Heavy quark-philic scalar dark matter with a vector-like fermion portal

TL;DR

This paper investigates a scalar dark matter model interacting with up-type quarks via a vector-like fermion portal, analyzing constraints from experiments and the implications for dark matter detection and top quark flavor-changing processes.

Contribution

It introduces a novel scalar dark matter model with a vector-like fermion portal focusing on heavy quark interactions and evaluates experimental constraints and phenomenological implications.

Findings

XENON1T strongly constrains Yukawa couplings y_1 and y_3.

Combined data nearly exclude m_S < m_t/2 for certain Yukawa benchmarks.

FCNC top decays are suppressed below current experimental sensitivities.

Abstract

In this work we consider a real scalar dark matter $S$ interacting only with $SU(2)_L$ singlet Up-type quarks $U_i=u_R,c_R,t_R$ via a vector-like fermion $\psi$ which has the same quantum number as $U_i$. The DM-nucleon scattering can proceed through both $h$-mediated Higgs portal (HP) and $\psi$-mediated vector-like portal (VLP), in which HP can receive sizable radiative corrections through the new fermions. We first study the separate constraints on the new Yukawa couplings $y_i$ and find that the constraints of XENON1T results are strong on $y_1$ from VLP scattering and on $y_3$ from its radiative contributions to HP scattering. Since both DM-light quark interactions and HP have been well studied in the existing literature, we move forward to focus on DM-heavy quark interactions. Since there is no valence $c,t$ quark inside nucleons at $\mu_{\rm had}\sim 1$ GeV, $y_2,y_3$ interactions are manifested in DM-gluon scattering at loop level. We find that renormalization group equation (RGE) and heavy quark threshold effects are important if one calculates the DM-nucleon scattering rate $\sigma^{\rm SI}_{p}$ at $\mu_{\rm had}\sim 1\, {\rm GeV}$ while constructing the effective theory at $\mu_{\rm EFT}\sim m_Z$. For the benchmarks $y_3=0.5, y_2=0.5, 1, 3$, combined results from $\Omega_{\rm DM} h^2\simeq 0.12$, XENON1T, Fermi-LAT, 13 TeV LHC data have almost excluded $m_S<m_t/2$ when only DM-$\{c,t\}$ interactions are considered. FCNC of top quark can be generated at both tree level $t\to \psi^{(*)}S \to cSS$ and loop level $t\to c+\gamma/g/Z$, of which the branching fractions are typically below $10^{-9}$ after passing the other constraints, which are still safe from the current top quark width measurements.