Gravitational wave probes of particle dark matter: a review

TL;DR

This review surveys how gravitational-wave interferometers and pulsar timing arrays can probe particle and macroscopic dark matter across an enormous mass range, from ultralight fields to DM spikes and dark atoms. It categorizes DM effects into two broad pathways: direct DM–detector couplings producing measurable strains, and DM sources or environments modifying GW emission (e.g., boson clouds, solitons, spikes). It details the theoretical models (axions, dilatons, vector/tensor DM, WIMPs), the expected signatures (quasi-monochromatic CWs, transients, dephasing), and the analysis methods (cross-correlation, BSD excess power, LPSD, Viterbi, Wiener filtering). Current constraints from LIGO/Virgo/KAGRA and pulsar timing arrays constrain couplings and DM properties in multiple regimes, while future ground- and space-based detectors (ET/CE/LISA Pathfinder successors) promise substantial improvements and potential detections. The work highlights the complementary nature of terrestrial and astrophysical probes in testing a broad DM landscape, including novel DM scenarios like macroscopic objects, solitons, and atomic DM, within the broader context of gravity and beyond-standard-model physics.

Abstract

Various theories of dark matter predict distinctive astrophysical signatures in gravitational-wave sources that could be observed by ground- and space-based laser interferometers. Different candidates-including axions, dark photons, macroscopic dark matter, WIMPs, and dark-matter spikes-may appear in interferometer data via their coupling to gravity or the Standard Model, altering the measured gravitational-wave strain in distinct ways. Despite their differences, these candidates share two key features: (1) they can be probed through their effects on gravitational waves from inspiraling compact objects, isolated black holes, and neutron stars, or via direct interactions with detectors, and (2) their signatures likely persist far longer than the seconds-long mergers detected today, necessitating new data analysis methods beyond matched filtering. This review outlines these dark matter candidates, their observational signatures, and approaches for their detection.

Figures (28)

• Figure 1: Landscape plot showing the various types of dark matter (DM) that can be directly probed using gravitational-wave (GW) interferometers, both presently and in future experiments. The flow of logic proceeds from left to right: type of DM $\rightarrow$ affected astrophysical system or detector component $\rightarrow$ observable effect $\rightarrow$ relevant detector(s). This framework spans a wide mass range of particle DM, from $\mathcal{O}(10^{-23}-10^{-21})$ eV with pulsar timing arrays to $\mathcal{O}(10^{-16}-10^{-13})$ eV with space-based detectors to $\mathcal{O}(10^{-13}-10^{-11})$ eV with ground-based interferometers, as well as [1,$10^9$] kg macroscopic DM. Matching colors link each DM type to its corresponding observable.
• Figure 2: Similar to \ref{['fig:summary-dm-direct']}. This plot shows the kinds of DM that can be probed via their generation of GWs, both now and in the future. The logic is as follows, moving from left to right: type of DM $\rightarrow$ astrophysical source affected $\rightarrow$ observable $\rightarrow$ which detector(s) the signal could be seen in. We note that the mass range of particle DM probed here ranges many orders of magnitude, from $\mathcal{O}(10^{-23}-10^{-21})$ eV with pulsar timing arrays to $\mathcal{O}(10^{-16}-10^{-13})$ eV with space-based detectors to $\mathcal{O}(10^{-13}-10^{-11})$ eV with ground-based interferometers, and includes $[10^3,10^9]$ GeV WIMP DM, probes of large-scale DM structure via solitons and DM spikes, and atomic DM. Matching colors imply that the source would induce those physical signatures.
• Figure 3: Taken from Miller:2020vsl. We show the strain time series $h(t)$ (left) and the Fourier transform of it (right) of a simulated dark photon DM signal. The structure in the amplitude spectrum density in the right-hand plot arises because we have taken the length of the Fourier transform to exceed that of the coherence time of the signal. This dark photon DM signal is a superposition of $1000$ dark photons traveling with distinct Maxwell-Boltzmann-distributed velocities, which cause small frequency deviations away from the minimal frequency $f_0=740.436$ Hz ($m_A=3.06\times 10^{-12}$ eV). This plot shows why it is important to choose $T_\text{FFT}\sim T_{\text{coh}}$ for the analysis of DM signals. Though we have simulated a dark photon signal, the power spectrum will appear similarly for the other kinds of DM interaction signals. Here, $T_\text{obs}\sim 10^5$ s, meaning the frequency resolution is $\delta f=10^{-5}$ Hz. The coherence time and length of this signal are: $T_\text{coh}=4595$ s and $L_\text{coh}=5.28\times 10^8$ m; the coupling strength is $\epsilon=3\times10^{-21}$. The signal is actually simulated for $\sim 233$ days, though we only show the first day of its time evolution.
• Figure 4: Adapted from LIGOScientific:2021odm. Upper limits obtained from analyzing LIGO O3 data on the square of the coupling of dark photons to baryons $U(1)_{\rm B}$ in the LIGO mirrors. The physics behind this form of DM has been discussed in \ref{['subsec:vecdm']}. The limits derived from each of the methods discussed in \ref{['subsubsec:crosscorr', 'subsubsec:excesspower']} are show in red (cross-correlation) and cyan (BSD excess power), respectively. MICROSCOPE Berge:2017ovy and Eöt-Wash torsion balance upper limits are plotted as a comparison to the results here Schlamminger:2007ht. To produce limits on the square of the dark photon/baryon-lepton coupling, $U(1)_{\rm B-L}$ in the LIGO mirrors, these limits should be multiplied by four.
• Figure 5: Taken from Gottel:2024cfj. 95% confidence-level upper limits on $\Lambda_i^{-1}$, the coupling of dilaton DM to electrons or photons, as a function of mass and frequency from LIGO O3 data, which would have cause time-dependent oscillations of the sizes and indices of refraction of the beamsplitter and the LIGO mirrors. The physics behind this form of DM has been discussed in \ref{['subsubsec:size-dilaton']}.a) and b) show these constraints compared to existing ones on $\Lambda_e$ and $\Lambda_{\gamma}$, respectively. The results from the LIGO O3 search are shown by the thick blue line, constraints from direct experimental searches for DM Vermeulen:2021epaAiello:2021wlpAharony:2019iadSavalle:2020vgzKennedy:2020bacAntypas:2019qjiAntypas:2020rtgTretiak:2022ndxOswald:2021vtcCampbell:2020fvqBACON:2020ubhZhang:2022ewzFukusumi:2023kqdSherrill:2023zah are shown in thin grey, and constraints from searches for "fifth forces" Berge:2021yyeHees:2018fpg are depicted by the dashed red lines.