A new analysis indicates that two key quantum computing algorithms for chemistry problems have limited practical use, even with advancing hardware. Researchers suggest that calculating molecular energy levels may not justify the technology's investment as hoped. This challenges the view of quantum chemistry as a major application for quantum computers.
Quantum computers have seen rapid progress in recent years, raising questions about their most valuable applications. Among the leading candidates has been quantum chemistry, where the devices could calculate energy levels of molecules to aid biomedicine and industry, such as in drug development or agriculture. These calculations involve tracking multiple electrons simultaneously, aligning with quantum computers' strengths.
However, a study by Xavier Waintal at CEA Grenoble in France and his team argues that this promise may be overstated. They examined two prominent algorithms: the variational quantum eigensolver (VQE) for current, error-prone quantum hardware, and quantum phase estimation (QPE) for future fault-tolerant systems.
For VQE on noisy quantum computers, the team found that achieving accuracy comparable to classical methods requires suppressing errors to near-fault-tolerant levels. No such practical fault-tolerant quantum computer exists yet, though several companies target development within five years.
Even with fault-tolerant hardware, QPE faces the 'orthogonality catastrophe,' where success in finding a molecule's lowest energy level drops exponentially as molecule size grows. Thibaud Louvet at Quobly notes that this limits QPE to a narrow set of cases, viewing it more as a maturity benchmark than a routine tool for chemists.
Xavier Waintal expressed skepticism, stating, “My personal thinking is that it’s probably doomed, not proven doomed, but probably doomed,” regarding quantum computers for molecular energy calculations.
George Booth at King’s College London, not involved in the research, agrees on the challenges: “This study is clear to point out significant challenges for accurate molecular simulation, which will remain even in the ‘fault-tolerant era’, and cast doubt on whether quantum chemistry is really such a quick win for quantum computers.” He adds that quantum devices might still assist in other chemistry tasks, like simulating responses to perturbations such as laser light.
The findings appear in Physical Review B (DOI: 10.1103/hpt6-9tnk).