Photonic quantum computers, which use the unusual properties of particles of light, can solve problems that stump even the most powerful supercomputers. At Paderborn University (Germany), researchers have for the first time used the country's first scalable photonic quantum computer for this purpose - a so-called Gaussian boson sampler (a device that measures the distribution of photons at the output of a complex optical network).
A group led by Dr. Christine Silberhorn of the Institute for Photonic Quantum Systems (PhoQS) and the university's Department of Physics published the new results in the journal PRX Quantum. The work was carried out on the Paderborn Quantum Sampler (PaQS) setup.
The scientists did not merely demonstrate the operation of a large-scale quantum system. They investigated how individual quantum resources - for example, squeezed light (a special state of light with reduced noise), photon statistics and entanglement (a quantum connection between particles) - affect the computational capabilities of a Gaussian boson sampler.
To do this, they combined expertise in integrated photonics, computer science and theoretical quantum optics. The groups of Silberhorn, Dr. Jan Sperling and Dr. Tim Bartley took part in the collaboration. They created a flexibly programmable photonic quantum machine on scalable components, and also developed new methods of data analysis and a theoretical basis for interpreting the measurements.
"The results emphasize that building useful quantum computers is not just about scaling up the system. A precise understanding of how properties such as squeezed light, photon statistics and entanglement affect computational performance is necessary for developing the next generation of quantum technologies," explained Dr. Michael Stefszky, head of the "Quantum Computing" research group in Silberhorn's team, who coordinated the project.
PaQS measures at which points photons leave a complex photonic network. For ordinary computers, this task becomes extremely difficult as the number of photons and the complexity of the network grow. Simulation on a supercomputer for large systems could take many thousands of years. A key feature of PaQS is full programmability: the network can be reconfigured for different experiments, which makes the system a flexible research platform rather than a device for a single task.
The setup is now being expanded for more complex computations and the study of new applications. The authors emphasize that the data obtained do not yet mean that photonic quantum computers are already ready to solve practical problems. The work provides an understanding of which quantum resources will determine the performance of future photonic quantum computers.