X-ray powder diffraction is the workhorse technique in the characterization of crystalline materials. The product of this characterization technique, known as diffraction pattern, consist of a one-dimensional compressed image of the three-dimensional reciprocal space. As consequence, there is an inherent information loss due to the compression of the reciprocal space, which can make cumbersome tasks such as the structure determination of new crystalline materials. Convolutional neural networks can help us to circumvent the limitations in powder diffraction. In this tutorial, an introduction to the extension in the characterization capabilities of powder diffraction is provided, which is enabled using deep neural networks.

Target Audience

Physicists, chemists, and material scientists that are interested in the development of models to study crystalline materials.

Pre-requisites

Some knowledge in solid state physics/chemistry, as well as in deep neural networks.

Technical requirements

Internet connection, a Google account to work in colaboratory.