Towards quantum-centric simulations of extended molecules: sample-based quantum diagonalization enhanced with density matrix embedding theory

Summary

This study presents the first density matrix embedding theory (DMET) calculations using the sample-based quantum diagonalization (SQD) method as a subsystem solver. The DMET+SQD combination is tested on a hydrogen ring and cyclohexane conformers, demonstrating improved accuracy and reduced non-parallelity error compared to unfragmented calculations.

Highlights

• First DMET calculations using SQD as a subsystem solver
• Tested on a hydrogen ring and cyclohexane conformers
• Improved accuracy and reduced non-parallelity error
• DMET+SQD combination allows for larger subsystems and higher accuracy
• Reduced impact of quantum noise on SQD calculations
• Cyclohexane conformer energies accurately predicted
• Fragmentation approximation affects energy differences below 1 kcal/mol

Key Insights

• The DMET+SQD combination enables the treatment of larger subsystems and higher accuracy than previously possible, making it a promising approach for quantum-centric simulations of extended molecules.
• The use of SQD as a subsystem solver allows for improved accuracy and reduced non-parallelity error compared to unfragmented calculations, making it a valuable tool for quantum chemistry applications.
• The fragmentation approximation underlying DMET affects energy differences below 1 kcal/mol, highlighting the importance of careful fragmentation in DMET calculations.
• The DMET+SQD combination is able to accurately predict the energies of cyclohexane conformers, demonstrating its potential for applications in organic chemistry.
• The reduced impact of quantum noise on SQD calculations makes it a promising approach for near-term quantum computing applications.
• The study highlights the importance of continued development in error rates of quantum computers, error mitigation techniques, and construction and optimization of quantum circuits for SQD applications.
• The DMET+SQD combination serves as a proof-of-concept for the concerted use of quantum and classical computers, demonstrating the potential for innovative and promising modes of attack on correlated many-electron systems.

Mindmap

Citation

Shajan, A., Kaliakin, D., Mitra, A., Moreno, J. R., Li, Z., Motta, M., Johnson, C., Saki, A. A., Das, S., Sitdikov, I., Mezzacapo, A., & Merz, K. M. (2024). Towards quantum-centric simulations of extended molecules: sample-based quantum diagonalization enhanced with density matrix embedding theory (Version 2). arXiv. https://doi.org/10.48550/ARXIV.2411.09861