Dark Secrets of Baryons: Illuminating Dark Matter-Baryon Interactions with JWST

TL;DR

This paper uses JWST UV luminosity function measurements at $z\gtrsim 8$ to constrain non-gravitational dark matter–baryon interactions by linking IDM to the high-redshift UVLF through a Thesan-Zoom–based galaxy formation model. The authors solve IDM-modified Boltzmann equations to obtain the matter power spectrum, compute the halo mass function, and apply a calibrated galaxy–halo connection to predict the UVLF, fitting to JWST data with MCMC to derive 95% CL upper limits on the cross sections for various velocity dependences. The strongest constraint arises for DM–proton scattering with $n=-2$ in the sub-GeV mass range, with $\log_{10}(\sigma^{\chi p}_{-2}/\mathrm{cm^2}) \approx -33.5$ at $m_\chi \sim 1\ \mathrm{MeV}$, while other channels (DM–electron or other $n$) yield competitive limits relative to existing cosmological probes. The results demonstrate JWST’s potential as a novel probe of non-gravitational DM physics, robust to astrophysical modeling and with clear pathways for improvement as data and simulations advance.

Abstract

The James Webb Space Telescope (JWST) has discovered numerous bright galaxies at high redshifts ($z\approx$ 10 -- 14). Many astrophysical models and beyond the Standard Model physics scenarios have been proposed to explain these observations. We investigate, for the first time, the implications of dark matter (DM) scattering with baryons (protons and electrons) in light of the JWST UV luminosity function (UVLF) observations. These interactions suppress structure formation on galactic scales, which may have an observable effect on the UVLF measurements at high redshifts. Using a recent galaxy formation model designed to explain high-redshift observations, we obtain strong upper limits on DM-baryon scattering cross-sections and explore new regions of the parameter space. For DM-proton scattering with cross-section $\propto v^{-2}$ velocity dependence, we obtain the strongest limit for DM masses of $\sim$ 1 -- 500 MeV. For other cases that we study (DM-proton scattering cross-section $\propto v^{0},\,v^{-4}$, and DM-electron scattering cross-section $\propto v^{0},\,v^{-2},\,v^{-4}$), our limits are competitive with those obtained from other cosmological observables. Our study highlights the potential of JWST observations as a novel and powerful probe of non-gravitational interactions of DM.

Figures (8)

• Figure 1: 95% C.L. upper limits on the normalization of velocity-dependent ($\sigma\propto v^{-2}$) DM-proton elastic scattering cross-section as a function of the DM mass $m_\chi$. The green shaded region is ruled out by our analysis using the JWST UVLF spectroscopic data. Various dashed curves represent other cosmological constraints -- Lyman-$\alpha$ forest (cyan), Milky Way satellite count (pink), and CMB+BAO (purple) Buen-Abad2022Nguyen2021.
• Figure 2: Left: Comparison of the ratio $T^2(k) = P_{\rm IDM}(k)/P_{\rm CDM}(k)$ of power spectrum of IDM to that of non-interacting CDM. We consider IDM cases with different values of $n$: $n = 0$ (pink), $n = -2$ (green), and $n = -4$ (purple), for DM of mass $m_\chi = 1~\rm{MeV}$, interacting with protons. Cross-section for each case corresponds to the 95% C.L. upper limits obtained in this work: $\log_{10}(\sigma^{\chi p}_{n}/\mathrm{cm}^{2})$ for $n = 0,-2,-4$ cases are $-27.49$, $-33.48$, and $-40.33$, respectively. Suppression of power spectrum at the characteristic scales (gray shaded region) is visible. Right: Comparison of UVLFs at $z\approx 14$ for $\Lambda$CDM (dashed yellow) and IDM (same color code as left panel) for the same IDM scenarios. Data points represent UVLF measurements at $z \approx 14$ as derived in Harikane et al. Harikane2024b (blue squares) and from the MoM-z14 galaxy observation 2025arXiv250511263N (pink diamond). The choice of astrophysical parameters for $\Lambda$CDM correspond to the best-fit values from Shen:2025isu with $\log_{10}{M_{c}} = 10.5\,M_{\odot}$, while those for the interacting scenarios are fixed at the best-fit values obtained from our model by MCMC sampling.
• Figure 3: The magenta solid line shows SFE as a function of halo mass for the $\Lambda$CDM cosmology, using the best-fit SFE model parameters from Shen:2025isu at $z\approx 14$. The gray line, and dark gray and light gray shaded regions show respectively the median, 1$\sigma$, and 2$\sigma$ intervals for SFE (Eq. \ref{['eq:SFE']}) sampled from our posterior pdfs for $m_\chi=1~\rm{MeV}$, $n=-2$, $B=p$ as a function of $M_h$. The dashed line represents the $\epsilon_\star(M_{h})$ from the Thesan-Zoom best-fit model, extended beyond the cutoff $M_{c} = 10^{10.5}\,M_{\odot}$
• Figure S1: Left: Comparison of the ratio $T^2(k) = P_{\rm IDM}(k)/P_{\rm CDM}(k)$ of power spectrum of IDM to that of non-interacting CDM. We consider IDM cases with different values of $n$: $n = 0$ (pink), $n = -2$ (green), and $n = -4$ (purple), for DM of mass $m_\chi = 1~\rm{MeV}$, interacting with electrons. Cross-section for each case corresponds to the 95% C.L. upper limits obtained in this work: $\log_{10}(\sigma^{\chi e}_{n}/\mathrm{cm}^{2})$ for $n = 0,-2,-4$ cases are $-27.84$, $-31.14$, and $-35.42$, respectively. Suppression of power spectrum at the characteristic scales (gray shaded region) is visible. Right: Comparison of UVLF at $z\approx 14$ for $\Lambda$CDM (dashed yellow) and IDM (same color code as left panel) for the same IDM scenarios. Data points represent UVLF measurements at $z \approx 14$ as derived in Harikane et al. Harikane2024b (blue squares) and for the measurement of the MoM-z14 galaxy 2025arXiv250511263N (pink diamond). The choice of astrophysical parameters for $\Lambda$CDM correspond to the best-fit values from Shen:2025isu and $\log_{10}{M_{c}} = 10.5$, while those for the interacting scenarios are fixed at the best-fit values obtained from our model by MCMC sampling.
• Figure S2: 95% C.L. upper limits on the velocity-independent DM-baryon elastic scattering cross-section as a function of the DM mass $m_\chi$. The green shaded region in both panels is ruled out by the JWST UVLF spectroscopic data (our work). The various dashed curves represent other cosmological constraints -- Lyman-$\alpha$ forest (cyan), Milky Way satellite count (pink), and CMB+BAO (purple) Buen-Abad2022Nguyen2021. Left: Shows constraints on DM-proton interactions. Dark pink shaded region represents the parameter space excluded by cosmic ray observations Yin:2018yjn. Yellow shaded area shows the region ruled out by direction detection experiments CRESST:2019jnqCRESST:2019axxCRESST:2017uesXENON:2017vdwEDELWEISS:2019vjv. Some other constraints on $\sigma^{\chi p}_{0}$ are shown in Buen-Abad2022. Right: Shows constraints on DM-electron interactions. The celestial blue shaded region represents constraints imposed by cosmic ray observations Ema:2018bihCappiello:2019qsw. Yellow shaded region represents constraints from solar reflection of cosmic rays An:2017ojcEmken:2021lgc. Brown shaded region shows constraints obtained from gas cooling in Leo-T dwarf galaxy Wadekar:2019mpc. The shaded gray region shows constraints from BBN + CMB Sabti:2019mhn. The orange shaded region is constrained from XQC rocket Erickcek:2007jvMahdawi:2018euy. The black dashed line shows the constraints from weak lensing Zhang:2024mmg. The dark pink shaded region is constrained from LUX-ZEPLIN (LZ) collaboration Maity:2022exk. Some other constraints on $\sigma^{\chi e}_{0}$ are shown in Buen-Abad2022An:2021qdl.