBs, giving retired Si a second life and alleviating greenhouse gas emissions caused by Si production.

Nanomolding of Gold and Gold-Silicon Heterostructures at Room Temperature.

Nanofabrication techniques are limited by at least one of the required characteristics such as choice of material, control over geometry, fabrication requirements, yield, cost, and scalability. Our previously developed method of thermomechanical nanomolding fulfills these requirements, although it requires high processing temperatures. Here, we demonstrate low-temperature molding where we utilize the enhanced diffusivity on "eutectic interfaces". Gold nanorods are molded at room temperature using Au-Si alloy as feedstock. Instead of using alloy feedstock, these "eutectic interfaces" can also be established through a feedstock-mold combination. We demonstrate this by using pure Au as feedstock, which is molded into Si molds at room temperature, and also the reverse, Si feedstock is molded into Au molds forming high aspect ratio Au-Si core-shell nanorods. We discuss the mechanism of this low-temperature nanomolding in terms of lower homologous temperature at the eutectic interface. This technique, based on enhanced eutectic interface diffusion, provides a practical nanofabrication method that eliminates the previous high-temperature requirements, thereby expanding the range of the materials that can be practically nanofabricated.

Numerical forecast of redzone extension in cast silicon ingots in dependence on the purity level of crucible, coating and feedstock

The extension of the areas with extremely low minority carrier lifetime in cast silicon ingots for photovoltaic application is a crucial parameter for the industrial manufacturers determining the yield of the grown Si ingot material usable for solar cell fabrication. In order to make a forecast of this redzone extension in dependence on the purity levels of the consumables like crucible, Si3N4-coating and Si-feedstock, a numerical 2D model was used to investigate the diffusive iron (Fe) incorporation into the silicon during the directional solidification process. In order to set up this model, the diffusion parameters of Fe in crucible and Si3N4-coating were determined by 1D numerical analysis of annealing experiments. These values were used in a 2D model to simulate several G1 experiments with varying Fe impurity levels of all crucible consumables (SiO2-crucible and internal Si3N4-coating). The 2D model was validated through comparison of the calculated and experimentally observed redzone.The 2D model was scaled up to the industrial G6 dimensions and a forecast of redzone extension considering the use of the same consumables was set up. The results show that as long as the influence of conventional silica ceramic crucibles is present, the use of Si3N4-coatings or Si feedstock with higher purity will show no positive effect on the redzone extension. For setups without impact of the crucible, similar to current setups including a diffusion barrier, and high silicon feedstock purity, the Si3N4-coating becomes the dominating Fe contamination source. At the same time, the effect of high purity Si3N4-coatings becomes negligible, when the assumed Fe concentration level of the silicon feedstock exceeds values expected for aggressive blending scenarios or low grade feedstock scenarios, which could be motivated by an aggressive cost down attitude.

Composite Si–Ni nanoparticles produced by plasma spraying physical vapor deposition for negative electrode in Li-ion batteries

Si–Ni composite nanoparticles have been produced by a single and continuous plasma spray physical vapor deposition (PS-PVD) from Si and Ni powder feedstocks and their electrochemical performances as anode in lithium-ion batteries (LiB) are investigated. Si nanoparticles with 20–40 nm on which Ni is directly attached with Si/NiSi2 epitaxial interface are formed spontaneously through co-condensation of high temperature elemental gas mixtures during PS-PVD. When only a little amount of Ni is added to Si, the effect of the epitaxial Ni attachment on the Si nanoparticles becomes evident; the cycle capacity is appreciably improved to reach a 1.6 times higher capacity than that of the Si only cell after 50 cycles, due to reduced charge-transfer resistance and nanosized Si particle. In contrast, excessive Ni addition to Si feedstock leads to formation of various silicides as a result of the accelerated silicidation during PS-PVD, which results in a significant decrease in the cycle capacity due to reduction of the active Si phase amount despite reduced charge-transfer resistance.

Novel Application of Electroslag Remelting Refining in the Removal of Boron and Phosphorus from Silicon Alloy for Silicon Recovery

With the shortage of high-grade silicon (Si) feedstock due to the rapid development of photovoltaic, electronic information, and organic Si industries, there is an increasing demand to recycle Si waste economically and efficiently. However, the deep removal of B and P from Si has always been a significant challenge for Si waste recycling. A novel application of electroslag remelting (ESR) refining has been developed to remove B and P for the purification and recovery of Si. In addition, industrial tests were performed to recover the diamond wire saw Si powder. The design principle is put forward for the Si alloy electroslag system. A new 35% CaO–30% CaF₂–17.5% Al₂O₃–17.5% SiO₂ electroslag applicable for Si alloy ESR is designed to meet both physical and impurity removal requirements. Furthermore, the conductivity (2.67 Ω–¹·cm–¹), density (2.97 g/cm³), and viscosity (0.05 Pa·s) meet the requirements of ESR. The dynamic refining process of Si alloy passing through the slag layer in the droplet form is the core of ESR refining, and it is found that more than 80% of B and P removal is achieved in the dynamic refining process of the droplet penetration slag. In order to understand the dynamic refining process, the mathematical model of droplet movement and impurity removal model are proposed to study the droplet movement behavior in slag and the competitive influence mechanism of multiple parameters on B removal efficiency during the dynamic refining process. Mathematical models and experimental results show that there is an ultimate velocity of droplet entry into the slag, which increases linearly with the growth of droplet size under the same slag condition. In addition, the specific surface area of the droplet plays a dominant role in the removal of B. The successful industrial experiments of Si alloy ESR prove that the design principle of the Si alloy electroslag system and the size control mechanism of Si alloy are feasible in this work. The removal efficiency of B and P reached 80 and 65%, respectively, within 30 min. It is also believed that ESR will have an application in purification and recovery of Si waste.

Microstructure and formation mechanism of the Si-Cr dual-alloyed coating prepared by pack-cementation

In this study, two types of SiCr coatings (named as SiCr-1 and SiCr-2) were fabricated on AISI 5140 steels by pack-cementation via different Si feedstocks, aiming to strengthen the steel surface and avoid the carbon-poor layer caused by chromizing. Phases, microstructure, and composition were characterized by X-ray diffraction technique (XRD), secondary electron imaging (SEI) and backscattering electron imaging (BSEI), and energy dispersive spectroscopy (EDS), respectively. The results show that both the coatings mainly consist of α-Fe (with Si and Cr in concentration), Fe3Si, CrFe8Si, and CrC. However, there are some apparent differences between the two coatings including the morphology, element content, microhardness, and wear resistance, which is owing to the difference in the coating growth mechanism. The coating growth of SiCr-1 sample using Si powders as Si feedstock is controlled by the interdiffusion mechanism of the substrate and coating elements, while the coating growth of SiCr-2 sample using ferrosilicon powders as Si feedstock is mainly dominated by the reaction deposition mechanism. For both samples, the presence of micropores in the coatings is observed and their formation is mostly due to the non-equilibrium diffusion induced Kirkendall effect. A new gradient growth mode of micropores based on the Kirkendall effect was proposed.

Numerical analyses and experimental validations on transport and control of carbon in Czochralski silicon crystal growth

Czochralski silicon (CZ-Si) crystal growth is invariably accompanied by the generation and transport of impurities, such as carbon (C) and oxygen (O), from chemical reactions in the high-temperature range. Reduction of C contamination in the grown crystal is required for the production of a high-quality Si wafer. Therefore, we systematically performed the transient global simulations and in situ measurements of CO concentrations in argon (Ar) gas domain. Parametric studies on the furnace pressure, flow rate of argon gas, and gap width were conducted for the transport of CO and accumulation of C during the melting process of Si feedstock. SiO etching reactions on graphite and SiC coating are the major sources of CO production. The contact reaction between the crucible and susceptor is an additional source of CO and SiO. Furthermore, the control mechanisms of gas flow on the incorporation of the generated CO and the accumulation of C in Si feedstock were clarified by comparison of different parameter settings. According to the mechanisms of C transport, the final C content of the melting process depends on the contamination flux at the gas/melt interface which could be controlled by the Péclet number of the gas flow and diffusion distance of CO.

Phosphorus Diffusion Gettering Efficacy in Upgraded Metallurgical-Grade Solar Silicon

In the context of the continuous price reduction in photovoltaics (PV) in recent years, Si feedstock continues to be a relevant component in the cost breakdown of a PV module, highlighting the need for low-cost, low-capital expenditure (CAPEX) silicon technologies to further reduce this cost component. Upgraded metallurgical-grade silicon (UMG Si) has recently received much attention, improving its quality and even attaining, in some cases, solar cell efficiencies similar to those of conventional material. However, some technical challenges still have to be addressed when processing this material to compensate efficiently for the high content of impurities and contaminants. Adaptation of a conventional solar cell process to monocrystalline UMG Si wafers has been studied in this work. In particular, a tailored phosphorus diffusion gettering step followed by a low-temperature anneal at 700°C was implemented, resulting in enhanced bulk lifetime and emitter recombination properties. In spite of the need for further research and material optimization, UMG Si wafers were successfully processed, achieving efficiencies in the range of 15% for a standard laboratory solar cell process with aluminum back surface field.

Effect of PS-PVD production throughput on Si nanoparticles for negative electrode of lithium ion batteries

Silicon nanoparticles (Si-NPs) have been produced by plasma spray physical vapor deposition at throughput as high as 1 kg h−1 (17 g min−1) and the effect on the battery performance is investigated. When the Si powder feed-rate is changed from 1 to 17 g min−1, although the average primary particle size increases to 50 nm, the cycle capacity of the batteries using these Si-NPs is improved slightly owing to their less agglomerated structure. In contrast, when Ni is added to Si feedstock, the cycle capacity is improved at 1 g min−1 due to modified Si-NP structure having SiNi2 interface. Whereas, the batteries with the Si-NP produced at 17 g min−1 shows significant decrease in the cycle capacity because of the excess Ni silicide formation that is resulted from the elevated co-condensation point and the increased reaction area at high throughputs despite the constant Ni concentration in the feedstock.

Effect of controlled crucible movement on melting process and carbon contamination in Czochralski silicon crystal growth

In Czochralski silicon (CZ-Si) crystal growth, packed Si chunks experience collapse and volume shrinkage during the melting process. The axial movements of the crucible and the melting of the Si feedstock lead to dynamic thermal and flow fields and affect mass transport. To study the effect of crucible movement on the melting process and carbon (C) contamination, the cases of fixed and lifting crucible were investigated by the transient global simulation with dynamic mesh deformation. The gap width between the gas-guide and the top surface of Si feedstock was kept constant during the crucible lifting process. Impurity and species transport in Si feedstock and argon gas was investigated for the cases with the fixed and lifting crucibles. The comparison of C accumulation processes indicated that the lifting crucible case resulted in higher C contamination than that found in the fixed crucible case. Furthermore, lifting crucible cases with different gap widths were investigated to elaborate strategies for controlling the crucible movement for its effect on the melting process and C contamination in CZ-Si crystal growth. It was observed that the optimum gap width for C reduction results from the trade-off between the back diffusion and gas convection.

Effect of the packing structure of silicon chunks on the melting process and carbon reduction in Czochralski silicon crystal growth

Carbon (C) contamination in Czochralski silicon (CZ-Si) crystal growth mainly originates from carbon monoxide (CO) generation on the graphite components, which reaches a maximum during the melting stage. Loading a crucible with poly-Si feedstock includes many technical details for optimization of the melting and growth processes. To investigate the effect of the packing structure of Si chunks on C accumulation in CZ-Si crystal growth, transient global simulations of heat and mass transport were performed for the melting process with different packing structures of poly-Si. The heat transport modeling took into account the effective thermal conductivity (ETC) of the Si feedstock, which is affected by the packing structure. The effect of the chunk size on the melting process and C accumulation were investigated by parametric studies of different packing structures. The heat transport and melting process in the crucible were affected by the ETC and the emissivity of the Si feedstock. It was found that smaller Si chunks packed in the upper part could speed up the melting process and smooth the power profile. Decreasing the duration of the melting process is favorable for reduction of C contamination in the Si feedstock. Parametric studies indicated that optimization of the melting process by the packing structure is possible and essential for C reduction in CZ-Si crystal growth.

Comparison of Characterization Techniques for Measurements of Doping Concentrations in Compensated n-type Silicon

Nowadays, compensated silicon (Si) is used in photovoltaic (PV) processes, whether it is through intentional co-doping of resistivity-adjusted Czochralski ingots for high efficiency n-type Si solar cells, as a result of alternative Si purification processes for the production of low-cost Si feedstock, or as a result of recycling end-of-life materials. Whatever the origin of the compensated Si, the doping concentrations need to be accurately and quickly characterized in order to control such processes. In this work, a rapid and highly sensitive characterization technique based on low temperature Hall Effect measurements is described in scientific details and compared to three well-established chemical methods: Glow Discharge Mass Spectrometry (GDMS), Inductively-Coupled Plasma Mass Spectrometry (ICP-MS), and Secondary Ion Mass Spectrometry (SIMS). The characterized samples were extracted from the n-type top part of a casted solar grade Si ingot. A very good agreement is observed between the dopants densities extracted from the electrical method and from the standard methods. With the advantage of a very low detection limit combined with a short measurement time, the advanced Hall Effect technique is promising for the rapid and accurate characterization of dopant concentrations in compensated Si.

Selective Boron Elimination from B2H6–SiH4 Gas Mixtures for Purifying Si

A boron elimination method based on the difference in the pyrolysis temperatures of SiH4 and B2H6 is proposed for purifying metallurgical grade Si feedstock for photovoltaic applications through direct hydrogenation using high-pressure hydrogen plasma. A simple kinetic model is constructed to estimate the proposed boron elimination filter performance. The filter performance is experimentally evaluated. It is confirmed that there is a significant temperature difference between the B2H6 and SiH4 transmission behavior through the filter. However, when SiH4 and B2H6 gases are simultaneously supplied to the filter at the same flow rate of a few standard cubic centimeters per minute, the SiH4 transmission curve is affected by B2H6 and almost overlaps with the B2H6 curve. In contrast, by decreasing the B2H6 flow rate to one thousandth the SiH4 flow rate, B2H6 co-supply has little influence on the SiH4 transmission curve. The type of carrier gas (H2 and He) also affects the transmission curve. With increasing filter temperature, the B elimination effect of the filter becomes greater. The B concentration in Si films prepared by a single elimination process can be reduced to about 1/3000 of the value estimated from the concentration ratio in the unfiltered gas mixture.

Microstructural characterization of Si wafers processed by multi-wire sawing of hot pressed silicon powder based ingots

The purpose of this work is to verify the possibility to process highly doped Si supporting substrates using a 2-step process: (i) sintering of a low-cost Si powder based ingot using hot pressing and a ELKEM Silgrain material as a feedstock; and (ii) wafering of such ingots using multi-wire sawing technique similar to that, which is used for cast multi-Si or Cz-Si grown ingots. Moreover, the possibility to dope Si powder ingot at sintering temperatures below melting point of Si, using a mixture of Si and boron powders to fabricate highly conductive Si wafers is verified as well. The slurry technique has been chosen for multi-wire sawing of sintered Si powder based ingots. Surface properties of Si powder based substrates as well as their chemical composition have been studied by optical microscopy imaging and energy-dispersive X-ray spectroscopy (EDX). Although the overall concentrations of oxygen, carbon and possibly also metal impurities, which are initially present in a low-cost Si feedstock, are too high to achieve acceptable semiconducting properties, it is concluded, that sintered Si powder based wafers have high enough conductivity (resistivity ∼0.001 Ω cm) to serve as supporting substrates for low cost Si wafer equivalent structures.

Carbon nanotube surface modification with the attachment of Si nanoparticles in a thermal plasma jet

The nanocomposite of carbon nanotube and silicon is a promising anode material for a high-capacity lithium-ion battery, because it can improve electrode performance by decreasing the pulverization and swelling of the electrode. In this work, the nanocomposite of multiwalled carbon nanotube and silicon (MWCNT–Si) was synthesized by attaching silicon nanoparticles to the surface of MWCNT in a thermal plasma jet generated from a nontransferred arc plasma torch. The effects of the input power of plasma, the injection position of Si feedstock, and the mass ratio of silicon to MWCNT on the characteristics of nanocomposite were examined. The structure of the nanocomposite and the chemical species in the product were analyzed by electron microscopy and X-ray diffraction, respectively. Although a small amount of SiC was found, MWCNT–Si nanocomposite was synthesized at a relatively high input power. The size and content of silicon nanoparticles in MWCNT–Si nanocomposites were controllable by changing the mass ratio of silicon to MWCNT raw materials. On changing the silicon feedstock injection position, MWCNT–Si nanocomposite was found when the feed position of silicon was at the inside of plasma torch due to the high heat flux and the long residence time.

Removal of Phosphorus in Metallurgical Silicon by Rare Earth Elements

Removal of phosphorus in metallurgical silicon is one of the crucial steps for the production of solar grade Si feedstock. The possibility of doping rare earth elements for phosphorus removal has in this work been studied both theoretically and experimentally. Thermochemical properties of Ce, Nd, and Pr monophosphides have first been estimated by ab initio thermodynamic simulations based on density functional theory and the direct phonon method. The reliability of the first principles calculations was assessed by coupling with the phase diagram data of the Pr-P system. Equilibrium calculations confirmed the existence of stable rare earth monophosphides in solid silicon. Experimental investigations were then carried out, employing a high temperature resistance furnace. The Ce-doped silicon samples were examined by electron probe micro analyzer and inductively coupled plasma analysis. The efficiency of phosphorus removal by means of rare earth doping was discussed in detail in the paper.

Numerical investigation of carbon contamination during the melting process of Czochralski silicon crystal growth

Czochralski (CZ) growth of single silicon (Si) crystals is invariably accompanied by transport of impurities such as carbon (C), oxygen (O), and related compounds produced by reactions at high temperature. To study the generation and accumulation of C during the melting process, a transient global model was developed that included coupled O and C transport. Transport phenomena of C, O, and related compounds were predicted by considering five chemical reactions in the furnace. The dynamic behavior of impurities was revealed during the melting process of the Si feedstock. It was found that C contamination is activated once the melting front contacts argon gas. For accurate control of C contamination in CZ–Si crystals, the accumulation of C during the melting stage should be considered. Parameter studies of furnace pressure and gas flow rate were conducted on the accumulation of C during the melting stage. At the gas/melt interface, pressure and flow rate affected the C flux in different ways. The results suggest that increase in gas flow rate could reduce C contamination much more effectively than decrease in pressure.

Why did the price of solar PV Si feedstock fluctuate so wildly in 2004–2009?

Great attention has been paid to the origin of observed wild price fluctuations of solar PV Si feedstock in both contract and spot markets during 2004–2009. This paper sheds light on this issue and tries to resolve it by addressing the following questions: what kind of structural shock is underlying the price fluctuations of PV Si feedstock? How can we quantify the magnitude, timing and relative importance of these shocks? What are their dynamic effects on the real price of PV Si feedstock? By carefully studying development conditions, the structural decomposition of the real price of PV Si feedstock is proposed: exchange rate shocks, production cost shocks, aggregate demand shocks and demand shocks specific to feedstock markets. With a Structural Vector Autoregression model, the paper quantifies and verifies the impact of structural shocks on PV Si feedstock real price changes. Based on national data, an analysis is further taken to confirm the essential role of demand shocks specific to feedstock markets in determining sharper price fluctuations during 2004–2009. The results of this study have important implications for national solar PV development, which can be better promoted and administrated if structural shocks in feedstock markets can be carefully evaluated and understood.