Deep quantum Monte Carlo approach for polaritonic chemistry

Area of Science:

• Quantum Chemistry
• Computational Physics
• Materials Science

Background:

• Controlling matter properties via optical cavities is a growing research area.
• Hybrid light-matter states (polaritons) emerge from enhanced light-matter coupling.
• Accurate ab initio calculations for quantum electromagnetic and matter fields are computationally demanding.

Purpose of the Study:

• To introduce a novel deep learning variational quantum Monte Carlo method.
• To solve coupled electronic and photonic Schrödinger equations for molecules in optical cavities.
• To extend neural network wave function ansätze for joint electron-photon systems.

Main Methods:

• Deep learning variational quantum Monte Carlo approach.
• Extension of neural network wave functions for fermionic and bosonic systems.
• Application to hydrogen molecules in an optical cavity.

Main Results:

• Calculated ground and excited states of hydrogen molecules in a cavity.
• Assessed energy, dipole moment, and charge density shifts.
• Analyzed photonic field states and electron-photon entanglement.
• Achieved good qualitative agreement with conventional quantum chemistry methods.

Conclusions:

• The developed method accurately simulates molecules in optical cavities.
• This approach extends the applicability of machine learning in quantum chemistry.
• Enables tackling complex quantum light-matter interaction problems.