Multideterminant Wave Functions in Quantum Monte Carlo.

Area of Science:

• Computational Chemistry
• Quantum Mechanics
• Materials Science

Background:

• Quantum Monte Carlo (QMC) methods are powerful tools for solving the many-body Schrodinger equation in electronic systems.
• Their low scaling with particle number makes them suitable for large molecular and solid-state calculations.
• A major limitation of QMC is the difficult-to-control fixed-node error.

Purpose of the Study:

• To present a systematic application of large-scale multideterminant expansions within QMC.
• To evaluate the performance of this approach for first-row dimers and the G1 test set.
• To demonstrate the potential for reducing fixed-node error and achieving chemical accuracy.

Main Methods:

• Systematic application of large-scale multideterminant expansions in Quantum Monte Carlo.
• Testing on first-row dimers and the 55 molecules of the G1 test set.
• Comparison with traditional quantum chemistry methods (MP2, CCSD(T), DFT).

Main Results:

• Impressive performance of multideterminant expansions in QMC.
• Systematic reduction of fixed-node error in the wave function.
• Achievement of chemical accuracy in energy predictions, outperforming MP2, CCSD(T), and DFT approximations.
• Results comparable to explicitly correlated CCSD(T) with large basis sets.

Conclusions:

• Large-scale multideterminant expansions are a promising strategy for improving QMC accuracy.
• This approach effectively reduces fixed-node errors, leading to reliable energy predictions.
• QMC methods, with further development, are poised to become a benchmark for large electronic systems.