Variational Quantum Eigensolver Approach to Electronic Structure Simulation of Molybdenum Disulfide Quantum Dots: Towards Quantum-Accelerated Nanomaterial Design
Cyprian Moyo, Adeniyi Sunday Ogunlaja, Zenixole R. Tshentu
This poster was presented at NanoAfrica 2026 (Poster #31, 21 October 2026, Nelson Mandela University, Gqeberha, South Africa). It reports a pilot project of the NMU Quantum Computing Society. The project shows that a full hybrid quantum–classical chemistry workflow can be carried out at Nelson Mandela University, from classical electronic-structure calculations through to execution on real IBM quantum hardware. The poster is a workflow demonstration and a learning exercise. It makes no claim of methodological novelty or of quantum advantage. Workflow demonstrated Model systems: small stoichiometric (MoS₂)ₙ clusters (n = 1, 3) used as model fragments of MoS₂ quantum dots. Geometries were optimised at the PBE/def2-SVP level in PySCF. Classical references: PBE, PBE0 and HSE06 HOMO–LUMO gaps, Hartree–Fock, CASSCF and CASCI. CASCI in a (4e,4o) active space serves as the exact reference. Qubit Hamiltonian: Jordan–Wigner mapping in CASCI natural orbitals with PennyLane, giving 8 qubits, 193 Pauli terms and 44 measurement groups. Ansätze: UCCSD and GateFabric as benchmarks, plus ADAPT-VQE to obtain compact, hardware-friendly circuits. Noise-aware simulation: shot-based simulation with IBM Heron noise models, using readout correction (TREX) and zero-noise extrapolation (ZNE). Hardware execution: energy evaluation on the 156-qubit IBM Heron r2 processor ibm_kingston through the IBM Quantum Open Plan. Main observations For MoS₂, a compact ADAPT-VQE circuit with 6 two-qubit gates came within 0.70 mHa of the CASCI reference. UCCSD needed about 540 two-qubit gates. In the noise-model simulation, error mitigation reduced the error from 23.6 mHa to 0.6 ± 0.7 mHa. On real hardware (ibm_kingston), TREX + ZNE reduced the error from 123.1 ± 7.3 mHa to 25.8 ± 14.4 mHa. This shows that mitigation works on current devices, but the result is not yet at chemical accuracy (1.6 mHa). Scope and limitations: the clusters are small models, not full quantum dots. The errors quoted are total energies relative to CASCI in the same active space. Every circuit used is classically simulable. The work is intended as a reproducible starting point for quantum-computing-in-chemistry research and training at NMU. Acknowledgements: We acknowledge the Centre for High Performance Computing (CHPC), South Africa, for its support of computational research (Project: CHEM0864). We acknowledge the use of IBM Quantum services for this work. The views expressed are those of the authors and do not reflect the official policy or position of IBM or the IBM Quantum team. We also thank Nelson Mandela University, NITheCS, the NMU Quantum Computing Society and the African Quantum Consortium. Code and data: the analysis scripts, result files and figures are available in the accompanying MoS2_VQE_pipeline repository.
•Nov 11, 2026•0 citations