Freeze-In Production of FIMP Dark Matter

TL;DR

The paper develops thermal freeze-in as an IR-dominated mechanism for dark matter production via feebly interacting massive particles (FIMPs), whose relic density is largely set near the FIMP mass and is largely insensitive to the reheat temperature. It identifies motivated FIMP candidates from SUSY/string frameworks (moduli/modulinos), Dirac neutrino setups, kinetic-mixing hidden sectors, and very heavy extra-dimensional scenarios, and discusses collider and cosmological signatures including long-lived LOSPs, BBN implications, and potential warm DM components. The authors derive abundance phase diagrams and provide explicit calculations for direct freeze-in, decays of frozen-in FIMPs, and freeze-in via 2→2 scattering, while also addressing higher-dimension operators and multi-FIMP sectors. Overall, freeze-in offers a compelling, testable alternative to WIMP-like freeze-out with distinctive phenomenology across collider, astrophysical, and cosmological observations, and it naturally ties to UV frameworks such as string compactifications and GUT-scale physics.

Abstract

We propose an alternate, calculable mechanism of dark matter genesis, "thermal freeze-in," involving a Feebly Interacting Massive Particle (FIMP) interacting so feebly with the thermal bath that it never attains thermal equilibrium. As with the conventional "thermal freeze-out" production mechanism, the relic abundance reflects a combination of initial thermal distributions together with particle masses and couplings that can be measured in the laboratory or astrophysically. The freeze-in yield is IR dominated by low temperatures near the FIMP mass and is independent of unknown UV physics, such as the reheat temperature after inflation. Moduli and modulinos of string theory compactifications that receive mass from weak-scale supersymmetry breaking provide implementations of the freeze-in mechanism, as do models that employ Dirac neutrino masses or GUT-scale-suppressed interactions. Experimental signals of freeze-in and FIMPs can be spectacular, including the production of new metastable coloured or charged particles at the LHC as well as the alteration of big bang nucleosynthesis.

Figures (7)

• Figure 1: Log-Log plot of the evolution of the relic yields for conventional freeze-out (solid coloured) and freeze-in via a Yukawa interaction (dashed coloured) as a function of $x=m/T$. The black solid line indicates the yield assuming equilibrium is maintained, while the arrows indicate the effect of increasing coupling strength for the two processes. Note that the freeze-in yield is dominated by the epoch $x\sim 2-5$, in contrast to freeze-out which only departs from equilibrium for $x\sim 20-30$.
• Figure 2: Schematic picture of the relic abundances due to freeze-in and freeze-out as a function of coupling strength. The way in which the freeze-out and freeze-in yield behaviours connect to one another is model-dependent. As we show in detail in Section \ref{['phasediagrams']}, freeze-in and freeze-out are in fact two of the four basic mechanisms for thermal DM production, and we sketch the "abundance phase diagrams" of DM depending upon the strength and type of the DM-thermal bath interaction and the DM mass.
• Figure 3: Schematic representation of the four possible scenarios involving the freeze-in mechanism. The left-hand figures show the LOSP/FIMP spectrum with circles representing cosmologically produced abundances. The large (small) circles represent the dominant (sub-dominant) mechanism for producing the dark matter relic abundance, a dotted (solid) circle signifies that the particle is unstable (stable), and a filled (open) circle corresponds to production by freeze-in (freeze-out). The right-hand figures show the LOSP and FIMP abundances as a function of time. The initial era has a thermal abundance of LOSPs and a growing FIMP abundance from freeze-in. The LOSP and FIMP are taken to have masses of the same order, so that FIMP freeze-in ends around the same time as LOSP freeze-out. Considerably later, the heavier of the LOSP and FIMP decays to the lighter.
• Figure 4: In (a) we show the contours of $\Omega h^2$ as a function of the mass $m_X$ and coupling $\lambda$ for the case of a quartic interaction. The plane can be divided into two "phases": $X$ freeze-out, phase (I), occurs for large coupling and $X$ freeze-in, phase (III), occurs for weak coupling. In (b) we take a slice at fixed $m_X \sim v$ and plot the variation of $\Omega h^2$ as a function of the coupling $\lambda$.
• Figure 5: In (a) we show the abundance phase diagram with contours of $\Omega h^2$ as a function of the coupling $\lambda$ and the mass $m_X$ for the case of a Yukawa interaction, eq. (\ref{['eq:Yukawa']}). If $\lambda^2 > \sqrt{m_X/M_{Pl}}$, $X$ undergoes conventional freeze-out giving region I; while in region II, with $m_X/m_{Pl} < \lambda^2 < \sqrt{m_X/M_{Pl}}$, $X$ decouples from the bath giving a yield $Y_X \sim 1$. In region III, with $m_X/m_{Pl} > \lambda^2 > (m_X/M_{Pl})^2$, $Y_X$ never reaches unity and freeze-in provides the dominant contribution to dark matter. For $\lambda^2 < (m_X/M_{Pl})^2$ the dominant mechanism generating dark matter arises from $\psi_1$ freezing out and then decaying to $X +\psi_2$, giving region IV. In (b) we take a slice at $m_X=1\, {\rm TeV}$ and plot the variation of $\Omega h^2$ as a function of the coupling $\lambda$.