Precision Luminosity Determination at the CMS Experiment
Date Issued
August 2026
Author(s)
Karunarathna, Warusapperuma Don Nimmitha
The University of Tennessee
Advisor(s)
Stefan Spanier
Additional Advisor(s)
Sarah Cousineau
Tova Holmes
Robert Grzywacz
Abstract
The Compact Muon Solenoid (CMS) collaboration probes for new forces and particles at CERN's Large Hadron Collider (LHC). One approach searches for significant discrepancies in Higgs-boson properties relative to Standard Model predictions that arise from beyond-Standard Model processes.
Luminosity is essential for estimating background rates and for converting observed event yields into cross sections or decay branching ratios. This dissertation presents studies using the CMS detector aimed at achieving high-precision luminosity measurements during recent data-taking periods. It describes calibration studies for a dedicated luminosity detector, the Pixel Luminosity Telescope, including the van der Meer beam scan method, beam-related corrections, detector-response corrections, and nonlinearity studies.
Furthermore, an independent luminosity measurement was developed by counting reconstructed J/ψ-meson candidates decaying into muons in the large data set. Event yields are extracted from maximum-likelihood fits to di-muon invariant-mass distributions over short data-taking intervals. Event rates as a function of the LHC proton beam separation provide absolute calibration. This new method, together with corrections for pile-up and efficiency variations, yields a relative uncertainty of less than 2% with respect to reference measurements.
The dissertation presents a search for the rare Higgs boson decay into a Z boson and a J/ψ meson, with the Z decaying into a muon pair and the J/psi into an electron or muon pair. A machine-learning (ML) discriminant is trained to suppress background from misidentified event candidates. A two-dimensional unbinned maximum-likelihood fit using the 4-lepton invariant mass and the ML discriminant isolates the Higgs signal from background. According to Standard-Model predictions, no signal is expected in collision data taken during 2016–2018. Using the modified frequentist method, an expected upper limit on the decay branching ratio is estimated to be 10^-3 at 95% confidence. This corresponds to about 400 times the Standard-Model value.
Degree
Doctor of Philosophy
Major
Physics
File(s)![Thumbnail Image]()
Name
KARUNARATHNA-PRIMARY-2026.pdf
Size
10.24 MB
Format
Adobe PDF
Checksum (MD5)
7d63af7f77c9a66802e4cf57cd85d0fa