Topology-Based Approach for Rapid Identification of Unfeasible Candidates among Hypothetical Zeolite Structures

Zeolites are important functional materials in the chemical, energy, and environmental sectors, yet most computationally generated hypothetical frameworks remain synthetically inaccessible. Conventional screening methods rely on geometric optimization, which is computationally expensive and limits scalability. Here, we introduce two topology-based approaches, coordination sequence deviation (CSD) criteria and the number of ring types (NRT) criteria, for rapid identification of unfeasible candidates prior to geometric optimization. By using these topological descriptors, we bypass time-consuming geometric optimizations, enabling rapid elimination of unfeasible candidates at the preliminary screening stage. Application of these approaches to 25 million hypothetical zeolite structures successfully eliminated 82% of unfeasible structures, substantially enhancing the efficiency of subsequent optimization and validation. These topology-based approaches establish a reliable, high-throughput protocol for zeolite design, providing a streamlined pathway toward accelerated discovery of function-oriented crystalline materials.