X-ray reflectivity data analysis using Bayesian inference: The study of induced Pt magnetization in Pt/Co/Pt

Abstract

The induced Pt magnetization in a Pt/Co/Pt thin film structure is studied. The normally nonmagnetic Pt acquires a magnetic moment due to the magnetic proximity effect at the Co–Pt interfaces. Element specific Pt structural and magnetic properties are characterized by synchrotron-based resonant x-ray reflectivity and x-ray resonant magnetic reflectivity measurements. An advanced analysis method based on Bayesian inference is used for model fitting of the x-ray data. Using this method, we retrieve the best fit values of material parameters (e.g., thickness, interfacial roughness) from the data. Analysis of x-ray reflectivity data of this specific system shows that the Pt magnetization and Co–Pt interfacial roughness is significantly different between the top and bottom Pt layers, with both values being larger in the top Pt. The successful application of this Bayesian method to study the magnetic and structural properties of a thin film system demonstrates its effectiveness for x-ray reflectivity data analysis.