Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Exploring new physics in the dark sector at CMS

Kai Hong Law

TL;DR

The work addresses dark-sector questions by searching for hidden-valley dark showers, long-lived particles, and lepton-enriched semivisible jets in CMS data from Run 2 and Run 3. It leverages dedicated data streams (parking and scouting) to access challenging low-mass and displaced signatures and employs ML-based techniques such as BDTs and graph neural networks to enhance signal discrimination and background estimation. Across three benchmark models, the analyses set stringent 95% confidence level limits on Higgs decays to dark partons, on Z'–dark sector production, and on dark-hadron lifetimes, with complementary sensitivity across lifetimes and masses. The results extend experimental constraints on the dark sector and underscore the value of specialized data streams and ML-driven analyses for new-physics searches.

Abstract

A selection of new results from the CMS experiment is presented. These results focus on searches for dark-sector particles using Run 2 or Run 3 data. Dedicated data streams were utilised to explore the low-mass parameter space. Machine learning techniques were employed to discriminate between signal and background.

Exploring new physics in the dark sector at CMS

TL;DR

The work addresses dark-sector questions by searching for hidden-valley dark showers, long-lived particles, and lepton-enriched semivisible jets in CMS data from Run 2 and Run 3. It leverages dedicated data streams (parking and scouting) to access challenging low-mass and displaced signatures and employs ML-based techniques such as BDTs and graph neural networks to enhance signal discrimination and background estimation. Across three benchmark models, the analyses set stringent 95% confidence level limits on Higgs decays to dark partons, on Z'–dark sector production, and on dark-hadron lifetimes, with complementary sensitivity across lifetimes and masses. The results extend experimental constraints on the dark sector and underscore the value of specialized data streams and ML-driven analyses for new-physics searches.

Abstract

A selection of new results from the CMS experiment is presented. These results focus on searches for dark-sector particles using Run 2 or Run 3 data. Dedicated data streams were utilised to explore the low-mass parameter space. Machine learning techniques were employed to discriminate between signal and background.

Paper Structure

This paper contains 6 sections, 3 figures.

Table of Contents

  1. Introduction
  2. New Physics Results
  3. Search for low-mass hidden-valley dark showers with non-prompt muon pairs
  4. Search for long-lived particles decaying into muons using the CMS scouting data sets
  5. Search for the resonant production of lepton-enriched semivisible jets
  6. Conclusion

Figures (3)

  • Figure 1: Upper limits at $95\%$ CL on the branching fraction $\mathcal{B}\left(H\rightarrow\psi\overline{\psi}\right)$ for the vector portal as a function of the dark vector meson $c\tau$ (upper left), for Scenario A as a function of the dark photon $c\tau$ (upper right), and for Scenario B1 as a function of the dark pion $c\tau$ (lower). The limits are shown for an example mass hypothesis for each model CMS-PAS-EXO-24-008.
  • Figure 2: Upper limits at $95\%$ CL on the branching fraction $\mathcal{B}\left(H\rightarrow\psi\overline{\psi}\right)$ for Scenario A as a function of the dark photon $c\tau$ (left), and for Scenario B1 as a function of the dark pion $c\tau$ (right). The limits are shown for an example mass hypothesis for each model CMS-PAS-EXO-24-016. The dark blue solid line represents the observed upper limits set by Ref. CMS-PAS-EXO-24-008.
  • Figure 3: Upper limits at $95\%$ CL on $\sigma_{\textrm{Z}'}\mathcal{B}_{\textrm{dark}}$ for the SVJ$\ell$ model as a function of $m_{\textrm{Z}'}$ for $r_{\textrm{inv}}=0.3, 0.5, 0.7$ and $m_{\textrm{dark}}=16~\textrm{GeV}$. The red solid line represents the nominal $\textrm{Z}'$ cross section CMS-PAS-EXO-24-029.