Silicon Quantum Computing Launches Quantum Twins for Advanced Simulation
While the pursuit of universal quantum computers continues to face significant hurdles, a parallel approach – analog quantum simulation – is gaining momentum. This method doesn’t aim for the full control of individual qubits, but instead focuses on directly mimicking complex systems, such as molecules and materials, that are intractable for classical computers. What analog quantum simulation lacks in universal flexibility, it compensates for in near-term feasibility.
“Instead of using qubits, as you would typically in a quantum computer, we just directly encode the problem into the geometry and structure of the array itself,” says Sam Gorman, quantum systems engineering lead at Sydney-based startup Silicon Quantum Computing.
On , Silicon Quantum Computing unveiled its Quantum Twins product, a silicon quantum simulator now available to customers through direct contract. Concurrently, the company demonstrated its device, comprised of 15,000 quantum dots, can simulate a metal-insulator transition – a fundamental phenomenon in materials science. The findings were published this week in the journal Nature.
“We can do things now that we think nobody else in the world can do,” Gorman says.
The Power of Precision Manufacturing
The Quantum Twins product is the culmination of work that began following Silicon Quantum Computing’s establishment in , building on over 25 years of academic research led by the company’s founder, Michelle Simmons. The core technology is a manufacturing process for placing single phosphorus atoms in silicon with subnanometer precision.
“We have a 38-stage process,” Simmons explains, “for patterning phosphorus atoms into silicon. The process starts with a silicon substrate, which gets coated with a layer of hydrogen. Then, using a scanning-tunneling microscope, individual hydrogen atoms are removed, exposing the silicon underneath. The surface is then exposed to phosphine gas, which adheres only to the exposed silicon. A low-temperature thermal anneal incorporates the phosphorus atom into the silicon crystal, and layers of silicon are grown on top.”
“It’s done in ultrahigh vacuum, ensuring a very pure and clean system,” Simmons adds. “It’s a fully monolithic chip that we make with that subnanometer precision. In , we figured out how to create markers within the chip to locate the placed atoms for making contacts. These contacts are made at the same length scale as the atoms and dots.”
While the team places individual phosphorus atoms, they utilize clusters of 10 to 50 atoms to form registers for these application-specific chips. These registers function as quantum dots, preserving the quantum properties of the individual atoms. Gate voltages from contacts atop the chip control these registers, and interactions between registers are tuned by precisely controlling their distances.
Silicon Quantum Computing is also pursuing more traditional quantum computing with this technology, but realized the potential for useful simulations in the analog domain by placing thousands of registers on a single chip and measuring global properties, without needing to control individual qubits.
“The thing that’s quite unique is we can do that very quickly,” Simmons says. “We put 250,000 of these registers [on a chip] in 8 hours, and we can turn a chip design around in a week.”
Simulating the Physical World
In , the team simulated a molecule of polyacetylene using a previous iteration of this technology. This molecule, composed of carbon atoms with alternating single and double bonds, exhibits drastically different conductivity depending on whether the chain is cleaved at a single or double bond. Accurately simulating these bonds required controlling the distances between registers to subnanometer precision. By tuning the gate voltages of each quantum dot, the researchers replicated the observed conductivity jump.
The current demonstration focuses on the metal-insulator transition in a two-dimensional material. The polyacetylene simulation used 10 registers, while the new model employs 15,000. This transition is particularly challenging for classical computers because, at the extremes – fully metallic or fully insulating – the physics can be simplified. However, in the intermediate state, the quantum complexity of each electron becomes crucial, rendering the problem classically intractable. “That is the part which is challenging for classical computing. But we can actually put our system into this regime quite easily,” Gorman says.
The metal-insulator model served as a proof of concept. Gorman states the team can now design a quantum twin for almost any two-dimensional problem.
“Now that we’ve demonstrated that the device is behaving as we predict, we’re looking at high-impact issues or outstanding problems,” says Gorman. The team plans to investigate areas like unconventional superconductivity, the origins of magnetism, and materials interfaces found in batteries.
While initial applications are expected in scientific research, Simmons is optimistic about the potential for industrial applications, such as drug discovery. “If you look at different drugs, they’re actually very similar to polyacetylene. They’re carbon chains, and they have functional groups. So, understanding how to map it [onto our simulator] is a unique challenge. But that’s definitely an area we’re going to focus on,” she says. “We’re excited at the potential possibilities.”
