Error free Quantum computers with perfect photons

Quantum computers are extremely sensitive to errors. Previous methods required hundreds of physical photons to make a single reliable qubit. The Twente researchers’ method can replace part of the required error correction, requiring far fewer photons. This ultimately makes quantum computers cheaper and more accessible.

Filtering without knowing the problem

Whereas previous methods focused on error correction afterwards, the UT researchers tackled the problem at the source. They designed a new component that can conveniently distil a single high-quality photon from a brew of imperfect photons. To do this, they devised an optical circuit consisting of programmable light guides and detectors. By exploiting the quantum properties of light, they created a ‘Schrödinger’s cat’-like state in which photons are more likely to assume ‘good’ rather than ‘bad’ properties. As in Schrödinger’s famous thought experiment, measurement ultimately determines whether a photon has retained the desired properties.

Instead of accepting imperfect photons and correcting them later, the new method directly filters perfect photons. Although filtering each perfect photon requires sacrificing several imperfect photons, this reduces the total number of photons needed. This saves an enormous amount of computing power and makes quantum computers cheaper and more efficient. By using an optical circuit with programmable switches, the researchers can filter out bad photons without having to know exactly what causes the error. “Normally, you have to decide in advance what you are filtering, such as a colour filter that only lets red light through. We can now filter without knowing in advance what the problem is,” says PhD student Frank Somhorst.