Search

In the present study, the effects of the solidification parameters and magnetic field on the enrichment of primary Si via separation of a hypereutectic Ti–85 wt.% Si alloy during directional solidification were experimentally investigated. The results indicated that increasing the pull-down rate from 5 to 15 μm/s increased the temperature gradient and growth rate at the same position of the ingot, although the Si content decreased. At a certain pull-down rate (e.g., 5 μm/s), with the increasing thickness of the Si-rich layer, the temperature gradient, growth rate, and Si content all decreased. In presence of a magnetic field, the solidification direction produced an electromagnetic force, which acted on the tip of the dendrite and changed the shape of the separation interface. Meanwhile, the electromagnetic stirring enhanced mass transfer at pull-down rates of 5, 10, and 15 μm/s, while dendritic breakdown processes maintained corresponding enrichment layer thicknesses of 45 mm, 35 mm, and 15 mm, respectively. The results could be helpful for expanding the applications of Ti–Si alloys and solving the problems presented by the solid waste of Ti-BF slag, i.e., environmental pollution and wastage of Ti resources.

In this work a new type of contamination-free Si x N y seed particles for the directional solidification of HPM silicon ingots is introduced and evaluated. Several G1 crystallization experiments have been carried out to examine the influence of the seed particle size, the introduced mass and the influence of the composition of the underlying Si 3 N 4 coating at the crucible bottom. Finally, multi-PERC cells over the ingot height from one of the most promising Si x N y seeded ingot have been prepared. The results show that an oxygen contamination is not occurring by the new particles. A minimum seed particle density of around 120/cm 2 at the crucible bottom is sufficient to generate a HPM like grain structure with random grain boundary fractions R of 65%±5%, which is very close to typical values of classical silicon seeded HPM. A further increase of the particle density doesn’t lead to a further increase of the R value due to the competition between the very close packed seeds. While lower oxygen levels of the Si 3 N 4 -coating promote significant higher R values in crucible setups without seed particles, the Si 3 N 4 -coating composition has almost no influence on the resulting grain structure properties when Si x N y seed particles are used. This could be shown even for small seed densities of ~30 cm −2 and a remaining coating area fraction of 80%. The cell analysis shows same efficiencies of the ingots seeded on the Si x N y particles and on the silicon feedstock layer, but a higher wafer yield can be achieved in the case of the Si x N y seeding method since a full melt process can be used.

An elliptical region was observed on the wafer cut from the ingot along (1 1 1) face after etching and the ingot was grown by the Vertical Bridgman method. The wafers with the elliptical region were characterized with zinc concentration distribution scanner and infrared transmission microscope. Infrared transmission images revealed the morphology evolution of Te inclusions from irregular droplets to regular tetrahedrons and different stages from the Te-rich threads to the pearl-string Te inclusions inside the elliptical region. Te atoms preferentially diffused and aggregated in the dislocation direction, which resulted in the formation of Te-rich droplets and Te-rich threads. The Zn component map indicated that Zn concentration varied significantly at the elliptical boundary, the Zn concentration inside the ellipse was lower than other normal region. Based on the above experiments, it was considered that such an elliptical region created because of the occurrence of constitutional supercooling, in other words, it is caused by a drastic change of the Zn concentration gradient in the liquid phase at the front of the solid-liquid interface. The segregation of the Zn concentration in the liquid phase at the front of the interface was related to the growth rate. During crystal growth, when approaching the region where the Zn concentration changed drastically, the growth rate should be appropriately lowered to meet the normal Zn concentration gradient to avoid the generation of elliptical region.

Directional solidification process is the main process for producing the multi-crystalline silicon ingots. It is economically beneficial to grow the good quality mc-Si ingots with lower power consumption. Here we have carried out numerical simualtion for analysing the impact of addition of an insulation block on the DS furnace. The addition of the insulation block has been implemented experimentally and the quality of the grown ingot was evaluated by the minority carrier lifetime measurement. Though the temperature gradient and hence the growth rate are higher in the insulation block added DS furnace, the generated dislocation density has not increased beyond the acceptable limit 108 1/m2. The power used by the DS furnace has been reduced considerably by the addition of the insulation block. The good quality mc-Si ingot with lower power consumption has been obtained as a result of insulation block addition.

High performance multi-crystalline silicon (HPM-Si) for the use in photovoltaics is characterized by a very fine grain structure and a high content of random grain boundaries, finally resulting in a low dislocation density and consequently in a high material quality. Typically, the grain size increases and the fraction of random grain boundaries decreases over ingot height due to annihilation mechanisms, especially in the first 150 mm. One approach for further material improvement is to further increase the initial random grain boundary fraction and to maintain it as high as possible over the complete ingot height. In this work, several theoretical approaches to achieve these points were evaluated by experiments in G1 scale. Firstly, the influence of the silicon seeding material on the initial grain structure was investigated regarding the effect of extremely fine Si particles in the µm to nm range and the bulk density of the particle layer. Secondly, the effect of the initial geometrical grain boundary arrangement in the seed layer was evaluated. For that purpose, special seed alignments similar to the Quasimono approach were tested. Finally, the process parameter growth rate was varied in a wide range to investigate its influence on the evolution of the grain boundaries during growth. The results show that the optimum for the initial random grain boundary fraction is already reached by the existing methods/commonly used seed materials. Concerning the decrease of the random grain boundary fraction over ingot height, some technical aspects were identified which are able to keep the amount of random grain boundaries at a high level. However, the practical realization within an industrial setup might be difficult.

It is well known that manufactures of power electronic devices rely on FZ silicon wafers of highest possible purity. The manufactures however also have a demand for wafers with lower and lower radial resistivity variation (RRV) across the wafer. The current product on the 200 mm wafer market with the lowest RRV is based on neutron transmutation doping (NTD) where a RRV value lower than 5% is guaranteed for all wafers cut from an ingot. This work presents a novel concept called Advanced NTD which can provide 200 mm wafers with RRV lower than 2%. The radial resistivity profile of an NTD wafer is influenced by the neutron flux in the nuclear reactor. The result is a resistivity that is lower at the edge compared to centre of the wafer. By engineering a gas phase doped profile to be the inverse of the NTD profile, and simultaneously carefully controlling the required NTD doping factor, an almost flat resistivity profile can be achieved by combining the developed gas phase doping process with the NTD process. The capability of the Advanced NTD process is demonstrated in a pilot production of 11 ingots.

In the production of monocrystalline silicon by the Czochralski process, structure loss frequently reduces yield and increases production time. Structure loss means the transition from monocrystalline to multicrystalline structure. It is typically preceded by generation of dislocations and slip. Previous work concluded that a pinhole (i.e. a cavity from gas bubble in the melt) was the main cause of structure loss in an n-type Czochralski ingot. In the present study, seven industrial n-type Czochralski silicon ingots were investigated to assess whether pinholes frequently contribute to cause structure loss. Pinholes were found within 5 cm above the position where dislocations were generated in four out of the seven ingots. Slices were cut with the pinholes at one surface and etched with Sopori etchant. The pinholes did not appear to have affected the arrangements of etch pits in these samples and investigation of pinholes did not reveal the cause of structure loss. Furthermore, there were not significantly higher pinhole frequency before structure loss compared to ingots without structure loss among 4041 industrial ingots. Modeling of thermal stresses showed stress concentration around a pinhole, and it gives new insight into the level of stress the ingot can endure without generating dislocations and slip.

Oxygen-related thermal donors generated generally form less effectively in multicrystalline than in Czochralski grown monocrystalline silicon. In addition to a lower interstitial oxygen concentration, carbon and extended crystallographic defects are possible candidates that can account for this feature. To isolate the potential effect of extended defects, we investigate the formation of thermal donors in a carbon-lean Czochralski-grown multicrystalline ingot, resulting from a structure loss during growth. The grain orientations and the densities of dislocations are characterized and are shown to be representative of typical multicrystalline silicon. Eventually, we demonstrate that the macroscopic thermal donor formation kinetics are not affected either by grain boundaries or by dislocations, in contrast with previous findings from the literature. Our results indirectly support the role of carbon as the main thermal donor inhibitor in multicrystalline silicon.

We present a technique for modeling three-dimensional (3D) anisotropic crystal shape evolution in meniscus defined growth systems. In particular we apply this model to a simplified low gradient Czochralski growth process used to grow bismuth–germanium–oxide–like crystals. Our Lattice Boltzmann Model (LBM) based approach involves a two-dimensional unstructured mesh (embedded within the 3D LBM grid) which accurately tracks the crystal/melt and melt/gas interfaces while accounting for anisotropic interface attachment kinetics as well as capillarity, involving the enforcement of either isotropic or anisotropic growth angles (GA) at the crystal/melt/gas triple phase line (GA-TPL). The robustness of the algorithm is demonstrated by a series of tests involving mesh refinement, comparisons with results from an older study and an evaluation of the impact on accuracy of a time-saving artificial pull-rate speedup scheme described within the manuscript. Physically significant results (which do not involve diameter control) include a demonstration of the impact of changes in kinetic parameters on resultant ingot shapes, calculations showing how (for conditions considered here) the diameter of the growing crystal decreases with increasing isotropic growth angle values, and growth simulations demonstrating how anisotropic growth angles can impact the shape of the evolving ingot. The dependence of crystal diameter on isotropic growth angle values is reproduced in a local sense in the case of anisotropic growth angles. A specific point on the GA-TPL will tend to extend inwards or outwards (respectively exhibiting a decreased or increased local radial position value) when the growth angle value is locally (and respectively) increased or decreased. As a result, considering crystals whose non-circular cross-sectional shapes are due to kinetic anisotropy as a baseline, growth angle anisotropy acts to either decrease or increase shape anisotropy when the extended (rough) parts of the GA-TPL respectively support large or small local growth angle values as compared to reduced-radius (faceted) parts of the GA-TPL. It should finally be mentioned that our results, concerning the sensitivity of the (local or average) radial position value of the GA-TPL to the growth angle, are consistent with other computational results exhibited in the literature (concerning solidifying semiconductor drops and floating zone growth of oxides).

The influence of different melt streams on the distribution of phosphorus in multi-crystalline silicon ingot was studied. Phosphorus-doped multi-crystalline silicon (mc-Si) ingots were directionally solidified using travelling magnetic fields (TMF) to alter axial phosphorus profiles. Resistivity distributions in the crystallized n-type ingots were measured along the ingot length. Different Lorentz forces were applied in order to enhance melt stirring and with that to transport phosphorus more rapid towards the melt surface. A new rectangular setup was developed, which enables simultaneous directional solidification of 4 G0-sized mc-Si ingots (80 x 80 x 60 mm3) under the influence of TMF. 900 g ingots with different initial level of phosphorus doping were grown, and dopant concentrations in ingots were related to stirring intensities. The phosphorus evaporation rate significantly affects axial dopant profile of mc-Si material, thus this approach can be used as a powerful tool to control and tailor resistivity distribution along phosphorus-doped mc-Si ingots.

Single crystals of PrCo2 have been grown from flux starting from Pr0.44Co0.56 that was synthesized by arc melting of stoichiometric mixtures of the elements. The Pr0.44Co0.56 ingot was loaded in a Boron Nitride (BN) crucible and sealed inside a quartz tube under vacuum. The quartz ampoule was heated to 1203 K, and then slowly cooled down to 923 K at a rate of 2 K/h. The obtained crystals have octahedral shapes and are of cubic C15 Laves phase structure confirmed by room temperature powder and single-crystal X-ray diffraction measurements. Magnetic susceptibility data show the Curie temperature TC = 46.7 K. The experimental value of saturation magnetization Ms = 3.2 μB/f.u is consistent with our theoretical calculations.

The noncontact crucible (NOC) method is defined as a growth method to intentionally establish a distinct low-temperature region inside a Si melt. For the NOC growth of a uniform large Si single ingot without contact with the crucible wall, a large and deep low-temperature region must be created in the upper central part of a Si melt to allow natural crystal growth inside it. The novel method to create the low-temperature region has not been disclosed until now. In this paper, the concept and method are clearly shown on the basis of controlling the heat flow from the bottom heater into the Si melt using an insulating plate locally set under the crucible bottom. The effectiveness of such heat-flow control is proved by the growth of several Si single ingots with a large diameter ratio using the NOC method. The effects of the size and thickness of the insulating plate is also shown by the growth of Si single ingots.

Thermal stress computations during sapphire growth are compared between a resistive Czochralski furnace and a Kyropoulos inductive furnace. 2D – axisymmetric global simulations are performed to compute the thermal field in the furnace and the convection in the melt. Temperatures carried out from global modeling at a given stage of the growth process, are used for thermal stress computations in the crystal. Three-dimensional stress analysis, which takes into account the anisotropic elastic constants of sapphire, shows nearly axisymmetric von Mises stress distribution in the crystal. It is shown that applying 2D – axisymmetric modeling of thermal stress may result in significant errors. Computations performed for a crystal of 10 cm in diameter grown in a Kyropoulos furnace show only a thin region of 2–3 mm with high thermal stress located at the crystal periphery. The model predicts very low thermal stresses in the central part of the crystal. Simulations of an ingot grown by Czochralski technique, show higher thermal stresses in almost the whole volume of the crystal. Numerical computations are in agreement with our previous measurements of dislocation density in sapphire crystals grown by the Kyropoulos method. The present numerical results can explain the experimental observations showing that sapphire crystals grown by Czochralski technique have a much higher dislocation density than Kyropoulos grown ingots.

Analysis of nucleation sites in multicrystalline silicon (mc-Si) grown by directional solidification is required for further grain refinement to reduce dislocation density. In this study, Voronoi diagrams were utilized to analyze nucleation sites of mc-Si grown by single-layer Si beads (SLSB)-seeding method. The grain distribution at the bottom of the ingot was almost reproduced by the weighted Voronoi diagram with a relaxation method to optimize the positions of generating points of the diagram, which correspond to the nucleation sites, and the difference of nucleation timing of each crystal grain. Comparison of the generating points with the optical image indicated that the nucleation started at the remarkably deep portions of the nucleation layer. Further grain refinement by SLSB-seeding method could be achieved by suppressing the formation of the remarkably deep portions of the nucleation layer.

CaMoO4 crystal is a promising scintillator for investigating the neutrinoless double beta decay process and is commonly grown by the Czochralski method. In this study, CaMoO4 single crystals were grown by the vertical Bridgman method using both spontaneous crystallization and seed crystal approaches. The CaMoO4 crystal grown using the spontaneous crystallization approach exhibited yellow, light blue, and deep blue colours (from top to middle part of the ingot). The thermal expansion coefficients, αa and αc, were calculated to be 6.11 × 10−5/°C, and 23.58 × 10−5/°C, respectively. The optical transmission spectra and photo- and thermally stimulated luminescence spectra of the crystals before and after thermal annealing were also investigated. The yellow cubic CaMoO4 crystal showed the highest x-ray-stimulated luminescence intensity and the best light output under the excitation of a 137Cs radioactive source.

The potential for producing a large refined equiaxed zone by ultrasonic treatment (UST) of high purity zinc was investigated in order to improve the mechanical performance and formability. The macrostructure of cast ingots changed from large columnar grains without UST to three zones of fine columnar grains adjacent to the mould walls of the ingot, a refined equiaxed zone and a zone of a mixture of coarse equiaxed and columnar grains. A small zone of equiaxed grains was obtained when UST was applied during cooling from 440 °C to 419 °C for 2 min. The size of the equiaxed zone increased from about 20% of the casting's cross section to 50% when UST was applied for 3 or 4 min. In contrast, the application of UST for a longer time from a higher temperature (450–419 °C for 4 min) resulted in a smaller equiaxed zone of 18% indicating that a specific combination of UST time and temperature is required for the formation of a large equiaxed zone. The factors affecting the formation of the equiaxed zone throughout the solidification cycle are described.

The reduction of SiC, Si3N4 and transition metals impurities in directionally solidified Si ingots poses one of the crucial challenges in the solar cells production. Particularly strong contamination comes from the graphite parts in the hot zone. Therefore, we selected three massive ceramic materials to replace graphite, developed the novel design of the crucible support and cover and compared the crystals grown in them with ingots from the standard graphite design. The experiments were performed for phosphorus n-doped silicon of G0 size. The ingots were compared with respect to O- and C-content, metal impurities, resistivity and lifetime. The superior performance of TiC relative to other ceramics was observed, particularly due to the lower concentration of substitutional carbon in Si ingot (up to 2.6 times) and the higher minority carrier lifetime of (up to 4.4 times) with narrow red zones.

In this study, the power ratio between the top and side heaters and the moving velocity of the side insulation are designed to control the shape of the crystal-melt interface during the growth process of a 1600 kg multi-crystalline silicon ingot. The power ratio and insulation gap are adjusted to ensure solidification of the melt. To ensure that the crystal-melt interface is slightly convex in relation to the melt during the entire solidification process, the power ratio should be augmented gradually in the initial stages while being held to a constant value in the middle stages. Initially the gap between the side and the bottom insulation is kept small to reduce thermal stress inside the seed crystals. However, the growth rate will be slow in the early stages of the solidification process. Therefore, the movement of the side insulation is fast in the initial stages but slower in the middle stages. In the later stages, the side insulation gap is fixed. With these modifications, the convexity of the crystal-melt interface in relation to the melt can be maintained during the growth process with an approximately 41% reduction in the thermal stress inside the growing ingot and an 80% reduction in dislocation density along the center line of the ingot compared with the original case.

The accelerated crucible rotation technique (ACRT) was introduced in the traveling heater method (THM) growth process of detector-grade CdZnTe (CZT) crystals to regulate the convection in the melt and to modify the growth interface morphology. Several ingots with the diameter of 53 mm were grown by THM with/without ACRT. The ingots were quenched during the growth to show both macroscopic and microscopic morphologies of the growth interfaces. The results show that by using ACRT the growth interface can be changed from a concave one to the flat or even convex one depending on the ACRT parameters, which is favorable for reducing nucleation in the melt to get larger CZT grains. Meanwhile, by using ACRT in THM process, the microscopic interface was changed from a diffused one to cellular or even planar one (at suitable ACRT parameters), through which the trapped Te inclusions was decreased for one order. An ingot grown by THM with constant rotation rate of 40 rpm was also grown, which have also reduced the interface curvature in macro-scale and Te inclusions to some extent, but the effects are not as significant as ACRT with high crucible rotation rate.

Oxygen is one of the highest-concentration impurities in single crystals grown by the Czochralski (CZ) process, and seriously impairs the quality of the Si wafer. In this study, computer simulations were applied to design a new CZ system. A more appropriate thermal field was acquired by optimization of the heater structure. The simulation results showed that, compared with the conventional system, the oxygen concentration in the newly designed CZ system was reduced significantly throughout the entire CZ process because of the lower crucible wall temperature and optimized convection. To verify the simulation results, experiments were conducted on an industrial single-crystal furnace. The experimental results showed that the oxygen concentration was reduced significantly, especially at the top of the CZ-Si ingot. Specifically, the oxygen concentration was 6.19 × 1017 atom/cm3 at the top of the CZ-Si ingot with the newly designed CZ system, compared with 9.22 × 1017 atom/cm3 with the conventional system. Corresponding light-induced degradation of solar cells based on the top of crystals from the newly designed CZ system was 1.62%, a reduction of 0.64% compared with crystals from the conventional system (2.26%).

A GaSe: Al (0.13 wt%) single crystal was grown using the Bridgman method combined with a crucible rotation technique. A 35 × 19×5.5 mm 3 sample was cut from the as-grown crystal ingot. The GaSe: Al (0.13 wt%) crystal has an indentation hardness of 2.27 GPa, which is 2.6 times harder than the pure GaSe crystal. In particular, a sample with a thickness of 5.5 mm has an infrared transmission of approximately 60%. The absorption coefficient of this sample is as low as 0.1 cm −1 over the range of 0.83 to −14 μm, which demonstrates its high optical quality. A crystal growth method with the described procedures may be suitable to grow other doped GaSe crystals.

Ce-doped Cs2LiYCl6 (CLYC) has widespread applications in homeland security owing to its excellent neutron/gamma discrimination ability. Ce3+ ions, as luminescent centers of CLYC, have a significant difference in effective ionic radius compared with Y3+, hinting at the severe inhomogeneous distribution of Ce3+ in CLYC. Still, only the relation between scintillation properties and nominal doping concentrations in CLYC crystals has been reported. The segregation behavior of Ce3+ ions along the ingot are still open for research. Herein, we measured the actual Ce3+ doping concentration along the CLYC ingot. Gamma-ray energy spectrum and powder XRD patterns of different samples were utilized to study the effects of Ce segregation on CLYC crystals. The effective segregation coefficient of Ce in CLYC crystals, keff = 0.22, was calculated. In most part of the ingots, the concentration of Ce3+ ions was lower than the nominal doping amount and rose rapidly at the tail. The scintillation properties of CLYC crystals along the ingot were affected by the distribution of Ce concentrations. This work may guide the optimization of properties for large-size CLYC crystals by modifying the distribution of Ce3+ ions.

Two inches in diameter and four inches in length quaternary alloy Cd1−xZnxTe1−ySey (CZTS) ingots have been grown by the Vertical Gradient Freeze Method (VGF) for two different Se concentrations (y = 0.02 and 0.07). The compositional and crystallographic properties, as well as defect morphology of the ingots have been analysed using EDX, TXRF, XRD and IR microscopy. Despite of using these two extreme Se values proposed in the literature, CZTS continues to maintain the proper zinc blende structure of CZT. However, the addition of selenium has the effect of shifting the diffraction peaks to higher angles, or what is the same, to reduce the Cd1−xZnxTe (CZT) constant lattice. It is observed that the addition of Se also favours the reduction of bubble formation along the ingots due to cadmium deficiency and change in Zn concentration along the growth CZTS ingot direction respect to CZT ingot. The VGF growth method provides great confidence and reproducibility since both ingots share similar features in terms of dimensions and defects formation.

We grew large single crystals of metal nickel with a maximum diameter of 8 in. using the Czochralski (CZ) method. Characteristics of the metal material, namely, the seed elongation and intense cooling radiation at high temperatures during crystal growth, are the main challenges when growing ingots with large diameters. These problems were resolved by optimizing the parameters related to crystal growth, such as reduction of crucible and ingot rotation speed, and adjustment of shoulder formation angle of ingot shape. The produced ingots were confirmed to be single crystals by the Laue back-reflection method. Wafers were cut from single crystal ingots using a multi-wire saw with diamond abrasive grains. A smooth, mirror-like surface was obtained on the wafers by using an ordinary mechanical polishing method. The full width at half maximum values of the X-ray rocking curve of the metal nickel crystal surface significantly depends on the surface treatment method. It was found that ordinary mechanical mirror polishing introduces strain on the surface and non-uniformity of the crystal orientation.

In this paper, the effects of Te inclusions on the optical and electrical properties of CdMnTe and CdMgTe crystals grown by the modified vertical Bridgman method were investigated. For CdMnTe ingot, the average size and density distribution of Te inclusions along axial direction were in the range of 12 ~ 29 μm and 1.53 × 104 cm−2 ~ 2.81 × 105 cm−2, respectively. For CdMgTe ingot, the average size and density distribution of Te inclusions were in the range of 6 ~ 20 μm and 8.17 × 103 cm−2 ~ 2.42 × 104 cm−2, respectively. The density and size of Te inclusions increased gradually from the first-to-freeze part to the last-to-freeze part for CdMnTe ingot, while that of CdMgTe ingot first decreased and then increased. For both CdMnTe and CdMgTe ingots, the resistivity, IR transmittance and FWHM of (D0, X) peak first increased and then decreased from the first-to-freeze part to the last-to-freeze part. The middle-to-freeze parts of both ingots had the best crystal quality. For both CdMnTe and CdMgTe detectors, the energy resolution and (μτ)e value were first increased and then decreased from the first-to-freeze part to the last-to-freeze part. The detectors fabricated by the middle-to-freeze parts of both crystals had the best performance and can could meet the requirements of room temperature radiation detection. PACS: 78.30.Fs; 78.55.Et; 81.40.Ef.

The present study is focused on the preparation of single crystalline LaPd 2 Al 2 by the Czochralski method. Differential scanning calorimetry (DSC) and energy dispersive X-ray spectroscopy (EDX) analyses reveal that LaPd 2 Al 2 is an incongruently melting phase which causes difficulties for the preparation of single crystalline LaPd 2 Al 2 by the Czochralski method. Therefore several non-stoichiometric polycrystalline samples were studied for its preparation. Finally the successful growth of LaPd 2 Al 2 without foreign phases has been achieved by using a non-stoichiometric precursor with atomic composition 22:39:39 (La:Pd:Al). X-ray powder diffraction, EDX analysis and DSC were used for the characterisation. A single crystalline sample was separated from the ingot prepared by the Czochralski method using the non-stoichiometric precursor. The presented procedure for the preparation of pure single phase LaPd 2 Al 2 could be modified for other incongruently melting phases.

The growth of mono-like ingot by directional solidification has suffered serious problems in defect control. We proposed a simple approach by using seed partitions, and the grown crystal had much lower defects and better orientation uniformity. Furthermore, the partitions allowed the much easier seed preparation, which had a significant advantage in production. The concept was demonstrated by a G1 experiment, and the detailed defect analyses were carried out. The wafers after gettering had the best lifetime of more than 1 ms after surface passivation. The color mismatch in the appearance of the solar cells made from the wafer was also significantly mitigated.

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.

When continuous silicon ingot casting is performed using directional solidification—a metallurgical refining method for producing high-purity silicon ingots—the thermal stress and microstructure inside the ingots must be monitored for ensuring the desired yield and purity. This study determined the temperature gradient and thermal stress during silicon ingot casting by applying various controlled parameters to the process simulation. Our results indicate that the temperature gradient influences the thermal stress generated during casting; reducing the temperature gradient can help reduce the thermal stress. The ingot exhibited the lowest thermal stress at an electron beam (E-beam) power of 6 kW, an ingot growth rate of 0.02 mm/min, and a molten silicon level of −10 mm. Quasi-single crystalline silicon ingot was produced via crystal growth simulation under thermal stress-minimizing conditions with single-crystalline seed. The quasi-single crystalline silicon ingot maintained the [0 0 1] grain orientation and the purity increased from 2N to 7N6.

The emergence of the high-performance multi-crystalline silicon (HP mc-Si) in 2011 has made a significant impact to photovoltaic industry. In addition to the much better ingot uniformity and production yield, HP mc-Si also has better material quality for solar cells. As a result, the average efficiency of solar cells made from HP mc-Si in production increased from 16.6% in 2011 to 18.5% or beyond in 2016. More importantly, the efficiency distribution became much narrower; the difference from various producers became smaller as well. Unlike the conventional way of having large grains and electrically-inactive twin boundaries, the crystal growth of HP mc-Si by directional solidification is initiated from uniform small grains having a high fraction of random grain boundaries. The grains developed from such grain structures significantly relax thermal stress and suppress the massive generation and propagation of dislocation clusters. The gettering efficacy of HP mc-Si is also superior to the conventional one. Nowadays, most of commercial mc-Si is grown by this approach, which could be implemented by either seeded with silicon particles or controlled nucleation, e.g., using nucleation agent coating. The future improvement of this technology will also be discussed in this review.

The noncontact crucible (NOC) method was proposed for obtaining Si single bulk crystals with a large diameter and volume using a cast furnace and solar cells with high conversion efficiency and yield. This method has several novel characteristics that originate from its key feature that ingots can be grown inside a Si melt without contact with a crucible wall. Si ingots for solar cells were grown by utilizing the merits resulting from these characteristics. Single ingots with high quality were grown by the NOC method after furnace cleaning, and the minority carrier lifetime was measured to investigate reduction of the number of impurities. A p -type ingot with a convex growth interface in the growth direction was also grown after furnace cleaning. For p -type solar cells prepared using wafers cut from the ingot, the highest and average conversion efficiencies were 19.14% and 19.0%, respectively, which were obtained using the same solar cell structure and process as those employed to obtain a conversion efficiency of 19.1% for a p -type Czochralski (CZ) wafer. Using the cast furnace, solar cells with a conversion efficiency and yield as high as those of CZ solar cells were obtained by the NOC method.

In this study, a global transient numerical simulation of silicon growth from the beginning of the solidification process until the end of the cooling process is carried out modeling the growth of an 800 kg ingot in an industrial seeded directional solidification furnace. The standard furnace is modified by the addition of insulating blocks in the hot zone. The simulation results show that there is a significant decrease in the thermal stress and dislocation density in the modified model as compared to the standard one (a maximal decrease of 23% and 75% along the center line of ingot for thermal stress and dislocation density, respectively). This modification reduces the heating power consumption for solidification of the silicon melt by about 17% and shortens the growth time by about 2.5 h. Moreover, it is found that adjusting the operating conditions of modified model to obtain the lower growth rate during the early stages of the solidification process can lower dislocation density and total heater power.

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

Dislocation structure of the Ge single crystals grown by Czochralski method with low thermal gradient has been studied. The selective etching technique and the X-Ray transmission and reflection topography were used. Clearly defined non-uniform dislocation distribution over the crystal cross - section is revealed. Helical dislocations and sets of prismatic dislocation loops are registered. Helical dislocations perpendicular to the ingot axis are situated near the boundary between the regions with low and high dislocation densities (102 and 103 cm−2, respectively). Their length can be as much as several millimeters. Dislocation formations lying at a 35.3° to the crystal axis along directions are also observed. These formations have the shape of prism confined by {111} planes.

The relationship among lattice constant a , Al content, and acoustic properties were experimentally examined using a plate specimen perpendicular to Y -axis prepared from Ca 3 Ta(Ga 0.75 Al 0.25 ) 3 Si 2 O 14 [CTGAS(0.25)] single crystal grown by Czochralski method. As the acoustic properties, leaky surface acoustic wave (LSAW) velocities with different propagation directions, X - and Z -propagations, and longitudinal wave velocity propagating along Y -axis direction were measured by the line-focus-beam/plane-wave ultrasonic-material-characterization (LFB/PW-UMC) system. The measured results of LSAW velocity distributions revealed inhomogeneity in radial direction of the crystal ingot exhibiting lower velocity area at the center of the ingot. In addition, the distributions of lattice constant a and chemical composition (especially Al content) were measured along the radial direction. Abnormal changes suggesting existence of residual stresses concentrated on the central part of the crystal ingot other than the effect of chemical composition change were detected from the relationships among the measured parameters.

Transient global simulations were carried out to study the effects of the furnace pressure on oxygen (O) and carbon (C) coupled transport in an industrial directional solidification (DS) furnace for silicon ingots. The global simulations of impurity transport, taking into account the O and C impurity segregation, chemical reaction on the graphite surface, evaporation of SiO and dissolution of CO at the melt-gas surface, were based on a fully coupled calculation of the thermal and flow fields of the DS furnace. It was found that the furnace pressure affects the intensity and structure of argon flow above the melt surface, which further influences the O and C impurities transport in the DS furnace. The O concentration in the grown ingot decreases with the decreasing furnace pressure. While the C concentration first decreases, and then increases when the pressure is smaller than 200 mbar due to the strong diffusion effect of the CO through the gap between crucible and cover. The numerical predictions of O and C concentrations agree well with the experimental measurements.

The grain structure of high-performance (HP) multicrystalline silicon (mc-Si) is characterized by a small initial grain size with randomly oriented grains and a high length fraction of random grain boundaries. However, the remaining unmelted feedstock at the ingot bottom used as seeding layer for achieving the HP mc-Si properties in the standard crystallization procedure causes yield loss. To overcome this disadvantage, the influence of wetting angle, and surface roughness of non-Si nucleation layers at the crucible bottom on the grain structure properties of mc-Si ingots with a weight of 14.5 kg was investigated and compared to classical HP mc-Si. For that purpose, SiC and SiO 2 nucleation layers realized by spraying and embedding of particles with different sizes resulting in different surface morphologies and wetting angles were studied. Nucleation on rough layers of both materials with a root mean square roughness value greater than 100 µm yielded an initially fine grain structure comparable to HP mc-Si. This did not necessarily result in a random orientation distribution and high length fraction of random grain boundaries. Nucleation on SiC layers caused random grain boundary length fractions between 20 and 30% and non-uniform grain distributions. But, nucleation on SiO 2 layers yielded increased random grain boundary length fractions between 50 and 70% and homogenous grain distributions, both values are similar to HP mc-Si. These differences are discussed in terms of the thermal conductivity of the different nucleation layers.

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 40–30 nm 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.

The In-doped Cd 0.9 Mn 0.1 Te (CdMnTe:In) crystal was grown by the Travelling Heater Method. The crystallity of the CdMnTe:In wafers extracted from different part of the CdMnTe:In ingot was studied including the impurity contents, the Te inclusions and the Mn composition. The point defects in the CdMnTe:In wafers were revealed by the Photo-Induced Current Transient Spectroscopy (PICTS). The ionization energy, the capture cross-section and the concentration of the defect traps in different parts of the ingot were identified. The PICTS result showed the concentration of the V Cd decreases from the top part to the tail part along the bulk while the concentration of A-center increase. The wafers from the middle part of the bulk were proved to have less deep level point defects compared to other parts of the crystal bulk, and had higher electronic performance with the resistivity up to 3.07 × 10 10 Ω cm and the μτ value up to 1.45 × 10 −3 cm 2 V −1 . The concentration of the deep level point defects plays an important role in the electrical properties, especially for the mobility-lifetime product.

Vanadium (V) doped Cd 0.9 Mn 0.1 Te (CdMnTe:V) crystals were grown with a nominal Vanadium concentration of 1×10 17 atoms/cm 3 from excess 10 at% Te solution that was carried out by the vertical Bridgman method with accelerated crucible rotation technique(ACRT). The as-grown crystals display a high resistance characteristic of 4.123×10 1 °Ω·cm in the wafer cutting from the middle part of the ingot. The infrared microscopy images show that the planar density of Te inclusions/precipitates in the crystals is between 1.4×10 3 to 6×10 5 cm −2 . The measured highest IR transmission in the middle part of the ingot of 63% is near the theoretical limit of 65%. Moreover, the PL spectra show a sharp (D°, X) peak, a flat D complex peak and a low DAP peak. The (D°, X) peak has a FWHM of 4.36 meV, indicating a high quality of crystallization, while a flat D complex peak means very low dislocations and defects relative to the vacancies of Cd in the middle part of this ingot. The low DAP peak with a relative intensity of I DAP /I (D ° , X) of 0.045 demonstrates low impurity concentration in this crystal.

Photoluminescence (PL) images and micro-PL maps were taken on Cu- or Ni-doped monocrystalline silicon wafers, to investigate the distribution of the metal precipitates. Several n-type and p-type wafers were used in which Cu or Ni were introduced in the starting melt of the ingots and precipitated during the ingot cooling (as opposed to surface contamination). The micro-PL mapping allowed investigation of the metal precipitates with a higher spatial resolution. Markedly different precipitation patterns were observed in n- and p-type samples: in both Cu- and Ni-doped n-type samples, circular central regions and edge regions were observed. In these regions, particles were distributed randomly and homogeneously. In the p-type Cu-doped and Ni-doped samples, by contrast, the precipitates occurred in lines along orientations. The difference in the precipitation behaviour in n- and p-type samples is conjectured to be caused by different concentrations of self-interstitials and vacancies remaining in the crystal during the ingot cooling: there are more vacancies in the n-type ingots but more interstitials in the p-type ingots. The dopant effects on the intrinsic point defect concentrations in silicon crystals and possible precipitation mechanisms are discussed based on the findings in this work and the literature.

The influence of Sn doping on improvement of minority carrier lifetime (MCL) of Fe contaminated directionally solidified p-type multi-crystalline Si ingots is studied. The macrostructure and resistivity distribution of the Si ingots indicate that no significant difference exists with and without Sn doping. The average MCL increase by 26.2%, 31.8%, 8.1% with 20 ppmw, 40 ppmw, 60 ppmw Sn doping, respectively. The MCL was improved evidently due to the reduction of formation of interstitial Fe, FeB. The doping of Sn promotes formation of vacancies, which also contributes to passivation of interstitial Fe and FeB. A calculation of constitutional supercooling is carried out, which shows that Fe have great influence on the solidification interface stability, and Sn have little influence on the interface stability.

The nucleation process of high-performance multicrystalline silicon (HP mc-Si) growth seeded by fused quartz particles (FQP) through directional solidification is crucial for the ingot quality. To determine the optimal density of FQP and obtain a better nucleation process and the grain growth, we cast ingots using four different densities of FQP fixed on the bottom of the four quartz crucibles and covered them with a certain thickness of Si 3 N 4 coating. FQP sizes of 30–50 mesh were used, and the influence of the fused quartz particle density on the nucleation mechanism, initial grain uniformity, grain size, density of dislocation clusters, and cell efficiency were analyzed. Compared with the ingots seeded with other three densities of FQP, the 220 particles/cm 2 of FQP seeded ingot showed better uniformity of nucleation and initial grains. A large number of small uniform Si grains with lower density of dislocation clusters in the bottom of the ingot were observed. The average conversion efficiency of p-type solar cells manufactured with the 220 particles/cm 2 seeded ingot (18.28%) was 0.19% higher than that manufactured with the 120 particles/cm 2 seeded ingot (18.09%).

In this work the dislocation formation in the seed crystal induced by feedstock indentation during the growth of quasimono (QM) silicon ingots for photovoltaic application was investigated. It could be shown by special laboratory indentation experiments that the formed dislocations propagate up to several millimeters deep into the volume of the seed crystal in dependence on the applied pressure of the feedstock particles on the surface of the seed crystal. Further, it was demonstrated that these dislocations if they were not back-melted during the seeding process grow further into the silicon ingot and drastically reduce its material quality. An estimation of the apparent pressure values in a G5 industrial crucible/feedstock setup reveals that the indentation phenomenon is a critical issue for the industrial production of QM silicon ingots. Therefore, some approaches to avoid/reduce the indentation events were tested with the result, that the most promising solution should be the usage of suitable feedstock particles as coverage of the seed.

A recovery process for effective separation of silicon, silicon carbide micro powders and polyethylene glycol from the wire sawing slurry is proposed. The separation between silicon and silicon carbide is based on their size difference and surface charging state. The aim of this work is the study of the solid phase and liquid phase separation of silicon carbide and Silicon. Some methods applied for this purpose are the centrifugation process, phase-transfer separation as well as liquid–liquid extraction followed by the regeneration of polyethylene glycol by the distillation process. It is verified experimentally that silicon and silicon carbide micro powder can be effectively separated by phase transfer separation, centrifugation, chemical cleaning, filtering and distillation. In this study, the liquid–liquid extraction was used to separate a particle from a powder mixture. The removal of liquid component from a solution via a solvent separated considerably larger silicon carbide particles. The optimal results showed that Si content can reach 82% in the Si-rich powder and 3.8% wt% in the silicon carbide-rich powder. By separating the mixed powder the content reached 31.2 wt% of silicon, 63.3 wt% of silicon carbide as the raw material, 5.3 wt% of the iron fragment, and 0.2 wt% of other impurities.

A new method for evaluation of homogeneity of Ca 3 Ta(Ga 0.5 Al 0.5 ) 3 Si 2 O 14 (CTGAS) single crystals was established based on leaky surface acoustic wave (LSAW) velocity measurements performed by the line-focus-beam ultrasonic material characterization (LFB-UMC) system. Three plate specimens cut perpendicular to X -, Y -, and Z -axes were prepared from the CTGAS crystal ingot and LSAW velocity distributions were examined for these specimens. LSAW velocity changes due to Al-substitution effect were successfully extracted by using a relationship between two LSAW velocities propagating along different directions for Ca 3 TaGa 3 Si 2 O 14 (CTGS) and Al-substituted CTGS. Comparison of measured LSAW velocities and the results of chemical composition analysis performed by electron probe microanalysis (EPMA) demonstrated that LSAW velocity is mainly affected by Al-content change in CTGAS. Maximum velocity variation was observed in radial direction of the crystal ingot through the Z -axis propagating LSAW velocity measurements for Y -cut CTGAS specimen corresponding to Al-content change of 0.226 mol%. Accuracy of evaluation of Al content by velocity measurement for Y -cut Z -propagating LSAW is estimated to be ±0.0047 mol% and is superior to that by EPMA.

Numerical simulations were applied to analyze the effects of the power ratios between top and side heaters on the crystal–melt (c‐m) interface and the thermal stress distribution during the solidification process. The simulation results showed that among the different increase velocities of the power ratio, increasing the power ratio uniformly provided a most favorable solidification process: a slightly convex c–m interface shape and low thermal stress were obtained. The optimized design was implemented in casting experiments, which showed that the high-performance multi-crystalline silicon ingot had a vertical columnar structure and a lower dislocation density. The average conversion efficiency of solar cells was about 0.08% higher with this design (18.24%) than with the original design (18.16%).

In this paper, transient heat and mass transfer numerical simulations have been done to analyse the variation in thermal stress induced dislocations in 7 kg, 40 kg and 330 kg mc-silicon ingots grown using Directional Solidification (DS) furnace. The thermal parameters such as temperature distribution, principal stress, shear stress and von Mises stress distribution are examined. The effect of ingot size on temperature parameters, solid/melt interface, stress distribution and dislocation generation are investigated to obtain essential information needed for system design. It is observed that with respect to increase in size of the ingot the convexity of solid/melt interface decreases and the uniformity in radial and axial distribution of temperature and thermal stress also increase thereby decreasing the dislocation generation in the grown mc-silicon ingot. It shows that furnace design and ingot size play a crucial role in generation of thermal stress induced dislocation and defects in the grown ingot.

The transitional structures of 7075 aluminum alloy ingot after switching off a low-frequency electromagnetic field (LFEF) in the horizontal direct chill casting was investigated. The results showed that the equiaxed grains with the LFEF are firstly transformed to columnar grains and then to feathery grains without the LFEF. Also, there exists a transitional region consisting of relatively fine equiaxed grains between the equiaxed grains and the columnar grains. Crystallographic analysis indicated that only the relatively fine equiaxed grains with ideal orientations can grow into columnar grains. Moreover, there exist multiple twin relationships between some adjacent feathery grains, which have a near-common or common 〈1 1 0〉 direction. The equiaxed grains with the LFEF result from the forced convection. The relatively fine equiaxed grains and the columnar grains are attributed to the weakening of the forced convection when the LFEF is switched off. For the feathery grains, it is of interest to find that the quasicrystal-enhanced nucleation mechanism is responsible for their formation under high thermal gradient although it has been reported for twinned equiaxed grains under nearly nil thermal gradient in the literature.

The noncontact crucible (NOC) method has a large and deep low-temperature region in the upper central part of a Si melt to grow a uniform large Si single ingot without contact with the crucible wall. The temperature distribution in the NOC ingot is quite different to that in an ingot grown by the Czochralski (CZ) method because the NOC ingot is grown inside the Si melt. However, the distributions of point defects such as vacancy and interstitial Si atom are not known for the NOC ingot. To clarify the distributions of point defects, a simulation model is proposed to calculate the distribution of point defects in the NOC ingot. The effect of accumulation of diffusion fluxes on the concentrations of both vacancy and interstitial Si atom is considered for NOC growth because these diffusion fluxes are expected to continuously affect such concentrations in the ingot grown with the moving interface under the relatively low temperature gradient inside the melt. This point is one of the large merits of NOC growth. The concentrations of vacancy and interstitial Si atom were also calculated after the dynamic equilibrium pair-annihilation process between the diffused point defects during growth. They were calculated as a function of distance from the growing interface. By selecting the growth conditions, the concentrations of vacancy and interstitial Si atom were found to be very close to each other all over the ingot. The cross point where the concentrations of vacancy and interstitial Si atom were the same was found to strongly depend on G, the temperature gradient in the crystal.