The Compact Muon Solenoid (CMS) experiment collects petabytes of proton-proton collision data using extremely sophisticated electronics and data acquisition systems. Experimental particle physicists harvest this data to perform statistical analyses to hunt for signatures that could explain open questions in particle physics.
Image from this article
From Wikipedia
The Standard Model of particle physics is a prescription of all the fundamental particles that make up everything around us. With the discovery of the Higgs boson in 2012, all the fundamental particles that constitute the Standard Model had been discovered. However, several mysteries remain unexplained even with the help of these fundamental particles. For eg.: experimental results from the LHCb experiment show discrepancies in the way that a W boson interacts with tau particles versus electrons - this is unexpected according to the “Standard Model” of particle physics. Theorists have therefore spent decades hypothesizing the existence of unique, exotic particles that could “complete” the Standard Model.
Our project focuses on one such set of particles called “leptoquarks”. This result is the first among major particle physics experiements to search for specific leptoquarks called “nonresonant first-generation scalar and vector leptoquarks”. Leptoquarks have piqued the curiosity of physicists for ~40 years for their unique properties that could explain such mysteries. Physicists have built several theoretical and statistical models to identify the unique signatures of these leptoquarks, and detect their production in particle detectors.
Image from Fermilab Today, 2009
Image from this article
Protons colliding in the Large Hadron Collider (LHC) convert energy into mass using the famous Einstein relation $E=mc^2$ and produce millions of new particles. These particles are detected using detailed instrumentation techniques, which ultimately lead to a determination of the particle properties like their masses, momenta etc. This data is then distributed to various servers around the globe. Scientists create Monte Carlo simulations of their physics models, and use the data to perform statistical inference.
There are several families of leptoquarks, each family theorized to solve a unique physics puzzle. Leptoquarks can be produced in proton-proton collisions in several ways - one at a time, in pairs, or “virtually”. Virtual leptoquarks are termed as nonresonant, and can be thought of as transient energy blips that quickly produce other particles. These particles require special handling while designed a statistical model. Our search is the first among major collider experiments to look at virtual leptoquarks that interact with first and second generation Standard Model particles.
In my thesis, I created a detailed statistical model to check whether the collisions data shows signs of leptoquark existence. The probability distributions of leptoquark production were modeled using histograms, which were then fit to collisions data to extract the signal strength for potential leptoquark presence. Data selection and histogramming was performed using ROOT, C++ and Python, and maximum likelihood estimation was performed using the Combine package to extract the signal strengths for 8 types of leptoquarks. The results showed that for the data collected by CMS in 2016-2018, the hypothesis that virtual leptoquarks interact with first and second generation SM particles can be rejected with high probability. The final result in the paper is quoted in the form of confidence limits on the signal strength as a function of the mass of the leptoquark. Compared to previous searches, we exclude leptoquarks upto 4 times more massive, at 95% confidence level. The limits on the signal strength are also 30% better than theoretical estimates. The Analysis Summary is available publicly on the CERN Document Server, and the final paper will be submitted to the Journal of High Energy Physics in October 2024. This result was presented at ICHEP 2024 and at multiple conferences and workshops across the US, including DPF-PHENO 2024.
