A quantum calculation cracking one of the most common forms of data encryption has been demonstrated on a silicon chip.
The study demonstrates that complex quantum circuits can be built relatively easily out of silicon and silica – a significant milestone on the road to full-blown quantum computing.
Fifteen years ago, Peter Shor, a computer scientist at the Massachusetts Institute of Technology, predicted that quantum computers could beat even the most powerful supercomputers and crack the widely used RSA encryption algorithm.
RSA relies on a mathematical asymmetry: it's easy to calculate the result of multiplying together two large prime numbers, but hard to work backwards from the result and find the two factors. RSA encryption uses the product of two large primes to make a public "key", safe in the knowledge that only those authorised to know the factors used to make it can decode the message.
There is a way to crack the code, however: by carrying out a relatively basic calculation again and again a computer can search for patterns and break the code. But in reality the pattern will be so large that a conventional computer would take an impossibly long time to find it.
Shor predicted that a quantum computer could do it much faster, though. Thanks to quantum entanglement, all the numbers spewed out by a quantum computer are interconnected, so just a few of them hint at the code-breaking pattern. Shor wrote an algorithm for future quantum computers that would allow them to take a shortcut to decoding confidential material.
Now a proof-of-principle version of Shor's idea has been demonstrated on a silicon chip for the first time.
The 26-millimetre-long chip was designed and built using standard fabrication processes by Jeremy O'Brien, Jonathan Matthews and Alberto Politi at the University of Bristol, UK. It can run Shor's algorithm in cut-down form – confirming that 3 and 5 multiply to form 15.
Unlike the silicon chips inside conventional computers, the Bristol team's chip uses light rather than electricity. Light-transmitting silica on a silicon wafer guides photons with entangled quantum properties around, an approach first demonstrated by the same team last year.
"We've made rapid progress, which is testimony to the strength of this approach," says O'Brien.
Dan Browne, a quantum physicist at University College London who was not involved in the study, agrees. He worked on one of the first quantum circuits to run Shor's algorithm in 2007: a table-top setup that sent photons travelling through the air rather tiny guides on a chip.
"A free-space optics experiment looks a mess, with many mirrors and lenses for an experiment with say, 4 or 5 photons," Browne says. "Imagine being the poor grad student who has to align all those mirrors."
The new chip saves time, says O'Brien. "It's almost as simple as stamping the design out onto the chip and it's there and working," he says.
Andrew White, a quantum physicist at the University of Queensland in Brisbane, Australia, is impressed with the progress towards shrinking quantum-circuit size.
"It is very important to shrink the circuitry, and the Bristol group has shown the quantum community that this can be done using well-established techniques from classical photonics," he says.
But White points out that the technology used to generate individual photons to feed into the chip, and to detect them as they emerge, is not efficient, fast or compact enough yet. Although the new chip is only 26 mm long, it has to be surrounded by a whole table top of that equipment.
Journal reference: Science, DOI: 10.1126/science.1173731