A machine condition transfer function approach to run-to-run and machine-to-machine reproducibility of III-V compound semiconductor molecular beam epitaxical 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 p roposed to meet the challenge of run-to-run and machine-to-machine reproduc ibility of optimized growth conditions for III-V compound semiconductor mol ecular beam epitaxical growth. The variation in the intensity of the specul ar beam in reflection high-energy electron diffraction from compound semico nductor surface, measured as a function of the temperature and anion pressu re prior to growth, is used as the intrinsic surface property for the gener ation of MCTF's. The mathematical methodology for realizing MCTF's, includi ng the sensor response surface reconstruction and optimal minimization in c ombination with a composite statistical design for initial estimation, is p resented. Numerical evaluation of the developed methodology shows its promi se in practice.