A machine condition transfer function approach to run-to-run and machine-to-machine reproducibility of III-V compound semiconductor molecular beam epitaxial growth

A concept of machine condition transfer function (MCTF) based on the use of in-situ sensor response to intrinsic surface property of the material is proposed to meet the challenge of run-to-run and machine-to-machine reproducibility of optimized growth conditions for III-V compound semiconductor molecular beam epitaxical growth. The variation in the intensity of the specular beam in reflection high-energy electron diffraction from compound semiconductor surface, measured as a function of the temperature and anion pressure prior to growth, is used as the intrinsic surface property for the generation of MCTF´s. The mathematical methodology for realizing MCTF´s, including the sensor response surface reconstruction and optimal minimization in combination with a composite statistical design for initial estimation, is presented. Numerical evaluation of the developed methodology shows its promise in practice