A quantum leap for industry

(From "Horizons" no. 110 September 2016)​​​

​Quantum computing is a beguiling idea. Whereas classical computers encode data in the form of digital bits, the strange laws of the microscopic world allow quantum bits (or 'qubits') to exist as a '0' and '1' at the same time and also to be ' entangled' with one another. These properties mean that quantum computers could, in principle at least, operate simultaneously on all the possible values held by a set of qubits, so making them exponentially faster than today's devices when processing certain problems.

When it comes to fundamental research on these computers – or on other quantum technology, such as cryptography and sensing – Switzerland has become a prominent country. In the latest available international citation ranking on quantum science, published by the consultancy Technopolis in 2011, Switzerland came out top along with Austria. And for the past five years, groups at universities across the country have been linked together as a National Centre of Competence in Research in Quantum Science and Technology (QSIT) run by the SNSF, joint publishers of this magazine.

Quantum people next door

The country's strength is its breadth of research, both across and within individual institutions, argues Klaus Ensslin of ETH Zurich, who heads NCCR QSIT. "If I walk ten metres from my office, I meet quantum people who work on a variety of physical systems", he says, "whereas I think elsewhere in Europe centres are more specialised".

However, Ensslin and many others in the field say that Swiss institutions are less good at converting this scientific knowledge into commercial products. Daniel Loss of the University of Basel notes that universities in a number of other countries – including the Netherlands, Denmark, Japan and Australia – receive funding specifically targeted to the construction of a quantum computer. Such funding in Switzerland, he says, is "somewhat lacking".

The country already has one clear quantum success story: ID Quantique, a company spun off from the University of Geneva, sells cryptographic equipment and single- photon detectors. The former allows confidential messages to be encrypted and decrypted using a secret 'key' comprising the quantum states of a series of photons. Quantum mechanics dictates that anyone trying to eavesdrop will automatically reveal their presence by changing the key, making such encryption in principle uncrackable. Set up in 2001, the company now sells its technology for a profit to banks, multinationals and governments around the world.

Other companies, however, have yet to follow suit. One of the founders of ID Quantique, Nicolas Gisin of the University of Geneva, points out that while quantum technology is being developed by the US companies Google, Microsoft and IBM and by the Japanese firm Toshiba, no large Swiss firm is doing likewise. He hopes this will be changed by a new EUR 1 billion ' flagship' project to develop and commercialise such technology, announced by the European Union in April. "Quantum information will revolutionise computing and communications over the next two decades", he says. "It is no longer time to wait". Loss agrees, hoping that Switzerland participates in the new flagship, or sets up its own analogous programme, or does both.

Size matters

For most of the time since the field's inception about three decades ago, research on quantum computers has been largely academic. But in recent years physicists have made major improvements to the vital error-correction schemes needed to compensate for the destruction of delicate quantum states by outside interference. Researchers are now on the verge of making logic gates that operate reliably enough so that errors don't spiral out of control as more qubits are added to a device, so opening up the prospect of scaling up today's tiny quantum computers – which consist of no more than about a dozen qubits – to ones containing hundreds, thousands or millions of qubits.

Physicists are also currently investigating a variety of different kinds of qubit. Loss carried out pioneering work on one of the leading candidates, known as spin qubits, having proposed in 1998 to encode data in the spin of electrons embedded in nanometre- sized pieces of semiconductor – systems known as 'quantum dots' (spin is a quantum-mechanical property describing the intrinsic rotation of a particle). Loss believes that these qubits are well-suited to building full-scale quantum computers because, he says, they are small and speedy, and also because they could exploit existing semiconductor manufacturing techniques.

Another solid-state technology is being investigated by Andreas Wallraff and colleagues at ETH Zurich. In this case qubits are encoded in the direction of electrical currents travelling around superconducting circuits - whether those currents travel clockwise, anti-clockwise or in both directions at the same time.

100 million qubits

However, solid-state qubits are not the only game in town. A group led by Jonathan Home, also at ETH Zurich, traps atomic ions in electric fields and then places the ions in superposition states using laser beams. This technology currently holds the record for the most reliable logic gates and the highest number of entangled qubits, and Home argues that because ions are identical to one another, it simplifies scaling and allows the use of error correction based on symmetry.

Whichever technology wins out, however, commercialising it will be a mammoth task. A quantum computer large enough to crack today's Internet security by factorising the long numbers used to encrypt communications – one of the most widely advertised and feared applications of these devices – would require more than 100 million qubits, according to estimates by John Martinis of the University of California, Santa Barbara, USA. Scaling up to that kind of level, says Wallraff, will be less a question of mastering physics and more a question of overcoming major technical hurdles, such as supplying enough laser beams or cooling the qubits. "We will have to take a more engineering-based approach to build these systems", he says.

How quickly these challenges can be overcome depends on how much industry is willing to invest, says Loss. "It is tough if you are a small team with just one or two postdocs on temporary contracts", he notes. "But if you have a large permanent staff working on these problems, it is obviously much easier to make progress".

Secret keys

Quantum cryptography by comparison has been relatively easy to commercialise because it involves sending and detecting just one photon at a time, rather than entangling multiple quantum particles. While today it is mainly used in a stand-alone capacity to connect organisations' main computer centres and back-up systems, Gisin envisages that quantum links in future could be established between Switzerland's largest cities. Internet users, he says, could then choose whether to connect via cheap but relatively insecure classical protocols or instead using quantum cryptography.

Measuring at the limit

Another technology being primed for markets is that of quantum sensing. Patrick Maletinsky and colleagues at the University of Basel place individual electron spins (created by adding nitrogen atoms to diamond) on the tip of an atomic force microscope (AFM) in order to detect any weak magnetic fields close to the tip. The rate of spin precession is proportional to the strength of that field, allowing very sensitive quantitative imaging on scales of nanometres.

Maletinsky says the technique could be used to map tiny spatial variations in the stray fields of thin magnetic films, which are important for data storage. Alternatively, he says, it could be employed to look at vortices in superconductors, relevant for technological applications such as MRI machines. In the life sciences, meanwhile, the technique could potentially determine the structure of individual protein molecules (which contain nuclear spins that create very small magnetic fields). Maletinsky says his group should have set up a company to commercialise the AFM diamond tips by the end of this year, then have its sensor on the market "within the next year or two".

The Holy Grail

In fact, even quantum computers have entered the market place. In 2007, the Canadian company D-Wave unveiled a 'quantum annealer' that can run optimisation programs and which, in its latest incarnation, boasts 1,000 superconducting qubits. The firm has leased machines for a million dollars apiece to NASA, Google and defence giant Lockheed Martin, but its technology has been and remains controversial. Many have doubted the extent to which it really uses quantum mechanics, while Matthias Troyer of ETH Zurich and other researchers in 2014 showed that it could operate no quicker than classical computers.

The first bona fide quantum computer that can carry out useful operations impossible to perform on a classical device is expected in about ten years, says Wallraff. That computer, he says, would contain perhaps a few hundred qubits and would likely be used to carry out simulations of small molecules and other quantum systems.

But Loss says that the 'holy grail' of quantum information science remains the construction of a fully-fledged 'universal' quantum computer capable of advanced operations like factorising large numbers. Now that industry is interested, he argues, that aim can at last be realised. He just hopes that Swiss and other European companies will join their American counterparts in the chase..

Based in Rome, Edwin Cartlidge writes for Science and Nature.