Australian scientists were able to reduce the error rate in semiconductor qubits, the unit cells of a quantum computer, to a level of 0.04%. This paves the way for the creation of universal computers, physicists say in the journal Nature Electronics.
For several years now, Dzurak and his university colleagues have been developing the components needed to assemble a full-fledged solid-state quantum computer. So, in 2010 they created a quantum single-electron transistor, and in 2012 - a full-fledged silicon qubit based on the phosphorus-31 atom.
In 2013, they assembled a new version of the qubit, which made it possible to read data from it with almost 100% accuracy and remained stable for a very long time. In October 2015, Dzurak and his team took the first step towards creating the first silicon quantum computer by combining two qubits into a module that performs a logical OR operation.
There was only one step left - to learn how to combine similar qubits using the same semiconductor technologies as the cells of quantum memory themselves. It was extremely difficult to do this, since “ordinary” semiconductor qubits can only interact with each other at a short distance.
Having solved this problem two years ago, Australian scientists thought about how to "glue" qubits into a single whole and learn how to "print" them the way electronics manufacturers do when creating microcircuits. The fruit of these reflections was the first plans for the creation of quantum "microcircuits", presented by the Dzurak team in December 2017.
These ideas, as noted by Dzurak, his team managed to put into practice last fall, using the so-called CMOS technology - one of the most common and proven methods of manufacturing microcircuits. Scientists have used it to "print" all the components of the qubits, as well as microwave emitters, quantum dots and transistors required to correctly write new data into a quantum memory cell.
Having solved this problem, physicists thought about the next big step - to create a truly universal quantum computer, they needed to make their qubits work almost perfectly, making mistakes no more than 1% of the time. In this case, other problems in their work can be eliminated using special error correction algorithms and logical, rather than physical, qubits.
As the researcher notes, there are two ways to improve the accuracy of such devices - by improving the design of the memory cells themselves and changing the way information is read and written into them. Australian physicists took the second path, using algorithms and techniques developed by their theoretical colleagues at the University of Sydney.
Promotional video:
They helped Zuraku and his team change the structure of the microwave control pulses in such a way that the number of errors when reading or writing data was reduced by several orders of magnitude. As a result, scientists not only stepped over the “barrier of error correction”, but also bypassed superconducting and “atomic” qubits, which were previously considered more promising for creating complex quantum machines.
In the near future, both groups of researchers plan to carry out similar measurements on combinations of several qubits and microcircuits that have already been created by Dzurak and his team in the past. Scientists hope they will be able to reduce the overall error rate to a level that will allow the creation of a full-fledged quantum computer in the coming years.