Quantum computers take a step into the science of physical materials

Studying and designing quantum mechanics is a fundamental application of quantum mechanics. Chemists, material scientists, and physicists focus on the subtle interactions of quantum materials and to reveal them they rely on sophisticated experimental and experimental techniques. Computer simulations that link tiny quantum interactions to measurable properties complement experimental data to link structure to function — but older computers can struggle to simulate those properties. Fortunately, scientists today have a new tool in their toolbox: quantum computers.

In a new publication, a team of researchers from Oak Ridge National Lab’s (ORNL’s) Quantum Science Center (QSC), Purdue University, Los Alamos Laboratory, University of Illinois at Urbana-Champaign, University of Tennessee, and IBM used quantum simulation to combine the energy of the KCustrum’s magnetic field.$_{3}$

The quantum simulations used the IBM Quantum Heron processor, while the experimental data were obtained from neutron sources at the Spallation Neutron Source (SNS) at ORNL and the Rutherford Appleton Laboratory in the United Kingdom. This work serves as another realization of Richard Feynman’s vision: the use of a well-controlled, well-organized quantum system to simulate the properties of the quantum system of interest.

“This is the most fascinating match I’ve seen between experimental data and a qubit simulation, and it really raises the bar for what can be expected from supercomputers,” said co-author Allen Scheie, a physicist at Los Alamos National Laboratory. “I’m really excited about what this means for science.”

By blasting the KCuF sample$_{3}$

However, because the experimentally determined values ​​involve the energies of many entangled spins, they can be very difficult to calculate classically. “There is a lot of neutron scattering information about magnetic materials that we don’t fully understand because of the limitations of classical techniques,” Banerjee said.

Quantum computers have long been expected to make hardware simulation a challenge for traditional methods. Despite advances in quantum computing and hardware, it is not yet clear whether modern fault-tolerant computers with finite-valued gates are capable of simulating real devices. Simulating the energy from neutron scattering was a particularly good candidate, because the interaction of spins in matter and neutrons can be easily mapped to quantum circuits.

“A spin is a qubit is a spin,” Banerjee said. “Quantum computing provides observables such as neutron scattering.” And despite this effective map, the feasibility of such a comparison with current equipment is still a question mark. “When we started this project, it was not clear to us how many qubits and gates would be needed for this simulation,” said IBM research scientist Bibek Pokharel, one of the study’s lead authors.

But as the study showed, advances in the scale and quality of quantum processors were critical to the accuracy of the simulations achieved in this work. “These results are really made possible by the low error rates in all fifty qubits used for the simulation,” said co-author Abhinav Kandala, principal research scientist at IBM. The hardware development was also complemented by a powerful algorithm and the use of old computer hardware in the Illinois Campus Cluster to reduce the circuit depth of quantum circuits. This approach is consistent with IBM’s broader vision of quantum-centric supercomputing: Combining high-performance computing (HPC) with quantum devices will prove more capable and useful for scientific problems than the technology itself. “All in all, it was very impressive to finally see that you can use a quantum computer as a new computational tool that now has a sufficient resolution to handle parts of real experimental data,” Kandala noted.

Although quantum computers are well suited to simulating spin Hamiltonians, with the right encodings they can simulate a wider class of Hamiltonians associated with many quantum devices. This creates a single quantum processor, with universal gates, capable of simulating multiple devices. Travis Humble, director of the QSC at ORNL said: “Many simulations of real types of devices and their experimental analysis is a great indicator of the impact of quantum computing on the flow of scientific work. Indeed, the researchers used the same processor and universal gate set to simulate the properties of another family of cobalt-based devices with complex interactions.

This work also confirms that quantum computers can find applications that work as reliable quantum simulators, even before the advent of fault-tolerant quantum computing.

Going forward, the researchers plan to apply this type of simulation to quantum devices with higher levels and more complexity than KCuF.$_{3}$