Electrical engineers at the University of New South Wales trying to develop a silicon quantum computer have cleared one of the last hurdles to building a simple device. The researchers have reported this missing piece in the journal Science Advances.
“Once you have demonstrated all the parts, then it's like a Lego box – you can start building up a large architecture by piecing its components together,” project leader Andrea Morello says.
In their quest to build a silicon quantum computer, Morello and his colleagues have so far been perfecting its basic element, the “quantum bit”. This is a single phosphorus atom entombed in a silicon crystal. Using a carefully tuned magnetic field, the researchers can manipulate the atom’s quantum “spin”, flipping it up or down.
A quantum computer is “not just a 'faster' computer,” Morello says. “They are the equivalent of a jet plane to a bicycle.”
That phosphorus atom is equivalent to a transistor in an ordinary computer. A transistor is on or off, which is how it represents the 1s and 0s of the binary code the computer uses to process instructions. A quantum bit is more complex. It can be spin-up, spin-down or in a “superposition” of both: 1 and 0 at the same time. Theoretically, this should enable a quantum computer to weigh multiple solutions to a complex problem at once, and solve it at phenomenal speed.
A quantum computer is “not just a 'faster' computer,” Morello says. “They are the equivalent of a jet plane to a bicycle.”
Last year the UNSW team showed they can write, read and store the spin of a single quantum bit with better than 99.99% accuracy using a magnetic field. But to carry out complex calculations, a quantum computer needs thousands, or even millions of quantum bits, that can all be individually controlled. And for that, the high frequency oscillating magnetic fields Morello has been using to master the control of a single quantum bit are not suitable.
For a start, the magnetic field generators Morello and his team used are around $100,000 a pop. If they had to use one for each quantum bit in a large array, the cost would be astronomical. There is also a practical problem. Magnetic fields spread, making it impossible to control one quantum bit in an array without inadvertently affecting all its neighbours.
In their latest work, carried out by experimental physicist Arne Laucht, Morello and his team found a way to control each quantum bit using a simple electrical pulse. Instead of each phosphorus atom having a dedicated magnetic field generator to control it, their new design floods the whole device with a single magnetic field.
This field is broadcast at a frequency the phosphorus atoms are not tuned in to, and so they don't feel its magnetic tug. But when a precise electrical pulse is applied to the quantum bit, the electron orbiting the phosphorus atom feels a strong force, stretching its orbit. This distortion to the electron's orbit works like twisting a tuning knob on a radio – the phosphorus atom is tuned in to the frequency of the magnetic field being broadcast around it, which then causes the quantum bit to flip.
By timing their electrical pulses, the team can tune the phosphorus atom in and out of the oscillating magnetic field, and so flip the phosphorus atom’s spin into any position they want – up, down or an intermediate superposition – without affecting its neighbours.
This idea of combining electric and magnetic fields to control individual quantum bits in an array, called “A-gate” control, has been around since 1998. Bruce Kane, an American quantum physicist who was then working at UNSW, proposed it in a paper in Nature that Morello calls “visionary”. Now, 17 years later, technology has caught up with Kane’s ideas as we can now routinely make structures at the scale needed to build his design.
Kane – now at the University of Maryland and not directly involved in Morello’s research – says he’s been impressed by the “outstanding” work on the design done at UNSW in recent years. The devices work even better than he anticipated. Back in 1998, Kane worried that imperfections in the materials would prevent the device from working as it should. But, he says, the recent work at UNSW, such as the demonstration of an A-gate, proves material imperfections “will not be a show-stopper for silicon quantum computing”.
Kane cautions that we are still a long way from large-scale quantum computing in silicon, as the challenges that remain, such as moving quantum information around and controlling interactions between large numbers of spins, are daunting. “I continue to believe that large-scale silicon quantum computing will become a reality, but there is still a long, steep road ahead of us,” he says.
The group is already at work on these challenges. Morello is confident they will have all the elements in place to build a small-scale test-system within 10 years.
And as for a large-scale quantum computer capable of making useful calculations? Here, Morello is more coy: “To quote Niels Bohr, ‘It's hard to make predictions, especially about the future’.”
More on this topic from Cosmos: The quantum spinmeister
Can physics protect us from Big Brother's snooping?
Quantum computing? Yes, no and maybe
