The best way to conference proceedings by Francis Academic Press

Web of Proceedings - Francis Academic Press
Web of Proceedings - Francis Academic Press

Research on Temperature Control of Single Crystal Furnace Based on Neural Network

Download as PDF

DOI: 10.25236/AISCT.2019.062


Xudong Zheng, Yilang Tu, Xudongzheng, Xutao Mo

Corresponding Author

Xutao Mo


Silicon material has excellent semiconductor characteristics. It is the raw material of electronic devices, photovoltaic power generation and other industries. The growth of silicon mainly includes Czochralski method and zone melting method, among which monocrystalline silicon grown by Czochralski method is the most popular in the market. The growth environment of monocrystalline silicon is very complex, which requires very high temperature and precise control. The current development of monocrystalline silicon towards large diameter and high quality requires higher temperature control in the furnace. However, the single crystal furnace used for the growth of monocrystalline silicon has the characteristics of large lag, and cannot adapt to the growth of large diameter monocrystalline silicon. Therefore, in order to ensure the sTable growth of monocrystalline silicon, the requirements for temperature control in the furnace are higher.


Single crystal silicon Neural network: PID: Temperature control; Thermal field