Dark Matter Deficient Galaxies in the IllustrisTNG100-1 Cosmological Simulation
Claire Riggs, University of Oklahoma, Astrophysics Major
Mentored by Dr. Bharat Ratra
While galaxies that lack dark matter have been predicted in the 𝛬CDM cosmological model, they had not been observed until 2018, when Van Dokkum et al. 2018 claimed that the ultra-diffuse galaxy NGC10252-DF2 has a ratio of dark matter to baryonic matter of effectively zero (Van Dokkum et al. 2018). A similar object, NGC 1052-DF4, was subsequently discovered the following year (Van Dokkum et al. 2019). Although the validity of these results is still being discussed, the existence of such objects should not come as a surprise if the standard model is correct, since, in a 𝛬CDM cosmology, the dual dwarf galaxy theorem must be true; that is, both dark matter dominated dwarf galaxies formed primordially and dark matter free dwarf galaxies formed via galaxy interactions must simultaneously exist (Kroupa 2012).
The standard, spatially-flat, 𝛬CDM model (Peebles 1984) is consistent with current observational constraints such as the cosmic microwave background (CMB) anisotropies, baryon acoustic oscillation distances, supernova Type Ia apparent magnitudes, and Hubble parameters (Yu et al. 2019). Thus, we assume that the standard 𝛬CDM model is accurate, implying that the dual dwarf galaxy theorem must be true and that dark matter deficient galaxies must exist. We utilize the IllustrisTNG100-1 simulation to probe certain properties of these dark matter deficient galaxies with the ultimate goal of providing insight into how to better observationally detect these galaxies and gaining a clearer understanding of how they form, as observationally discovering more of these galaxies can provide significant insight to our understanding of dark matter.
IllustrisTNG100-1 (www.tng-project.org, Nelson et al. 2019a; Pillipech et al. 2018b; Springel et al. 2018; Nelson et al. 2018a; Naiman et al. 2018; Marinacci et al. 2018) is a rerun of the older Illustris-1 simulation, with updated physics described in Pillipech et al. 2018, along with the implementation of magnetic fields. The Illustris TNG simulations utilize AREPO code (Springel 2010) to account for gravity, hydrodynamics, magnetic fields, along with other effects such as stellar evolution, metal advection, galactic winds, and black hole feedback. Additionally, the IllustrisTNG100-1 simulation assumes a spatially flat 𝛬CDM cosmology with parameter values coming from the Planck Collaboration 2015 data: (Ωm, Ω𝛬, Ωb, σ8, ns, h) = (0.3089, 0.6911, 0.02230, 0.8159, 0.968, 0.6774). The simulation consists of 18203 dark matter particles, 18203 baryonic particles, and 2 x 18203 tracer particles in a box of comoving length (75 Mpc/h)3. Each dark matter particle is 7.5 x 106 solar masses, while the smallest resolvable gas particle is 1.4 x 106 solar masses.
The IllustrisTNG100-1 simulation contains 4,371,211 initial SUBFIND subhalos at redshift z=0, which are gravitationally bound structures that contain stars, gas, and/or dark matter. Initially, we select subhalos with a nonzero stellar mass and check the 'Parent' property of each subhalo, which indicates which subhalo is the most massive in a certain friends-of-friends (FoF) group. However, there are instances where the 'Parent' property is unable to resolve whether or not the subhalo is a substructure in a more massive subhalo, so additional criteria are applied. Thus, for each subhalo, we look at the separation distance d between the subhalo under consideration and its host subhalo. If d is greater than two times the host's stellar half mass radius, then we say that the subhalo is a galaxy. We also ignore any subhalos with less than 20 star particles and a stellar mass less than 107 solar masses, since any subhalos with values smaller than these are more prone to errors associated with resolution. After applying these criteria, we find a total of 79,068 galaxies. The mass distribution of these galaxies is shown in Figure 1.
Figure 1. Each point represents a galaxy, and the dark matter mass and stellar mass are the values from 'SubhaloMassType'. Each galaxy located below the green dashed line is a dark matter deficient galaxy candidate. Additionally, the red points represent galaxies with a ratio of dark matter to stellar mass equal to zero, moved here artificially in order to visualize them.
In order for a galaxy to be dark matter deficient, we look at its ratio of dark matter mass to stellar mass, which are determined by 'SubhaloMassType'. If the ratio of dark matter to stellar mass is less than 1, then this is considered a dark matter deficient galaxy candidate. However, because the 'SubhaloMassType' method may not detect dark matter that may be present in a diffuse halo around the galaxy, we look at the individual particle data associated with each subhalo and integrate out to a distance of 3 times the stellar half mass radius of the galaxy. If, after integrating, the ratio of dark matter to stellar mass is still less than one, we say that this is a dark matter deficient galaxy.
One interesting result that we are still trying to interpret is the increased number of dark matter deficient galaxies found in IllustrisTNG100-1 when compared to Illustris-1. The results are summarized in Figure 2.
Figure 2. Illustris results are from Yu et al. 2019. Although the IllustrisTNG100-1 simulation has less overall galaxies, the probability of finding a dark matter deficient galaxy in the TNG simulation increases by over four percent.
In order to determine if the galaxies in IllustrisTNG100-1 are truly dark matter deficient versus the result of some error on our part, we analyze certain properties of these dark matter deficient galaxies in order to ensure that they match our theoretical predictions about these galaxies.
For example, galaxies that lack dark matter are expected to form as a direct result of galaxy interaction, while dark matter dominated, primordial galaxies tend to form early on in the universe. Primordial galaxies form when cold dark matter collapses into overdense regions, or halos, and the resulting gravitational potential attracts baryonic matter that settles into these halos. However, as these primordial galaxies interact and merge, tidal forces can cause the galaxies to eject gas and stars, which end up forming overdensities in regions surrounding the host galaxy (Halsbauer et al. 2018, Kroupa et al. 2012, Ploeckinger et al. 2017). As a result, the metallicities of these dark matter deficient galaxies should tend to be higher than dark matter dominated galaxies, since they are formed from pre-existing galaxies that have already had time to enrich their gas and stars (Halsbauer et al. 2018, Kroupa et al. 2012). Furthermore, these galaxies tend to be younger when compared to dark matter dominated galaxies since they are formed exclusively from interactions between already existing galaxies (Halsbauer et al. 2018).
Figure 3. The blue histogram shows the number of dark matter dominated galaxies per bin. The TDGCs are galaxies that contain no dark matter, while the DM-poor DGs have a dark matter to stellar mass ratio greater than zero but less than one.
To analyze these galaxies, we split the dark matter deficient galaxies into two groups: the first, tidal dwarf galaxy candidates (TDGCs), have a ratio of dark matter to stellar matter equal to zero. The second are dark matter poor, dark matter containing dwarf galaxies (DM-poor DMC DGs). When we compare the metallicities of the dark matter rich galaxies in our simulation versus the metallicities of the dark matter deficient galaxies, we find that the dark matter deficient galaxies tend to have higher gas and stellar metallicities (see Figure 3). While most of the dark matter deficient galaxies have higher metallicities than the dark matter dominated objects, there are a few galaxies with low metallicity. However, this does not necessarily imply that these objects are not dark matter deficient, as they may have been formed from the outer, metal-poor region of a gas-rich galaxy, implying that they may begin with a lower metallicity (Kroupa 2012).
Figure 4. This shows the distribution of dark matter dominated galaxies versus the distribution of dark matter deficient galaxies, split into TDGCs and dark matter poor DMC DGs. As expected, the dark matter rich galaxies tend to be much older than the dark matter deficient galaxies, suggesting that the dark matter rich galaxies are formed primordially early on in the universe, while the dark matter deficient galaxies are formed later on through galaxy interactions.
Looking at the time distribution of dark matter rich galaxies versus dark matter deficient galaxies (see Figure 4), it is clear that the dark matter deficient galaxies are younger than the dark matter dominated galaxies, which matches our predictions. Additionally, after analyzing the distance between the stellar half mass radius of the DM-deficient galaxy's host and itself and found that both populations of dark matter deficient galaxies tend to fall close to their host galaxy (see Figure 5).
Figure 5. Distance distribution of the two populations of dark matter deficient galaxies (TDGCs and DM-poor DMCs). We expect dark matter deficient galaxies to be near their host since they are the result of galaxy interactions.
I will be continuing this research throughout the year, with the results to be submitted for publication. As of now, the main goal is to determine if there are any significant differences between the two types of galaxies, as well as determine if different selection criteria significantly changes the results.
Hai, Y., Ratra, B., & Wang, F. 2018 [arXiv:1809.05938]
Haslbauer, M., et al., Galaxies lacking dark matter in the Illustris simulation. Astronomy and Astrophysics, 2019. 626. [arxiv: 1905.03258]
Kroupa, P., The Dark Matter Crisis: Falsification of the Current Standard Model of Cosmology. Publications of the Astronomical Society of Australia, 2012. 29: p. 395-433. [arXiv:1204.2546]
Marinacci, F., et al., First results from the IllustrisTNG simulations: radio haloes and magnetic fields. Monthly Notices of the Royal Astronomical Society, 2018. 480: p. 5113-5139. [arXiv: 1707.03396]
Naiman, J.P., et al., First results from the IllustrisTNG simulations: a tale of two elements - chemical evolution of magnesium and europium. Monthly Notices of the Royal Astronomical Society, 2018. 477: p. 1206-1224. [arXiv:1707.03401]
Nelson, D., et al., First results from the IllustrisTNG simulations: the galaxy colour bimodality. Monthly Notices of the Royal Astronomical Society, 2018. 475: p. 624-647. [arXiv:1707.03395]
Nelson, D., et al., The IllustrisTNG simulations: public data release. Computational Astrophysics and Cosmology, 2019. 6. [arXiv:1812.05609]
Peebles, P.J.E., Dark matter and the origin of galaxies and globular star clusters. The Astrophysical Journal, 1984. 277: p. 470-477.
Pillepich, A., et al., First results from the IllustrisTNG simulations: the stellar mass content of groups and clusters of galaxies. Monthly Notices of the Royal Astronomical Society, 2018. 475: p. 648-675. [arXiv: 1707.03406]
Springel, V., E pur si muove: Galilean-invariant cosmological hydrodynamical simulations on a moving mesh. Monthly Notices of the Royal Astronomical Society, 2010. 401: p. 791-851. [arXiv:0901.4107]
Springel, V., et al., First results from the IllustrisTNG simulations: matter and galaxy clustering. Monthly Notices of the Royal Astronomical Society, 2018. 475: p. 676-698. [arXiv: 1707.03397]
van Dokkum, P., et al., A galaxy lacking dark matter. Nature, 2018. 555: p. 629-632. [arXiv:1803.10237]
van Dokkum, P., et al., A Second Galaxy Missing Dark Matter in the NGC 1052 Group. The Astrophysical Journal, 2019. 874. [arXiv:1901.05973]
I would like to thank my mentor, Dr. Bharat Ratra and grad students Hai Yu and Harish Murali for their help and guidance on this project. I would also like to thank the TNG Collaboration for providing my data, the National Science Foundation for funding, Kansas State University, and Dr. Bret Flanders, Dr. Loren Greenman, and Kim Coy for coordinating this REU program.