Angelina Vypritskaia

The arrangement of atoms plays a crucial role in shaping the properties of materials, ranging from new material design to life-saving drug development. While X-ray crystallography has traditionally been the preferred method, growing suitable crystals can pose significant challenges. Fortunately, electron diffraction offers a powerful alternative that allows us to investigate the atomic arrangements in small crystals, particularly those that are too small or disordered for X-rays. However, the sensitivity of the sample to the electron beam presents a hurdle in obtaining complete information. As part of the NanED project funded by the EU, I am working on enhancing the crystal finding algorithm for serial electron diffraction (SerialED) by applying machine learning and image recognition algorithms. My goal is to train computer models to automatically detect and select crystals, thereby improving the analysis of delicate materials, accelerating structure solution, and discovering new substances.