So far we have had to use supercooled materials to try out ultra-fast quantum logic. Now silicon, already standard in electronics, has gained that talent.
The ingredients for superfast computers could be nearly in place. For the first time, researchers have demonstrated that two silicon transistors acting as quantum bits can perform a tiny calculation. Now it’s just a question of using these as the building blocks of a larger quantum computer – taking advantage of the very material that is ubiquitous in conventional electronics.
Where conventional computing uses bits, quantum computing uses qubits, which can take the values 0, 1 or various combinations of these, instead of being stuck at either 0 or 1. This means they could exponentially shrink the time it takes to solve problems, transforming fields like encryption and the search for new pharmaceuticals.
Previously qubit calculations had been made using ultra-cold superconductors, which are easier to couple together into a basic calculator – but never with user-friendly silicon. In silicon, the qubits are isolated to keep them stable, which is a barrier to making two qubits interact with each other.
Now, a team led by Andrew Dzurak of the University of New South Wales in Sydney, Australia, has achieved that feat. Their device looks at the spin of two electrons and follows instructions: if the first one is spinning in a particular direction, flip the spin of the second electron. If not, do nothing.
This is an example of a logic gate, a fundamental unit in a computer.
Repetition of that same humble logic by creating sequences of gates can enable more and more complex calculations. Dzurak’s team says it has patented a design for a chip containing millions of such qubits.
“This is a seminal breakthrough in the world of quantum computer development – with some caveats,” says Thomas Schenckel of the Lawrence Berkley National Laboratory in California. Although easier to scale up, “silicon-based qubits are still way behind superconducting qubits”, he says.
But that doesn’t diminish the potential of the work. “Nothing beats what we can do in silicon in terms of economical scaling and large-scale integration,” Schenkel says.
Journal reference: Nature, DOI: 10.1038/nature15263