Taking an analog approach to complex chemical simulation

Simulating the intricate dynamics of chemical reactions remains one of the toughest challenges in quantum chemistry, with the potential to unlock new insights into molecular behaviour and materials design.

While quantum computers hold great promise for tackling this problem, current algorithms demand vast numbers of logical qubits and gates, far exceeding today’s technological limits.

An innovative 2024 EQUS study combined theoretical and experimental expertise to introduce an analog approach that dramatically reduces these resource requirements, bringing chemical simulation within reach.

The novel approach simulates chemical reactions using bosonic systems and introduces a hybrid analog-digital method that could redefine how quantum simulations tackle the complexities of molecular dynamics.

“We could improve understanding of processes that rely on molecules interacting with light, such as how smog is created, how the ozone layer is damaged, or how to improve photovoltaic cells,” says EQUS Associate Investigator Prof Ivan Kassal at the University of Sydney.

A bosonic approach to complex chemical simulations

The prospect of simulating natural processes, such as chemical phenomena, on quantum computers was the idea that launched quantum computing back in 1982.

However, such ultrafast photochemical reactions are notoriously difficult to model on classical computers.

The simplifying approximations used within such simulations break down when modelling these reactions, because electronic and vibrational states in such molecules are so strongly coupled.

Traditional computational methods require immense resources to account for this complexity, but the innovative approach presented in the 2024 EQUS study leverages the unique properties of ‘bosons’—particles capable of occupying the same quantum state as each other—to drastically reduce computational overhead.

The vibrational degrees of freedom of a molecule are encoded in the oscillatory bosonic degrees of freedom of trapped ions, offering a natural and efficient mapping.

“The research explores how quantum computers can help simulate chemical reactions, particularly those involving complex interactions between electrons and atomic vibrations,” explained theoretician Ivan.

“Using a trapped ion quantum system, we theoretically determined and then experimentally demonstrated a way to study these reactions with far fewer resources than traditional methods,” explained EQUS Associate Investigator Dr Ting Rei Tan, who led the experimental component of the work, also at Sydney.

Accessing additional degrees of freedom

Quantum computing relies on qubits, the fundamental units of quantum information that can encode a superposition of values.

However, simulating chemical reactions demands the simulation of interactions with a molecule’s vibrations, which are bosons, which can take infinitely many values. Mapping these bosonic states onto qubits is highly resource-intensive, often requiring many qubits per boson.

“We can quickly see that representing these bosons with qubits isn’t very efficient, as we’d need many qubits per boson and complex interactions,” says lead author Dr Tomas Navickas, former EQUS Research Associate (USYD).

Instead, the researchers took advantage of the bosonic modes inherent in their trapped ion hardware to represent (encode) the bosons involved in molecular dynamics.

This experimental platform allows for precise control over both the internal (qubit-like) and external (bosonic) degrees of freedom of individual charged atoms.

This hybrid encoding of information is the cornerstone of the EQUS team’s approach. “Using this system, with just one trapped ion— including one qubit and two bosonic modes—we can simulate molecules that would otherwise require 11 ‘perfect’ logical qubits and 300,000 noise-free gates on a traditional quantum computer,” said coauthor Dr Henry Nourse.

“Current best qudit-only quantum computers have only shown a handful of logical qubits, which were far from perfect, and each logical qubit required dozens of actual physical qubits.”

Unlike previous efforts, which primarily encoded information in qubits, the team leveraged the full, broad/multi-dimensional Hilbert space of their trapped ion system, allowing them to encode more molecular degrees of freedom in a single trapped ion, crucial for modelling complex chemical dynamics.

“Analog quantum simulations are an exciting new prospect for tackling difficult chemistry problems,” coauthor Vanessa Olaya-Agudelo noted. “With just a few dozen trapped ions, we could potentially simulate problems that are not only classically intractable, but also difficult under conventional quantum modelling methods.”

Challenges and future directions

EQUS’ Designer Quantum Materials research theme seeks to harness the unique properties of quantum systems towards scalable, resource-efficient simulations.

While the results of this study are promising, scaling up these simulations to handle more complex molecules, and incorporating additional environmental factors, remain significant challenges.

The next steps for the research team include extending their method to higher-order vibronic couplings and demonstrating the scalability of their approach with more trapped ions.

“The main challenge is scaling up these simulations while maintaining adequate control and incorporating more realistic environmental interactions,” coauthor Dr Christophe Valahu, EQUS Research Affiliate (USYD) explained.

“However, what we learned from this demonstration makes us confident we can apply these insights to more complicated systems.” The new work offers a new experimental framework for studying chemical dynamics with programmable noise. In the long term, it could shift the quantum computing community’s focus toward hybrid analog-digital approaches for solving classically intractable problems.

Elevating the use of bosons in quantum hardware could prove to be a game-changer for both fundamental science and practical applications.

“If sufficient work is carried out and we successfully scale to more complex molecules, we could be performing classically intractable simulations in the near term,” Ting Rei said.

This story was first published in the 2024 EQUS annual report.