Redox-Controlled Modulation of Correlation Energy in Transition-Metal Oxide Clusters: A Statistical VQE Study

Redox-Controlled Modulation of Correlation Energy in Transition-Metal Oxide Clusters: A Statistical VQE Study Why This Matters Modern quantum chemistry often feels abstract: orbitals, correlation, Hilbert spaces. But at its core, it asks a simple question: How sensitive is matter to the addition or removal of a single electron? In this work, we explored that question using quantum algorithms on transition-metal oxide clusters. By holding everything constant except electron count, we were able to observe how redox state alone reshapes the accessible quantum correlations of the system. What we found was striking: electron removal consistently amplifies correlation energy, and in one metal (Ru), produces two entirely distinct quantum basins under identical conditions. This sensitivity could inform the design of redox-tunable catalysts, battery materials, or quantum simulation benchmarks. Sometimes, changing one electron changes everything. Understanding that sensitivity matters. Redox processes lie at the heart of catalysis, energy storage, corrosion, battery materials, and biological electron transfer. If electron count can systematically reorganize quantum correlation structure, then redox state is not just a bookkeeping variable—it becomes a tunable parameter. By learning how correlation responds to that control knob, we move closer to designing materials and catalytic systems where electronic behavior can be deliberately steered rather than passively observed. Description This dataset reports the first systematic study of redox-dependent correlation energy modulation across multiple transition metals using Variational Quantum Eigensolver (VQE) methods. Through 65 independent calculations on Cu₅MO₁₂ clusters (M = Ni, Ru, Pd), we demonstrate that electron count acts as a systematic control parameter governing the magnitude of VQE-accessible correlation energy within fixed computational frameworks. KEY SCIENTIFIC FINDING: Across all three metals studied (3d Ni, 4d Ru, 4d Pd), a consistent redox ordering emerges: Cation ≫ Neutral ≫ Anion Hole doping (cation formation) activates large correlation energy (-126 to -337 kcal/mol), while electron addition (anion formation) results in negligible VQE-accessible correlation (< 0.2 kcal/mol) within this computational framework. This pattern is reproducible across independent random initializations and holds across chemically distinct 3d and 4d transition metals. Because all non-redox parameters are held constant (geometry, basis set, ansatz, optimizer), the observed differences can be attributed directly to electron count variation within the defined VQE-UCCSD framework. STATISTICAL VALIDATION RESULTS Main campaign: 45 independent VQE calculations (3 metals × 3 charge states × 5 statistical seeds) Correlation Energy (Mean ± Std, kcal/mol, n=5) System Anion (-1) Neutral (0) Cation (+1) Cu₅NiO₁₂ -0.17 ± 0.06 -12.38 ± 0.10 -137.66 ± 8.03 Cu₅RuO₁₂ -0.11 ± 0.20 -6.77 ± 1.63 -299.67 ± 76.08 Cu₅PdO₁₂ +0.06 ± 0.16 -56.33 ± 0.99 -126.42 ± 7.66 Key observations: Anions exhibit negligible correlation energy relative to cations within this framework 4d Ru shows 2-3× larger cation activation than 3d Ni Standard deviations < 10 kcal/mol for Ni and Pd demonstrate excellent reproducibility Ru cation high variance (±76 kcal/mol) motivated extended investigation SECONDARY FINDING: Ru Cation Bimodal Optimization Landscape Extended validation: 20 additional Ru(+1) restarts to characterize high variance Result: Two distinct energy basins discovered Deep Basin (12/20 runs = 60%): -336.45 ± 1.94 kcal/mol Shallow Basin (8/20 runs = 40%): -141.82 ± 2.91 kcal/mol Basin Separation: ΔE ≈ 195 kcal/mol (100× larger than intra-basin variance) Independent restarts under identical computational settings access the deeper basin in ~60% of runs. This demonstrates that Ru under hole doping exhibits a rugged correlation landscape with well-separated optimization basins. Metal specificity: Ni and Pd do not exhibit comparable bimodality across equivalent sampling (n=5). TECHNICAL ACHIEVEMENTS Zero checkpoint contamination: Verified across all 65 runs (0 RECOVERY events) Novel checkpoint isolation protocol prevents cross-run contamination Critical for statistical validity in multi-system VQE campaigns Statistical validation framework: Multi-seed approach (n=5-20) establishes reproducibility Standard deviations demonstrate tight convergence Bimodal distribution resolved through targeted extended sampling Controlled computational experiment: Identical conditions across all systems Single variable: electron count (redox state) Enables direct attribution to redox effects Large-scale correlation recovery: Up to -337 kcal/mol for Ru cation 20-qubit active space on consumer GPU (NVIDIA L40S 48GB) METHODOLOGICAL CONTRIBUTIONS Redox State as Correlation Control Parameter Holding geometry, basis set, ansatz, and optimizer constant, redox variation alone produces order-of-magnitude changes in recovered correlation energy (from near-zero in anions to >300 kcal/mol in Ru cations). This establishes redox state as a tunable computational control parameter within the Cu₅MO₁₂ cluster model. Checkpoint Isolation Protocol Novel wrapper methodology ensures complete checkpoint cleanup between runs, eliminating cross-contamination that plagued earlier multi-system campaigns. Validated by zero RECOVERY events across 65 independent calculations. Basin Characterization Through Extended Sampling Systematic restart protocol (n=20) resolves bimodal distributions that would be obscured by standard 5-seed sampling. Reveals metal-specific optimization landscape structure not accessible through single-run or limited-seed approaches. COMPUTATIONAL PROTOCOL Quantum Method: HF-anchored Variational Quantum Eigensolver (VQE) Ansatz: UCCSD (Unitary Coupled-Cluster Singles and Doubles), depth 6 Parameters: 534 variational parameters Active Space: 12 electrons in 10 orbitals (20 qubits) Optimizer: Adam (learning rate 0.02) Convergence: < 1 kcal/mol threshold Basis Sets: LANL2DZ (ECP) for transition metals (Ni, Ru, Pd, Cu) 6-31G for oxygen Hardware: NVIDIA L40S 48GB GPU VRAM utilization: ~37 GB for 20-qubit systems Runtime: ~5-15 minutes per system Systems: Cu₅MO₁₂ hexagonal lattice clusters M = Ni (3d⁸), Ru (4d⁷), Pd (4d⁹) Charge states: neutral (0), cation (+1), anion (-1) SCIENTIFIC CONTEXT These results demonstrate that within a fixed VQE framework, electron removal (cation formation) consistently increases accessible correlation energy, with metal-dependent magnitude ranging from -126 to -337 kcal/mol. Electron addition (anion formation) results in negligible correlation energy across all metals studied. The discovery of bimodal optimization landscapes in Ru cations reveals that metal-specific electronic structure interacts with redox perturbation in non-trivial ways, producing qualitatively different optimization behavior between 3d and 4d metals. This study does NOT claim: Physical existence of metastable electronic states Identification of true ground-state wavefunctions Universal catalytic design principles Mechanisms of biological systems or superconductivity Generalization beyond Cu₅MO₁₂ clusters Appropriate interpretation: Results strictly describe VQE-recovered correlation energy behavior under controlled redox variation within the UCCSD ansatz framework for Cu₅MO₁₂ cluster geometries. This is a computational control experiment demonstrating systematic redox-dependent modulation within a defined quantum chemical framework. DATASET CONTENTS Main Campaign Logs (45 files) logs_seed/Cu5NiO12__seed0.log through Cu5NiO12__seed4.log (neutral Ni, 5 seeds) logs_seed/Cu5NiO12_cation__seed0.log through seed4.log (Ni cation, 5 seeds) logs_seed/Cu5NiO12_anion__seed0.log through seed4.log (Ni anion, 5 seeds) (Same structure for Cu5RuO12 and Cu5PdO12) Ru Cation Extended Validation (20 files) logs_seed_restarts/Cu5RuO12_cation__restart1.log through restart20.log Statistical Summaries data/summary_statistics.csv — Main results table with mean ± std for all systems data/ru_cation_bimodal_analysis.csv — Deep/shallow basin characterization data/individual_systems/ — Per-system detailed breakdowns (9 files) Metadata & Documentation README.txt — Comprehensive technical documentation metadata/computational_parameters.json — Full VQE configuration Analysis Scripts Parsing and statistical analysis code (Python) Reproducibility instructions REPRODUCIBILITY What is provided: Complete convergence logs with iteration-by-iteration energies Final VQE energies and HF reference values Molecular geometries (atom coordinates) Computational parameters (basis sets, active spaces, spin multiplicities) Hamiltonian specifications (qubit counts, Pauli term counts) Optimizer settings (learning rate, convergence criteria) Statistical analysis methodology What is NOT provided: Source code is intentionally withheld as proprietary Quantum-Clarity technology. How to reproduce: Results can be independently reproduced using standard open-source tools: PySCF: Quantum chemistry integrals, SCF, active space construction OpenFermion: Jordan-Wigner transformation PyTorch + TorchQuantum: VQE implementation Key parameters (extracted from logs): Basis: LANL2DZ (ECP) for metals, 6-31G for oxygen Active space: 12 electrons, 10 orbitals Ansatz: UCCSD-inspired, depth 6, HF initialization Optimizer: Adam, lr=0.02, threshold 1×10⁻⁶ Ha Hardware: NVIDIA GPU with ≥48 GB VRAM recommended Statistical distributions will reproduce within standard error across independent implementations given identical computational parameters. DATASET USAGE This is a reference dataset providing validated statistical benchmarks for redox-controlled correlation energy in transition metal oxide clusters. Intended uses: