Solute transport in Czochralski (CZ) crystal growth and in magnetic Czochralski (MCZ) with a vertical applied field are simulated numerically, and compared with published experimental data on gallium segregation in silicon growth. In the MCZ cases, the low segregation coefficient of 0.008 leads to a steep concentration gradient under the s/1 interface due to a boundary layer buildup of rejected gallium. This required a high resolution computation and a segregation model accounting for the melt drift toward the interface. Also, the relatively high density of gallium compared with silicon melt made it necessary to include solutal convection. With these improvements to the simulation code, good agreement with the data was achieved in the CZ growth and in the MCZ growth at 0.1 and 0.3 tesla.
Computer simulation leading to controlled large diameter Czochralski crystal growth is discussed. A simple mathematical model, which describes the different crystal growth phases including neck-in, fast flat top, r011-over to constant diameter bulk growth, and tail-off is presented. This model, in conjunction with a computer-implemented simulator, is used to simulate silicon crystal growth. Good agreement between siniulation results and experimental crystal growth is obtained.Large diameter silicon single crystals free of dislocations are routinely grown from the melt by using the Czochralski (CZ) technique. This paper presents a simple mathematical model which describes the behavior of the crystal radius during all phases of 125 m m CZ silicon crystal growth. This model is the basis for a computer simulation of the crystal growth process. The simulation includes the initial neck-in growth after seeding, which represents a standard procedure for elimination of dislocations; then fast flat growth, also called shoulder growth, roll-over to constant diameter bulk growth, and finally, the tailoff procedure to terminate crystal growth. The simulation results are compared to actual growth of 125 m m diam silicon crystals. This modeling effort is an integral part of a larger program directed toward complete interactive digital computer control of the CZ crystal growth process. Modeling of 125 mm Diameter Silicon Crystal GrowthFluctuations of crystal diameter during growth are caused by variations both in crystal pull rate and thermal conditions. In principle, it should be possible to .obtain insight into a cause and effect relationship of crystal diameter variations during growth through thermal analysis. To the best of our knowledge, such an analysis has not yet been performed. Thermal analyses of the CZ crystal growth process have been made by several investigators. However, the efforts reported are concerned with other aspects of crystal growth, such as temperature distributions in the crystal a n d / o r at the solid-liquid interface (1-7), and crystal pull rate limits (8,9). This paper provides insight into crystal diameter variations during crystal growth, through development of a mathematical model in conjunction with a computer simulation comprising all phases of the growth process. This insight is gained through a comparison between simulated and actual crystal growth data. The simulation i:s based on the model described in the following sections.Mathematical model.--The sequence of large diameter CZ silicon crystal growth is schematically illustrated in Fig. 1. The radius (R) of the growing crystal is a function of the crystal pull rate (V) and of the temperature (T) of the system. T is defined here as the difference between the crucible wall temperature and the melting point (I412~ of silicon. Experimentally, it is found that R is inversely proportional to V and T. The behavior of R prevails during neck-in, constant diameter bulk growth, a n d tailoff. During fast fiat top growth, R is assumed to increase exp...
Effect of a relatively high axial magnetic field false(Bfalse) on the segregation behavior of oxygen in Czochralski silicon crystal growth is investigated. B=4 normalkG was applied continually during crystal growth after initial growth of 50 mm length at B=0 . Well‐defined growth conditions were used without rotating either crystal or crucible. The oxygen concentration was measured by a high resolution FTIR technique using a small sampling area of 100×100 μnormalm in the longitudinal sections. At B=0 , significant fluctuations are observed in the oxygen concentration Cnormals , which is indicative of substantial convections in the melt. On applying 4 kG, the oxygen level decreased drastically, after a temporary rise in the middle of the initial transient, to 3.0 from an average value of 15.0 ppma at B=0 in 8.4 mm crystal length or about 8 min. This value increased subsequently to 5.0 ppma in 70 mm growth or about 70 min when the crystal was detached and the melt quenched. 5.9 ppma is measured in the quenched melt. The effective segregation coefficient knormaleff at 4 kG is 0.85. The closeness of the knormaleff to unity indicates a substantial suppression of the melt convection at 4 kG. The transient profile and the drastic decrease of Cnormals on applying 4 kG indicate that the interfacial distribution coefficient knormali is significantly larger than one.
A computer model for the fluid flow and dopant transfer/segregation in Czochralski crystal growth (CZ) and in Czochralski crystal growth in an axial magnetic field (MCZ) is applied to the simulation of the oxygen source, transport, and segregation in CZ and MCZ silicon growth. To model the oxygen source, which is ablation of the silica crucible, the oxygen concentration at the melt/crucible interface is assumed to be at an equilibrium concentration which is dependent on temperature as described by Arrhenius kinetics. The oxygen sink is mostly evaporation of silicon monoxide from the meltfree surface; a small percentage of the dissolved oxygen is incorporated into the growing crystal. Good agreement is established between the simulated oxygen radial profiles and experimental results in CZ/MCZ silicon growth, as well as in the comparison of the simulated ablation rate of the crucible with the experimental data reported to date.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.