Everyone agrees that the first working quantum computer will represent a quantum leap into a higher-tech future, and the race is on to build it. Now, researchers at an Australian company, Silicon Quantum Computing (SQC), believe they have taken us a step closer.
Just like regular (classical) computers, quantum computers use transistors to encode information. But, unlike classical computers, a quantum computer’s transistor is on the quantum scale – down to the size of a single atom. While classical computers use bits – zeros and ones – quantum transistors encode quantum information using zeroes, ones, or a mix of zero and one.
Engineers can make use of the quantum effects of a single-atom transistor to perform calculations. But in the quantum world, things get strange.
Particles are said to exist in a “superposition” of states – their position, momentum and other physical properties aren’t defined by single values but are expressed as probabilities. With superpositions, a quantum bit (or “qubit”) can store multidimensional computing data of much greater complexity than an ordinary bit.
This is why quantum computers are expected to be thousands, even millions of times faster than classical computers, and perform calculations far more efficiently than even the most powerful classical computers.
They also have other magic tricks up their sleeves.
When superpositions are extended over multiple systems – or atoms – you can have an “entangled state” where one qubit is correlated to another. Changes to one qubit can affect the other when they are entangled. This has the potential for unhackable encryption technologies.
Which all sounds very cool, but the same quantum effects which make quantum computers such a lip-licking prospect for physicists and computer scientists make them extremely difficult to produce and turn into useful machines.
More on quantum computing: First demonstration of universal operations on an error-free quantum computer
Above all, the probabilistic nature of quantum systems means they are highly susceptible to errors. So, a major challenge when creating quantum machines is to make them “coherent” to reduce the noise in the signals.
It is this problem that the SQC team believe they have cracked.
Lead researcher and senior author of a Nature paper published today, Michelle Simmons, spoke with Cosmos about the team’s research.
Building on classical computer architecture, SQC is at the forefront of quantum computing using good old-fashioned silicon. Simmons, professor of physics at the University of New South Wales (UNSW), says this allows her team to “map out” their place in the context of the history of computing. The first transistor was invented in 1947, followed by the integrated circuit chip in 1958. In the 1960s, calculators were built on the silicon technology, before engineers created the first industrial computers.
“We want to now create a quantum computer,” says Simmons. “The key difference is we have to make things at the atomic scale. It’s got to be much smaller so we can access the quantum states and have them coherent and fast.”
Simmons’ team built the world’s first single-atom transistor in 2012, and the first integrated circuit built at the atomic scale ahead of schedule in 2021. “What we’re looking at is the next device up – some kind of commercially relevant algorithm to be solved before we make a computer that people can use. When we started out, we didn’t know what we would demonstrate in that circuit.”
The team chose to tackle polyacetylene – a carbon-based molecular chain with chemical formula (C2H2)n where the n represents a repeating pattern of two hydrogen and two carbon atoms.
Atoms in polyacetylene are bound by covalent bonds – strong molecular bonds where atoms share outer-shell electrons. A single bond means one outer-shell electron is shared between the two bonded atoms. A double bond indicates two shared electrons. The alternating single- and double-bonds between the carbon atoms in the polyacetylene chain make the molecule an interesting study in physical chemistry.
The Su-Schrieffer-Heeger (SSH) model is a well-known theoretical representation of the molecule that takes the interactions between the atoms and their electrons and explains the physical and chemical properties of the compound. “It’s a well-known problem that you can solve with a classical computer,” says Simmons. “It’s few enough atoms that a classical computer can look at all the interactions. But now we’re doing it in a quantum system.”
How did the SQC team model polyacetylene on their quantum device?
“What we’re doing is making the actual processor itself mimic the single carbon-carbon bonds and the double carbon-carbon bonds,” Simmons explains. “We literally engineered, with sub-nanometre precision, to try and mimic those bonds inside the silicon system. So that’s why it’s called a quantum analog simulator.”
Using the atomic transistors in their machine, the researchers simulated the covalent bonds in polyacetylene.
According to the SSH theory, there are two different scenarios in polyacetylene, called “topological states” – “topological” because of their different geometries.
In one state, you can cut the chain at the single carbon-carbon bonds, so you have double bonds at the ends of the chain. In the other, you cut the double bonds, leaving single carbon-carbon bonds at the ends of the chain and isolating the two atoms on either end due to the longer distance in the single bonds. The two topological states show completely different behaviour when an electrical current is passed through the molecular chain.
That’s the theory. “When we make the device,” Simmons says, “we see exactly that behaviour. So that’s super exciting.”
Dr Charles Hill, senior lecturer in quantum computing at the University of Melbourne, agrees.
“One of the most promising potential uses of quantum technology is to use one quantum system to simulate other quantum systems,” Hill says. “In this work, the authors considered a chain of ten quantum dots and used them to emulate the so-called SSH model.
“This is a remarkable piece of engineering. The quantum devices used for this demonstration were fabricated with sub-nanometre accuracy. This experiment paves the way for larger and more complex quantum systems to be emulated in future.”
The advantage of the complex production process, Simmons says, is that the team is “not creating new materials that you have to invent and figure out how to manufacture”.
“We literally have atomic sub-nanometre precision,” she adds. “The atoms themselves are in a silicon matrix, so you’re building a system in a material that has been used in the semiconductor industry.
“There are only two atoms – phosphorus and silicon – in our whole device. So, we get rid of everything else, all the interfaces, the dielectrics, all the stuff that causes problems in other architectures, and we just have those two atoms. It’s simple conceptually, but, obviously, challenging to make. It’s a nice, clean, physical, scalable system.
“The challenges were how do you put an atom in place, and then how do you know it’s there? It took us a literal decade to figure out the chemistry of getting phosphorous atoms to go into a silicon matrix so that it’s protected. (One of) the technologies we used was a scanning tunnelling microscope (STM), a lithographic tool.”
After placing a silicon slab in a vacuum, the team first heat the substrate to 1,100°C, before gradually cooling to around 350°C, creating a flat two-dimensional silicon surface. The silicon is then covered in hydrogen atoms, which can be selectively removed individually using the STM tip. Phosphorous atoms are placed in the newly formed gaps in the hydrogen layer, before the whole thing is covered in another layer of silicon.
“It means we’re making one device at a time,” Simmons admits. “But my analogy is it’s like a Swiss watch. It’s very precise and handmade. My view is, to make a scalable system, you need that precision. It’s very hard to build a qubit state when you don’t have precision, because you don’t know what you’ve got. So, our view is, yes, it’s slower, but you know what you’ve got.”
Once the device is fabricated, Simmons says their choice of algorithm to test it has historical significance.
“The simulation algorithm is the dream of Richard Feynman from the 1950s,” Simmons explains. “If you want to understand how nature works, you’ve got to build it at that length scale. With this kind of sub-nanometre precision accuracy, can we mimic the single and double bonds of the carbon molecule?
“Instead of using a single atom to mimic the carbon atom, we actually found that we use 25 phosphorus atoms.”
The team found that they were able to control the flow of electrons along the chain.
“So, you’ve got individual and local control and extended control,” says Simmons. “We’ve shown that we can do that with just six electrodes for a 10-dot chain. So, a lot fewer electrodes than the actual number of dots. And that’s great for scaling. Because fundamentally, in a quantum computer, you want to have that lower number of gates compared to the active elements, otherwise it scales badly.”
Not only has their device matched the SSH theory, Simmons believes that quantum computers will soon begin simulating problems beyond even our best theories.
“It’s opening a door into the kinds of things that we’ve never imagined before. It’s kind of terrifying and exciting at the same time,” she says.
The device has similar drawbacks as other quantum computers – in particular, the expensive and energy-intensive requirement for enormous refrigerators to keep the operating temperature extraordinarily low, approaching absolute zero.
For commercial confidentiality, Simmons is tight-lipped about the projects the SQC is tackling after this initial demonstration. But she does say: “We want to apply it to as many different things as we can and see what we can discover.”
“The fact that we can get the electrons coherently across the whole chain tells us that this is a very quantum coherent system,” she says. “It gives me confidence the physical system we’re using is really stable.
“It’s a demonstration of the purity of the system. There are lots of different ways we can go now. Making bigger physical systems is definitely one. Looking at the spin state rather than the charge states is another one.”
Simmons describes this kind of research as “a journey”. She particularly appreciates the interdisciplinary character of it, with quantum physicists, chemists, engineers and software engineers all involved.
“For young people, this is such an exciting field to be in,” she says. “It’s the evolution of a fundamental research project into something practically useful.”
Interested in having science explained? Listen to our new podcast.