There's another quantum computer to keep track of in this Wild West era of quantum computing research we're in. And it uses some parts you might already be familiar with.
A new silicon wafer delivered to TU Delft researchers to hold the spin qubits (Image: Intel)
Researchers from two teams now working with Intel have reported advances in a new quantum computing architecture, called spin qubits, in a pair of papers out yesterday. They obviously aren't the full-purpose quantum computers of the future. But they have a major selling point over other quantum computing designs.
"We made these qubits in silicon chips, similar to what's used in classical computer processes," study author Thomas Watson from TU Delft in the Netherlands told me. "The hope is that by doing things this way, we can potentially scale up to larger numbers needed to perform useful quantum computing."
Quantum computers, for the uninitiated, turn the rules of computers on their head like the Wizard of Oz going from black-and-white to colour. Classical computers perform all of their calculations by converting data into binary code. Each zero or one is represented by some physical two-choice bit. Quantum computers instead use "qubits" - quantum bits that take on the two values simultaneously during calculations. Pairs of qubits talk to one another using the rules of quantum mechanics. They output regular bit values once the user needs an answer. You can read this if you need more info.
There are a lot of ways to physically construct qubits. It requires building a collection of two-state systems that operate and communicate via the rules of quantum mechanics. Google and IBM use tiny pieces of supercooled, superconducting electronics. IonQ hopes to use atoms trapped by lasers, with two different internal states representing the two qubit states. Microsoft hopes to use some pretty out-there, still unobserved physics. But there are other ways.
Yesterday, a research group at TU Delft, called QuTech, announced that they'd successfully tested two "spin qubits" on hardware supplied by researchers from the University of Wisconsin, Madison. These qubits involve the interaction of two confined electrons in a silicon chip. Each electron has a property called spin, which sort of turns it into a tiny magnet, with two states: "Up" and "down". The researchers control the electrons with actual cobalt magnets and microwave pulses. They measure the electron's spins by watching how nearby electric charges react to the trapped electrons' movements.
Those researchers, now working in partnership with Intel, were able to perform some quantum algorithms, including the well-known Grover search algorithm (basically, they could search through a list of four things), according to their paper published yesterday in Nature. Additionally, a team of physicists led by Jason Petta at Princeton reported in Nature that they were able to pair light particles, called photons, to corresponding electron spins. This just means that distant spin qubits might be able to talk to one another using photons, allowing for larger quantum computers.
"It frees you up from having to have a nearest-neighbour interaction," Petta told me. "You can couple an electron spin to a photon and connect that photon to any other spin or circuit."
There are some advantages to these systems. Present-day semiconductor technology could create these spin qubits, and they would be smaller than the superconducting chips used by IBM. Additionally, they stay quantum (meaning they can maintain their ability to hold simultaneous values) longer than other systems.
"Both of these papers report on research done in dilution refrigerators, similar to those used for superconducting qubits," UC Berkeley postdoc Sydney Schreppler, who was not involved in the studies, told me. "But there may be a future where these operate at room temperature, unlike superconducting qubits. You can also contrast this with ion-based quantum computers, which require ultra-high vacuum and multiple control lasers to operate."
There are drawbacks. Since these qubits are so isolated, said Schreppler, "it's very difficult to measure these spins, and even more difficult to get them to interact with each other. That's why gate times have been historically slow for these systems." She also mentioned that the qubits needed to be really close to each other, which is why she was especially excited about Petta's team's work. "That will enable longer-range interactions," she said, such as qubits talking to ones further away on the same chip or even on another chip.
We're still in a sort of foggy early era of quantum computers, where systems of less than a thousand qubits can only make pretty limited, noisy calculations, and scientists are still working on physical qubit architectures. So, you can get excited - but not too excited.