Neural network applications for X-ray photoelectron spectroscopy data analysis

Abstract

Surface analysis techniques are an integral part of high innovation sectors related to nanotechnologies, biotechnologies, nanoparticle characterization, lightweight materials development and energy storage. There are several techniques available to analyze surface properties and the behaviour of materials in their dedicated environment including XPS, AES, SIMS, LEED, EXAFS, STM and AFM. These instruments allow for the study of the physical topography, the chemical composition, the chemical structure, the atomic structure, the electronic state, and the bonding behaviour of surface molecules. In this doctoral work we will focus on X-ray photoelectron spectroscopy (XPS), a surface sensitive (<10nm) characterisation technique used to investigate material properties including chemical composition, chemical depth distribution and electronic structure. Particularly in recent years, XPS has become a reliable and advanced experimental technique across various disciplines of science and engineering resulting in the generation of large spectral datasets. Extracting quantitative information from these datasets has traditionally required trained spectroscopists to perform empirical peak-fitting routines for each individual spectrum. For example, to obtain the atomic fraction of a particular element in a sample, the expert must determine the integrated peak area from non-trivial fitting routines based on the inelastic scattering background and zero energy loss line shapes and normalize it according to relative sensitivity factors. In response to the increasing demand for reliable and instantaneous spectral analysis, we propose an automated quantitative X-ray photoelectron spectrum analysis pipeline by updating the in-house developed Simulation of Electron Spectra for Surface Analysis (SESSA) software with different deep learning models ranging from standard neural networks to more advanced active learning techniques. SESSA serves as an important tool in the field of XPS either as a database for material parameter retrieval or as a Monte Carlo-based simulation software for quantitative interpretation of XP spectra or Auger electron spectra (AES) for a variety of materials (bulk, nanostructures, layered spheres, etc.). Finally, we plan to apply the simulation environment in practice to the material characterisation of a variety of experimental spectra measured by different laboratories and research institutions to evaluate our method. In addition to our efforts in X-ray photoelectron spectroscopy an additional part of this work will also be the investigation of low energy electrons and multiple electron scattering processes in solids by using coincidence electron spectroscopy. Low energy electrons play a crucial role in many domains of science ranging from electron beam lithography, spacecraft surface charging to electron cloud formation in charged particle storage rings. Here the main focus is on the investigation of insulator materials (PMMA, SiO2) used in the semiconductor industry and the derivation of their optical properties as well as of 2D materials (graphene, graphite) to study the basic principles behind secondary electron emission processes.