Constructing quantum computers Dr Euan Allen is the 2020 Rosalind Franklin Award Lecture winner for physical sciences and mathematics. This Award is in recognition of their cutting-edge work and committed public engagement efforts. Full details of the 2020 Award Lecture winners can be found here: www.britishscienceassociation.org/news/introducing-our-2020-award-lecturers Written by Alan Barker, freelance writer Working at the Quantum Engineering Technology Labs at the University of Bristol, Euan works in silicon photonics – and is investigating how to apply it to the construction of quantum computers. This emergent technology could offer unprecedented computing power. But how easy is it to develop, and who will reap the benefits? First things first. What is a quantum computer? To build a classical computer, you need bits and gates. A bit is normally current in a wire, which can either be on or off: 0 or 1. And gates are the operations that you can perform on your bit: you could think of gates as systems of switches. To build a quantum computer, you need qubits and gates. A qubit is like a bit – you could write it as a 0 or a 1 – but it can also inhabit other, intermediate or exotic states – we call them superposition states. So the obvious next question: what are superposition states and how do you create them? At Bristol, we’re working with photons. We do something called dual-rail encoding, where the photon can be in either the top or bottom paths – or rails – of our device, denoting if it’s a 1 or a 0. We start with one photon in the top rail, and we fire it through two beam splitters in sequence. Beam splitters are half-mirrored devices that split light: each beam can either stay on the top rail or move to the bottom rail. So you’re randomly deflecting the photon, and you’re doing that twice. You’d expect a stream of photons to divide half and half each time, and you’d end up with 50% on one path and 50% on the other. But that’s not what happens. As it passes through the two beam splitters, each photon behaves both as a particle and as a wave; and, as a wave, it ends up cancelling itself out in the bottom rail. So the photon always comes out on in the top path. It’s like flipping a coin twice and always getting heads on the second flip. That wave-particle duality is a quantum effect – it’s the thing that most people might know about quantum mechanics. It’s odd, but it happens and we can – and do – measure it. What about gates? We use different kinds of gates to play with this effect. A Hadamard gate, for instance, generates the superposition state and is our beam splitter; phase gates – we call them phase shifters – delay the path of the photon through the system and can give you a range of different outputs. If you build a system using these gates, you get a much wider range of possible outputs from your inputs than you would in a classical system. And that’s part of the way to giving you a great deal more computing power. Are there other ways of creating a quantum computer? There are three other main lines of research. You can use superconducting rings or wires, which need to be at very low temperatures. Then, you can trap ions in an electric field and fire lasers at them, which changes their quantum state. And the third way is to dope materials with other materials to create solid state spin systems using electrons. So why have you chosen photonics? Because we think we can make it work at scale. To build a useful quantum computer, you’d need millions of beam splitters and phase shifters. Photonic components historically are large. We have to shrink them to something like the width of a human hair, and fit hundreds of thousands of components onto a silicon chip that’s smaller than a fingernail. It’s the equivalent, really, of replacing valves with transistors in classical computers. We can do that: Bristol was the first lab to demonstrate that it’s possible. And we’re still world-leading in that field. We’re going with a technology that’s already scalable and making it quantum – whereas others are perhaps taking quantum tech and trying to make it scalable. What challenges do you face? One of the main challenges – and it’s the challenge all the researchers face in this field – is getting usable qubits. Physical qubits are noisy. To get a logical qubit – a qubit we can use – we have to collect together maybe 1000 physical qubits, which we then have to error-correct. A useful algorithm would need, maybe, 200 logical qubits, which means you’d need at least 200,000 qubits and the ability to error-correct them. Lots of research in Bristol involves working towards these useable, logical qubits, and then manipulating them so that we can compute with them. What are the main applications for quantum computers? The applications are probably going to be very specialised to begin with: simulations of quantum systems like chemical interactions or the ways electrons move within materials. Then, in a second phase, we envisage using these computers to deal with problems that require very large amounts of computing power. Things like large fluid simulations of air around a plane wing or finding the fastest way to deliver lots of parcels. I don’t think you should expect to see a quantum computer sitting on your desk any time soon. It’s much more likely that a quantum computer will be a specialised unit inside a classical computer – a bit like a graphics card, which does just a few things very well. Does this new technology bring any dangers with it? In the early stages, particularly, these computers are going to be big, complicated and expensive. They’re going to be based in institutions or large companies. We’re seeing the likes of Google buying up people, expertise and resources, taking them out of academia into the private sector. We could see enormous amounts of computing power in private hands. I think we need to be careful about this power becoming too centralised, and make sure that everyone gets the benefits of this technology. Things seems to be going well so far, and the interaction of industry with academia has actually been quite fruitful. Alan Barker is a writer, trainer and coach specialising in communication skills. He has been working with the British Science Association since 2015. Alan’s webinar, Storytelling for Scientists, is on the 3M YouTube channel.