Insider Brief
- A new study in The Journal of Physical Chemistry B shows that quantum computers can now simulate solvated molecules, marking a step toward solving realistic chemical problems in biology and industry.
- The research extended the SQD method to include solvent effects using IEF-PCM, and tested it on IBM quantum hardware for water, methanol, ethanol, and methylamine.
- Results matched classical benchmarks within chemical accuracy, demonstrating that hybrid quantum-classical models are becoming viable for complex molecular simulations.
Quantum computers may be able to simulate molecules in realistic environments, a key step that could bridge a gap that has long hindered quantum chemistry from addressing biologically and industrially relevant problems, according to a new study in The Journal of Physical Chemistry B.
The research team, led by Kenneth Merz Jr., of Cleveland Clinic’s Center for Computational Life Sciences, have successfully extended the sample-based quantum diagonalization (SQD) method — originally limited to gas-phase simulations — to account for solvent effects using an implicit model. The work integrates the integral equation formalism polarizable continuum model (IEF-PCM) — a well-established technique in classical chemistry that treats the liquid around a molecule as a smooth, invisible material to simplify molecular interactions — into quantum simulations run on real IBM quantum devices.
“This study is a significant stride towards practical quantum chemistry on quantum computers,” Dr. Merz said in a statement. “Quantum hybrid models are still largely unexplored and very few are tested on quantum hardware. By testing this model on quantum hardware, we are demonstrating its abilities to advance chemical research using quantum computers.”
From Dry Molecules to Realistic Chemistry
The inclusion of solvent effects in quantum chemistry is critical for studying phenomena like protein folding, drug binding and catalytic reactions. Yet, solvent modeling has mostly remained a classical computing task due to the immense complexity involved. In most previous quantum computing studies, molecules were modeled in isolation, ignoring how they interact with their environment.
This study takes a step in changing that. Using IEF-PCM, the team modeled water, methanol, ethanol and methylamine — common polar molecules in biochemistry — within an aqueous solution. The method approximates the solvent as a continuous medium, simplifying the many-body interactions without adding thousands of individual water molecules.
The SQD-IEF-PCM technique combines this classical solvent model with quantum-generated samples, achieving chemical accuracy while limiting computational cost. The researchers found their results were in tight agreement with established high-accuracy methods, like complete active space configuration interaction (CASCI), which are notoriously expensive to run on classical computers.
How It Works
The method begins by generating electronic configurations from a molecule’s wavefunction using quantum hardware. These samples, affected by hardware noise, are corrected through a self-consistent process (S-CORE) that restores key physical properties like electron number and spin. The corrected configurations are used to construct a smaller subspace of the full molecular problem, one that’s manageable to solve classically.
What sets this work apart is the use of IEF-PCM to include the influence of the solvent in this quantum-classical workflow. Once the solvent effect is added as a perturbation to the molecule’s Hamiltonian — the operator that describes the system’s total energy — the process becomes an iterative one, updating the molecular wavefunction until solvent and solute are mutually consistent.
This hybrid approach was tested on IBM quantum computers with 27 to 52 qubits. Despite the modest size of the quantum devices and the inherent noise of current quantum hardware, the simulations produced solvation free energies that closely matched classical benchmarks. For example, the solvation energy of methanol differed by less than 0.2 kcal/mol between the quantum and classical approaches, well within the threshold of chemical accuracy. (Kilocalories per mole is a measurement used to express energy differences in chemistry and is a standard way to measure chemical accuracy.)
What the Results Show
The accuracy of SQD-IEF-PCM improved with the number of samples used in each simulation. Even in complex molecules like ethanol, where the full quantum configuration space is enormous, the method identified and focused on the most critical regions, reaching good accuracy with only a fraction of the full data.
In all four molecules studied, energy convergence to the classical CASCI-IEF-PCM reference improved with sample size, and solvation energies stayed within 1 kcal/mol of both CASCI and experimental values from the MNSol database.
The approach demonstrated scalability, robustness to hardware noise, and versatility across different chemical systems. More importantly, it showed that chemically relevant problems — previously off-limits to quantum computing — are becoming more and more accessible as methods and machines improve.
Limitations and Next Steps
Despite the impressive advances, the authors acknowledge some limitations. The current method is most suitable for neutral molecules, and further work is needed to assess its performance on charged systems. They also highlight the need for better parameterization of the quantum circuits to reduce the number of samples required for accurate results.
Additionally, while the implicit solvent model works well for electrostatics, it doesn’t capture all solute–solvent interactions, such as hydrogen bonding and dispersion. These effects would require future extensions to include explicit solvent molecules or more advanced hybrid models.
To improve efficiency, the team plans to integrate a parallel eigensolver and explore the use of optimized ansatz circuits. In simple terms, researchers will use multiple computers working together to quickly find important energy values in a molecule — and that could potentially enable the simulations of even larger systems or more precise results with fewer samples. They are also considering adapting the method to incorporate other solvent models and benchmarking it against classical methods like heat-bath configuration interaction (HCI), which has not yet been implemented with implicit solvents.
Implications for Drug Design and Quantum Advantage
This work has immediate relevance for pharmaceutical and materials science research. Being able to simulate molecules in solution with quantum algorithms opens new possibilities in understanding how drugs behave in biological environments or how catalysts function in industrial processes.
While the long-promised “quantum advantage” in chemistry remains elusive, the authors argue that sample-based diagonalization could be among the most promising routes to get there. Unlike variational algorithms that require repeated evaluations and are vulnerable to noise, SQD avoids these pitfalls by reducing the quantum workload to sampling and outsourcing most of the heavy lifting to classical algorithms.
No current quantum algorithm has yet beaten classical ones in real-world chemistry problems. But with methods like SQD-IEF-PCM becoming more accurate and scalable, that moment may be closer than expected.
In addition to Merz, the research team included Danil Kaliakin, Akhil Shajan and Fangchun Liang, all of the Cleveland Clinic.
0 Comments