Annu. Rev. Astron. Astrophys. 2017. 55:343-387 Copyright © 2017 by Annual Reviews. All rights reserved

4. CURRENT FRONTIERS

The tremendous progress in identifying and characterizing faint stellar systems in the Local Group has led to a variety of new questions. For one, these discoveries have blurred what was previously a clear difference between dwarf galaxies and star clusters, leading to the question, “what is a galaxy?” (Willman & Strader 2012). DES has identified several new satellite galaxies, many of which appear to be clustered around the Large Magellanic Cloud (LMC; Drlica-Wagner et al. 2015). The putative association of these satellites with the LMC is intriguing (Jethwa, Erkal & Belokurov 2016, Sales et al. 2017), as the nearly self-similar nature of dark matter substructure implies that the LMC – which is likely to be hosted by a halo of M_peak ∼ 10¹¹ M_⊙ (Boylan-Kolchin et al. 2010) – could itself contain multiple dark matter satellites above the mass threshold required for galaxy formation. Satellites of the LMC and even fainter dwarfs will be attractive targets for ongoing and future observations to test basic predictions of ΛCDM (Wheeler et al. 2015).

The 800 pound gorilla in the dwarf discovery landscape is the Large Synoptic Survey Telescope (LSST). Currently under construction and set to begin operations in 2022, LSST has the potential to expand dwarf galaxy discovery space substantially: by the end of the survey, co-added LSST data will be sensitive to galaxies ten times more distant (at fixed luminosity) than SDSS, or equivalently, LSST will be able to detect galaxies that are one hundred times fainter than SDSS at the same distance. This means that LSST should be complete for galaxies with L_⋆ ≳ 2 × 10³ L_⊙ within ∼ 1 Mpc of the Galaxy, dramatically increasing the census of very faint galaxies beyond ∼ 100 kpc from the Earth.

One of the unique features of LSST data sets will be the ability to explore the properties of low-mass, isolated dark matter halos (i.e., those that have not interacted with a more massive system such as the Milky Way), thereby separating out the effects of environment from internal feedback and dark matter physics. Given the predictions discussed in Sec. 3.1.1, any new discoveries with M_⋆ ≲ 10⁶ M_⊙ at ∼ 1 Mpc from the Milky Way and M31 will be attractive targets for discriminating between baryonic feedback and dark matter physics. At this distance, spectrographs on 10m-class telescopes will not be sufficient to measure kinematics of resolved stars; planned 30m-class telescopes will be uniquely suited to this task.

In addition to hosting surviving satellites, galactic halos also act as a graveyard for satellite galaxies that have been disrupted through tidal interactions. These disrupted satellites can form long-lived tidal streams; more generally, the stars from these satellites are part of a galaxy’s stellar halo (which may also encompass stars from globular clusters or other sources). Efforts are underway to disentangle disrupted satellites from other stars in the Milky Way halo via chemistry and kinematics (see Bland-Hawthorn & Gerhard 2016 for a recent review).

4.2. Dwarfs beyond the Local Group

An alternate avenue to probing deeper within the Local Group is to search for low-mass galaxies further away (but still in the very local Universe). The Dark Energy Camera (DECam) and Subaru/Hyper Suprime-Cam are being used by several groups to search for very faint companions in a variety of systems (from NGC 3109, itself a dwarf galaxy at ∼ 1.3 Mpc, to Centaurus A, a relatively massive elliptical galaxy at ≈ 3.8 Mpc (Sand et al. 2015b, Crnojević et al. 2016, Carlin et al. 2016). Searches for the gaseous components of galaxies that would otherwise be missed by surveys have also proven fruitful, with a number of individual discoveries (Giovanelli et al. 2013, Sand et al. 2015a, Tollerud et al. 2016).

Recently, the rediscovery of ultra-diffuse dwarf galaxies (Impey, Bothun & Malin 1988, Dalcanton et al. 1997, Koda et al. 2015, van Dokkum et al. 2015) has led to significant interest in these odd systems, which have sizes comparable to the Milky Way but luminosities comparable to bright dwarf galaxies. Ultra-diffuse dwarfs have been discovered predominantly in galaxy clusters, but if similar systems – perhaps with even lower luminosities – exist near the Local Group, they could have escaped detection. Understanding the formation and evolution of ultra-diffuse dwarfs, as well as their dark matter content and connection to the broader galaxy population, has the potential to alter our current understanding of faint stellar systems.

4.3. Searches for starless dwarfs

Very low mass dark matter halos must be starless, should they exist. Detecting starless halos would represent a strong confirmation of the ΛCDM model (and would place stringent constraints on the possible solutions to problems covered in this review); accordingly, astronomers and physicists are exploring a variety of possibilities for detecting such halos.

A promising technique for inferring the presence of the predicted population of low-mass, dark substructure within the Milky Way is through subhalos’ effects on very cold low velocity dispersion stellar streams (Ibata et al. 2002, Carlberg 2009, Yoon, Johnston & Hogg 2011). Dark matter substructure passing through a stream will perturb the orbits of the stars, creating gaps and bunches in the stream. Although many physical phenomena may produce similar effects, and the very existence of gaps themselves remains a matter of debate, large samples of cold streams would likely provide the means to test the abundance of low-mass (M_vir ∼ 10⁵⁻⁶ M_⊙) substructure in the Milky Way. We note that the streams from disrupting satellite galaxies discussed above are not suitable for this technique, as they are produced with large enough stellar velocity dispersions that subhalos’ effects will go undetectable. Blind surveys for HI gas provide yet another path to searching for starless (or extremely faint) substructure in the very nearby Universe. Some ultra-compact high-velocity clouds (UCHVCs) may be gas-bearing “mini-halos” that are devoid of stars (e.g., Blitz et al. 1999).

Most of the probes we have discussed so far rely on electromagnetic signatures of dark matter. Gravitational lensing is unique in that it is sensitive to mass alone, potentially providing a different window into low-mass dark matter halos. Vegeti et al. (2010, 2012) have detected two relatively low-mass dark matter subhalos within lensed galaxies using this technique. The galaxies are at cosmological distances, making it difficult to identify any stellar component associated with the subhalos; Vegetti et al. quote upper limits on the luminosities of detected subhalos of ∼ 5× 10⁶⁻⁷ L_⊙, comparable to classical dwarfs in the Local Group. The inferred dynamical masses are much higher, however: within 300 pc, Milky Way satellites all have M₃₀₀ ≈ 10⁷ M_⊙ (Strigari et al. 2008), while the detected subhalos have M₃₀₀ ≈ (1−10) × 10⁸ M_⊙. It remains to be seen whether this is related to the lens modeling or if the substructure in lensing galaxies is fundamentally different from that in the Local Group.

More recently, ALMA has emerged as a promising tool for detecting dark matter halo substructure via spatially-resolved spectroscopy of lensed galaxies. This technique was discussed in Hezaveh et al. (2013), and recently, a subhalo with a total mass of ∼ 10⁹ M_⊙ within ∼ 1 kpc was detected with ALMA (Hezaveh et al. 2016). At present, the detected substructure is significantly more massive than the hosts of dwarf galaxies in the Local Group: the velocity dispersion of the substructure is σ_DM ∼ 30 km s⁻¹ as opposed to σ_⋆ ≈ 5−10 km s⁻¹ for Local Group dwarf satellites. This value of σ_DM is indicative of a galaxy similar to the Small Magellanic Cloud, which has M_⋆ ∼ 5× 10⁸ M_⊙ and M_vir ∼ (5−10) × 10¹⁰ M_⊙. The discovery of additional lens systems, and the enhanced resolution and sensitivity of ALMA in its completed configuration, promise to reveal lower-mass substructure, perhaps down to scales similar to Local Group satellites but at cosmological distances and in very different host galaxies.

4.4. Indirect signatures of dark matter

If dark matter is indeed a standard WIMP, two dark matter particles can annihilate into Standard Model particles with electromagnetic signatures. This process is exceedingly rare, on average; as discussed in Section 1.6, the freeze-out of dark matter annihilations is what sets the relic density of dark matter in the WIMP paradigm. Nevertheless, the annihilation rate is proportional to the local value ρ_DM², meaning that the centers of dark matter halos are potential sites for annihilations. While the brightest source of such annihilations in the sky should be the Galactic Center, foregrounds make unambiguous detection of annihilating dark matter toward the Galaxy challenging. Dwarf spheroidal galaxies have somewhat lower predicted annihilation fluxes owing both to their greater distances and lower masses, but they have the significant advantage of being free of foreground contamination. The Fermi γ-ray telescope has surveyed MW dwarfs extensively, with no conclusive evidence for dark matter annihilation products. The upper limits on combined dwarf data from Fermi are already placing moderate tension on the most basic “WIMP miracle” predictions for the annihilation cross section for wimps with m ≲ 100 GeV (Ackermann et al. 2015). Searches for annihilation from starless dark matter subhalos within the Milky Way via the Fermi point source catalog have not yielded any detections to date (Calore et al. 2016).

On cosmic scales, dark matter annihilations may contribute to the extragalactic gamma-ray background (Zavala, Springel & Boylan-Kolchin 2010). The expected contributions of dark matter depend sensitively on the spectrum of dark matter halos and subhalos, as well as the relation between concentration and mass for very low mass systems. These relations can be estimated by a variety of methods (though generally not simulated directly, owing to the enormous range of scales that contribute), with uncertainties being grouped into a “boost factor” that describes unresolved annihilations.

If dark matter is a sterile neutrino rather than a WIMP-like particle, self-annihilation will not be seen. Sterile neutrinos decay radiatively to an active neutrino and a photon, however; for all of the relevant sterile neutrino parameter space, this decay is effectively at rest and a clean signature is therefore a spectral line at half the rest mass energy of the dark matter particle, E_γ = m_DM / 2. While there is no a priori expectation for the mass of the sterile neutrino, arguments from Section 3.2.1 point to E_γ ≳ 1 keV, so searches in the soft X-ray band are constraining. The most promising recent result in this field is the detection of a previously unknown X-ray line near 3.51 keV in the spectra of individual galaxy clusters, stacked galaxy clusters, and the halo of M31 (Bulbul et al. 2014, Boyarsky et al. 2014). X-ray observations and satellite counts in M31 rule out an oscillation (Dodelson & Widrow 1994) origin for this line if it indeed originates from sterile neutrino dark matter (Horiuchi et al. 2014), leaving heavy scalar decay and possibly resonant conversion as possible production mechanisms (Merle & Schneider 2015). A definitive test of the origin of the 3.5 keV line was expected from the Hitomi satellite, as it had the requisite energy resolution to see the thermal broadening of the line due to virial motions (i.e., the line width from a halo with mass M_vir should be ∼ V_vir / c). With Hitomi's untimely demise, tests of the line's origin may have to wait for Athena.

4.5. The high-redshift Universe

While studies of low-mass dark matter halos are most easily conducted in the very nearby Universe owing to the faintness of the galaxies they host, there are avenues at higher redshifts that may provide alternate windows in to the spectrum of density perturbations. One potentially powerful probe at z ∼ 2−6 is the Lyman-α forest of absorption lines produced by neutral hydrogen in the intergalactic medium between us and high-redshift quasars (see McQuinn 2016 for a recent review and further details). This hydrogen probes the density field in the quasi-linear regime (i.e., it is in perturbations that are just starting to collapse) and can constrain the dark matter power spectrum to wavenumbers as large as k ∼ 10 h Mpc⁻¹. Any model that reduces the power on this scale relative to ΛCDM expectations will predict different absorption patterns. In particular, WDM will suppress power on these scales.

Viel et al. (2013) used Lyman-α flux power spectra from 25 quasar sightlines to constrain the mass of thermal relic WDM particle to m_{WDM, th} > 3.3 keV at 95% confidence. This translates into a density perturbation spectrum that must be very close to ΛCDM down to M ∼ 10⁸ M_⊙ (Schneider et al. 2012) and would rule out the possibility that free-streaming has direct relevance for the scales of classical dwarfs (and larger-mass systems). The potential complication with this interpretation is the relationship between density and temperature in the intergalactic medium, as pressure or thermal motions can mimic the effects of dark matter free-streaming.

Counts of galaxies in the high-redshift Universe also trace the spectrum of collapsed density perturbations at low masses, albeit in a non-trivial manner. The mere existence of galaxies at high redshift places an upper limit on the free-streaming length of dark matter (so long as all galaxies form within dark matter halos) in much the same way that the existence of substructure in the local Universe does (Schultz et al. 2014). Menci et al. 2016 have placed limits on the masses of thermal relic WDM particles of 2.4 keV (2.1 keV) at 68% (95%) confidence based on the detection of a single galaxy in the Hubble Frontier Fields at z ∼ 6 with absolute UV magnitude of M_UV = −12.5 (Livermore, Finkelstein & Lotz 2017). While this stated constraint is very strong, and the technique is promising, correctly modeling faint, high-redshift galaxies – particularly lensed ones – at can be very challenging. Furthermore, the true redshift of the galaxy can only be localized to Δ z ∼ 1; the rapid evolution of the halo mass function at high redshift further complicates constraints. With the upcoming James Webb Space Telescope, the high-redshift frontier will be pushed fainter and to higher redshifts, raising the possibility of placing strong constraints on the free-streaming length of dark matter through structures in the early Universe.

The challenge of detecting “empty” dark matter halos The detection of abundant, baryon-free, low-mass dark matter halos would be an unambiguous validation of the particle dark matter paradigm, would strongly constrain particle physics models, and would eliminate many of the dark matter candidates for the origin of the small-scale issues described in this review. Why is this such a challenging task? The answer lies in the densities of low-mass dark matter halos compared to other astrophysical objects. From Equation (4), the average density within a halo’s virial radius is 200 times the cosmic matter density. For the most abundant low-mass halos in standard WIMP models – those just above the free-streaming scale of ∼ 10−6 M⊙ – the virial radius is approximately 0.1 pc. This is the equivalent of the mass of the Earth spread over a a distance that is significantly larger than the Solar System (the mean distance between Pluto and the Sun is ∼ 2 × 10−4 pc). Even the lowest-mass, earliest-collapsing CDM structures are incredibly diffuse compared to typical astrophysical objects. Although there may be (1017) Earth-mass dark matter subhalos within the Milky Way's ≈ 300 kpc virial radius, detecting them is a daunting challenge.