Industrial Doctorate: Artificial Intelligence Recognition of Atomic Force Microscopy Molecular Images

Quasar Science Resources announced on the 31st of October that it was looking for candidates for an opportunity to request and Industrial Doctorate (Convocatoria 2017 para la concesión de ayudas para la realización de Doctorados Industriales en la Comunidad de Madrid) to do a PhD Thesis in the field of nanomechanics and the theory of microscopic forces (Atomic Force Microscopy, AFM). The proposal was submitted in collaboration with the Department of Theoretical Condensed Matter Physics at the Universidad Autónoma de Madrid. We are happy to announce that our 3 year project has been accepted and we plan to start working on this project in February 2018.

Atomic Force Microscopy (AFM), one of the key tools in Nanotechnology, uses the force between an atomically sharp tip and the sample under study to image and manipulate matter at the nanoscale level. The aim of the proposed project is to provide a reliable, robust and efficient method to identify the structure and chemical composition of individual molecules from information (2D images and 3D force maps) that can be gathered with the AFM using functionalized tips. Our approach to accomplish this ambitious goal combines two key elements: (i) the deep understanding of the mechanisms that control the contrast in AFM images taken with functionalized tips, and (ii) the use of advanced computational techniques to store, classify and process experimental data using machine learning tools in order to enhance the predictive power of the identification method.

To achieve this, we have identified a series of tasks:

  • We will put together an extended data set of AFM images of molecules
  • We will develop and apply computational techniques to analyse the AFM images
  • We will develop and apply machine learning and A.I. algorithms to recognize and correlate the features of the AFM images with the molecule structure
  • We will implement the automatic recognition in the AFM images of known molecules on the substrate in different adsorption configurations
  • We will implement the automatic discrimination of different molecules absorbed on a substrate in the AFM images
  • We will implement the molecular identification of molecules in the AFM images.
  • We will extend the work to nonplanar molecules
  • And last, we will develop a user-friendly platform/package to count, discriminate and, when possible, identify molecules in AFM images measured with functionalized tips



Leave a comment

Your email address will not be published. Required fields are marked *