Exploring the Discovery Reach for a 95 GeV Scalar in Future e + ​ e − e^{+}e^{-} Collisions Bruce Mellado Abstract The observed indications for a new scalar resonance with a mass around 95 GeV, initially reported by LEP and supported by CMS and ATLAS in di-photon, τ ​ τ \tau\tau , and W + ​ W − W^{+}W^{-} channels, motivate exploring its discovery potential at future electron-positron colliders. This study focuses on the production of the new scalar ( S S ) via e + ​ e − → Z ​ S e^{+}e^{-}\rightarrow ZS with Z → μ + ​ μ − Z\rightarrow\mu^{+}\mu^{-} and S → b ​ b ¯ S\rightarrow b\bar{b} and optimizes the signal recognition using the recoil-mass method. By employing deep neural networks for signal-background discrimination, we demonstrate that a 95 GeV scalar, mixing with the Standard Model Higgs by an angle of ∼ \sim 0.1, can be observed with a 5 σ \sigma significance at s \sqrt{s} = 250 GeV or 200 GeV with 5 ab -1 of integrated luminosity. [a]Mukesh Kumar 1 Introduction The scalar particle with a mass of 125 GeV observed by the CMS [ 1 ] and ATLAS [ 2 ] experiments at the Large Hadron Collider (LHC) exhibits properties consistent with the ones of the predicted by the Standard Model (SM). While this discovery represents a major triumph for the SM, the underlying symmetry principles do not preclude the existence of additional Higgs bosons. Intriguingly, the CMS [ 3 ] and ATLAS [ 4 ] experiments have reported local excesses with significances of 2.9\sigma and 1.7\sigma , respectively, near 95 GeV in the diphoton final state. Furthermore, a global significance of 3.4\sigma is obtained within a combined analysis including the \tau\tau [ 5 ] and the WW channels [ 6 , 7 ] . Motivated by these hints for a new scalar boson, we investigate the discovery potential of future e^{+}e^{-} colliders – including the Circular Electron–Positron Collider (CEPC), the International Linear Collider (ILC), the Compact Linear Collider (CLIC), and the Future Circular Collider (FCC-ee). The signal process considered is e^{+}e^{-}\rightarrow Z^{*}\rightarrow ZS , with subsequent decays Z\rightarrow\mu^{+}\mu^{-} and S\rightarrow b\bar{b} . The dominant background arises from the Standard Model process e^{+}e^{-}\rightarrow Zb\bar{b} , with Z\rightarrow\mu^{+}\mu^{-} . To enhance the signal over background sensitivity, we use the recoil mass method and a Deep Neural Network (DNN) – based classifier. 2 Signal and background analysis The dominant production mechanism for the scalar boson S in the center-of-mass energy range of 220 GeV to 350 GeV is the Higgs-strahlung process, e^{+}e^{-}\rightarrow Z^{*}S . The SZZ coupling is parametrized by a factor \kappa_{Z} within the \kappa -framework and can be probed using the recoil mass of the Z boson, with the decay channel Z\rightarrow\mu^{+}\mu^{-} offering excellent sensitivity due to the precise reconstruction of muon four-momenta. The recoil mass, \displaystyle M_{\rm recoil}=\sqrt{s+M_{\mu^{+}\mu^{-}}^{2}-2(E_{\mu^{+}}+{E_{\mu^{-}}})\sqrt{s}}\,, (1) where M_{\mu^{+}\mu^{-}} denotes the invariant mass of the muon pair and E_{\mu^{\pm}} represent the energies of the individual muons. The signal and background events are generated using MadGraph at center-of-mass energies of \sqrt{s}=200\,\mathrm{GeV} and 250\,\mathrm{GeV} with the following generation-level cuts: transverse momentum of muons p_{T_{\mu}}>10\,\mathrm{GeV} , pseudo-rapidity of muons |\eta_{\mu}|<2.5 , and angular distance between muons \Delta R_{\mu^{+}\mu^{-}}>0.4 . The parton-level events are then processed through Pythia8 for showering and hadronization, followed by Delphes to simulate detector effects. For the optimization of signal over background, we require at least two b -tagged jets and two muons with energies E_{b,\mu}> 5 GeV. The recoil mass distribution M_{\rm recoil} is computed and shown in Figure 1 . We then apply selection cuts on the recoil mass M_{\rm recoil}<120\,\mathrm{GeV} and on the invariant mass of the b -jet pair M_{b\bar{b}}<100\,\mathrm{GeV} for both signal and background events. As presented in Table 1 , the signal significance improves from 2.4\sigma to 3.4\sigma at an integrated luminosity of \mathcal{L}=1~\mathrm{ab}^{-1} for \sqrt{s}=250\,\mathrm{GeV} and \kappa_{Z}^{2}=0.1 . Figure 1 : Recoil-mass distributions for the SM background (dashed) and the combined signal plus background (solid) for a scalar mass of m_{S}=95.5\,\mathrm{GeV} at \sqrt{s}=200\,\mathrm{GeV} (blue) and \sqrt{s}=250~\mathrm{GeV} (red) with an integrated luminosity of \mathcal{L}=500\,\mathrm{fb}^{-1} . Kinematic cuts \mathcal{N}_{S} \mathcal{N}_{B} significance At least 2 b -tagged jets and 2 muons 188 (219) 2926 (1876) 2.4 \sigma (3.8 \sigma ) E_{b,\mu}> 5 GeV 188 (219) 2924 (1874) 2.4 \sigma (3.8 \sigma ) M_{b\bar{b}}< 100 GeV 187 (217) 2223 (1864) 2.9 \sigma (3.8 \sigma ) M_{\text{recoil}}< 120 GeV 184 (216) 1732 (1851) 3.4 \sigma (3.8 \sigma ) 93.5 GeV