Big-data reflection high energy electron diffraction analysis for understanding epitaxial film growth processes

Area of Science:

• Materials Science
• Surface Science
• Data Science

Background:

• Reflection High Energy Electron Diffraction (RHEED) is standard for in situ monitoring of film growth.
• Current RHEED applications primarily analyze specular spot intensity oscillations, discarding significant data.
• Advancements in data acquisition and computation enable sophisticated analysis of RHEED image sequences.

Purpose of the Study:

• To develop and demonstrate a multivariate statistical analysis approach for RHEED image sequences.
• To extract comprehensive information about fundamental film growth processes.
• To improve control and quality of epitaxial films through advanced data mining.

Main Methods:

• Multivariate statistical analysis (Principal Component Analysis, k-means clustering) of RHEED image sequences.
• Fourier analysis to isolate signal components and identify growth characteristics.
• Application to La(x)Ca(1-x)MnO(3) film growth on SrTiO(3) substrates.

Main Results:

• Identification of statistically significant patterns in growth behavior.
• Characterization of transitions between disordered and ordered growth.
• Separation of growth components, revealing deviations from ideal layer-by-layer growth.

Conclusions:

• Big data approaches significantly enhance insights into epitaxial film growth from RHEED data.
• Multivariate statistical analysis provides a universal method for understanding complex growth dynamics.
• Forward prediction methods based on this analysis can improve control over film quality.