Accurate Chemical Reaction Modeling on a Noisy Intermediate-Scale Quantum Computer with an Active Space-Based Workflow

Quantum computing offers a compelling route to ab initio simulations of chemical reactions, while noisy intermediate-scale quantum (NISQ) hardware demands resource-aware protocols. In this study, we introduce an efficient workflow using a noise-resilient hardware-adaptable ansatz (HAA) in an active space combined with driven similarity renormalization group (DSRG) downfolding. The active space is obtained via a new selection algorithm, many-body-expanded correlation-energy active space (MBECAS), based on orbital ranking by first-/second-order correlation-energy increments. We validate the MBECAS-DSRG-HAA workflow on reactions with up to tens of atoms, reproducing Diels-Alder barriers within a few millihartrees after error mitigation. By combining an efficient active space selector, effective-Hamiltonian downfolding, and a low-depth variational circuit, our approach delivers highly accurate reaction energetics with modest quantum resources and motivates extensions to larger systems and excited-state dynamics.