Search

A contactless, non-destructive approach to measure the geometrical parameters of the growth ridge, based on surface topography, is presented and established. It allows a systematic, large scale analysis of growth ridges of single crystals of almost any type. Here, it is applied to Czochralski-grown silicon crystals. Based on the measurement results, Voronkov’s theory of the shape of the growth ridge is verified. This theory is used to calculate the temperature gradient at the growth ridge from the geometrical parameters. The presented method gives an easy, direct experimental access to the thermal conditions, both qualitative and quantitative, at the solid-liquid interface during the growth process.

In the photovoltaic industry a total of 100,000 tons of silicon is lost as waste per year. This waste is originating from several cropping and sawing steps of the high purity silicon blocks and ingots during the solar cell wafer production, resulting in a silicon containing suspension. Among different approaches to recycle the silicon from this waste is the utilization of hydrocyclones, which can be used to separate or classify particles by weight and size. In this work the use of a hydrocyclone was evaluated to upgrade the silicon fraction from a typical sawing waste. A potential field of use for the recycled silicon particles might be as anode material for next generation lithium ion batteries.

We studied the impact of melt removal on the formation of the top redzone in directionally solidified silicon. Solidification experiments in G1-scale were carried out during which the residual melt was removed before full solidification by dumping. The impact of this procedure on the extension of the top redzone was evaluated by lifetime mappings and lifetime profiles extracted therefrom. A lineshape analysis of these profiles enabled to distinguish between cooldown time effects and the impact of the iron contamination at the ingot top. The experiments show that the contamination and thereby the size of the top redzone can be slightly reduced by the melt removal. We found indications that a contamination via the gas phase is partially or fully responsible for the remaining top redzone and corroborated this by analysis of an annealing experiment in the crystallization furnace.

Lumped-parameter models are derived from boundary layer and other physical arguments to describe oxygen concentration levels during the Czochralski (CZ) growth of silicon. These models are assessed against predictions from a detailed, high-fidelity 2D-3D numerical simulation of the entire CZ puller, whose solutions are realistic but require intense computational effort. Comparisons of predictions show that the lumped-parameter model captures the correct trends of melt oxygen levels influenced by melt height, crucible rotation, and crystal rotation. A simple fitting of coefficients provides reasonably good quantitative predictions by the lumped-parameter model, and its near-instantaneous computations make it an interesting candidate for real-time growth optimization and control. Possible model improvements and extensions are discussed.

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.

The impurity level of cast silicon ingots for photovoltaic application has become more and more crucial over the last years with respect to the competitiveness to the high quality monocrystalline Czochralski material. Typical contaminants diffusing into the silicon melt and ingot during the casting process are metals, dopants and light elements like C, O and N. They originate from the consumables like crucible, crucible coating and silicon feedstock as well as the furnace atmosphere. If high quality electronic grade or solar grade silicon feedstock is used, it exhibits several orders of magnitude lower impurities than the other consumables and therefore is mostly neglected during optimization processes in industrial environment. However, the pure feedstock is exposed to the furnace atmosphere during the heating process for long timescales at high temperatures even before melting and solidification. This can lead to a pronounced contamination of the feedstock particles, which negatively influences the resulting material quality of the silicon ingots. In this work, this phenomenon is investigated in laboratory scale by carrying out several two-step G1 experiments containing a pre-annealing step of the feedstock in different crucible setups followed by a regular crystallization step in a virgin crucible. Through lifetime, interstitial iron, resistivity and C-/O-measurements, the effect of feedstock contamination by the surrounding environment on the ingot quality was evaluated. The results show that a significant feedstock contamination by boron, iron and carbon species takes place if the vacuum phase during heating phase is too long. In contrast, no significant oxygen contamination could be observed. Further, the main sources for iron and boron contamination via gas phase diffusion were identified to be the silica crucible and the Si3N4 coating, respectively. In contrast, the contamination by the furnace itself (e.g. graphite parts) plays only a minor role concerning these elements, but is mainly responsible for carbon contamination.

Three different SiO2-based barrier layers, used in industry for prevention of metallic impurity diffusion into directionally solidified silicon ingots, were investigated in detail by G1 experiments in order to find out the most relevant barrier properties responsible for the blocking behaviour. Minority carrier lifetime and interstitial iron measurements on the grown silicon ingots show significant differences between the barriers regarding the red zone extension as well as the maximum lifetime/minimum iron content in the ingot centre. Structural analysis of the barriers by optical microscopy and Raman spectroscopy reveal a clear correlation to the porosity and thickness of the barrier layers. Further, indications were found that the barriers become effective only during the process and that this should occur as early as possible to obtain an optimal impact of the barrier. Only one of the tested barriers shows an overall positive impact on the lifetime, coming close to using a high purity quartz glass SiO2 crucible instead of the standard ceramic SiO2 crucible.

Gas bubbles in crucibles for Czochralski (Cz) silicon growth are both necessary and detrimental: In the outer, bubble-containing (BC) layer of the crucible, they are required for mechanical stability, while in the inner, bubble-free (BF) layer, bubbles can cause the release of particles from the crucible into the melt which may disrupt the single-crystalline growth. In this work, a vacuum bake-out test (VBT) procedure was set up for unused crucible parts and a microscopic characterization routine was developed to systematically investigate bubble formation and growth. Longer process time, higher temperature, and lower atmospheric pressure lead to an increased bubble growth in both, the BC and BF layer. During the VBT, no new bubbles form in the BF layer, while existing bubbles grow. The comparison to experimental data from crucibles used in an industrial Cz process indicates that VBTs can simulate this process. This allows the prediction of the gas-bubble formation in Cz crucibles using a cost-effective and less time-consuming analyzation method.

The growth of silicon crystals from a melt contained in a granulate crucible significantly differs from the classical growth techniques because of the granulate feedstock and the continuous growth process. We performed a systematic study of impurities and structural defects in several such crystals with diameters up to 60 mm. The possible origin of various defects is discussed and attributed to feedstock (concentration of transition metals), growth setup (carbon concentration), or growth process (dislocation density), showing the potential for further optimization. A distinct correlation between crystal defects and bulk carrier lifetime is observed. A bulk carrier lifetime with values up to 600 μs on passivated surfaces of dislocation-free parts of the crystal is currently achieved.

The contamination of silicon by the furnace atmosphere is investigated based on lab-scale annealing experiments in a carbon-based furnace. The analysis comprises measurements of charge carrier lifetime and interstitial iron as well as elemental analysis. An approach to obtain information on the concentration of the metal species involved in the contamination process based on charge carrier lifetime profiles is established. A significant reduction of the charge carrier lifetime below the sample surface is observed. This reduction is governed by iron and at least one (cobalt and/or nickel) faster diffusing species. The contributions from the furnace parts and the purging gas are investigated and possibilities to reduce the gas phase contamination are elaborated. The results indicate a possible relevance of gas phase contamination in crystallization experiments.

A setup for annealing experiments was developed which allows the detailed investigation of metal impurity solid state diffusion into silicon crystals. In order to simulate the impurity diffusion into multicrystalline or mono-cast silicon ingots during directional solidification processes, small cuboidal Czochralski silicon ingots with extremely low metal contamination, placed on various substrates, were annealed for several hours at temperatures just below the melting point of silicon. As substrates, different crucible materials, partly including SiO2 diffusion barrier layers as well as different Si3N4-coatings with widely varying content of metals (Fe, Cr, Ti, Cu, Co and Ni) were used. After ingot annealing, vertical samples were prepared out of the ingot center and the red zone dimensions were determined by lifetime measurements. Additionally, highly sensitive neutron activation elemental analysis was carried out on the silicon samples. In combination with the concentrations of transition elements of various crucible materials and silicon nitride powders the solid state diffusion process of these impurity species into and through the silicon was investigated. The results indicated, that the red zone area mainly contains the fast and intermediate diffusing elements Fe, Co, Cr and Cu. By variation of certain metal element concentration (Fe, Cr, Co, Ni, Cu and Ti) in the substrates it turned out that Fe and Co have the major impact on red zone formation and charge carrier lifetime reduction. Whereas for increasing Fe concentrations in the substrate material (>1017 at/cm3) no further increase of red zone can be observed due to the formation of precipitates in the lower areas of the red zone, fast diffusing Co atoms are able to reduce also the charge carrier lifetime in the center of the silicon samples.

During the growth of [0 0 1]-oriented, heavily n-type doped silicon crystals by the Czochralski (CZ) method dislocation formation occurs frequently which leads to a reduction of the crystal yield. In this publication the evolution of the solid-liquid interface and the formation of the {1 1 1} edge facets are analyzed on a microscopic scale as possible reason for dislocation formation in heavily n-type doped [0 0 1]-oriented CZ crystals. A correlation between the length of the {1 1 1} edge facets and the curvature of the interface is found. They ultimately promote supercooled areas and interrupted growth kinetics, which increase the probability for dislocation formation at the boundary between the {1 1 1} edge facets and the atomically rough interface.

In this work one high performance multi-crystalline silicon ingot and one conventional multi-crystalline silicon ingot, each with an extraordinary ingot height of 710 mm, 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 350 mm, 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 350 mm. From these results it is concluded that the advantage of high performance silicon material is limited to the first grown 350 mm of the ingot.