Axion-Like Particle Searches at Muon Colliders: Photon-Neutrino Signatures and Heavy-Neutral-Lepton Feasibility

Abstract

We study the mono-photon signature ${\mu }^{+}{\mu }^{-}\to \gamma a\to \gamma \nu \bar{\nu }$ at $3$ and $10\TeV$ muon colliders, with integrated luminosities of $1$ and $10{\mathrm{ab}}^{-1}$, in a two-coupling axion-like-particle (ALP) simplified model. Five benchmarks spanning $m_a=100$--$1000\GeV$ are generated at $g_{aγγ}=0.01\TeV^{-1}$ and flavor-universal pseudoscalar neutrino couplings $y_{\nu }=0.1$ or $0.3$, giving invisible branching fractions above 99.995\% and mono-photon cross sections of $0.00854$--$0.01217\mathrm{fb}$. Assuming a 1\% background normalization uncertainty, a cut-based photon-line selection gives projected 95\% sensitivity to $g_{a\gamma \gamma }\simeq 0.039$--$0.045\TeV^{-1}$ at $3\TeV$ and $0.0137$--$0.0140\TeV^{-1}$ at $10\TeV$. A ROOT/TMVA gradient-boosted decision tree with validation-fixed thresholds reaches held-out AUC values of $0.9967$--$0.9993$ and $g_{95}=0.0192$--$0.0400\TeV^{-1}$ at $3\TeV$ and $0.0122$--$0.0144\TeV^{-1}$ at $10\TeV$. Because only 2-41 irreducible-background test events populate the optimized classifier tails, the mass--coupling contours are reported as projected expected sensitivities rather than profile-likelihood or CL${}_s$ exclusions. A separate appendix records a phenomenological, parton-level HNL diphoton feasibility study; it is not used to claim detector-level sensitivity or exclusion.} \keywords{Axion-like particles, muon collider, mono-photon, diphoton, heavy neutral leptons, neutrinos, multivariate analysis, effective field theory

arXiv comment32 pages, 22 figures, 11 tables