Interactive quantum chemistry: a divide-and-conquer ASED-MO method.

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Molecular Modeling

Background:

• Traditional quantum chemistry simulations are computationally intensive, limiting their application to smaller systems.
• Semiempirical methods offer a balance between accuracy and speed but often lack interactivity for complex systems.

Purpose of the Study:

• To develop an interactive quantum chemistry simulation method for large molecular systems.
• To implement and validate a divide-and-conquer approach for atom superposition and electron delocalization molecular orbital (ASED-MO) theory.
• To enable real-time feedback for molecular design and prototyping.

Main Methods:

• Utilized a divide-and-conquer (D&C) algorithm for non-self-consistent semiempirical quantum chemistry.
• Employed direct algorithms for all computational steps, ensuring controllable and non-iterative calculations.
• Analyzed D&C approach errors empirically and theoretically using analytically solvable toy models.

Main Results:

• Achieved linear complexity with respect to the number of atoms and efficient scaling with the number of cores.
• Demonstrated interactive quantum chemistry simulations for systems up to a few hundred atoms on a desktop computer.
• Showcased immediate, intuitive feedback on chemical structures during molecular editing.

Conclusions:

• The D&C approach provides an accurate and efficient method for interactive quantum chemistry simulations.
• Interactive simulations at the ASED-MO level are feasible for systems of hundreds of atoms.
• This technology holds significant potential for accelerating the understanding, design, and prototyping of molecules, devices, and materials.