Simultaneous qubit operation on a quantum chip
By simultaneously operating multiple spin qubits on the same quantum chip, researchers from the University of Copenhagen (UCPH) have overcome a key obstacle on the road to the supercomputer of the future. Their result, published in the journal Physical Review X Quantum, bodes well for the use of semiconductor materials as a platform for solid-state quantum computers.
One of the engineering headaches in the global marathon towards a large functional quantum computer is the control of many basic memory devices — qubits — simultaneously. This is because the control of one qubit is typically negatively affected by simultaneous control pulses applied to another qubit. UCPH quantum physicists Federico Fedele and Assistant Professor Anasua Chatterjee, collaborating with researchers from Purdue University, have now managed to overcome this obstacle with the use of semiconductor qubits, known as spin qubits.
“Broadly speaking, they consist of electron spins trapped in semiconducting nanostructures called quantum dots, such that individual spin states can be controlled and entangled with each other,” explained Fedele, formerly a PhD student and now a postdoc at UCPH.
The brain of the quantum computer that scientists are attempting to build will consist of many arrays of qubits, similar to the bits on smartphone microchips, that will make up the machine’s memory. While an ordinary bit can either store data in the state of a 1 or 0, a qubit can reside in both states simultaneously — known as quantum superposition — which makes quantum computing exponentially more powerful.
Spin qubits have the advantage of maintaining their quantum states for a long time. This potentially allows them to perform faster and more flawless computations than other platform types. They are also so minuscule that far more of them can be squeezed onto a chip than with other qubit approaches; the more qubits, the greater a computer’s processing power.
The UCPH team fabricated and operated four qubits in a 2x2 array on a single chip. The next step was getting them to communicate with each other, said Asst Prof Chatterjee.
“Now that we have some pretty good qubits, the name of the game is connecting them in circuits which can operate numerous qubits, while also being complex enough to be able to correct quantum calculation errors,” she said. “Thus far, research in spin qubits has gotten to the point where circuits contain arrays of 2x2 or 3x3 qubits. The problem is that their qubits are only dealt with one at a time.”
It is here that the physicists’ quantum circuit, measuring no larger than the size of a bacterium, makes all the difference. The four spin qubits in the chip are made of the semiconducting material gallium arsenide. Situated between the four qubits is a larger quantum dot that connects the four qubits to each other, which the researchers can use to tune all of the qubits simultaneously.
“The new and truly significant thing about our chip is that we can simultaneously operate and measure all qubits,” said Asst Prof Chatterjee. “This has never been demonstrated before with spin qubits, nor with many other types of qubits.”
Being able to operate and measure simultaneously is essential for performing quantum calculations. Indeed, if you have to measure qubits at the end of a calculation — that is, stop the system to get a result — the fragile quantum states collapse. Thus, it is crucial that measurement is synchronous, so that the quantum states of all qubits are shut down simultaneously. If qubits are measured one by one, the slightest ambient noise can alter the quantum information in a system.
“To get more powerful quantum processors, we have to not only increase the number of qubits, but also the number of simultaneous operations, which is exactly what we did,” said Associate Professor Ferdinand Kuemmeth, who directed the research. The realisation of the new circuit thus marks a milestone on the long road to a semiconducting quantum computer.
At the moment, one of the main challenges is that the chip’s 48 control electrodes need to be tuned manually, and kept tuned continuously despite environmental drift, which is a tedious task for a human. That’s why the research team is now looking into how optimisation algorithms and machine learning could be used to automate tuning. The researchers have also begun working with industrial partners to fabricate the next generation of quantum chips, to allow fabrication of even larger qubit arrays.
Researchers have combined two traditional semiconductor doping methods to achieve new...
Pure germanium was bonded with aluminium in a way that atomically sharp interfaces are created,...
The fabrication process, involving placing an atom-thick semiconductor between two mirrors, marks...