In 2011, computer scientist Scott Aaronson of the University of Texas outlined the fundamentals of what a light-based quantum computer would be, but, given the opinion of fellow physicists that it would never work, shelved the project – to see it come to fruition. nine years later. “I didn’t think it would happen anytime soon. I take my hat off to them. ” Aaronson was referring to physicists Chao-Yang Lu and Jian-Wei Pan, who made it possible for the quantum computer Jiuzhang to achieve quantum supremacy this Friday (4).
This expression refers to the power of a quantum computer to solve problems that classical computers would not be able to – at least, not in a timely manner. In 2019, Google claimed for itself the feat, with Sycamore, which took just three minutes to do the same calculation that would take three days for a supercomputer to achieve the same result.
While Sycamore uses superconductors to form qubits, in Jiuzhang the photons themselves are the quantum bits. To perform the same calculation as the quantum computer did now, its Chinese colleague Sunway TaihuLight, the third most powerful supercomputer in the world, would take 2.5 billion years.
Built at the China University of Science and Technology (USTC) in Shanghai, the Jiuzhang has a complicated setup: a laser in the structure hits 25 crystals made of titanyl potassium phosphate – when hit, each releases two photons in opposite directions. These are then sent along 100 different paths, covering trails made with 300 prisms and 75 mirrors. Finally, they fall into 100 different channels, where they are detected.
Not that big
THE Wikipedia teaches us that “a quantum computer is a device that performs calculations making direct use of properties of quantum mechanics, such as overlap and interference.”
Realizing that a universal photonic quantum computer would require millions of lasers, mirrors and other optical devices, the scientific community buried the idea – until Aaronson and physicist Alex Arkhipov, at the time his student, came up with the idea of using the concept of boson: the quantum computer would have only a few lasers, mirrors, prisms and photon detectors.
To try to understand how boson sampling works in Jiuzhang, we can use the Galton board.
Simulating the distribution of the balls through the channels is not a complicated job for a computer. In the case of a quantum, boson sampling uses photons instead of balls; mirrors and prisms are the pins through which they run.
Sips of molecules
The laser’s photons bounce off mirrors through prisms until they fall into the channels to be detected. Unlike classical balls, the quantum properties of the photon lead to an exponentially increasing number of possible distributions.
Boson sampling schemes do not withstand a low number of photons. To preserve the sensitive quantum arrangement, they must remain indistinguishable. Those who travel the narrow paths within Jiuzhang cover a distance of 22 meters and their positions do not stand out from each other by more than 25 nanometers. “This is the equivalent of 100 horses traveling 100 kilometers and crossing the finish line with at most one hair between them,” said Chao-Yang Lu.
For the success achieved this Friday, photons need to be generated in the purest form possible, produced one after the other. “It is easy for anyone to drink water in measured sips, but it is difficult to drink just one molecule at a time,” said Pan.