In the CMS detector at the Large Hadron Collider, protons smash together with enormous energy, producing a shower of particles. Most of them decay almost instantly, right at the collision point. But some hypothetical particles, if they exist, could live much longer — traveling several centimeters or even meters before decaying. Such "long-lived" particles leave an unusual signature in the detector: their decay products appear not from the center but off to the side, like a branch sprouting from a tree trunk.
That's exactly what physicists on the CMS experiment were looking for. They were hunting for the superpartner of the tau lepton — a particle predicted by some theories that extend the Standard Model. If it exists, it could decay into an ordinary tau lepton and an elusive neutral particle. The tau lepton itself is also unstable and quickly decays into other particles, but if the decay happens away from the center, the detector records a characteristic offset.
To pick out such events from the massive data stream, the researchers used a graph neural network for the first time, specifically trained to identify displaced tau decays. The network analyzed millions of events recorded from 2016 to 2018, selecting those where particle tracks did not converge at a single point. This significantly improved the search sensitivity compared to previous methods.
No new particles were found, but that's not an empty result. The absence of a signal allowed them to set stricter limits on the possible masses and lifetimes of tau superpartners. For example, if such a particle has a mass of 200 GeV, its lifetime cannot fall within a certain range — otherwise it would have been seen already. These constraints narrow the region where new physics might still be hiding and guide future searches.
This approach, using neural networks to hunt for non-standard signatures, is becoming an increasingly important tool in high-energy physics. As data accumulates and algorithms improve, the chances of catching something truly new grow. And even if tau superpartners don't exist, the methods developed to find them will prove useful for studying other rare processes that might be hiding in the Large Hadron Collider's data.
as written by the authors arXiv:2601.17576
Gauge-mediated supersymmetry-breaking models provide a strong motivation to search for a supersymmetric partner of the tau lepton (stau) with a macroscopic lifetime. Long-lived stau decays produce tau leptons that are displaced from the primary proton-proton interaction vertex, leading to an unconventional signature. This paper presents a search for the direct production of long-lived staus decaying within the CMS tracker volume in proton-proton collisions at $$\sqrt{s}$$ = 13 TeV, performed for the first time with an identification algorithm based on a graph neural network dedicated to displaced tau leptons. The data sample, corresponding to an integrated luminosity of 138 fb$$^{-1}$$, was recorded with the CMS experiment at the CERN LHC between 2016 and 2018. This search excludes, at 95% confidence level, stau masses, $$m_\tilde{\tau}$$, in the 126$$-$$260 (90$$-$$425) GeV range for a proper decay length of 50 mm in the maximally mixed (mass-degenerate) scenario, while for $$m_\tilde{\tau} $$ = 200 GeV, stau proper decay lengths are excluded in the range 21$$-$$94 (6$$-$$333) mm. These results improve the exclusion limits compared to previous searches, and extend the parameter space explored in the context of supersymmetry.
The tau lepton is the heaviest of the charged leptons: it's 3500 times heavier than the electron and lives only 0.3 picoseconds, but at high energies it manages to travel a few millimeters before decaying.
In Liu Cixin's novel "The Three-Body Problem," supersymmetric particles are used as an energy source, and their search at colliders is key to unraveling the mysteries of the universe.