Atomically Thin Materials Significantly Shrink Qubits

Quantum IT is diabolically complex technology, with many technical obstacles having an impact on its development. Among these challenges, two critical problems are distinguished: miniaturization and quality of the qubit.

IBM has adopted the superconductive qubit map of the damage of a processor of 1,121 qubit by 2023, leading to the expectation that 1,000 qubits with the qubit factor today are possible. However, current approaches will require very large chips (50 millimeters on one or more side) on the scale of small plates, or the use of chiplets on multi-chien modules. Although this approach will work, the goal is to get a better way to scalability.

Now MIT researchers have been able to reduce the size of the qubits and have done so in a way that reduces interference that occurs between neighboring qubits. MIT researchers have increased the number of superconductive qubits that can be added to a device by a factor of 100.

“We approach both the miniaturization and the quality of the qubit,” said William Oliver, director of the Center for Quantum Engineering at MIT. “Unlike the scale of conventional transistors, where only the number is really important, for qubits, many are not sufficient, they must also be high performance. Sacrifying the performance for the number of qubit is not a useful exchange for quantum calculation. They must go hand in hand. “

The key to this large increase in qubit density and the reduction in interference is summed up with the use of two -dimensional materials, in particular the 2D hexagonal boron nitride (HBN). MIT researchers have shown that a few atomic monocouches of HBN can be stacked to form the island in the capacitors of a superconductive qubit.

Like the other capacitors, the capacitors of these superconductive circuits take the form of a sandwich in which an insulating material is sandwich between two metal plates. The big difference for these capacitors is that the superconductive circuits can only operate at extremely low temperatures-less than 0.02 degrees above absolute zero (-273.15 ° C).

Supervisive qubits are measured at temperatures as low as 20 millikelvin in a dilution refrigerator.Nathan Fiske / MIT

In this environment, insulating materials that are available for work, such as PE-CVD silicon oxide or silicon nitride, have a lot of defects that are too compatible for quantum computer applications. To get around these material gaps, most of the superconductive circuits use what are called coplanar capacitors. In these capacitors, the plates are positioned laterally to each other, rather than each other.

Consequently, the intrinsic silicon substrate under the plates and, to a smaller extent, the void above the plates serves as a dielectric capacitor. The intrinsic silicon is chemically pure and therefore has little defects, and the large size dilutes the electric field with plate interfaces, which leads to a capacitor with low loss. The lateral size of each plate in this open design ends up being quite large (generally 100 over 100 micrometers) in order to reach the required capacity.

In an effort to move away from the great lateral configuration, MIT researchers have launched a search for insulation that has very little defects and is compatible with superconductive capacitor plates.

“We have chosen to study HBN because it is the most used insulator in research on 2D materials due to its cleanliness and its chemical inertia,” said Colead author Joel Wang, researcher in the engineering group of quantum systems of MIT Laboratory for Electronics.

On each side of the HBN, the MIT researchers used the 2D superconductive material, the Niobium desélenide. According to Wang, one of the most delicate aspects of the manufacturing of the capacitors was to work with the Niobium desélenide, which oxidizes in seconds when exposed to air, according to Wang. This requires that the capacitor assembly occurs in a glove box filled with Argon gas.

Although this apparently complicates the scale of production of these capacitors, Wang does not consider this as a limiting factor.

“What determines the quality factor of the capacitor is the two interfaces between the two materials,” said Wang. “Once the sandwich is done, the two interfaces are” sealed “and we do not see any notable degradation over time when it is exposed to the atmosphere.”

This lack of degradation is due to the fact that around 90% of the electric field is contained in the sandwich structure, so that the oxidation of the external surface of the Niobium desélenid no longer plays a significant role. This ultimately makes the capacitor’s imprint much lower, and this explains the reduction in the counterpart between neighboring qubits.

“The main challenge to increase manufacturing will be growth on the level of the HBN and 2D superconductors as well as [niobium diselenide]And how can we make a stack on the scale of the slice of these films, “added Wang.

Wang believes that this research has shown that 2D HBN was a good insulating candidate for the superconductive qubits. He says that the bases that the MIT team made will serve as a roadmap to use other 2D hybrid materials to build superconductive circuits.