How to build a quantum computer

Andrea Morello is Scientia Professor of Quantum Engineering at UNSW and a Program Manager in the ARC Centre of Excellence for Quantum Computation and Communication Technology and a 2024 Australian Research Council (ARC) Laureate Fellow. He believes quantum computing is at a fork in the road.

“To build a quantum computer, you still need to encode information in a binary form – a zero and a one, just the way you do with a standard, classical computer. Your mobile phone, for example, has a chip that contains billions of tiny transistors in silicon – a transistor is like an electronic switch, either open or closed. By applying a voltage to the switch, you make your zero or one.

For a quantum computer, you need to first decide what physical object you code with the zeros and ones – but in this case, you want them to be quantum systems, where the zero and one are represented by the quantum states of the object. The most natural places to find such quantum states is in the realm of atoms, at the scale of nanometres.

There are many ways to do it. You can do it with single electrons, for example. You can make a very small transistor which holds just one electron, and have it held in two possible places – on the left to encode a zero, on the right to encode a one.

Another way is to use the spin of an electron or a nucleus. What’s a spin?

Just as every elementary particle has an electrical charge (electrons negative, protons positive), every electron, every proton, and even every neutron, is slightly magnetic. It creates a magnetic dipole, a magnetic field, like the needle of a compass at an atomic scale. Your zero and one can be the spin, depending on whether the magnetic dipole points north or south.

Now you may wonder, what’s the point of going through the effort of making zeros and ones out of subatomic particles when we have perfectly fine objects – transistors – that do the same job cheaply? The point is that quantum bits are not confined to being zero or one: they can be both at the same time. And if I take two quantum bits, I can put them in a state where they are neither zero nor one, but they are the opposite of each other. Their only existential feature is their mutual correlation. This is called an entangled state, and is like having an extra digital code available to program the computer. The number of entangled states I can create grows exponentially with the number of qubits available. This can make the quantum computer solve certain problems exponentially faster than a classical one.

Now you’ve decided to build a quantum computer where information is encoded in the spin of some elementary particle, how do you manipulate that? How do you hold it in place? How do you read and write quantum information on it?

Again, there’s many ways to do it, but the way that was decided upon here in Australia was to do it in silicon, using the same kind of nano-electronic devices that are used in classical computers.

Why do it this way? This is because silicon microelectronics is the most incredible thing humankind has ever developed. The amount of work and ingenuity that has gone into it is more impressive than going to the Moon. Everything pales compared to the science and technology that has gone into the silicon chip that you probably hold in your hand right now.

Over the last 60 years, a trillion-dollar industry has been developed that can build a chip that contains several billion devices, each one of which is about 100 atoms across, all working perfectly and reliably. And they sell them to you for a few dollars.

The next quantum computer revolution

Wouldn’t it be great if we could recycle that amazing technology to make quantum computers as well?

That was the idea behind the work of Bob Clark and Andrew Dzurak back in the 1990s, here in Australia. These were people with extensive expertise in semiconductor microelectronics. They had ideas on how to capture a single electron, and how to hold the quantum information inside the silicon chip, but it hadn’t worked yet.

This is where my expertise in spin physics and microwave engineering came in. You need to combine sending radio frequency signals to control the spins with the ultra-low temperature physics necessary to detect them.

I came to Australia in 2006, and my group, together with Andrew Dzurak, helped make it work for the first time. We did it by implanting an atom of phosphorus inside the silicon chip, in collaboration with the team of David Jamieson. This was not a crazy idea – phosphorus atoms are used all the time to change the physical behaviour of silicon. In a classical transistor, there’s many phosphorous atoms, implanted at high concentration in a little blot. The goal here was to take only one of them. Each atom hosts a nuclear spin, which is a quantum bit, and an electron spin, which is another quantum bit. So one atom hosts two quantum bits. It’s very convenient, high-density quantum information.

Our breakthrough paper, published in Nature in 2010, was the first time that quantum computer technology was demonstrated on the silicon model. We had one atom of phosphorus in the chip, and we were able to watch, in real time, one by one, the direction of the spin – the north or south of a single electron bound to that atom.

That opened the floodgates.

The efficient quantum computer

Since then we have been busy coming up with new ideas how to make these operations very efficient, but now we are at a fork in the road. This is because building an actual quantum computer that can solve useful problems is beyond the realm of academic research – it will cost billions of dollars to develop.

I am a professor at the University. I teach, I supervise students, I do basic research. My job is to come up with good ideas and demonstrate that they work, and if they are useful they get picked up by industry and become the basis for their engineering and manufacturing development.

I do my work in world-class research facilities, which look almost artisan compared to the $10-billion silicon foundries where commercial chips are manufactured. The important point is that I do basic research – curiosity driven research – but I do it on a physical platform that is compatible with the most sophisticated industrial manufacturing, so that the transition to useful products can be as quick and as painless as possible.”

As told to Graem Sims

Also in this series 2024 ARC Laureate Fellows

Making mathematics count Professor Yihong Du

Keeping watch on natural disasters from near space Professor Jeffrey Walker

Can we predict how pests respond to climate change: Professor Michael Kearney