Probing the Higgs Portal to a Strongly-Interacting Dark Sector at the FCC-ee

TL;DR

This paper demonstrates the FCC-ee’s potential to probe Higgs-portal dark sectors that yield semi-visible jets, by combining a cut-based strategy for high invisible fraction with a LundNet graph neural network tagger to recover sensitivity in the low-invisible fraction regime. It develops a concrete dark-sector model with $SU(N_c)_d$, $N_c=3$, $N_f=2$, and a Higgs-mediated production of dark quarks, parameterized by $r_{inv}$, $\Lambda$, $p_v$, and $p_\eta$, and evaluates the signal and background using LO Monte Carlo with Pythia8 HV and Delphes IDEA. The study shows that BR$(H\to q_d\bar{q}_d)$ can be constrained at the permille level in the inclusive case and down to $\mathcal{O}(10^{-2})\%$ with GNN tagging across a wide parameter space, highlighting the FCC-ee’s clean environment as a powerful probe of Higgs-portal dark sectors. The proposed framework, including the LundNet architecture and parameter-mapping strategy, offers a scalable approach that can be extended to hadron colliders for complementary SVJ searches.

Abstract

This work explores exotic signatures from confining dark sectors that may arise in the e+e- collision mode at the Future Circular Collider. Assuming the Higgs boson mediates the interaction between the Standard Model and the dark sector, dark quarks can be produced in e+e- collisions. The ensuing strong dynamics may lead to semi-visible jet final states, containing both visible and invisible particles. We investigate semi-visible jets with different fractions of invisible states, and enriched in leptons and photons. When the invisible component is large, selections based on kinematic features, such as the missing energy in the event, already provide good signal-to-background discrimination. For smaller invisible fractions, the reduced missing energy makes these signals more similar to Standard Model events, and we therefore employ a graph neural network jet tagger exploiting differences in jet substructure. This machine learning strategy improves sensitivity and enhances the discovery prospects of Higgs boson-induced semi-visible jets at the Future Circular Collider. Our results show that the proposed strategy can effectively probe a wide parameter space for the models considered, and a variety of signatures, constraining the Higgs boson exotic branching ratios into dark quarks at the permille-level.

Figures (11)

• Figure 1: Sketch of the process studied in this work: $e^+e^-\to Z h$ followed by the decay $h\to q_d\bar{q}_d$ and dark showering and hadronization. The top panel corresponds to the specific model that we use in the high-$r_{\text{inv}}$ regime, where the dark shower contains $\eta_d'$, $\rho_d$ and $\pi_d$ and we consider the decays $\eta_d'\to\gamma\gamma$ and $\rho_d^0\to f\bar{f}$. The bottom panel corresponds to the parametric model that we use in the low-$r_{\text{inv}}$ regime.
• Figure 2: Signal-wise AUC scores in the low-$r_{\text{inv}}$ regime (left) and in the high-$r_{\text{inv}}$ regime (right).
• Figure 3: Expected sensitivities in the high $r_{\text{inv}}$ regime without (left) and with (right) the GNN tagger as a function of $\Lambda$ and $p_\eta$ for fixed $p_v=0.5$.
• Figure 4: Expected sensitivities in the low $r_{\text{inv}}$ regime without (left) and with (right) the GNN tagger as a function of $(\Lambda,r_{\text{inv}})$ and $r_{\mathrm{inv}}$.
• Figure 5: $r_{\mathrm{inv}}$ values obtained with simulation data changing $p_v$ and $p_\eta$ for different $\Lambda$ values.