Toward chemically accurate quantum simulations on quantum hardware: A quantum-HPC hybrid approach

We report one of the most comprehensive experimental demonstrations to date of quantum-computing-based molecular simulation on IQM's Sirius 24-qubit superconducting quantum processor, using up to 16 operational qubits. We employ Sample-based Quantum Diagonalization (SQD) with the Local Unitary Cluster Jastrow (LUCJ) Ansatz to compute ground-state energies of benchmark molecular systems, including H2, LiH, BeH2, H2O, and NH3. We further introduce the linear-CNOT unitary coupled-cluster singles and doubles Ansatz within the SQD framework, reducing the classical pre-computation overhead at the cost of increased quantum circuit depth and demonstrate its effectiveness by accurately recovering the ground-state energies of H2, LiH, and BeH2. A comparative analysis of the two Ansätze is presented, highlighting their respective merits, demerits, and practical use cases in near-term quantum simulations. In addition, we perform one-dimensional potential energy surface (1D-PES) scans for H2 and HeH+ using STO-3G and 6-31G basis sets and for LiH and BeH2 using STO-3G. We also report the first experimental generation of a full two-dimensional potential energy surface (2D-PES) map, consisting of a 32 × 32 grid in bond length and bond angle, for the water molecule in the STO-3G basis on IQM quantum hardware. Beyond small-molecule benchmarks, we integrate Density Matrix Embedding Theory (DMET) with SQD(LUCJ) to obtain chemically accurate 4-electron-in-4-orbital active-space energies for eight ligand-like molecules, as well as for the pharmacologically relevant amantadine molecular system, representing the first of such experimental demonstrations. Across all studies, the majority of quantum-computed energies agree with reference full configuration interaction results and with DMET-complete active space CI energies for embedded systems, to within chemical accuracy for the chosen basis sets. These results demonstrate the robustness of quantum sampling-based classical diagonalization approaches while highlighting the promise of hybrid embedding strategies combined with near-term Ansätze for advancing quantum chemistry toward classically intractable molecular systems and establishing their practicality on current quantum hardware.