Constraining off-shell Higgs boson production and the Higgs boson total width using $WW\to \ellν\ellν$ final states with the ATLAS detector

TL;DR

This work addresses the challenge of constraining the Higgs boson total width by exploiting off-shell Higgs production in the $H^{*}\to WW$ channel, where the rate depends on couplings but not on $\Gamma_H$. The authors implement a detailed event-categorised analysis using neural-network-based discrimination and a proxy mass variable $V_{31}$ to exploit interference with non-resonant $WW$ backgrounds, and they combine off-shell and on-shell information to extract a bound on $\Gamma_H$ at the level of a few MeV. Their most robust result is an observed upper bound on the Higgs width of $13.1$ MeV (95% CL) after combining with the on-shell measurement, with the off-shell signal strength constrained to $\mu_{\text{off-shell}}=0.3^{+0.9}_{-0.3}$. The methodology and precision achieved provide a stringent test of the SM Higgs sector and improve sensitivity to BSM scenarios that modify Higgs couplings or introduce custodial-symmetry-breaking effects, thereby enhancing our capability to probe new physics through Higgs phenomenology.

Abstract

A measurement of off-shell Higgs boson production is performed in the $H^{*} \rightarrow WW$ channel. The measurement uses a proton-proton collision dataset with an integrated luminosity of 140 fb$^{-1}$ collected at a centre-of-mass energy of 13 TeV by the ATLAS detector at the Large Hadron Collider. Final states in which both $W$ bosons decay leptonically are targeted, and events are categorised based on the flavour of the final-state leptons, the jet multiplicity, and the output of neural-network-based classifiers. The data are found to be compatible with the Standard Model expectation. An observed (expected) upper bound on the 95% symmetric confidence level interval is set on the rate of off-shell Higgs boson production at a value of 3.4 (4.4) times the Standard Model prediction. These results are combined with the results from the measurement of on-shell Higgs boson production in the same final states to obtain an observed (expected) upper bound at 95% confidence level on the Higgs boson total width of 13.1 (17.3) MeV.

Figures (4)

• Figure 1: Example diagrams of leading-order gluon-induced $WW$ production (ggF) (a) including and (b) not including a Higgs boson, and quark-induced $WW$ production (EW) (c) including and (d) not including a Higgs boson.
• Figure 2: Summary of the post-fit $V_{31}$ distributions in the high DNN-score category of the SRs. The top panel shows the event yield divided by the bin width in each of the measured categories for both measured data (dots) and expected post-fit predictions (histograms). Overflow is included in the last bin of each distribution. The bottom panel shows the ratio of the data to the post-fit expected predictions. The sum in quadrature of statistical and systematic post-fit uncertainties are depicted by a grey band. Two alternative hypotheses are included, where the nuisance parameters are fixed to their best-fit values, $\hat{\theta}$, and the value of the parameter of interest $\mu_{\mathrm{off\hbox{-}shell}}\xspace$ = 0 (solid line) and $\mu_{\mathrm{off\hbox{-}shell}}\xspace$ = 3.4 (dashed line). The former represents the null hypothesis whereas the latter represents the upper bound of the symmetric 95% CL interval of $\mu_{\mathrm{off\hbox{-}shell}}\xspace$.
• Figure 3: (a) Values of the test statistic $t_{\mu_{\mathrm{off-shell}}}$ as a function of the off-shell Higgs boson signal strength, $\mu_{\mathrm{off\hbox{-}shell}}\xspace$. The 68% and 95% CL are shown as gray dotted lines. (b) Contours of the values of the profile likelihood ratio $t_{\mu^{\mathrm{ggF}}_{\mathrm{off-shell}}\xspace,\mu^{\mathrm{EW}}_{\mathrm{off-shell}}\xspace}$ as a function of $\mu^{\mathrm{ggF}}_{\mathrm{off-shell}}\xspace$ and $\mu^{\mathrm{EW}}_{\mathrm{off-shell}}\xspace$. The SM predicted value and best-fit value are indicated by a star and a diamond, respectively. The observed (expected) results are shown in black (blue) for scenarios with and without systematics. For (a) the 68% and 95% CL are calculated using an explicit Neyman construction, whereas in (b) they are calculated via the asymptotic approximation.
• Figure 4: Values of the test statistic $t_{\kappa_H}$ as a function of $\kappa_H = \Gamma_{H}\xspace / \Gamma_{H}^{\mathrm{SM}}\xspace$ for the combined on-shell and off-shell result assuming $R_{gg}=R_{VV}=1$. The observed (expected) results are shown in black (blue) for scenarios with and without systematics. The 68% and 95% CL are shown as gray dotted lines. The 68% and 95% CL were calculated using an explicit Neyman construction.