While the scientific community holds its breath for a large-scale quantum computer that could carry out useful calculations, a team of IBM researchers has approached the problem with an entirely different vision: to achieve more and better results right now, even with the limited quantum resources that exist today.
By tweaking their method, the scientists successfully simulated some molecules with a higher degree of accuracy than before, with no need for more qubits. The researchers effectively managed to pack more information into the mathematical functions that were used to carry out the simulation, meaning that the outcome of the process was far more precise, and yet came at no extra computational cost.
“We demonstrate that the properties for paradigmatic molecules such as hydrogen fluoride (HF) can be calculated with a higher degree of accuracy on today’s small quantum computers,” said the researchers, at the same time priding themselves on helping quantum computers “punch above their weight”.
Car manufacturer Daimler, a long-term quantum research partner of IBM’s, has shown a strong interest in the results, which could go a long way in developing higher-performing, longer-lasting and less expensive batteries.
Since 2015, Daimler has been working on upgrading lithium-ion batteries to lithium-sulfur ones – a non-toxic and easily available material that would increase the capacity and speed-of-charging of electric vehicles.
Designing a battery based on new materials requires an exact understanding of which compounds should come together and how. The process involves accurately describing all the characteristics of all the molecules that make up the compound, as well as the particles that make up these molecules, to simulate how the compound will react in many different environments. In other words, it is an incredibly data-heavy job, with infinite molecular combinations to test before the right one is found.
The classical methods that exist today fail to render these simulations with the precision that is required for a breakthrough such as the one Daimler is working towards. “This is a big problem to develop next-generation batteries,” Heike Riel, IBM Research quantum lead, told ZDNet. “Classical computers, and the models we’ve developed in physics and chemistry for many years still cannot solve those problems.”
But the task could be performed at speed by quantum computers. Qubits, and their ability to encode different information at the same time, enable quantum algorithms to run several calculations at once – and are expected, one day, to enable quantum computers to tackle problems that are seemingly impossible, in a matter of minutes.
To do that, physicists need quantum computers that support many qubits; but scaling qubits is no piece of cake. Most quantum computers, including IBM’s, work with less than 100 qubits, which is nowhere near enough to simulate the complex molecules that are needed for breakthroughs such as lithium-sulfur car batteries.
Some of the properties of these molecules are typically represented in computer experiments with a mathematical function called a Hamiltonian, which represents particles’ spatial functions, also called orbitals. In other words, the larger the molecule, the larger the orbital, and the more qubits and quantum operations will be needed.
“We currently can’t represent enough orbitals in our simulations on quantum hardware to correlate the electrons found in complex molecules in the real world,” said IBM’s team.
Instead of waiting for a larger quantum computer that could take in weighty calculations, the researchers decided to see what they could do with the technology as it stands. To compensate for resource limitations, the team created a so-called “transcorrelated” Hamiltonian – one that was transformed to contain additional information about the behavior of electrons in a particular molecule.
This information, which concerns the propensity of negatively charged electrons to repel each other, cannot usually fit on existing quantum computers, because it requires too much extra computation. By incorporating the behavior of electrons directly into a Hamiltonian, the researchers therefore increased the accuracy of the simulation, yet didn’t create the need for more qubits.
The method is a new step towards calculating materials’ properties with accuracy on a quantum computer, despite the limited resources available to date. “The more orbitals you can simulate, the closer you can get to reproducing the results of an actual experiment,” said the scientists. “Better modelling and simulations will ultimately result in the prediction of new materials with specific properties of interest.”
IBM’s findings might accelerate the timeline of events for quantum applications, therefore, with new use cases emerging even while quantum computers work with few qubits. According to the researchers, companies like Daimler are already keen to find out more about the breakthrough.
This is unlikely to shift IBM’s focus on expanding the scale of its quantum computer. The company recently unveiled a roadmap to a million-qubit system, and said that it expects a fault-tolerant quantum computer to be an achievable goal for the next ten years. According to Riel, quantum simulation is likely to be one of the first applications of the technology to witness real-world impacts.
“The car batteries are a good example of this,” she said. “Soon, the number of qubits will be enough to generate valuable insights with which you can develop new materials. We’ll see quantum advantage soon in the area of quantum simulation and new materials.”
IBM’s roadmap announces that the company will reach 1,000 qubits in 2023, which could mark the start of early value creation in pharmaceuticals and chemicals, thanks to the simulation of small molecules.
