Quantum Leap: Using Hybrid Computing to Uncover the Secrets of Complex Molecules

By UNIVERSITY OF CHICAGO

April 09, 2023

 A quantum computational solution for material engineering

Argonne researchers are investigating the use of a quantum computer to solve the electronic structures of complex molecules.

If you know the atoms that make up a particular molecule or solid material, you can compute the interactions between those atoms by solving quantum mechanical equations—at least if the molecule is small and simple. However, for complex molecules and materials, solving these equations, which are critical for fields ranging from materials engineering to drug design, requires prohibitively long computational times.

Researchers at the United States Department of Energy's (DOE) Argonne National Laboratory, as well as the University of Chicago's Pritzker School of Molecular Engineering (PME) and Department of Chemistry, have now investigated the possibility of solving these electronic structures using a quantum computer.

The study, which employs a number of novel computational methods, was published online in the Journal of Chemical Theory and Computation. Q-NEXT, a DOE National Quantum Information Science Research Centre led by Argonne, and the Midwest Integrated Centre for Computational Materials provided funding (MICCoM).

"This is an exciting step towards using quantum computers to solve difficult problems in computational chemistry," said Giulia Galli, who co-led the study with Marco Govoni, an Argonne staff scientist and member of the University of Chicago Consortium for Advanced Science and Engineering (CASE).

A computational problem

Predicting a material's electronic structure entails solving complex equations that determine how electrons interact, as well as modelling how different possible structures compare to one another in terms of overall energy levels.

Unlike traditional computers, which use binary bits to store information, quantum computers use qubits, which can exist in superpositions of states, allowing them to solve certain problems more easily and quickly. Computational chemists have argued about whether and when quantum computers will be able to solve the electronic structure problem of complex materials better than traditional computers. However, today's quantum computers are still small and generate noisy data.

Despite these shortcomings, Galli and her colleagues wondered if they could still make progress in developing the underlying quantum computational methods needed to solve electronic structure problems on quantum computers.

"We really wanted to address the question of what is possible with the current state of quantum computers," Govoni said. "We asked, 'Can the results of quantum computers be useful to solve interesting problems in materials science even if they are noisy?'"

It is an iterative process

Using IBM quantum computers, the researchers devised a hybrid simulation process. A small number of qubits—between four and six in their approach—perform part of the calculations, and the results are then processed using a traditional computer.

"We designed an iterative computational process that takes advantage of the strengths of both quantum and conventional computers," said Benchen Huang, the new paper's first author and a graduate student in the Galli Group.

The simulation process was able to provide the correct electronic structures for several spin defects in solid-state materials after several iterations. Furthermore, the team created a new error mitigation strategy to help control the inherent noise generated by the quantum computer and ensure the accuracy of the results.

Hints at what is to come

For the time being, the electronic structures solved using the new quantum computational approach can be solved using a standard computer. As a result, the long-standing debate over whether a quantum computer can outperform a classical one in solving problems of electronic structure remains unresolved.

However, the new method's results pave the way for quantum computers to address more complex chemical structures.

"We think we might have an advantage over conventional computers if we scale this up to 100 qubits instead of 4 or 6," Huang said. "However, only time will tell."

The research team intends to continue improving and scaling up their approach, as well as applying it to various types of electronic problems, such as molecules in the presence of solvents and molecules and materials in excited states.

Read original article...