RNAAS
Einasto gravitational potentials have difficulty to hold spherically
symmetric stellar systems with cores

Jorge Sánchez Almeida Instituto de Astrofísica de Canarias, La Laguna, Tenerife, E-38200, Spain Departamento de Astrofísica, Universidad de La Laguna
(Received June 19, 2024)
Abstract

It was known that an ideal spherically symmetric stellar system with isotropic velocities and an inner core cannot reside in a Navarro, Frenk, and White (NFW) gravitational potential. The incompatibility can be pinned down to the radial gradient of the NFW potential in the very center of the system, which differs from zero. The gradient is identically zero in an Einasto potential, also an alternative representation of the dark matter (DM) halos created by the kind of cold DM (CDM) defining the current cosmological model. Here we show that, despite the inner gradient being zero, stellar cores are also inconsistent with Einasto potentials. This result may have implications to constrain the nature of DM through interpreting the stellar cores often observed in dwarf galaxies.

Cold dark matter (265) — Dark matter (353) — Dark matter distribution (356) — Dwarf galaxies (416)
software: Scipy (Virtanen et al., 2020)

1 Discusion

Refer to caption
Figure 1: (a) Stellar density profile with a core (the black line) together with the mass density giving rise to a NFW potential (the red line) and an Einasto potential (the blue line). (b) Distribution function required from the stellar profile to reside in the corresponding NFW potential (the red line) and in the Einasto potential (the blue line). In both cases, f(ϵ)𝑓italic-ϵf(\epsilon)italic_f ( italic_ϵ ) becomes negative (the dashed lines). The parameters were chosen to represent a dwarf galaxy with M5×105Msimilar-to-or-equalssubscript𝑀5superscript105subscript𝑀direct-productM_{\star}\simeq 5\times 10^{5}\,M_{\odot}italic_M start_POSTSUBSCRIPT ⋆ end_POSTSUBSCRIPT ≃ 5 × 10 start_POSTSUPERSCRIPT 5 end_POSTSUPERSCRIPT italic_M start_POSTSUBSCRIPT ⊙ end_POSTSUBSCRIPT.

Sánchez Almeida et al. (2023) showed that an ideal spherical stellar distribution having an inner core, i.e., having

limr0dρdr=0,subscript𝑟0𝑑𝜌𝑑𝑟0\lim_{r\to 0}\frac{d\rho}{dr}=0,roman_lim start_POSTSUBSCRIPT italic_r → 0 end_POSTSUBSCRIPT divide start_ARG italic_d italic_ρ end_ARG start_ARG italic_d italic_r end_ARG = 0 , (1)

is not consistent with a NFW potential (named after Navarro et al., 1997). The symbol ρ𝜌\rhoitalic_ρ stands for the stellar mass volume density, with r𝑟ritalic_r the distance to the center of the gravitating system. This result would be only a curiosity if it were not for the fact that stellar cores are very common in dwarf galaxies (e.g., Moskowitz & Walker, 2020; Richstein et al., 2024; Montes et al., 2024). Even if it is an academic result for ideal systems, the existence of cores seems to suggest the gravitational potential to differ from the NFW potential characteristic of the CDM particles defining the current cosmological model. Deviations from the NFW potential are well known from kinematical measurements (e.g., Oh et al., 2015) and they are ascribed to the effect of stellar feedback processes modifying the overall gravitational potential (e.g., Governato et al., 2010). This baryon driven feedback is expected to be inoperative for galaxies with stellar mass M106Mmuch-less-thansubscript𝑀superscript106subscript𝑀direct-productM_{\star}\ll 10^{6}\,M_{\odot}italic_M start_POSTSUBSCRIPT ⋆ end_POSTSUBSCRIPT ≪ 10 start_POSTSUPERSCRIPT 6 end_POSTSUPERSCRIPT italic_M start_POSTSUBSCRIPT ⊙ end_POSTSUBSCRIPT (e.g., Chan et al., 2015; Jackson et al., 2021), therefore, the existence of stellar cores in tiny galaxies would reflect the need to go beyond the CDM paradigm.

Sánchez Almeida et al. (2023) proved the inconsistency using the Eddington Inversion Method (EIM; e.g., Binney & Tremaine, 2008) which provides the distribution function in the phase space f(ϵ)𝑓italic-ϵf(\epsilon)italic_f ( italic_ϵ ) needed to reproduce a stellar distribution in an assumed gravitational potential. For a cored stellar profile to reside in a NFW potential f(ϵ)𝑓italic-ϵf(\epsilon)italic_f ( italic_ϵ ) has to be negative, which is unphysical. The proof employs an intermediate step in the EIM formalism, which reads,

dρ/drdΨ/dr=2π20Ψf(ϵ)Ψϵ𝑑ϵ,𝑑𝜌𝑑𝑟𝑑Ψ𝑑𝑟2𝜋2superscriptsubscript0Ψ𝑓italic-ϵΨitalic-ϵdifferential-ditalic-ϵ\frac{d\rho/dr}{d\Psi/dr}=2\pi\sqrt{2}\,\int_{0}^{\Psi}\frac{f(\epsilon)}{% \sqrt{\Psi-\epsilon}}\,d\epsilon,divide start_ARG italic_d italic_ρ / italic_d italic_r end_ARG start_ARG italic_d roman_Ψ / italic_d italic_r end_ARG = 2 italic_π square-root start_ARG 2 end_ARG ∫ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT roman_Ψ end_POSTSUPERSCRIPT divide start_ARG italic_f ( italic_ϵ ) end_ARG start_ARG square-root start_ARG roman_Ψ - italic_ϵ end_ARG end_ARG italic_d italic_ϵ , (2)

with Ψ(r)Ψ𝑟\Psi(r)roman_Ψ ( italic_r ) the relative gravitational potential and ϵitalic-ϵ\epsilonitalic_ϵ the total energy per unit mass of a star. In the case of a NFW potential,

limr0dΨNFWdr0,subscript𝑟0𝑑subscriptΨNFW𝑑𝑟0\lim_{r\to 0}\frac{d\Psi_{\rm NFW}}{dr}\not=0,roman_lim start_POSTSUBSCRIPT italic_r → 0 end_POSTSUBSCRIPT divide start_ARG italic_d roman_Ψ start_POSTSUBSCRIPT roman_NFW end_POSTSUBSCRIPT end_ARG start_ARG italic_d italic_r end_ARG ≠ 0 , (3)

which together with Eq. (1) implies the integral in the right-hand-side of Eq. (2) to be zero when r0𝑟0r\to 0italic_r → 0. For the integral to be zero f(ϵ)𝑓italic-ϵf(\epsilon)italic_f ( italic_ϵ ) must be negative somewhere in the phase space, which is unphysical and evidences the incompatibility.

The CDM halos giving rise to a NFW potential are equally well represented by Einasto profiles (e.g., Wang et al., 2020), where the derivative is zero (e.g., Retana-Montenegro et al., 2012)

limr0dΨEinastodr=0subscript𝑟0𝑑subscriptΨEinasto𝑑𝑟0\lim_{r\to 0}\frac{d\Psi_{\rm Einasto}}{dr}=0roman_lim start_POSTSUBSCRIPT italic_r → 0 end_POSTSUBSCRIPT divide start_ARG italic_d roman_Ψ start_POSTSUBSCRIPT roman_Einasto end_POSTSUBSCRIPT end_ARG start_ARG italic_d italic_r end_ARG = 0 (4)

and, consequently, for which the above argument of incompatibility does not hold. The question arises as to whether the incompatibility of stellar cores is specific to the NFW potential or if it more generally affects all CDM-produced halos. Recall that the NFW potential stems from a mass density profile scaling as r1superscript𝑟1r^{-1}italic_r start_POSTSUPERSCRIPT - 1 end_POSTSUPERSCRIPT, therefore, with an artificial singularity when r0𝑟0r\to 0italic_r → 0. The singularity is not present in the Einasto profile which scales as exp[(r/h)α]superscript𝑟𝛼\exp\left[\left(r/h\right)^{\alpha}\right]roman_exp [ ( italic_r / italic_h ) start_POSTSUPERSCRIPT italic_α end_POSTSUPERSCRIPT ] with α𝛼\alphaitalic_α and hhitalic_h two positive constants. This Note is meant to show that the inconsistency still remains in the case of an Einasto potential. In terms of Eq. (2), the Einasto profile has

dΨEinastodr0,𝑑subscriptΨEinasto𝑑𝑟0\frac{d\Psi_{\rm Einasto}}{dr}\not=0,divide start_ARG italic_d roman_Ψ start_POSTSUBSCRIPT roman_Einasto end_POSTSUBSCRIPT end_ARG start_ARG italic_d italic_r end_ARG ≠ 0 , (5)

for values of r0𝑟0r\not=0italic_r ≠ 0 where

dρdr0,similar-to-or-equals𝑑𝜌𝑑𝑟0\frac{d\rho}{dr}\simeq 0,divide start_ARG italic_d italic_ρ end_ARG start_ARG italic_d italic_r end_ARG ≃ 0 , (6)

and this condition is enough to force the need of f(ϵ)<0𝑓italic-ϵ0f(\epsilon)<0italic_f ( italic_ϵ ) < 0.

The argument is illustrated with the simulation in Fig. 1, which shows a cored stellar profile (the black line; a Plummer profile) with size and central density chosen to mimic a dwarf galaxy. The red line shows the NFW profile used to compute the potential giving rise to the distribution function shown in Fig. 1b (the red line). It was evaluated with the numerical tools developed by Sánchez Almeida et al. (2023). As expected, there is a range of values for which f(ϵ)<0𝑓italic-ϵ0f(\epsilon)<0italic_f ( italic_ϵ ) < 0. The same exercise is carried out with an Einasto profile having the same characteristic radius and density as the NFW profile, as in CDM simulations (e.g., Wang et al., 2020). Firstly, we note that the corresponding f(ϵ)𝑓italic-ϵf(\epsilon)italic_f ( italic_ϵ ) also has negative parts (the blue dashed line in Fig. 1b), making also the Einasto potential inconsistent with the stellar core. Secondly, and in contrast with the NFW potential, the distribution function for the Einasto potential becomes positive near the largest energy. Such energies are reached only by the stars moving close to the center of the potential at all times. Thus, the incompatibility is not at r0𝑟0r\to 0italic_r → 0 but at a finite radius within the core (Eq. [6]) but where the potential keeps changing (Eq. [5]).

2 Conclusions

The original inconsistency happens for spherically symmetric systems having isotropic velocities, and representing the CDM with NFW profiles. However, none of these simplifying assumptions seem to be responsible for the incompatibility. Sánchez Almeida et al. (2023) showed that the incompatibility remains for other potentials with an inner slope much shallower than the value of 11-1- 1 characteristic of the NFW profiles. They also showed that the assumption of isotropic velocities can be relaxed since the incompatibility remains for radial orbits and for isotropic orbits that turn into radial in the outskirts, as expected from cosmological numerical simulations (e.g., El-Badry et al., 2017; Orkney et al., 2023). On the other hand, Sánchez Almeida et al. (2024) proved that the incompatibility persists for axi-symmetric systems, which implies that it goes beyond the assumption of spherical symmetry. This Note adds that the incompatibility remains for Einasto potentials, which are also a good representation of the CDM halos and, in contrast to the NFW profiles, do not diverge at the center of the system. Thus, the incompatibility is not artificially set by the mathematical singularity at r=0𝑟0r=0italic_r = 0 present in the NFW profiles.

Thanks are due to Ignacio Trujillo and Angel Plastino for many insightful discussions on how the properties of the gravitational potential are constrained by the observed starlight. The Spanish Ministry of Science and Innovation supported the work through project PID2022-136598NB-C31 (ESTALLIDOS).

References

  • Binney & Tremaine (2008) Binney, J., & Tremaine, S. 2008, Galactic Dynamics: Second Edition
  • Chan et al. (2015) Chan, T. K., Kereš, D., Oñorbe, J., et al. 2015, MNRAS, 454, 2981
  • El-Badry et al. (2017) El-Badry, K., Wetzel, A. R., Geha, M., et al. 2017, ApJ, 835, 193
  • Governato et al. (2010) Governato, F., Brook, C., Mayer, L., et al. 2010, Nature, 463, 203
  • Jackson et al. (2021) Jackson, R. A., Martin, G., Kaviraj, S., et al. 2021, MNRAS, 502, 4262
  • Montes et al. (2024) Montes, M., Trujillo, I., Karunakaran, A., et al. 2024, A&A, 681, A15
  • Moskowitz & Walker (2020) Moskowitz, A. G., & Walker, M. G. 2020, ApJ, 892, 27
  • Navarro et al. (1997) Navarro, J. F., Frenk, C. S., & White, S. D. M. 1997, ApJ, 490, 493
  • Oh et al. (2015) Oh, S.-H., Hunter, D. A., Brinks, E., et al. 2015, AJ, 149, 180
  • Orkney et al. (2023) Orkney, M. D. A., Taylor, E., Read, J. I., et al. 2023, MNRAS, 525, 3516
  • Retana-Montenegro et al. (2012) Retana-Montenegro, E., van Hese, E., Gentile, G., Baes, M., & Frutos-Alfaro, F. 2012, A&A, 540, A70
  • Richstein et al. (2024) Richstein, H., Kallivayalil, N., Simon, J. D., et al. 2024, ApJ, 967, 72
  • Sánchez Almeida et al. (2023) Sánchez Almeida, J., Plastino, A. R., & Trujillo, I. 2023, ApJ, 954, 153
  • Sánchez Almeida et al. (2024) —. 2024, A&A, submitted, …
  • Virtanen et al. (2020) Virtanen, P., Gommers, R., Oliphant, T. E., et al. 2020, Nature Methods, 17, 261
  • Wang et al. (2020) Wang, J., Bose, S., Frenk, C. S., et al. 2020, Nature, 585, 39