JWST Lensed Quasar Dark Matter Survey III: Dark Matter Sensitive Flux Ratios and Warm Dark Matter Constraint from the Full Sample

TL;DR

The study leverages JWST/MIRI multi-band flux ratios from 31 quadruply imaged quasars to probe low-mass dark matter halos, focusing on a turnover in the halo mass function driven by warm dark matter. A forward-modeling, Bayesian framework incorporating field and subhalo populations, tidal evolution, and complex lens macromodels yields constraints on the half-mode mass m_hm and the corresponding thermal relic mass m_therm. The warm-dust flux ratios are shown to be largely free of microlensing contamination, enabling robust inference of millilensing effects. The analysis sets tight limits on WDM scenarios, with priors based on galacticus and N-body simulations leading to m_therm > several keV, and establishes a scalable approach for testing a broad class of dark matter models in lensing systems.

Abstract

We present the full sample of measurements of the warm dust emission of 31 strongly-lensed, multiply imaged quasars, observed with JWST MIRI multiband imaging, which we use to constrain the particle properties of dark matter. The strongly lensed warm dust region of quasars is compact and statistically sensitive to a population of dark matter halos down to masses of $10^6$ M$_\odot$. The high spatial resolution and infrared sensitivity of MIRI make it uniquely suited to measure multiply imaged warm dust emission from quasars and thus to infer the properties of low-mass dark halos. We use the measured flux ratios to test for a warm dark matter turnover in the halo mass function. To infer the dark matter parameters, we use a forward modeling pipeline which explores dark matter parameters while also accounting for tidal stripping effects on subhalos, globular clusters, and complex deflector macromodels with $m=1, m=3, \text{ and } m=4$ elliptical multipole moments. Adopting a comparable prior on the projected density of substructure to our previous analyses, the data presented here provide a factor of 2 improvement in sensitivity to a turnover in the halo mass function. Assuming subhalo abundance predicted by the semi-analytic model galacticus we infer with a Bayes factor of 10:1, a half-mode mass $m_{\rm{hm}} < 10^{7.8} M_{\odot}$ (m>5.6 keV for a thermally produced dark matter particle). If instead we use a prior from N-body simulations, we infer $m_{\rm{hm}} < 10^{7.6} M_{\odot}$ (m>6.4 keV). This is one of the strongest constraints to date on a turnover on the halo mass function, and the flux ratios and inference methodology presented here can be used to test a broad range of dark matter physics.

Figures (5)

• Figure 1: Comparison of multi-band SED fitting result (filled bands) with the flux ratios in each filter for the 26 lenses used in the dark matter inference. Blue circles, orange crosses and black triangles correspond to B/A, C/A and D/A respectively for all lenses, with the exception of: PSJ0147 and J1251 (A/C, B/C, D/C), H1413 (A/B, C/B, D/B) and WFI2033 (A2/A1, B/A1, C/A1). For all lenses, with the exception of J1251, the flux ratio in the reddest filter is less than 2 $\sigma$ from the inferred warm dust flux ratio, consistent with minimal microlensing contamination. For systems with narrow-line flux ratio measurements, we show the narrow-line measurements as stars at an arbitrary wavelength for comparison.
• Figure 2: The results of the spectral energy distribution fitting for J0924 for two of the four lensed images. The extreme difference in the spectral energy distribution model between the lensed images reveals strong microlensing in this system. The smooth macromodel expectation based on the image positions is that image A and D should have the same flux. J0924 shows a significant deviation from this expectation.
• Figure 3: The inferred joint posterior probability distribution for the dark matter parameters for the full sample of 26 mid-IR plus 2 narrow-line lenses. Black contours show the posterior probability distributions for a uniform prior on $\Sigma_{\rm{sub}}$ with a factor of 100 uncertainty. The green and purple contours shows the results when a log Gaussian prior with 0.3 dex width is used, centered on the theoretical prediction for galacticus and $N$-body simulations respectively.
• Figure 4: A comparison of the inferred mass in subhalos with the amount of correction to the flux ratios due to the SED modeling, relative to using only the reddest filter flux ratios alone (F2550W or F2100W). Systems which required more correction from SED modeling were more microlensed, so this test is intended to explore the extent to which microlensing may affect our measurement. With the exception of J1251, the reddest filter (F2550W or F2100W) flux ratios are within two sigma of the SED-corrected flux ratios, and J1251 does not show a significant preference for additional structure.
• Figure 5: Deflector spectra and measured redshift for J0607-2152 (top) and J2017+6204 (bottom). The extracted and smoothed 1D spectrum of the deflector is plotted over the observed wavelength where the smoothing has been done using a boxcar filter of size around 4Å. Prominent stellar absorption lines are marked if present. The measured redshift of the deflector is mentioned in the lower right corner of the plot.