Quantum information: Computing with a single nuclear spin in silicon

A research team led by Australian engineers has created the first working quantum bit based on the nuclear spin of a single phosphorus atom in silicon, opening the door for dramatically improved data processing in ultra-powerful quantum computers of the future.

A landmark paper published today in the journal Nature, describes how to write and read quantum information with record-setting accuracy using the nuclear spin, or magnetic orientation, of a phosphorus atom in a silicon transistor – similar to silicon chips used in modern electronics. 

The nucleus of a phosphorus atom is a very, very weak magnet, and can be imagined as a compass needle that can point north or south. These north or south positions are equivalent to the zero and one of binary code, which governs classical computing. In this experiment, the researchers controlled the direction of the nucleus, in effect “writing” an arbitrary value onto its spin, and were then able to “read” the value out. They observed quantum oscillations of the spin between north and south, and all the quantum superpositions of those two directions – where the spin exists in both states simultaneously.

“We achieved a read-out fidelity of 99.8 per cent, which sets a new benchmark for qubit accuracy in solid-state devices,” says Scientia Professor Andrew Dzurak, from the UNSW School of Electrical Engineering & Telecommunications (EE&T).

The primary author on the paper is Jarryd Pla, a PhD student in the Quantum Spin Control group led by Associate Professor Andrea Morello, also at UNSW. Morello and Dzurak are program managers in the ARC Centre of Excellence for Quantum Computation and Communication Technology (CQC2T). The collaboration also includes researchers from the University of Melbourne, University College London and University of Twente in the Netherlands.

Nuclear qubit rivals ion trap for top qubit

The accuracy of this nuclear spin qubit means it’s near the level of what many people consider to be the most “perfect” quantum bit yet realised – a single atom isolated in an electromagnetic trap inside a vacuum chamber. The pioneers of this “Ion Trap” technology were awarded the 2012 Nobel Prize in physics.

“Our nuclear spin qubit is essentially at that level, but it’s not held in a vacuum chamber – it’s in a silicon chip that can be wired up and operated electrically like normal integrated circuits,” says Associate Professor Andrea Morello from EE&T.   

This is a huge advantage, says Morello. Silicon is a the dominant material in the microelectronics industry, and by working with this material, the UNSW-developed technology will be easier to scale up and incorporate within existing industry standards. 

The Nuclear Qubit vs. Electron Qubit

In September 2012, the same UNSW team reported in Nature the first functional quantum bit based on an electron bound to a phosphorus atom embedded in silicon, “writing” information onto its spin and then “reading” the spin state back out.

Electron spin qubits will likely act as the main “processor” bit for quantum computers of the future. These are the key focus moving forward, and will be coupled with other electron qubits to perform operations.

But the nucleus also provides intriguing possibilities. The nucleus is the core of an atom, containing most of its mass, and is roughly one million times smaller than the overall size of the atom, determined by the orbits of the electrons . It is also 2000 times less magnetic than the electron. This has two consequences: it’s very challenging to measure, but it is also nearly immune to magnetic noise or electrical interference from the outside world. As a result, the nuclear spin has an excellent coherence time – this is what determines the time during which delicate quantum operations can be performed with minimal errors.

Pioneering studies on nitrogen-vacancies in diamond have previously demonstrated the potential of nuclear spins as quantum bits, but this is the first time that one has been successfully realised in silicon.

“The key benefit of using the nuclear spin as a quantum bit is that information stored on it can last for a long time in comparison with the time required to do calculations, meaning that very few errors occur during computation,” says Dzurak. 

The nuclear spin qubit could also be integrated with electron qubits, serving a vital memory function, or assist the quantum logic operations between pairs of electrons.

The benefits of quantum computing

A functional quantum computer will provide much faster computation in three key areas: searching large databases, cracking most forms of modern encryption, and modelling atomic systems such as biological molecules and drugs. This means they’ll be enormously useful for finance and healthcare industries, and for government, security and defence organisations. Functional quantum computers will also open the door for new types of computational applications and solutions that are, at this stage, difficult to predict.

How quantum computers work

In current computing, information is represented by classical bits. These are always either a zero or a one – physically represented by a transistor device being switched on or off. For quantum computing you need a physical system that has two distinguishable quantum “states”.  In the UNSW design, the quantum data is encoded on the spin orientation of individual electrons, bound to single phosphorus atoms. With this new result, the team has shown that quantum data can be encoded on a nuclear spin as well, thus obtaining two fully functional quantum bits out of a single atom.

A  spin pointing “north” would represent a one and a spin pointing “south” would represent a zero – but in the quantum realm, particles have a unique ability to exist in two different states at the same time, an effect known as quantum superposition. This is one of the properties that gives rise to the unique ability envisioned for quantum computers to rapidly solve complex, data-intensive problems.

Multiple, coupled qubits can exist in states that have no classical analog, and they can be in many of such states at the same time. These special states are called “entangled states” because the information they contain tells you something about the correlations between the particles, but not the individual state of each particle. “The entangled states represent extra ‘codes’ in a quantum computer,  which classical computers do not possess. Their number grows exponentially with the number of qubits. As a consequence, with just 300 qubits it is possible to store as much information as there are atoms in the universe,” says Morello.

The silicon approach: UNSW leading the way

In recent years, scientists around the world have been developing completely new systems based on exotic materials or light to build a quantum computer. At UNSW, however, the approach has been to use silicon – the material currently used in all modern-day computer chips. Silicon offers several advantages: the material is cost-effective, already used in almost all commercial electronics, and its properties are very well understood – the result of trillions of dollars of investment into R&D by the computer and electronics industry. In addition to this, it turns out that silicon is also an outstanding solid-state environment to host quantum bits based on magnetic systems, such as electron and nuclear spins. This is because silicon can be purified to contain strictly no magnetic nuclei other than the ones deliberately introduced to carry quantum information. This highly purified silicon is now often called a “semiconductor vacuum”, since the spins hosted in there are as isolated from their environment as they would be if they were held in a true vacuum..

In 1998, former UNSW researcher Bruce Kane first proposed the idea of using silicon as a base material for quantum computing. In a paper in Nature he outlined the concept for a silicon-based quantum computer, in which single phosphorus atoms in an otherwise ultra-pure silicon chip define the qubits.

His visionary work spawned an international effort to develop a quantum computer in silicon, and this latest result represents the biggest achievement en route to realising that dream – a result, researchers say, that could perhaps one day be seen as comparable to the invention of the transistors used in conventional computers.

A functional quantum bit – or qubit

In order to employ an electron or nuclear spin qubit, a quantum computer needs both a way of setting the spin state (writing information) and of measuring the result (reading information). These two capabilities together form a quantum bit or qubit – the equivalent of the bit in a conventional computer.

In 2012, the research team, led by engineers from UNSW, completed both stages for a single electron bound to a phosphorus atom in silicon. In their latest result, published in Nature, they have achieved the same result – writing and reading information – but now using the spin of the atom’s nucleus, which is much weaker than that of the electron.

To write information on the nuclear spin the team used a technique known as “nuclear magnetic resonance”, which is the same phenomenon used in magnetic resonance imaging for brain scans – but in this case they were controlling and measuring the nuclear spin of just one atom, rather than many trillions of atoms.

The magnetic field generated by the phosphorus nuclear spin is one thousand times smaller than that of the electron spin, so this new type of quantum bit is, in principle, much harder to measure than in their previous work. However, the team used a new type of readout process, which involved using the electron as an intermediary to measure the nuclear spin, leading to a one billion-fold amplification. This allowed the UNSW team to read out the nuclear spin in real time with extremely high accuracy.

Further information

Timeline of the development of silicon quantum computing in Australia

1994: Peter Shor from Bell Labs (USA) shows that a quantum computer would be able to decrypt Public Key Encrypted codes (at the heart of modern secure communications) exponentially faster than today’s supercomputers. This triggers massive interest in quantum computing worldwide.

1994-1998: Various schemes proposed for making a quantum computer using different systems including photons, ion traps, superconductors, semiconductor quantum dots.

1998: Dr Bruce Kane, then a postdoctoral researcher at UNSW (now a researcher at the US Laboratory for Physical Sciences in Maryland), publishes a paper in Nature [Nature 393: p 133 (1998)] outlining the concept for a silicon-based quantum computer, in which the qubits are defined by single phosphorus atoms in an otherwise ultra-pure silicon chip.

The quantum information is encoded in either the spin of the electron or the spin of the P nucleus. This is the first such scheme in silicon – the material used for all modern day microprocessors.

Kane’s paper attracts great interest because: (i) Silicon is “industrially relevant”; (ii) Silicon electron “spins” have very long “coherence times” (hence, low error rates). This paper has now generated over 2000 citations.

Despite the great potential for a silicon quantum computer, the task of building devices at the single atom level is considered almost science fiction back in 1998.

2000: Prof Bob Clark establishes the ARC Special Research Centre for Quantum Computer Technology, headquartered at UNSW, to attempt to build a quantum computer.

The Centre has now expanded to become an ARC Centre of Excellence – with more than 150 researchers in Australia, and major collaborations world-wide.

The Centre now generates about 100 research publications a year.

Bob Clark retired from CQCT in 2008 to take up the role of Australian Chief Defence Scientist from 2008-2011. He now researches energy policy at UNSW.

Prof Andrew Dzurak (UNSW) and Prof David Jamieson (University of Melbourne) were founding chief investigators in CQCT and were charged with using ion implantation to build a silicon quantum computer.

Ion implantation is a technology for injecting “dopant” atoms (such as phosphorus) into modern silicon integrated circuits. As a “mass production” technology ion implantation is very attractive if we are ever to build a commercially viable large-scale silicon quantum computer processor chip.

2000-2009: Various researchers at CQCT contribute to ground-breaking single atom nanotechnologies – generating hundreds of papers.

Two main approaches:

  • “Top-down” approach - using ion-implantation [led by Dzurak and Jamieson]
  • “Bottom-up” approach – using scanning-probe lithography [led by Michelle Simmons

Despite the amazing progress in single atom nanotechnologies over the past decade, until now there has been no demonstration of a single electron spin quantum bit based on phosphorus donors, as first envisaged by Kane back in 1998.

Timeline for the ion-implanted (“top-down”) single-atom quantum bit

2006: Associate Professor Andrea Morello joins the team at UNSW, in the quest to build silicon qubits using the “top-down” approach. Morello brings a background in quantum spin physics, which is essential to the team.

Morello and Dzurak form a close partnership – aimed at realising the dream of a silicon quantum bit.

2008: The “Top-Down” team at UNSW conceives a new scheme to read out the spin of an electron on an implanted phosphorus donor.

Paper published in Physical Review B in 2009 [A. Morello et al., Phys. Rev. B 80, 081307 (2009)].

2009: Morello and Dzurak, and their experimental teams, see the first experimental evidence of the read out of the electron spin on a phosphorus atom in a qubit device.

2010: Paper published in Nature, describing the “Single shot readout of an electron spin in silicon” – the step of measuring a silicon qubit. [A. Morello et al., Nature 467: p687 (2010)]

2012: Paper published in Nature, describing “A single-atom electron spin qubit in silicon” – the final crucial step of writing information on an electron, to fully operate a silicon qubit. [J. Pla et al., Nature 489: p541 (2012)]

2013: Paper published in Nature, describing "High-fidelity readout and control of a nuclear spin qubit in silicon" -- or a functional qubit based on the nuclear spin of a single atom in silicon [J. Pla et al., Nature (2013) in press]

Key Research Team Members

Associate Professor Andrea Morello (UNSW) – Project Co-leader (with Dzurak). Joint conceptual design of experiment. Leader of the Quantum Spin Control experimental program at CQC2T and responsible for the quantum measurement infrastructure.
Prof Andrew Dzurak (UNSW) – Project Co-leader (with Morello). Joint conceptual design of experiment. Director of Australian National Fabrication Facility at UNSW and co-leader of “top-down” single-atom device engineering (with Jamieson) at CQC2T.
Mr Jarryd Pla (UNSW) –PhD student at the UNSW School of Electrical Engineering & Telecommunications in Dr Morello’s group, and lead-author on the Nature paper. Lead experimentalist.
Prof David Jamieson (University of Melbourne) – Co-leader of “top-down” single-atom device engineering (with Dzurak) at CQC2T.
Other Authors
Dr Kuan Yen Tan, UNSW (now at Aalto University, Finland)
Mr Juan-Pablo Dehollain, UNSW
Dr Wee Han Lim, UNSW
Dr John Morton, University of Oxford (now at University College, London)
Mr Floris Zwanenburg, UNSW

Key Stakeholders and Funding Bodies

1. Centre of Excellence for Quantum Computation and Computer Technology (CQC2T): Australian centre of research excellence, headquartered at UNSW, in which Dr Morello and Prof Dzurak are project leaders. Founded in January 2000.

2. Australian National Fabrication Facility (ANFF): Founded in 2006 under the Australian Government’s National Collaborative Research Infrastructure Scheme. Provides infrastructure and technical support at UNSW for fabrication of the qubit devices.

3. Australian Research Council (ARC): Major funder of CQC2T via the ARC Centres of Excellence Scheme (a funder since 2000).

4. US Army Research Office: Funder of the Silicon Quantum Computer Program at UNSW and the University of Melbourne since 1999.

5. Australian Government Department of Innovation, Industry, Science, Research and Tertiary Education (DIISRTE): Major funder of ANFF through the National Collaborative Research Infrastructure Scheme and Super Science – Future Industries programs.

6. NSW Government – Department of Trade & Investment, Office of Scientific Research: Provides significant co-funding to CQC2T (since 2003) and also to ANFF (since 2006).