Reexamining the search for light ALPs at flavor and forward accelerator experiments

TL;DR

This work addresses the search for light, long-lived axion-like particles (ALPs) by integrating flavor constraints with forward-physics LLP searches. The authors formulate a general ALP Lagrangian with both universal fermion couplings and flavor-violating terms, including one-loop electroweak effects and RG running, and compute hadronic ALP decays and exclusive production using updated hadronic modes. They derive updated bounds from heavy-meson decays and neutral-meson mixing, recasting $B\to K^{(\ast)}+inv$, $B\to K^{(\ast)}\mu^+\mu^-$, and kaon decays, and they implement production and decay calculations in the forward-physics framework FORESEE/SensCalc. The forward facilities FASER, FASER2, and FACET are shown to probe complementary regions of ALP parameter space, with sensitivities spanning $c_q/f_a$ down to $\sim 10^{-6}$–$10^{-8}\,\mathrm{GeV}^{-1}$ for quark couplings and loop-induced $c_{WW}/f_a$ down to $\sim 10^{-3}$–$10^{-5}\,\mathrm{GeV}^{-1}$ depending on $m_a$ and $\Lambda$, including regions suggested by recent Belle II anomalies. Overall, the study demonstrates the crucial complementarity between flavor measurements and forward LLP searches in constraining light ALPs and highlights scenarios where forward detectors can access parameter space beyond flavor constraints.

Abstract

The axion-like particle (ALP) is a well-motivated extension of the Standard Model. In this work, we revisit the sensitivity of forward accelerator experiments to light long-lived ALPs, and analyze flavor constraints. Our analysis incorporates recent measurements of the rare decays $B\to K + X$ and $K\to π+X$, which place stringent bounds on quark flavor violation of a light ALP. We consider the complete list of hadronic modes in the calculation of the ALP decay rate and exclusive production channels based on recent improvements. The analysis includes the discussion of tree-level quark flavor-violating couplings in addition to a universal flavor-conserving ALP coupling to fermions and the electroweak ALP couplings. Our results demonstrate the complementarity of heavy meson decays and forward accelerator facilities in probing light ALPs. The interplay between two ALP couplings is also investigated.

Figures (3)

• Figure 1: Left panels: the ALP couplings versus $m_a$ for decay lengths $c\tau=0.1~{\rm m}$ (red), $1~{\rm m}$ (blue), $5~{\rm m}$ (green). Right panels: ALP decay branching ratios as a function of $m_a$. We consider three choices for ALP couplings: $c_q\neq0,~c_{WW}=\tilde{c}_q=0$ (top), $c_{WW}\neq0,~c_q=\tilde{c}_q=0$ (middle) and $\tilde{c}_q\neq0,~c_q=c_{WW}=0$ (bottom). The ALP couplings $\tilde{c}_q$ and $c_{WW}$ are RG invariant. For $c_q\neq 0$ we show the results of $\Lambda=10^9$ GeV.
• Figure 2: The sensitivities of FASER (black), FASER2 (blue) and FACET (red) to ALP for the case $c_q\neq0,~c_{WW}=\tilde{c}_q=0$ with UV scale $\Lambda=10^3$ GeV (top left) and $\Lambda=10^9$ GeV (top right), case $c_{WW}\neq0,~c_q=\tilde{c}_q=0$ (bottom left), case $\tilde{c}_q\neq0,~c_q=c_{WW}=0$ (bottom right). The reachable contours are shown by solid lines. The flavor constraints from $B$ or $K$ decays are displayed using exclusion regions enclosed by dashed (decay to neutral mesons) or dotted (decay to charged mesons) lines. The preferred region by $B^+\to K^++{\rm inv}.$ ($K^+\to \pi^++{\rm inv}.$) [$K_L\to \pi^0 \gamma\gamma$] is shown by a purple (green) [magenta] meshed band. For $B^+\to K^++{\rm inv}.$ in particular, the light and dark purple regions represent the constraints from non-fitted He:2022ljo and fitted Fridell:2023ssf results, respectively. The regions pointed to by the arrows are within $1\sigma$ preferred parameter space by $K$ or $B$ meson mixing.
• Figure 3: The sensitivity of FASER (black), FASER2 (blue) and FACET (red) to the ALP couplings $c_q/f_a$ versus $c_{\tilde{q}}/f_a$ (top) and $c_q/f_a$ versus $c_{WW}/f_a$ (bottom) with UV scale $\Lambda=10^9$ GeV. For the ALP mass, two benchmarks $m_a=0.1$ GeV (left) and 2 GeV (right) are considered. The most stringent constraints from $B\to K^{(\ast)}+{\rm inv}.$ (purple and blue regions), $B\to K^{(\ast)}\mu^+\mu^-$ (gray and brown regions), $K_L\to \pi^0 +{\rm inv}.$ (green region), and the preferred regions from $B^+\to K^+ + {\rm inv}.$ (purple meshed band) and $K^+\to \pi^+ + {\rm inv}.$ (green meshed band) are also shown, as labeled in Fig. \ref{['fig:cma']}. For $B^+\to K^++{\rm inv}.$ in particular, the light and dark purple regions represent the constraints from non-fitted He:2022ljo and fitted Fridell:2023ssf results, respectively. The $1\sigma$ preferred regions by $K$ or $B$ meson mixing are arrowed in the parameter space.