Quantum Circuits Simulate Molecular Motion with Langevin Thermostat Precision

Scientists Masari Watanabe of Quemix Inc, and colleagues, have developed a new quantum-circuit framework for simulating finite-temperature molecular dynamics using a Langevin thermostat. The framework encodes classical molecular phase-space distributions as Koopman-von Neumann wave functions, enabling the preparation of canonical states and subsequent physical property readouts on quantum hardware. Analytical quantification and correction for temperature biases arising from the probabilistic implementation of diffusion are included, and the approach is validated through simulations of the H 2 molecules. These simulations successfully demonstrate both dynamical property estimation and static rate constant calculations. Connecting Langevin canonical state preparation to physical property calculations represents a key step towards achieving quantum, classical hybrid molecular dynamics on quantum computers.

Rapid canonicalisation and vibrational density readout via quantum nuclear dynamics simulation

Molecular dynamics simulations are fundamental to understanding chemical processes and material properties, but accurately modelling finite-temperature effects remains a significant computational challenge. Classical molecular dynamics relies on sampling the Boltzmann distribution, which becomes increasingly difficult for systems with high dimensionality or complex potential energy surfaces. Quantum computing offers a potential avenue to overcome these limitations by leveraging quantum phenomena to efficiently explore phase space. This research introduces a novel approach to finite-temperature molecular dynamics by encoding the classical nuclear phase-space distribution as a Koopman-von Neumann (KvN) wave function. This encoding allows the researchers to formulate canonical state preparation as a Langevin-type Fokker-Planck relaxation process, effectively mimicking the behaviour of a thermostat on a quantum computer. The Langevin thermostat introduces a frictional force and a random force, allowing the system to reach thermal equilibrium.

Initial Kullback-Leibler divergence, a measure of the difference between the evolving phase-space distribution and a stable canonical distribution, began at 30 and was reduced to approximately 0.05 nats after 48.4 femtoseconds. This three-order-of-magnitude improvement demonstrates successful relaxation towards the desired canonical reference distribution, something previously unattainable with purely classical methods, particularly for complex systems. The quantum-circuit framework consistently transfers potential energy surfaces and forces from electronic-structure calculations, typically derived from methods like density functional theory or coupled cluster theory, to the Koopman-von Neumann nuclear dynamics. This transfer is crucial for maintaining accuracy and relevance to real molecular systems. The successful preparation of the canonical phase-space state was further confirmed by a mean internuclear distance relaxing from an initial value of 1.82 Å to an equilibrium bond length of approximately 0.74 Å. This demonstrates the framework’s ability to accurately reproduce the structural properties of the molecule at the target temperature. Kinetic temperature initially peaked at 4.9 × 103 K before decreasing and stabilising close to the target physical temperature of 947 K, with a minor correction to 942 K applied for the cosine-filter approximation used in the simulation. The cosine-filter approximation is a numerical technique used to represent the random force in the Langevin thermostat, and the small correction indicates the accuracy of the implementation. While finite evolution time and anharmonicity necessitate further refinement for truly practical molecular dynamics simulations, these initial results demonstrate the potential for accurate simulations, particularly in capturing the vibrational behaviour of molecules. Anharmonicity, the deviation of a molecule’s vibrations from simple harmonic motion, is a key factor influencing molecular properties and requires careful consideration in simulations.

Quantum simulation of molecular vibrations using encoded classical motion

Realistic temperatures are crucial for simulating molecular behaviour, aiding understanding of chemical reactions and the design of new materials. Accurate modelling of temperature-dependent effects is essential for predicting reaction rates, understanding spectroscopic properties, and designing materials with specific thermal characteristics. Accurately representing these thermal effects on quantum computers, however, presents a considerable hurdle. Quantum computers operate based on the principles of quantum mechanics, which differ significantly from the classical mechanics governing molecular motion at finite temperatures. Bridging this gap requires innovative approaches to encode classical information into quantum states. This research offers a promising pathway by encoding classical molecular motion as a quantum wave function, currently demonstrated with the hydrogen molecule as a proof of concept. The choice of the hydrogen molecule is strategic, as it is a simple system that allows for clear validation of the framework before applying it to more complex molecules. A quantum-circuit framework capable of simulating molecular dynamics at finite temperatures has been established, representing a major step towards hybrid quantum-classical computation. Hybrid quantum-classical computation combines the strengths of both classical and quantum computers, leveraging classical resources for tasks like data pre-processing and post-processing, while utilising quantum computers for computationally demanding tasks like solving the Schrödinger equation.

Encoding classical molecular motion as a Koopman-von Neumann wave function and utilising a Langevin thermostat successfully prepared a canonical state on a quantum computer. The Koopman-von Neumann formalism provides a mathematical framework for mapping classical observables onto quantum operators, enabling the simulation of classical dynamics on a quantum computer. Dynamic properties and static rate constants were then extracted from the hydrogen molecule, validating the approach and linking initial state preparation to measurable outcomes. Specifically, the framework allows for the calculation of vibrational frequencies and the determination of reaction rates, providing valuable insights into molecular behaviour. This paves the way for exploration of more complex molecular systems and refinement of simulations beyond the limitations of purely classical methods. Future research will focus on extending the framework to larger molecules, incorporating more accurate potential energy surfaces, and developing more efficient quantum algorithms to reduce the computational cost of the simulations. The ultimate goal is to develop a quantum-classical hybrid molecular dynamics engine that can accurately and efficiently simulate the behaviour of complex molecular systems, enabling the design of new materials and the discovery of new chemical reactions.

The researchers successfully constructed a quantum-circuit framework for simulating molecular dynamics at finite temperatures using the H₂ molecule. This means classical molecular motion can now be encoded as a wave function and prepared in a canonical state on a quantum computer. The framework links the initial preparation of this state to the readout of measurable properties, such as vibrational frequencies and reaction rates, demonstrating a key step towards hybrid quantum-classical computation. The authors intend to extend this work to larger molecules and more accurate simulations.