FAIR Data Management in Scanning Electron Microscopy

This work describes the deployment and configuration of a modular software infrastructure designed to support the FAIR management of scientific data, with particular focus on electron microscopy (SEM) experiments conducted at CNR NANOTEC Lecce. The project involved adapting the lame_fair_by_design application, initially developed for a different technological stack at Area Science Park in Trieste by Nicola Perin and Elaheh Saadat, to a new local environment based on vSphere virtualization, Docker containerization, and Isilon storage. A Linux virtual machine was provisioned as a development and integration platform, hosting services deployed through Docker and Docker Compose. MinIO was configured as the object storage backend, and secure API credentials were generated to ensure controlled access. User authentication was integrated using Keycloak and the OpenID Connect (OIDC) protocol, enabling centralized and secure Single Sign-On (SSO) across connected platforms. A critical part of the work focused on electron microscopy (SEM) data conversion. SEM instruments typically produce heterogeneous files (e.g., TIFF images, EDX spectra) accompanied by metadata embedded in non-standard formats. A specialized Python pipeline was developed to extract metadata from TIFF files, normalize it, and map it to the NeXus NXem standard schema. This involved: • Parsing image headers to retrieve experimental parameters (magnification, accelerating voltage, working distance, etc.); • Structuring these parameters into a formalized metadata schema; • Generating a NeXus-compliant HDF5 file, embedding both the experimental image and the structured metadata. By converting raw SEM data into the NeXus NXem format, the project achieves important FAIR goals: making the datasets findable (indexed metadata), accessible (standardized APIs), interoperable (standardized data format), and reusable (full traceability of experimental conditions). This modular, extensible infrastructure is positioned to support both current research needs and future integration into larger FAIR data ecosystems, such as NOMAD and FAIRmat.