The quantum electrodynamic process of photon–photon scattering has for the first time been confirmed experimentally to a high degree of certainty.
CERN’s ATLAS collaboration, which involves hundreds of physicists from around the world, made the breakthrough after analysing a large dataset of candidate scattering events using a neural network. Their discovery could fuel new research into a variety of theories beyond the Standard Model of particle physics. In classical electrodynamics, photons cannot interact with each other because they have no charge. At the same time, however, quantum electrodynamics predicts that two photons can scatter off each other by exchanging virtual charged fermions or W bosons. Some theorized extensions to the Standard Model predict that these scattering events are sensitive to as-yet unconfirmed particles, including axions and magnetic monopoles. To test these theories, physicists at CERN’s Large Hadron Collider (LHC) have attempted to induce photon–photon scattering by firing heavy ions towards each other at relativistic speeds. As they pass closer and closer to each other, the ions exchange an increasing number of virtual photons. If scattering occurs between any two of these photons, the ion pair will lose a small amount of energy and emit a pair of real photons. These light flashes will then hit opposite sides of the detector, revealing the characteristics of the original scattering event. Colliding lead ions In 2017, both the ATLAS and CMS collaborations, based at the LHC, searched for evidence of these photon pairs during high-energy collisions between lead ions that they had recorded in 2015. From a total of 13 candidate events, the experiments reported photon–photon scattering to certainties of 4.4σ and 4.1σ respectively – which falls short of the widely accepted 5σ certainty threshold required to confirm an experimental discovery. Read more Light is seen to scatter off light The ATLAS collaboration has now repeated the experiment using a far larger dataset of 59 candidate scattering events, gathered from a further, more extensive series of lead-ion collisions carried out in 2018. In addition, they developed a neural network to more effectively distinguish the photon pairs that indicate these scattering events from all the background photons picked up by the detector. This allowed the team to greatly increase the certainty to 8.2σ, which is well beyond the accepted threshold. Several recent theories have predicted that measurements of photon–photon scattering could be sensitive to phenomena beyond the Standard Model. These include particles containing just one magnetic pole, which is forbidden by classical electrodynamics, as well as axions, which are theorized to solve the strong CP problem of quantum chromodynamics. The discoveries of the ATLAS collaboration could, therefore, inform future studies aiming to confirm and constrain these theories, potentially allowing much-anticipated updates to the Standard Model. The full results are described in Physical Review Letters