关键词: 温度补偿;拉速稳定性;模糊控制;遗传算法;提拉法

Abstract: For the purpose of reducing the pull rate fluctuation during the Czochralski crystal growth process of single silicon, the approach of controlling the temperature compensation rate was presented based on analyzing the energy and mass balance at the solid-liquid interface. A new fuzzy control method was proposed of which the membership function parameters were optimized using the genetic algorithm to tune the power input, thus the inner circumstance of the furnace was more suitable for the crystal growth. Experiment data verifies that with this approach, the precision of the crystal diameter was improved and the pull rate fluctuation was considerable reduced also.

Key words: For the purpose of reducing the pull rate fluctuation during the Czochralski crystal growth process of single silicon, the approach of controlling the temperature compensation rate was presented based on analyzing the energy and mass balance at the solid-liquid interface. A new fuzzy control method was proposed of which the membership function parameters were optimized using the genetic algorithm to tune the power input, thus the inner circumstance of the furnace was more suitable for the crystal growth. Experiment data verifies that with this approach, the precision of the crystal diameter was improved and the pull rate fluctuation was considerable reduced also.