Public defence in Engineering Physics, M.Sc. (Tech) Yashasvi Ranawat

Title of the doctoral thesis: Simulating molecular adsorption on dielectric surfaces with DFT and machine learning

Opponent: Professor Ruben Perez, Universidad Autónoma de Madrid / Condensed Matter Physics Center (IFIMAC), Spain
Custos: Professor Adam Foster, Aalto University School of Science, Department of Applied Physics

The public defence will be organised on campus.

The doctoral thesis is publicly displayed 10 days before the defence in the publication archive Aaltodoc of Aalto University

Electronic thesis

Public defence announcement: 

Several technological and natural process are governed by the surface interactions at the nano-scale. AFM imaging has heralded the characterisation of such interactions through the possibility of imaging at atomic and molecular resolutions. It has become an indispensable tool in understanding of protein folding – and other molecular biological -- processes, mineral dissolution and growth in geology, and various surface phenomenon, like corrosion. 

However, at such high resolution the interpretation of the image and control of the environment variable is challenging, especially at the solid-liquid interfaces. Here, the thesis aims to complement the AFM images with simulation techniques to better characterise the surface phenomena. The thesis had two-fold objective: first, to use simulation techniques to understand the surface reconstruction of calcite surface and the consequent effect of the reconstruction on surface chemistry, and secondly, to establish deep learning workflows to aid the solid-liquid interface characterisation.

For the first objective, the calcite surface is chosen for its crucial calcite-water interface, that plays vital role in bio-mineralisation, geochemical and environmental systems. Further, it is a benchmark surface in AFM studies. For the second objective, state-of-the-art simulation techniques exist which can simulate liquid interactions over a given surface. However, given an AFM image of liquid layers over a surface, the candidate surface structures can be in-numerous. At this rate, the simulation techniques become computationally intensive and prohibit a direct characterisation. This highlights an express need of deep learning tools to obviate the intensive simulation workflow. 

Through the thesis work, we successfully characterised the calcite surface reconstruction and paved the path to understand its role in the surface chemistry of this vital mineral surface. Additionally, the deep learning workflows introduced in the thesis establishes the role such tools can play in expediting the characterisation of solid-liquid interfaces, which is crucial for AFM’s success at higher resolution imaging.