New records for accuracy of silicon quantum bits
UNSW teams jump major hurdle on road to building powerful quantum computer
Two UNSW research teams have found distinct solutions to a critical challenge that has held back the realisation of super powerful quantum computers.
The teams at UNSW Australia have developed two different types of quantum bit, or “qubit” – the building block for quantum computers – that each process quantum data with an accuracy above 99 percent. The two findings were published simultaneously on October 12, 2014 in the journal Nature Nanotechnology.
"For quantum computing to become a reality we need to operate the bits with very low error rates," says Scientia Professor Andrew Dzurak, who is Director of the Australian National Fabrication Facility at UNSW, where the devices were made.
Dzurak explains that, “Even though methods to correct errors do exist, their effectiveness is only guaranteed if the errors occur less than 1% of the time. Our experiments are among the first in solid-state, and the first-ever in silicon, to fulfill this requirement."
"We've now come up with two parallel pathways for building a quantum computer in silicon, each of which shows this super accuracy,” adds Associate Professor Andrea Morello from the School of Electrical Engineering and Telecommunications at UNSW Australia.
Computing with single atoms and artificial atoms
The UNSW teams were already the first in the world to demonstrate single-atom spin qubits in silicon, reported in Nature Nanotechnology in 2012 and 2013.
Now the team led by Dzurak has discovered a way to create an “artificial atom” qubit with a device remarkably similar to the silicon transistors used in consumer electronics, known as MOSFETs.
Post-doctoral researcher Menno Veldhorst, lead author on the paper reporting the artificial atom qubit, says, “It is really amazing that we can make such an accurate qubit using pretty much the same devices as we have in our laptops and phones”.
Meanwhile, Morello’s team has been pushing the “natural” phosphorus atom qubit to the extremes of performance.
Dr Juha Muhonen, a post-doctoral researcher and lead author on the natural atom qubit paper, notes: “The phosphorus atom contains in fact two qubits: the electron, and the nucleus. With the nucleus in particular, we have achieved accuracy close to 99.99%. That means only one error for every 10,000 quantum operations.”
The high-accuracy operations for both natural and artificial atom qubits is achieved by placing each inside a thin layer of specially purified silicon, containing only the silicon-28 isotope. This isotope is perfectly non-magnetic and, unlike those in naturally occurring silicon, does not disturb the quantum bit. The purified silicon was provided through collaboration with Professor Kohei Itoh from Keio University in Japan.
Record storage time for quantum information
Morello’s research team also established a world-record “coherence time” for a single quantum bit held in solid state. "Coherence time is a measure of how long you can preserve quantum information before it’s lost," Morello says. The longer the coherence time, the easier it becomes to perform long sequences of operations, and thus more complex calculations.
The team was able to store quantum information in a phosphorus nucleus for more than 30 seconds. "Half a minute is an eternity in the quantum world. Preserving a ‘quantum superposition’ for such a long time, and inside what is basically a modified version of a normal transistor, is something that almost nobody believed possible until today," Morello says.
The benefits of quantum computing
A functional quantum computer will provide much faster computation in three key areas: searching large databases, solving complicated sets of equations, and modelling atomic systems such as biological molecules and drugs. This means they’ll be enormously useful for the 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 probably too premature to even conceive.
In current computing, information is represented by classical bits, which are always either a zero or a one. Physically, this corresponds to a transistor device being switched on or off. For quantum computing you need an equivalent: and in the UNSW design the data will be encoded on the spin – or magnetic orientation – of individual electrons, confined within a nanometer-scale electronic device. These are known as quantum bits, or qubits.
An electron with spin “up” would represent a 1 and a counter-clockwise (or downward) spin would represent a 0 – 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, which gives rise to the unique ability envisioned for quantum computers to rapidly solve complex, data-intensive problems.
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 microprocessors, or computer chips. Silicon offers several advantages: the material is cheap, 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. Silicon electron “spins” also have very long “coherence times” – this means data encoded on the spin can remain there for longer periods than it would in most materials, before it is scrambled and lost. This is important for performing successful calculations.
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 Nanotechnology 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. This latest results show for the first time that qubits made out of spins in silicon can be operated with the high accuracy necessary to guarantee that the computation proceeds without errors.
A functional quantum bit – or qubit
In order to employ the electron spin, a quantum computer needs both a way of changing the spin state (writing information) and of measuring that change (reading information). These two capabilities together form a quantum bit or qubit – the equivalent of the bit in a conventional computer.
In 2012, the UNSW researchers demonstrated both stages for an electron bound to a phosphorus atom in silicon, a result they published in Nature Nanotechnology that year. Important as this result was, the accuracy of the qubit was only around 50%, well below that required for a functioning quantum computer.
Now Morello’s group, in their latest Nature Nanotechnology paper (by Muhonen et al.), have improved the electron qubit accuracy (or “fidelity”) to a remarkable 99.6% by configuring the phosphorus atom in specially purified silicon, containing only the silicon-28 isotope.
Dzurak’s group, in their Nature Nanotechnology paper (by Veldhorst et al.), have also demonstrated an electron spin qubit with 99.6% fidelity, but in their case they have configured the electron in an “artificial atom”, using a silicon transistor remarkably similar to those used in everyday laptops and smart phones.
The researchers will now work to combine pairs of these devices to create a two-bit logic gate – the basic processing unit of a quantum computer. While building a full-scale quantum computer remains a daunting and time-consuming engineering challenge, the problem of realising a highly accurate quantum bit in silicon has now been solved.