A team of Columbia University researchers has developed a new algorithm that could help quantum computers calculate molecular energy and lead to the design of new materials. The algorithm uses the most quantum bits to date to calculate the ground state energy, which is the lowest energy state in a quantum mechanical system.
The new study was published in Nature.
Ground state energy calculation
The algorithm was developed by Columbia chemistry professor David Reichman and postdoc Joonho Lee, along with Google Quantum AI researchers. It reduces statistical errors produced by quantum bits in chemical equations and uses up to 16 qubits on Google’s 53-qubit Sycamore computer to calculate ground state energy, which is the energy state the lowest of a molecule.
“These are the largest quantum chemical calculations ever performed on a real quantum device,” Reichman said.
By being able to accurately calculate the ground state energy, chemists will be able to develop new materials. For example, the algorithm could be used to design materials that accelerate nitrogen fixation for agriculture. This is just one of many possible uses for durability, according to Lee, who is a visiting researcher at Google Quantum AI.
The algorithm is based on a quantum Monte Carlo, which is a system of methods for calculating probability when there are many unknown random variables. The researchers deployed the algorithm to determine the ground state energy of three types of molecules.
Many variables can influence the ground state energy, such as the number of electrons in a molecule, the direction of their spin, and the paths they take when orbiting a nucleus. Electronic energy is encoded in Schrodinger’s equation, which becomes extremely difficult to solve on a conventional computer as molecules get larger. That said, there are methods to facilitate this, and quantum computers could potentially circumvent this exponential scaling problem.
Handling larger and more complex calculations
According to the principle, it should be possible for quantum computers to handle larger and more complex computations since qubits take advantage of quantum states. Qubits can exist in two states simultaneously, which is not true for binary digits. At the same time, qubits are fragile, and as the number of qubits increases, the accuracy of the final answer decreases. Lee developed the new algorithm to leverage the combined power of classical and quantum computers to solve these complex equations more efficiently while minimizing errors.
“It’s the best of both worlds,” Lee said. “We leveraged tools we already had as well as tools considered cutting-edge in quantum information science to refine quantum computational chemistry,” Lee said.
The previous ground-state energy resolution record was based on 12 qubits and a method known as variational quantum eigensolution (VQE). The problem with VQE is that it did not take into account the effects of interacting electrons, which is crucial for calculating the ground state energy. According to Lee, virtual correlation techniques from classical computers could be added to help chemists deal with even larger molecules.
The new hybrid classical-quantum calculations demonstrated accuracy comparable to some of the best classical methods, suggesting that complex problems could be solved more accurately and faster with a quantum computer.
“The feasibility of solving larger, more difficult chemical problems will only increase over time,” Lee said. “This gives us hope that the quantum technologies being developed will be practically useful.”