Non-thermal production of heavy vector dark matter from relativistic bubble walls

Abstract

Heavy vector boson dark matter at the TeV scale or higher may be produced non-thermally in a first-order phase transition taking place at a lower energy scale. While the production of vector dark matter has previously been studied for bubble wall collisions, here we calculate production by bubble wall expansion in a plasma, which can be the dominant production mechanism. We compute the results numerically and provide an analytical fit for the vector dark matter density. The numerical fit is also validated for scalar dark matter production, obtaining results in agreement with past literature. We find that vector pair production leads to bubble wall friction with a novel boost factor scaling behaviour compared to transition radiation emission of a single vector. We conclude that TeV-scale WIMP vector dark matter can be efficiently produced non-thermally by first-order phase transitions in a wide region of parameter space where thermal freeze-out is inefficient. In this scenario, the phase transition scale is predicted to be in the sub-GeV to $\mathcal{O}(10)$ TeV range and could therefore be accessible to future gravitational wave detectors.

Figures (14)

• Figure 1: Schematic illustration of the light-to-heavy $1\rightarrow 2$ process in a FOPT. The grey surface represents the bubble wall of the expanding broken phase (right side of the wall) where $\langle\Phi\rangle\neq0$. A $\phi$ excitation in the plasma passing through the fast bubble wall decays into a pair of much heavier $\chi$ particles, taking $z$-momentum from the wall itself.
• Figure 2: Numerical results of $n_\chi/A$ as a function of $\gamma_w$ for fixed $m_\chi/T=50, 500, 1000$, respectively.
• Figure 3: Comparison between the numerical results presented in Fig. \ref{['fig:nphi']} with the fit function given in Eq. \ref{['eq:fit-function']} with $c_1=0.00175, c_2=22.7$.
• Figure 4: Numerical result of $n_\chi/A$ in the range $m_\chi/T\in [20,1000]$ for $\gamma_w=10^6$, as well as its comparison between with the fit function given in Eq. \ref{['eq:fit-function']} with $c_1=0.00175, c_2=22.7$.
• Figure 5: Comparison between the numerical results for different values of $L_w T$ and the fit using parameters given in Table \ref{['tab:fit-kappa']}.