Data-Driven Serendipitous Discovery of a Self-Organized Nanostructure for Enhanced Pinning in Superconducting Films.

In nanocomposite-coated conductors of superconducting YBa2Cu3O7, an improvement of the critical current density is required for high-field applications. After two decades of development, more effective nanostructures are still required to further enhance the vortex pinning. While researcher-dependent empirical approaches are less effective for complex phenomena, serendipity can lead to discoveries beyond established theories, existing knowledge, and trends. Serendipity has been considered an uncontrollable, lucky, or accidental event, and researchers often overlook such events. We propose the concept of serendipity engineering. Here, we regard fluctuation-induced anomalies in experiments as a potential driver for serendipity and detect such fluctuation-induced anomalies using an anomaly score defined by log likelihood. We implemented data-driven serendipity detection for the discovery of structures and physical mechanisms. By applying this concept to nanoscale self-organization of the YBa2Cu3O7 nanocomposite multilayer, we improved the critical current density. Instead of forming the designed multilayer structure, spontaneous structural reconstruction occurred across several layer units, resulting in weak or random correlations between nanorods. Spacer layers between nanorods, which enable flexible vortex configurations, play a key role in enhancing the vortex pinning in magnetic fields. Weakly correlated nanorods represent an unexpected optimal pinning structure, leading to pseudo-high-density vortex pinning and a 2.9-fold enhancement in the critical current density. This study demonstrates the potential of serendipity engineering for discovery at the nanoscale self-organization.