Silicon Ingots Research Articles

The crucible/silicon interface in directional solidification of photovoltaic silicon

Photovoltaic silicon ingots are currently grown in silica crucibles coated with a porous silicon nitride layer which acts as an interface releasing agent between the silicon and the crucible. The interactions between Si and the Si3N4 coating determine the infiltration and sticking phenomena occurring at the interface and also affect the pollution of Si by the components of the coating. In this investigation the interfacial interactions and microstructure are studied in crystallization experiments performed in crucibles involving high silicon masses (tens of kg) and long contact time between the silicon and the coated silica (tens of hours). It is shown that for long times, a dramatic change in the nature of the coating/Si interface takes place, with the formation of a self-crucible which prevents the direct contact between the silicon and the coating. The stability of the self-crucible is modeled taking into account the capillary and hydrostatic pressures. The influence of the self-crucible on different practical aspects of the photovoltaic silicon crystallization process is discussed.

Bulk Lifetimes up to 20 ms Measured on Unpassivated Silicon Discs Using Photoluminescence Imaging

With high-efficiency silicon solar cells approaching 25% efficiency in mass production, the requirements on the bulk lifetime and its uniformity across the wafer and the ingot increase dramatically. Since some cell architectures require these high lifetimes on starting material, the need arises for characterization methods to measure very high bulk lifetimes that are spatially resolved at an early stage before cell processing. A method based on the spectral ratio of two photoluminescence images is applied here on two unpassivated silicon discs from different positions within a Czochralski-grown phosphorous-doped n-type silicon ingot. The method allows the determination of spatially resolved bulk lifetime images on samples with adequate thickness and can be done within seconds and without the need to passivate surfaces. As-grown bulk lifetimes up to 20 ms are measured on the ingot's central disc, indicating recent improvements in crystallization technology, but are strongly reduced closer to the crown. Evidence suggesting the impact of thermal donors on the lifetime and effective doping concentration near the crown is found from combining spectral photoluminescence and infrared spectroscopy analyses. The technique could find applications in research and development activities, particularly in the optimization of Czochralski silicon crystal growth conditions.

Microstructure and resistivity of the silicon target material prepared by adding Al–B master alloy

A silicon target material with the purity of 99.999 wt% (99.999%) was prepared by adding Al–B master alloy in directional solidification. The segregation behavior of the dopant and the effect on the resistivity were studied in this work. It was revealed that the AlB2 particles in the Al–B master alloy will generate the clusters of [B] and [Al] in molten silicon at 1723 K spontaneously. The concentrations of B and Al were increasing gradually along the solidified fraction in the silicon ingot. The measured values of B were in good agreement with the curve of the Scheil’s equation below 85% of the solidified fraction. The measured values of Al were fitting well with the curve of the Scheil’s equation when the effective segregation coefficient is 0.00378. It was found that the resistivity of the silicon target material was regulated by B co-doped Al simultaneously in directional solidification.

Global simulation of coupled oxygen and carbon transport in an industrial directional solidification furnace for crystalline silicon ingots: Effect of crucible cover coating

For accurate prediction of oxygen (O) and carbon (C) distributions in the crystalline silicon ingots grown by the industrial directional solidification (DS) furnace, we first performed transient global simulations of heat transfer based on a fully coupled calculation of the thermal and flow fields. The phase change problem was handled by an enthalpy formulation technique based on a fixed-grid methodology. The coupled C and O transport in the DS furnace was then carried out, taking into account five chemical reactions. Special attention was devoted to modeling the O and C impurity segregation during the entire solidification process. It was found that the developed model can successfully simulate the impurity segregation at the growth interface and the obtained boundary layer thickness is similar to that estimated by analytical calculation. The effect of the crucible cover coating on the coupled O and C transport was investigated. The numerical results show that the C concentration in the grown silicon ingot can be reduced by about 60% if the pure graphite cover was replaced by a graphite cover with an inert material coating on it. However, the cover coating did not significantly affect the O concentration in the grown ingot. The numerical predictions of the C concentration showed satisfactory agreement with the experimental measurements.

Impact of local atomic stress on oxygen segregation at tilt boundaries in silicon

Using the atom probe tomography, transmission electron microscopy, and ab initio calculations, we investigate the three-dimensional distributions of oxygen atoms segregating at the typical large-angle grain boundaries (GBs) (Σ3{111}, Σ9{221}, Σ9{114}, Σ9{111}/{115}, and Σ27{552}) in Czochralski-grown silicon ingots. Oxygen atoms with a covalent radius that is larger than half of the silicon's radius would segregate at bond-centered positions under tensile stresses above about 2 GPa, so as to attain a more stable bonding network by reducing the local stresses. The number of oxygen atoms segregating in a unit GB area NGB (in atoms/nm2) is hypothesized to be proportional to both the number of the tensilely-stressed positions in a unit boundary area nbc and the average concentration of oxygen atoms around the boundary [Oi] (in at. %) with NGB∼50nbc[Oi]. This indicates that the probability of oxygen atoms at the segregation positions would be, on average, fifty times larger than in bond-centered positions in defect-free regions.

Investigation of dislocation cluster evolution during directional solidification of multicrystalline silicon

Dislocation clusters are the main crystal defects in multicrystalline silicon and are detrimental for solar cell efficiency. They were formed during the silicon ingot casting due to the relaxation of strain energy. The evolution of the dislocation clusters was studied by means of automated analysing tools of the standard wafer and cell production giving information about the cluster development as a function of the ingot height. Due to the observation of the whole wafer surface the point of view is of macroscopic nature. It was found that the dislocations tend to build clusters of high density which usually expand in diameter as a function of ingot height. According to their structure the dislocation clusters can be divided into light and dense clusters. The appearance of both types shows a clear dependence on the orientation of the grain growth direction. Additionally, a process of annihilation of dislocation clusters during the crystallization has been observed. To complement the macroscopic description, the dislocation clusters were also investigates by TEM. It is shown that the dislocations within the subgrain boundaries are closely arranged. Distances of 40–30nm were found. These results lead to the conclusion that the dislocation density within the cluster structure is impossible to quantify by means of etch pit counting.

Environmental influence assessment of China’s multi-crystalline silicon (multi-Si) photovoltaic modules considering recycling process

The environmental burden of multi-Si PV modules in China has been discussed in existing studies, however, their data are mostly from local enterprises, and none of their environmental assessment involves the decommissioning and recycling process. This study quantitatively assesses the life-cycle environmental impacts of Chinese Multi-crystalline Photovoltaic Systems involving the recycling process. The LCA software GaBi is applied to establish the LCA model and to perform the calculation, and ReCiPe method is chosen to quantify the environmental impacts. LCA of production process reveals that Polysilicon production, Cell processing and Modules assembling have relatively higher environmental impact than processes of Industrial silicon smelting and Ingot casting and Wafer slicing. Among the 14 environmental impact categories evaluated by ReCiPe methodology, the most prominent environment impacts are found as Climate Change and Human Toxicity. LCA including recycling process reveals that although recycling process has environmental impact, the recycling scenario has less environmental impact by comparing with the landfill scenario. Among the five manufacturing processes and recycling process, environmental impacts of polysilicon production, cell processing and modules assembling have relatively higher uncertainty, probably because that the environmental impact of these processes is high, and standard error of parameters such as electricity, aluminum and glass in the three processes are high. Findings of our study indicate that proper measures should be taken in the high pollution processes such as polysilicon production and cell processing. In addition, efforts should also be made to enhance the recovery rate and seek for more environmental friendly materials in the recycling process.

Thermal Simulation of Silicon Ingot in Wire-saw Slicing Process

Thermal Simulation of Silicon Ingot in Wire-saw Slicing Process

Growth of High-Quality Multicrystalline Silicon Ingot through the Cristobalite Seeded Method

Growth of High-Quality Multicrystalline Silicon Ingot through the Cristobalite Seeded Method

Design of model experiments for melt flow and solidification in a square container under time-dependent magnetic fields

This paper describes novel equipment for model experiments designed for detailed studies on electromagnetically driven flows as well as solidification and melting processes with low-melting metals in a square-based container. Such model experiments are relevant for a validation of numerical flow simulation, in particular in the field of directional solidification of multi-crystalline photovoltaic silicon ingots. The equipment includes two square-shaped electromagnetic coils and a melt container with a base of 220×220mm2 and thermostat-controlled heat exchangers at top and bottom. A system for dual-plane, spatial- and time-resolved flow measurements as well as for in-situ tracking of the solid-liquid interface is developed on the basis of the ultrasound Doppler velocimetry. The parameters of the model experiment are chosen to meet the scaling laws for a transfer of experimental results to real silicon growth processes. The eutectic GaInSn alloy and elemental gallium with melting points of 10.5°C and 29.8°C, respectively, are used as model substances. Results of experiments for testing the equipment are presented and discussed.

Numerical analysis of steady and transient processes in a directional solidification system

Manufactures of multi-crystalline silicon ingots by means of the directional solidification system (DSS) is important to the solar photovoltaic (PV) cell industry. The quality of the ingots, including the grain size and morphology, is highly related to the shape of the crystal-melt interface during the crystal growth process. We performed numerical simulations to analyze the thermo-fluid field and the shape of the crystal-melt interface both for steady conditions and transient processes. The steady simulations are first validated and then applied to improve the hot zone design in the furnace. The numerical results reveal that, an additional guiding plate weakens the strength of vortex and improves the desired profile of the crystal-melt interface. Based on the steady solutions at an early stage, detailed transient processes of crystal growth can be simulated. Accuracy of the results is supported by comparing the evolutions of crystal heights with the experimental measurements. The excellent agreements demonstrate the applicability of the present numerical methods in simulating a practical and complex system of directional solidification system.

Surface chemical-bonds analysis of silicon particles from diamond-wire cutting of crystalline silicon

The recycling of the Si powder resulting from the kerf loss during silicon ingot cutting into wafers for photovoltaic application shows both significant and achievable economic and environmental benefits. A combined x-ray photoelectron spectroscopy (XPS), attenuated total reflection (ATR)-Fourier transform infrared (FTIR) and micro-Raman spectral analyses were applied to kerf-loss Si powders reclaimed from the diamond wire cutting using different cutting fluids. These spectroscopies performed in suitable configurations for the analysis of particles, yield detailed insights on the surface chemical properties of the powders demonstrating the key role of the cutting fluid nature. A combined XPS core peak, plasmon loss, and valence band study allow assessing a qualitative and quantitative chemical, structural change of the kerf-loss Si powders. The relative contribution of the LO and TO stretching modes to the Si-O-Si absorption band in the ATR-FTIR spectra provide a consistent estimation of the effective oxidation level of the Si powders. The change in the cutting media from deionized water to city water, induces a different silicon oxide layer thickness at the surface of the final kerf-loss Si, depending on the powder reactivity to the media. The surfactant addition induces an enhanced carbon contamination in the form of grafted carbonated species on the surface of the particles. The thickness of the modified surface, depending on the cutting media, was estimated based on a simple model derived from the combined XPS core level and plasmon peak intensities. The effective nature of these carbonated species, sensitive to the water quality, was evidenced based on coupled XPS core peak and valence band study. The present work paves the way to a controlled process to reclaim the kerf-loss Si powder without heavy chemical etching steps.

Towards optimized nucleation control in multicrystalline silicon ingot for solar cells

Single layer silicon beads (SLSB) and chunks (SLSC) coated with silicon nitride were used as seed layer to grow high-quality multicrystalline silicon (mc-Si) ingots. The mc-Si ingots were grown using various sizes of the Si beads and chunks in the range of <1mm up to 4mm as seed layer. Grain size, grain orientation and grain boundary density were analyzed over ingot heights. It was found that SLSB and SLSC seeded mc- Si wafers have uniform small grains, more randomly orientated grains and high density of random grain boundary in comparison to the conventional mc-Si wafers. The density of dislocation cluster was significantly lower in SLSB and SLSC seeded mc-Si wafers than conventional mc-Si wafers. The reason responsible for dislocation clusters reduction in SLSB and SLSC seeded mc-Si wafers was discussed on the basis of the difference in the grain size and density of random grain boundary. However, density of dislocation cluster was slightly high in SLSC seeded mc-Si wafers than SLSB seeded mc-Si wafers. The indentation effects and irregular shape of Si chunks to cause inhomogeneous stress that may account for dislocation generation and/or multiplication in SLSC mc-Si wafers.

Numerical investigation of an experimental Kyropoulos process to grow silicon ingots for photovoltaic application

Kyropoulos crystal growth seeding process for silicon was studied with 3D modeling of the full furnace. The process had three heating zones (top, bottom and side) to improve the control of the horizontal and vertical thermal gradients in the melt; and, a square crucible to reshape the final crystal into a square horizontal section. Modeling of the starting configuration, where crystal growth from the seed was not possible experimentally, explained that the ascendant melt flow in the center is bringing hot melt to the seed. The addition of a ceramic tube as a radiation shield around the seeding zone and the increase of side heating were modeled numerically as solutions to reverse the melt flow direction. The applied solutions on the experimental furnace allowed the growth from the seed. The process was sensitive to symmetry loss. The asymmetric growth of the crystal towards one preferential direction was explained by a coupling between the sudden increase of emissivity due to solidification of silicon and the local decrease of convection on the side of the preferred growth direction. A symmetric growth was obtained in the model by lowering radiative cooling at the top of the melt surface during the first stage of the growth, and by homogenizing the melt using crystal rotation. As a result, symmetrical circular crystals were obtained experimentally using the same growth conditions without contact with the crucible and with no upward pulling.

3D numerical study of coupled crystallization and carbon segregation during multi-crystalline silicon ingot solidification

During the solidification of silicon ingot, the carbon concentration may play a critical role with regard to the final silicon microstructure. If the carbon concentration passes the solubility limit, SiC particles precipitate and it is possible the grain structure changes from columnar to equiaxed grains. Furthermore, carbon atoms are confined inside these equiaxed grains during solidification, which may decrease the carbon concentration inside the liquid silicon, and the equiaxed grains then become columnar again. Therefore, both solidification and carbon concentration will impact each other. A 3D multiscale model is developed to simulate the microstructural evolution of Multi-Crystalline Silicon (mc-Si) ingots coupled with the carbon concentration during the casting process. Our microstructural model is based on the simulation of nucleation and crystal growth. Besides modelling the nucleation due to silicon-crucible interaction, this model can successfully represent the occurrence of equiaxed grains observed ahead of a planar faceted interface due to carbon segregation during solidification. In addition, by coupling concentration and crystallization, the transition from columnar to equiaxed growth and equiaxed to columnar growth can be modeled. Several comparisons are made between the mc-Si ingot's microstructure obtained by our model and the experimental data in order to evaluate our model. These comparisons show good agreement between our model and the experimental data.

Numerical analysis of the relation between dislocation density and residual strain in silicon ingots used in solar cells

We have developed a three dimensional Haasen-Alexander-Sumino model to investigate the distribution of dislocation density and residual strain in Si crystals and compared the calculation results with experimental data performed in mono-like and multicrystalline silicon ingots. The results show that the residual strain in a multicrystal is lower than in a mono-like crystal, whereas the dislocation density in the multicrystal is higher than that in the mono-like crystal. This phenomenon is due to the relation between dislocation density and residual strain caused by the difference of activated slip systems in a mono-like crystal and a multicrystal.

The development of a purification technique of metallurgical silicon to silicon of the solar brand

The experimental results that show the possibility of obtaining silicon of the solar brand by the recrystallization of metallurgical silicon in fusible metals, e.g., tin, and growing monocrystalline silicon ingots from the obtained silicon scales by the Czochralski method are presented. The experiments on purifying fusible metal (tin) after ending a cycle to obtain silicon scales for the purpose of its repeated use were made. Tin purification was carried out by the method of vacuum decontamination of tin melt, its filtration, and then zone recrystallization. The qualitative and quantitative analysis of the initial materials (silicon and tin) and their structure after various stages of the technological process was carried out by the method of the roentgen-fluorescent analysis on the Elvax light device. The structural features of the obtained silicon scales were considered by means of raster electronic microscopy on the REMMA106I device. The conductivity type and the specific electric resistance of the obtained silicon monocrystalline ingot were measured by the fourprobe method on the PIUS-1UM-K device. It was shown that the grown monocrystalline silicon ingot has silicon content of at least 99.999% weight, n-type conductivity, and specific electric resistance of at least 2.0 Оhm сm. The described parameters correspond to the silicon the solar brand and confirm the possibility of obtaining it from metallurgical silicon by recrystallization in fusible metals, for example, tin.

A thermodynamic model for the formation of bubble defects in multicrystalline silicon ingot

Preparation of multicrystalline silicon for solar cells is significant to photovoltaic industry. Up to now, the formation of bubble defects during directional solidification of multicrystalline silicon has not been explored clearly. In the present experiment, spherical bubble defects distributed randomly at ingot top, and the surface of bubble defects was covered with SiC or Si3N4 particle. A thermodynamic model was proposed to illustrate the formation of bubble defects. Bubble defects were caused by gaseous SiO in silicon melt. SiO was the reaction product of melt silicon and oxygen. Oxygen was dissolved by quartz crucible, diffusing through broken Si3N4 coatings into silicon melt. Gaseous SiO in the silicon melt nucleated and grew into bubbles against insoluble particles, such as SiC and Si3N4, and finally enveloped in solid silicon, producing bubble defects. According to the principle of thermodynamics, the formation process was discussed in detail.

Mechanism of material removal in abrasive electrochemical multi-wire sawing of multi-crystalline silicon ingots into wafers

Abrasive electrochemical multi-wire sawing is a comprehensive process that consists mainly of mechanical grinding complemented by anodic oxidation (erosion). The mechanism of material removal of this method is different from that of traditional multi-wire sawing. In this study, the electrochemical reaction of silicon in ethylene glycol was investigated using an electrochemical workstation (PGSTAT 302N). The oxidation peak of the current flow was achieved at a potential of 5 V versus Ag/AgCl during a cyclic voltammetry test. The occurrence of anodic oxidation of silicon was also confirmed. The properties of the anodic oxide were then analyzed. The oxide layer reached a depth of 90 nm after 5 min at 20 V. A few localized porous features were generated on the surface of the anodic oxide layer because of the occurrence of anodic oxidation (erosion) on the original surface. The oxide with loose and porous structure was easily removed. Such feature is beneficial for minimizing cutting load and improving slicing efficiency. Mechanical grinding is thus favorable for continuous anodic oxidation (erosion). Notably, the removal of material is achieved by the interaction of grinding and anodic oxidation. Future research will focus on fixed abrasive slicing, which remains a bottleneck for the texturing of sliced multi-crystalline silicon wafers.

Evolution of grain structure and recombination active dislocations in extraordinary tall conventional and high performance multi-crystalline silicon ingots

In this work one high performance multi-crystalline silicon ingot and one conventional multi-crystalline silicon ingot, each with an extraordinary ingot height of 710mm, were replicated by the successive growth of eight G1 ingots to evaluate the potential advantage of extraordinary tall HPM ingots in industrial production. By analyzing different grain structure parameters like mean grain size, grain orientation and grain boundary type distribution as well as the recombination active dislocation area over the complete ingot height, it was observed that the material properties strongly differ in the initial state of growth for the two material types. However, at ingot heights above 350mm, the difference has vanished and the grain structure properties for both materials appear similar. It is shown that the evolution of the grain structure in both material types can be explained by the same grain selection and grain boundary generation/annihilation mechanisms whereas the current grain structure determines which mechanisms are the most dominant at a specific ingot height. Since the grain structure directly influences the dislocation content in the silicon material, also the recombination active dislocation area becomes equal in high performance and conventional multi-crystalline silicon material at ingot heights above 350mm. From these results it is concluded that the advantage of high performance silicon material is limited to the first grown 350mm of the ingot.