Scientists at Yale University and Brookhaven National Laboratory have increased the runtime of superconducting quantum devices with a new approach to chip design and materials selection. The new paradigm enables the coherence time of qubits to be extended to one millisecond. The results are published in the journal Nature Communications.

Scientists at Yale University and Brookhaven National Laboratory have increased the runtime of superconducting quantum devices with a new approach to chip design and materials selection. The new paradigm enables the coherence time of qubits to be extended to one millisecond. The results are published in the journal Nature Communications.

Today, quantum computers are created on superconducting materials. Quantum computing is based on the use of qubits, which are information storage units that can obtain the values ​​zero, one, and the superposition of these values, both zero and one at the same time. In this paradigm, calculations are significantly accelerated. The materials used to create quantum computers also allow the use of the phenomenon of superconductivity, ensuring the lossless flow of electric current. However, not all problems can be minimized thanks to superconductivity.

The issue of energy dissipation is critical for quantum computers. It prevents qubits from remaining in their operating mode. Coherence (the state in which qubits can operate at maximum performance) must be maintained, particularly by reducing the computing power of the entire computer. Therefore, it is important for the development of quantum computing to find ways to maintain the coherent state for as long as possible.

Scientists focused on studying the energy loss mechanisms in superconducting quantum circuits. It was known that the use of tantalum allowed the coherence of qubits to be maintained for three-tenths of a millisecond. The researchers found that the combination of a calcined sapphire substrate and tantalum significantly reduced energy losses at the surface and in the volume of the dielectric. The use of tantalum provides high-quality transitions between the layers that make up the qubits, improves the quality of the metal surface and, as a result, interfaces with other materials. Annealing sapphire substrates at 1200 °C with a constant oxygen source leads to a significant reduction in dielectric losses in the volume of the finished qubit. Experimental data obtained for tantalum and aluminum structures confirmed the theoretical calculations.

The researchers also optimized the geometry of the devices. The qubit consisted of three thin-film superconducting strips deposited on a substrate. The strips were positioned in such a way that it was possible not only to measure the energy loss, but also to determine where it occurred. Thanks to the chosen architecture, it was possible to accurately distinguish between surface losses and bulk dielectric losses.

As a result, a structure was created that allows several qubits with improved characteristics to be placed in a single microcircuit within the framework of existing technical processes. The resulting memory elements have a time interval between signals in the coherence detection process in the range of 2.0 to 2.7 milliseconds, which is limited by an energy relaxation time of 1.0 to 1.4 milliseconds. These results significantly surpass previous achievements in the field of quantum memory in thin-film devices. The new approach made it possible to triple the coherence state time from three tenths of a millisecond to one millisecond.

The loss characterization studies presented in this work have shown clear and realistic ways to increase coherence in superconducting qubits. The development of more specific architectures and processes or the use of materials with lower losses in a well-defined qubit region are critical to improving the coherence of the system. In addition, the reduction of surface losses should be accompanied by the optimization of volumetric dielectric losses, which makes it possible to achieve the design of microcircuits taking into account energy losses.