Unprecedented photon interactions: another step towards quantum computing?

Broadly speaking, there are two categories of objects in the world around us: matter, which is made up of atoms, and waves. Atoms easily influence one another. For example, if you throw two balls at each other, their trajectories will be altered. 

Waves, on the other hand, can pass through each other. The ripples, i.e. the waves, created by throwing a stone into water can therefore pass through other ripples. 

However, certain electromagnetic waves, specifically microwave photons - or photons - which are generated by the movement, or vibration, of electrons, can be considered as particles capable of interacting with one another in the field of quantum physics. In the visible spectrum, for example, it is possible to merge two laser beams inside a nonlinear crystal. 

Thanks to the crystal's unique properties, photons can interact within this space, with one influencing the state of another. Recently, researchers have also succeeded in inducing interactions between two or three photons within giant atomic clouds. 

Yes, superconductors. Our laboratory has been working with them for a long time, and one key advantage is that they can be ‘sculpted’ into miniature circuits, like those found in conventional computers, making it easier to imagine adapting them to quantum computers. In these electromagnetic circuits, electrons, grouped in pairs, circulate with almost no friction. 

By adding tiny metallic objects known as resonators into the circuit, electrons can be put into vibration, thereby creating photons and inducing their potential interaction. To do this, the resonator must include another device, ten times thinner than a strand of hair, known as a Josephson junction. This device consists of two superconducting needles, separated by just a dozen oxide molecules. Under normal conditions, electrons cannot pass through this oxide layer, but, thanks to quantum tunnelling, pairs of electrons are able to cross it. The scientific community has been aware of this for several decades, but until recently the results have been relatively modest: either the photons did not interact with each other at all, or, at best, their interactions were weak. 

We initially studied Josephson junctions when we were conducting research for the development of a protected qubit. The qubit is the unit of information storage in quantum computing, with the advantage of being able to exist in two states simultaneously. This superposition of states is one of the key benefits of quantum computing, as it allows calculations to be performed much faster than conventional computers. The challenge? Qubits are not stable over time, and when they begin to deteriorate, errors can occur. This is why we came up with the idea of using an assembly of Josephson junctions as a filter, designed to allow only pairs of electron pairs to pass through certain points in a circuit. As a result, if two states differ by just a single pair of electrons, the qubit cannot transition between them, and therefore remains stable.

We designed a sort of interferometer, a small circuit comprising around a hundred interconnected junctions, through which a magnetic flux of a very precise value passes. We called it the ‘Kinetic Interference coTunneling Element’ or KITE. And when we realised that it could also be used to induce interactions between microwave photons, we decided to test it out. 

First, we had to build our KITE, and this process of nanomanufacturing was a challenge because it is only a few micrometres long and a few hundred nanometres wide. We used electron-beam lithography, a sort of etching using electron beams, similar to the procedure used in the design of conventional processors.

But the biggest challenge was figuring out how to prove, through experimentation, the interaction between photons. We opted for spectroscopy, which allows us to visualise the frequency required to introduce a photon into the resonator.

In fact, to create a photon, electrons must be put into motion using microwave light of a specific frequency. If, when I attempt to introduce a second photon, this frequency changes, it proves that the presence of the first photon modifies the system and therefore that there is an interaction. This has been observed in the circuits with Josephson junctions tested so far, in that for each photon introduced, the required frequency was slightly lower. 

We discovered a previously unknown pattern: to introduce a second photon, then a third and a fourth, sometimes the frequency had to increase, sometimes it had to decrease, and the amplitudes were much greater than in conventional systems. This is proof that the interactions between photons were much more powerful than those observed up until then! 

We stopped the experiment at four photons mainly for technical reasons (superconducting materials must be placed in refrigerators where the temperature is close to absolute zero!), but these results are sufficient to demonstrate how our system works, and theory can predict the rest. It's both a wonderful surprise to see that our idea works and to have been able to demonstrate it so well!

First and foremost, this discovery could lead to the observation of new states of light. In the same way that matter is created by an arrangement of atoms, we can imagine a system, based on our KITE, that condenses many photons to produce photonic agglomerates or crystals. This would already be considered a major breakthrough in quantum optics from a fundamental perspective. Once this has been achieved, it will be necessary to study the possible applications, but these are still difficult to imagine for the moment. 

On another level, our results also open up new possibilities for research in quantum computing: information could be encoded in such a way that, for example, a photon only passes if another photon is present or, conversely, absent, which can only be achieved if the photons are able to detect each other and therefore interact. By multiplying the interactions between photons, it could be possible to create more complex quantum logic gates (which calculate a result based on input data) and therefore perform calculations even faster.

Not immediately, because, although impressive, this result is not related to our research on qubit protection. We will therefore be returning to our initial focus, especially since this study has allowed us to make progress on the subject. We now have a solid theoretical foundation for the KITE, a better understanding of how it works, better knowledge of the manufacturing process, etc. All of this will enable us to move forward with qubit protection, to carry on making progress in research on quantum computers... and contribute to making this revolution a reality.