In the race to build large-scale quantum computers, two contrasting strategies – one based on trapping ions, the other on more conventional technology – have drawn neck-and-neck. Both sides can now create simple devices that run multiple varieties of quantum software.
Early endeavours had focused on trying to run a single quantum algorithm, such as Shor’s algorithm for factoring numbers. A large enough quantum device running these algorithms should massively outperform ordinary computers. But the strategy is limited in scope – if ordinary computers were designed like this, you’d need a different laptop for every app you wanted to run.
That’s why attention has now turned to creating programmable quantum computers. In May this year, IBM announced it was making such a device available for anyone to use over the internet. Its computer has five quantum bits, or qubits, so can only handle relatively small problems – but it’s programmable just like a regular PC.Researchers at Google have developed a similar device, although have not made it accessible to the public.
Both of these computers use superconducting qubits built using techniques from the conventional computer chip industry. Now, a team at the University of Maryland has succeeded with its own quite different approach to making a programmable five-qubit computer.
Their qubits are made from ytterbium ions held in place by magnetic fields and lasers, a technology with its origins in atomic clocks. “Ions are nature’s quantum units,” says team member Shantanu Debnath. “If you have a bunch of them in a processor, all of them are identical, and that is a significant advantage.”
Trapped-ion qubits have another edge over the superconducting variety in being able to communicate with each other at a distance, thanks to the weird property of quantum entanglement. This allows the computer to process data more easily. “Any ion can interact within any other,” says Debnath. “Quantum entanglement is at the heart of parallel processing and speed-up.”
In contrast, superconducting qubits can only swap data with their nearest neighbour, meaning two distant qubits have to slog through all the ones in between in order to communicate. “That is something they are going to pay for in the long term,” says Debnath. The upside is that existing chip fabrication technology should make it easier to manufacture superconducting qubits in bulk.
Ion traps and superconductors are the two most advanced quantum hardware techniques around, says Simon Devitt of the RIKEN Center for Emergent Matter Science in Saitama, Japan. Which one will pull ahead remains to be seen.
“Quantum information technology is certainly going through a second renaissance, and technological advances and investment are increasing at a faster and faster pace,” says Devitt. “Results like this are great to see and I hope they keep coming.”