Entanglement-Minimized Orbitals Enable Faster Quantum Simulation of Molecules

Area of Science:

• Quantum chemistry
• Computational physics

Background:

• Quantum computation, particularly quantum phase estimation (QPE), promises faster molecular simulations.
• Efficient initial state preparation is crucial for QPE speedup, but challenging for strongly correlated systems like iron-sulfur clusters.
• Overlap with the true ground state often decays exponentially with system size in these molecules.

Purpose of the Study:

• To develop an efficient classical algorithm for preparing accurate initial states for quantum simulations.
• To improve the overlap between the initial state and the true ground state for complex molecules.
• To reduce the computational resources required for quantum simulations of strongly correlated systems.

Main Methods:

• Introduced an efficient classical algorithm to identify entanglement-minimized orbitals (EMOs).
• Utilized spin-adapted low-bond-dimension matrix product states (MPS) for orbital optimization.
• Applied the EMO basis for more compact ground-state representation and easier initial state preparation.

Main Results:

• The EMO basis significantly enhances initial state overlap for challenging molecules.
• Achieved nearly an order of magnitude improvement for a four-iron cluster compared to previous methods.
• Demonstrated substantial overlap enhancements (O(10^2) to O(10^5)) for large systems like the P-cluster and FeMo-cofactor.

Conclusions:

• The entanglement-minimized orbital approach drastically eases initial state preparation in quantum computation.
• This method is scalable to large, complex molecules with multiple transition metal centers.
• Quantum simulations of challenging systems require fewer resources than previously estimated.