Cosmological observations have proven that dark matter (DM) not only exists but constitutes 27% of the universe. DM is expected to be electrically neutral, have mass, interact weakly with Standard Model (SM) particles, and be stable. DM cannot be directly detected in collider experiments, so if DM particles are produced via a mediator when protons collide, the DM itself would be invisible to the detector. However, it can be searched for in other ways. 

At the Large Hadron Collider (LHC), hadrons (groups of quarks, such as protons) are collided at velocities approaching the speed of light, and several layers of detectors reconstruct the particles and jets that emerge. Jets are sprays of particles that result from quarks and gluons at very high energies radiating more particles, essentially creating a cascade of particles in a narrow cone. Some heavier particles, such as Z-bosons, are highly unstable and decay immediately, in which case only the stable decay products are directly detected.

DM particles cannot be directly measured; however, they can be detected indirectly via lack of momentum conservation. Since in LHC collisions, the net momentum perpendicular to the beam is zero, any momentum imbalance in the transverse plane must be due to missing particles. Neutrinos, which are already in the SM, also cannot be detected because they have zero electric or color charge and thus don’t interact via either the electromagnetic or strong nuclear force. So the challenge is then discriminating neutrinos from DM, both invisible. For my research project this summer working with my faculty mentor Elliot Lipeles, who is part of the ATLAS collaboration at the LHC, we investigated one method of doing so by using correlations between the mass of the invisible and visible system.

Since in proton collisions, neutrinos are predominantly produced via an intermediate Z-boson, the mass of the invisible system in neutrino events will be near the Z-boson mass, while potential DM events will have masses at least twice the DM mass (because they are produced in pairs) and generally higher. In order to quickly investigate potential correlations, I used simulated samples of 2 muons (µ) instead of 2 neutrinos (𝜈) , because the physics of Z→µµ  very nearly matches that of Z → 𝜈𝜈, but the interactions involving muons are readily available and well-known.

I specifically looked at the properties of jets produced in Z → µµ interactions, and their relationships with the dimuons produced. In order to do so, I generated files containing simulated particle data of the Z → µµ interaction, and then ran codes on them using ROOT, a C++ based CERN framework, to plot the variables of interest for the simulated data in the samples.