Arxiv:2007.02213V2 [Astro-Ph.CO] 15 Apr 2021 Ple Assumptions and Are Subject to Signiﬁcant Uncertain- While We Do Not Discuss These Eﬀects in Detail Within Our Ties

IPMU20-0076 RIKEN-iTHEMS-Report-20 Constraining Primordial Black Holes with Dwarf Galaxy Heating

Philip Lu,1, ∗ Volodymyr Takhistov,1, 2, † Graciela B. Gelmini,1, ‡ Kohei Hayashi,3, 4, § Yoshiyuki Inoue,5, 6, 2, ¶ and Alexander Kusenko1, 2, ∗∗ 1Department of Physics and Astronomy, University of California, Los Angeles Los Angeles, California, 90095-1547, USA 2Kavli Institute for the Physics and Mathematics of the Universe (WPI), UTIAS The University of Tokyo, Kashiwa, Chiba 277-8583, Japan 3Institute for Cosmic Ray Research, The University of Tokyo, Kashiwa 277-8582, Japan 4Astronomical Institute, Tohoku University, Sendai 980-8582, Japan 5Department of Earth and Space Science, Graduate School of Science, Osaka University, Toyonaka, Osaka 560-0043, Japan 6Interdisciplinary Theoretical & Mathematical Science Program (iTHEMS), RIKEN, 2-1 Hirosawa, Saitama 351-0198, Japan (Dated: April 16, 2021) Black holes formed in the early universe, prior to the formation of stars, can exist as dark matter and also contribute to the black hole merger events observed in gravitational waves. We set a new limit on the abundance of primordial black holes (PBHs) by considering interactions of PBHs with the interstellar medium, which result in the heating of gas. We examine generic heating mechanisms, including emission from the accretion disk, dynamical friction, and disk outﬂows. Using the data from the Leo T dwarf galaxy, we set a new cosmology-independent limit on the abundance of PBHs 7 in the mass range O(1)M − 10 M , relevant for the recently detected gravitational wave signals from intermediate-mass BHs.

INTRODUCTION has the desired properties, such as a low baryon velocity dispersion, making it a sensitive probe of PBH heating. Primordial black holes (PBHs) can form in the early While constraints on BHs interacting with surrounding Universe through a variety of mechanisms and can ac- stars have been extensively discussed [48–50], gas heat- count for all or part of the dark matter (DM) (e.g. [1– ing has not been considered in detail. Other constraints 26]). PBHs surviving until present can span many orders focused on the X-ray emission, but not the heating of 15 10 of magnitude in mass, from 10 g to well over 10 M , the surrounding gas [51, 52]. ISM heating has been used and they can account for the entirety of the DM in the to constrain particle DM candidates [53–55], which have −16 −10 mass window ∼ 10 − 10 M , where there are no different heating mechanisms with a different velocity de- observational constraints [27–29]. PBHs with sublunar pendence compared to PBHs. masses can play a role in the synthesis of heavy elements, Additionally, emission from PBH accretion can result production of positrons, as well as other astrophysical in other observables and affect cosmological history, in- phenomena [30–33]. PBHs with larger masses can ac- cluding modifications to reionization and recombination. count for some of the gravitational wave events detected As discussed in [56], ionization from X-rays due to PBH by LIGO (e.g. [34–36]) as well as seeds of supermassive accretion will increase the amount of molecular hydrogen 3 black holes [37–39]. The mass window of 10 − 10 M and the resultant cooling will enhance early star forma- is particularly intriguing in connection with signals ob- tion. PBH emission will also result in spectral distor- served by LIGO [39–45]. Very recently, LIGO announced tions of the cosmic microwave background (CMB). Fur- first detection of an intermediate-mass 150M BH [46]. thermore, heating and ionization of intergalactic medium While a variety of constraints exist for this PBH mass arising from PBH emission will modify 21 cm power spec- range (see Ref. [47] for review), they often rely on multi- trum and will be probed by upcoming experiments [57]. arXiv:2007.02213v2 [astro-ph.CO] 15 Apr 2021 ple assumptions and are subject to significant uncertain- While we do not discuss these effects in detail within our ties. work, they provide complementary probes of PBHs. In this work, we set new cosmology-independent constraints on PBH abundance based on the lack of gas heating from PBH interactions with the interstellar medium (ISM). We consider several generic heating mechanisms, including dynamical friction, accretion disk emission as BLACK HOLES IN INTERSTELLAR MEDIUM well as mass outflows/winds from accretion disk. We then apply our analysis to dwarf DM-rich galaxies, focusing on Leo T. Leo T is a transitional object between The accretion of gas onto freely floating BHs has been a dwarf irregular galaxy and a dwarf spheroidal galaxy analyzed in [58] and applied to PBHs in [52]. Bondi- that has been well studied and modeled theoretically. It Hoyle-Lyttleton accretion results in the mass accretion 2 rate1 [23, 60, 61] GAS HEATING

2 2 ˙ 2 4πG M nµmp For a gas system in thermal equilibrium, the total M = 4πrBvρ˜ = , (1) v˜3 amount of heating by PBHs of mass M is

2 where M is the PBH mass, rB = GM/v˜ is the Bondi H (M) = N (M)H(M) (4) radius, µ is the mean molecular weight, n is the ISM gas tot PBH Z nmax Z vmax 2 dfn dfv number density, mp is the proton mass andv ˜ ≡ (v + = dndv H(M, n, v) , 2 1/2 dn dv cs) . Here, v is the PBH speed relative to the gas nmin vmin and cs is the temperature-dependent sound speed in gas, where df /dn is the gas density distribution, df /dv is which we take to be cs ∼ 10 km/s [52]. n v The accretion rate can be related to the bolometric the PBH relative speed distribution and H(M, n, v) is emission luminosity as L = (M˙ )M˙ , where (M˙ ) is the the amount of heat deposited into the system from a sin- radiative efficiency which scales with accretion rate. The gle PBH. Here, H represents the cumulative contribution Eddington accretion rate, assuming a characteristic ra- from all heating processes. For photon emission and outflows we perform an additional integration to treat the diative efficiency of 0 = 0.1, is defined in terms of 2 absorption efficiency. For gas of approximately constant the Eddington luminosity M˙ Edd = LEdd/0c . A con- venient parameter for characterizing the accretion flow is density, one can replace dfn/dn by a delta function. First, we consider gas heating due to photon emission m˙ = M/˙ M˙ Edd. With a sufficient angular momentum, the infalling gas from accretion. Emission in the X-ray band generally can form an accretion disk around the BH. The angular constitutes the dominant contribution and it becomes momentum necessary for a disk formation can be sup- more efficient at high mass accretion rates. plied by perturbations in the density or the velocity of Photon emission from accretion depends on the accre- the accreting gas. For a Schwarzschild BH, the inner tion flow. To characterize the accretion flow, we fol- radius of the disk is taken to be the innermost stable cir- low the scheme outlined in [62] and assume that the accretion flow results in a (geometrically) thin disk for cular orbit (ISCO) of a test particle rISCO = 3rs, where 2 m˙ > M˙ = 0.07α. The thin α-disk is the so-called stan- rs = 2GM/c is the Schwarzschild radius. Following the arguments of [52, 58], we have confirmed that an accre- dard disk [63], where α ∼ 0.1 is a phenomenological tion disk always forms for our parameters of interest. parameter describing viscosity. A thin disk is optically thick and efficiently emits blackbody radiation. Thin disk If PBHs constitute a fraction fPBH of the DM, the total number of PBHs of mass M within a volume V is emission allows for a fully analytic description, and we employ the scaling characterization of [64]. ρDMV For accretion rates with efficiency below the thin disk N (M) = f , (2) PBH PBH M regime, accretion is described by the advection dominated accretion flow (ADAF) [62, 65]. Here, the heat where ρDM is the DM density, assumed to be approx- generated by viscosity during accretion is not efficiently imately constant. We assume a monochromatic PBH radiated out, and much of the energy is advected via mat- mass function for definiteness and for presenting our re- ter heat capture into the BH event horizon along with sults in the form of a differential exclusion plot. The ve- the gas inflow. In contrast with the thin disk, the ADAF locity of PBHs contributing to the DM can be described “disk” is geometrically thick and optically thin. by a Maxwell-Boltzmann distribution An ADAF disk results in a complicated multi- component emission spectrum. We consider three com- r 2 2 2 v v ponents of the ADAF spectrum, arising from electron fv(v) = 3 exp − 2 , (3) π σv 2σv cooling: synchrotron radiation, inverse Compton scat- tering and bremsstrahlung. To describe the ADAF spec- where σv is the velocity dispersion in a given system. trum, we employ approximate analytic expressions ob- A distribution in gas number density fn(n) can also be tained in [66] in combination with the updated values for introduced, as in [52, 58]. the phenomenological input parameters consistent with recent numerical simulations and observations [62]. We take the ratio of direct viscous heating to electrons and ions δ = 0.3, and the ratio of gas pressure to total pres- 1 Recent 3D hydrodynamical simulations show that the accretion sure β = 10/11. rate at high Mach number would be limited to ∼ 10–20% of the We do not consider a slim disk or other solutions for canonical Bondi-Hoyle-Lyttleton accretion rate of Eq. (1) [59]. Since this depends on the assumed conditions and the dominant near- or super-Eddington accretion,m ˙ ∼ 1, because such constraints in our study rely on low-velocity regime, throughout high accretion rates are not achieved for PBH masses and this work we adopt the canonical rate. gas densities that we discuss. 3

Emitted photons heat the ISM. Hydrogen gas is op- where rout is the outer radius and the real index s, tically thin to radiation below the ionization threshold 0 ≤ s < 1 is limited by energy and mass conserva- of Ei = 13.6 eV, and the velocity dispersion is not high tion. There is a significant uncertainty in the description enough for a significant Doppler broadening of the emis- of the outflows. We vary the exponent s in the range sion spectra. Thus, we ignore the absorption of photons 0.5 − 0.7 [62, 74], in agreement with numerical simula- with energies less than Ei. tions. Furthermore, we consider the outer radius rout If the medium is optically thick, the photons are over a wide range of values, from 100rs [75] to rB [76]. absorbed, and most of their energy is deposited as The resulting outgoing wind has a velocity that is a frac- heat. For absorption of photons with E > Ei, we tion fk ' 0.1−0.2 [74, 77, 78] of the Keplerian velocity at p use the photo-ionization cross-section [67, 68], σ(E) = the radius at which it is ejected, i.e. v(r) ' fk GM/r. −4 − 3 1 We note that additional considerations regarding details σ y 2 1 + y 2 , where y = E/E , E = 1/2E and 0 0 0 i of accretion may reduce emission efficiency (e.g. feed- −16 2 σ0 = 6.06 × 10 cm . The optical depth of a gas sys- back), but we do not expect this to be very significant. tem of size l and density n is τ(n, E) = σ(E)nl. Above To evaluate how much energy is deposited into the gas 30 eV, we use the combined attenuation length data from system from streaming outflow protons, we convolve the Fig. (32.16) of [69]. The resulting heating power is proton emission with the heat generated per proton ∆E. The total heat deposited in the gas system is Z Emax H (M, n, v) = L (M, n, v)f 1 − e−τ dν , phot ν h Z rout fh∆E dM˙ out Ei H = dr , (8) (5) out rin µmp dr where Lν (ν) is the luminosity for the corresponding pho- where ton emission process and fh is the fraction of energy deposited as heat that we estimate to be ∼ 1/3 [70]. Z dE ∆E = dx ' min(E, nS(E)r ) (9) For both the ADAF and thin disk regimes, the emis- dx max sion spectrum is exponentially decreasing at high energies, and we evaluate the integral up to the maximum takes into account energy losses due to the proton stop- ping power dE/dx = nS(E) adopted from [79] (see their energy Emax = ∞. The second contribution to gas heating that we con- Fig. 9). Here, rmax is taken to be the size of the gas sider is dynamical friction due to gravitational interac- system. tions of traversing PBHs with the surrounding medium. Dynamical friction can be described as work done by the TARGET SYSTEM: LEO T “gravitational drag” force Fdyn (see e.g. [71, 72]). The resulting power deposited as heat is We demonstrate our analysis by applying it to dwarf 4πG2M 2ρ galaxies, focusing on the Leo T dwarf galaxy. We stress, H = F v = − I, (6) dyn dyn v however, that our methods are general and can be readily applied to other systems as well. To constrain the PBH where G is the gravitational constant, ρ is the gas den- mass fraction fPBH, we consider the balance between the sity and I is a velocity-dependent geometrical factor that heating and cooling processes of the gas system. Our differs if the medium is collisionless or not [71, 72]. We approach to set the limits is similar to that used for par- have confirmed that the effect of the dynamical friction ticle DM [53–55], but the heating mechanisms and the on the PBH velocity is small. preferred gas systems are different in our case. As a third heating component, mass outflows (winds) For simplicity, we ignore the contribution of natural composed of protons can also contribute and they are ex- heating sources (e.g. stellar radiation), and hence our pected to be significant for hot accretion flows [62]. In bounds are conservative. Requiring thermal equilibrium, 2 contrast to jets , the outflows are not highly relativis- we only consider gas systems that are expected to be tic and cover a wider angular distribution. The out- approximately stable on sufficiently long timescales τsys. flows reduce the accretion rate at smaller radii and can Hence, the characteristic time over which the gas sys- be approximately modelled by a self-similar power-law tem remains steady must be greater than the cooling form [73] timescale of the gas τtherm, i.e. τsys τtherm = 3nkT/2C˙ , ˙ r s where k is the Boltzmann constant and C is the gas cool- M˙ (r) = M˙ (r ) , (7) ing rate per volume. We note that presence of magnetic out in out r out fields in the ISM can also affect emitted protons that we consider. However, the strength, orientation and distribution of magnetic fields in Leo T is highly uncertain 2 As jets are typically associated with Kerr black holes, they would and very poorly known. Hence, we do not consider such require a separate treatment and we do not consider them here. effects. 4

100 100

1 1 I 10− 10− X/R

2 2 10− 10− S Ly-α LSS allowed allowed

f 3 f 3 10− 10− XRB

DF 4 4 10− 10− P

5 5 10− 1 0 1 2 3 4 5 6 7 8 10− 1 0 1 2 3 4 5 6 7 8 10− 10 10 10 10 10 10 10 10 10 10− 10 10 10 10 10 10 10 10 10 M(M ) M(M )

FIG. 1. Left: Constraints from Leo T on the fraction of DM in PBHs, for a monochromatic mass function, derived from considerations of only photon emission (red), dynamical friction (green), mass outﬂows (blue), as well as combined heating (dashed black). The reach of the constraints is bounded by the diagonal (solid black) line from the condition of Eq. (12). The uncertainty in the emission and outﬂow input parameters leads to the uncertainty in the corresponding constraint (upper and lower dashed black lines). Right: Constraints from the Leo T dwarf galaxy on the PBH gas heating are shown in blue. The light blue shaded band denotes the variation in the PBH emission parameters. Other existing constraints (see Ref. [47]) are shown by dashed lines, including Icarus [80] (I) caustic crossing in purple, Planck [81, 82] (P) in yellow, X-ray binaries [52] (XRB) in green, dynamical friction of halo objects (DF) in red, Lyman-α [83] (Ly-α) in maroon, combined bounds from the survival of astrophysical systems in Eridanus II [84], Segue 1 [85], and disruption of wide binaries [86] (S) shown in magenta, large scale structure [21] (LSS) in cyan, and X-ray/radio [56] (X/R) in brown.

3 Gas temperature exchange is a complex process, and a fPBH ρDM(4πrsys/3)/M < 1, such a system cannot be detailed analysis involving a full chemistry network can used for our purposes. We, therefore, set a limit only as be performed using numerical methods [87]. For the pa- long as rameters of interest, we employ approximate results ob- tained in [55]. For hydrogen gas, the cooling rate is given 3M fbound > 3 . (12) by 4πrsysρDM ˙ 2 [Fe/H] C = n 10 Λ(T ) , (10) The gas in the inner region of Leo T, at a radius r 350 pc from its center, is dominated by atomic hy- where [Fe/H]≡ log10(nFe/nH)gas − log10(nFe/nH)Sun is . the metallicity, and Λ(T ) ∝ 10[Fe/H] is the cooling func- drogen, while the gas outside is highly ionized [88]. Since tion. Fitting numerically to the results of [87] li- the free electrons in the ionized region cool very eﬃ- brary, one can obtain Λ(T ) = 2.51 × 10−28T 0.6, valid ciently [87], we limit our analysis to the central region for 300 K < T < 8000 K [55]. of Leo T. From the model of [88], the hydrogen gas density is found to vary from ∼ 0.2 cm−3 in the center to The total PBH heating in the cloud of gas Htot = ∼ 0.03 cm−3 at r = 350 pc. Both the cooling and heating NPBHH(M) = fPBHρDMVH(M)/M given by Eq. (4), 2 where H(M) is the average heat generated from one PBH rates scale roughly as n , so we approximate the gas den- −3 of mass M, should be less than the total cooling CV˙ . sity to be a constant n = 0.07 cm in the inner region. This yields a condition on the PBH abundance that we Similarly, we approximate the DM mass density to be a 3 use to set our limits: constant value of 1.75 GeV/cm . The hydrogen gas has a dominant non-rotating warm component with a velocity MC˙ fPBH < fbound = . (11) dispersion of σg = 6.9 km/s and T ' 6000 K [88, 89] ρDMH(M) and also a sub-dominant cold component that we ig- We note that gas heating can be used to set a limit nore. The DM is expected to have the same velocity on the PBH abundance only if it is statistically likely dispersion as the gas, σv = σg. The sound speed is for the gas system to harbor PBHs. If the PBH num- taken to be cs = 9 km/s from the adiabatic formula with ber density is so low that, on average, a gas sys- T ' 6000 K. Combining the radius and number density, tem of the size rsys contains fewer than one PBH, i.e., the column density of hydrogen gas in the central region 5

19 −2 of Leo T is nrsys = 7.56 × 10 cm . We adopt the [3] B. J. Carr and S. W. Hawking, Mon. Not. Roy. Astron. gas metallicity to approximately follow the stellar one3, Soc. 168, 399 (1974). [Fe/H] ' −2 estimated by stellar spectra [90]. Using the [4] J. Garcia-Bellido, A. D. Linde, and D. Wands, Phys. above parameters in Eq. (10), the resulting Leo T’s cool- Rev. D54, 6040 (1996), arXiv:astro-ph/9605094 [astro- ˙ −30 −3 −1 ph]. ing rate is taken to be C = 2.28 × 10 erg cm s . [5] M. Yu. Khlopov, Res. Astron. Astrophys. 10, 495 (2010), In Fig.1 we display the resulting limits from gas heat- arXiv:0801.0116 [astro-ph]. ing in Leo T on PBHs contributing to DM, along with [6] P. H. Frampton, M. Kawasaki, F. Takahashi, and T. T. other existing constraints. Yanagida, JCAP 1004, 023 (2010), arXiv:1001.2308 [hep-ph]. [7] M. Kawasaki, A. Kusenko, Y. Tada, and T. T. Yanagida, Phys. Rev. D94, 083523 (2016), arXiv:1606.07631 [astro- SUMMARY ph.CO]. [8] B. Carr, F. Kuhnel, and M. Sandstad, Phys. Rev. D94, In summary, we have presented a new constraint 083504 (2016), arXiv:1607.06077 [astro-ph.CO]. on the abundance of PBHs over a broad mass range [9] K. Inomata, M. Kawasaki, K. Mukaida, Y. Tada, and 7 T. T. Yanagida, Phys. Rev. D 95, 123510 (2017), of O(1)M − 10 M . This parameter space covers the detected stellar and the very recently observed arXiv:1611.06130 [astro-ph.CO]. [10] S. Pi, Y.-l. Zhang, Q.-G. Huang, and M. Sasaki, JCAP intermediate-mass BHs, as well as seeds for supermas- 1805, 042 (2018), arXiv:1712.09896 [astro-ph.CO]. sive black holes. PBH interactions with ISM result in [11] K. Inomata, M. Kawasaki, K. Mukaida, Y. Tada, and the heating of gas, which we applied to dwarf galaxy Leo T. T. Yanagida, Phys. Rev. D 96, 043504 (2017), T to set the limit. We considered several generic heating arXiv:1701.02544 [astro-ph.CO]. mechanisms, including the photon emission from accre- [12] J. Garcia-Bellido, M. Peloso, and C. Unal, JCAP 1709, tion, dynamical friction, and mass outﬂows/winds. This 013 (2017), arXiv:1707.02441 [astro-ph.CO]. type of a constraint has not been previously considered [13] J. Georg and S. Watson, JHEP 09, 138 (2017), arXiv:1703.04825 [astro-ph.CO]. for PBHs. Our limit does not depend on the cosmologi- [14] K. Inomata, M. Kawasaki, K. Mukaida, and cal history, which makes it an attractive independent test T. T. Yanagida, Phys. Rev. D 97, 043514 (2018), of PBHs in the IMBH mass-range. Our analysis can be arXiv:1711.06129 [astro-ph.CO]. readily applied to other systems. [15] B. Kocsis, T. Suyama, T. Tanaka, and S. Yokoyama, The work of G.B.G., A.K., V.T., and P.L. was sup- Astrophys. J. 854, 41 (2018), arXiv:1709.09007 [astro- ported in part by the U.S. Department of Energy (DOE) ph.CO]. [16] K. Ando, K. Inomata, M. Kawasaki, K. Mukaida, grant No. DE-SC0009937. This work was supported in and T. T. Yanagida, Phys. Rev. D 97, 123512 (2018), part by the MEXT Grant-in-Aid for Scientiﬁc Research arXiv:1711.08956 [astro-ph.CO]. on Innovative Areas (No. 20H01895 for K.H.). A.K. was [17] E. Cotner and A. Kusenko, Phys. Rev. Lett. 119, 031103 also supported by Japan Society for the Promotion of Sci- (2017), arXiv:1612.02529 [astro-ph.CO]. ence (JSPS) KAKENHI grant No. JP20H05853. Y.I. is [18] E. Cotner and A. Kusenko, Phys. Rev. D96, 103002 supported by JSPS KAKENHI grant No. JP18H05458, (2017), arXiv:1706.09003 [astro-ph.CO]. JP19K14772, program of Leading Initiative for Excel- [19] E. Cotner, A. Kusenko, and V. Takhistov, Phys. Rev. D98, 083513 (2018), arXiv:1801.03321 [astro-ph.CO]. lent Young Researchers, MEXT, Japan, and RIKEN [20] M. Sasaki, T. Suyama, T. Tanaka, and S. Yokoyama, iTHEMS Program. A.K., V.T., and Y.I. are also sup- Class. Quant. Grav. 35, 063001 (2018), arXiv:1801.05235 ported by the World Premier International Research [astro-ph.CO]. Center Initiative (WPI), MEXT, Japan. [21] B. Carr and J. Silk, Mon. Not. Roy. Astron. Soc. 478, 3756 (2018), arXiv:1801.00672 [astro-ph.CO]. [22] U. Banik, F. C. van den Bosch, M. Tremmel, A. More, G. Despali, S. More, S. Vegetti, and J. P. McK- ean, Mon. Not. Roy. Astron. Soc. 483, 1558 (2019), ∗ [email protected] arXiv:1811.00637 [astro-ph.CO]. † [email protected] [23] F. Hoyle and R. A. Lyttleton, Proceedings of the Cam- ‡ [email protected] bridge Philosophical Society 35, 405 (1939). § [email protected] [24] E. Cotner, A. Kusenko, M. Sasaki, and V. Takhis- ¶ [email protected] tov, JCAP 1910, 077 (2019), arXiv:1907.10613 [astro- ∗∗ [email protected] ph.CO]. [1] Y. B. Zel’dovich and I. D. Novikov, Sov. Astron. 10, 602 [25] A. Kusenko, M. Sasaki, S. Sugiyama, M. Takada, (1967). V. Takhistov, and E. Vitagliano, Phys. Rev. Lett. 125, [2] S. Hawking, Mon. Not. Roy. Astron. Soc. 152, 75 (1971). 181304 (2020), arXiv:2001.09160 [astro-ph.CO]. [26] M. M. Flores and A. Kusenko, Phys. Rev. Lett. 126, 041101 (2021), arXiv:2008.12456 [astro-ph.CO]. [27] A. Katz, J. Kopp, S. Sibiryakov, and W. Xue, JCAP 12, 005 (2018), arXiv:1807.11495 [astro-ph.CO]. 3 This is accurate to factor of few. [28] N. Smyth, S. Profumo, S. English, T. Jeltema, K. McK- 6

innon, and P. Guhathakurta, Phys. Rev. D 101, 063005 [56] M. Ricotti, J. P. Ostriker, and K. J. Mack, Astrophys. (2020), arXiv:1910.01285 [astro-ph.CO]. J. 680, 829 (2008), arXiv:0709.0524 [astro-ph]. [29] P. Montero-Camacho, X. Fang, G. Vasquez, M. Silva, [57] O. Mena, S. Palomares-Ruiz, P. Villanueva-Domingo, and C. M. Hirata, JCAP 1908, 031 (2019), and S. J. Witte, Phys. Rev. D 100, 043540 (2019), arXiv:1906.05950 [astro-ph.CO]. arXiv:1906.07735 [astro-ph.CO]. [30] G. M. Fuller, A. Kusenko, and V. Takhistov, Phys. [58] E. Agol and M. Kamionkowski, Mon. Not. Roy. Astron. Rev. Lett. 119, 061101 (2017), arXiv:1704.01129 [astro- Soc. 334, 553 (2002), arXiv:astro-ph/0109539 [astro-ph]. ph.HE]. [59] M. Guo, K. Inayoshi, T. Michiyama, and L. C. Ho, [31] V. Takhistov, Phys. Lett. B789, 538 (2019), Astrophys. J. 901, 39 (2020), arXiv:2006.08203 [astro- arXiv:1710.09458 [astro-ph.HE]. ph.HE]. [32] V. Takhistov, Phys. Lett. B782, 77 (2018), [60] H. Bondi and F. Hoyle, Mon. Not. Roy. Astron. Soc. 104, arXiv:1707.05849 [astro-ph.CO]. 273 (1944). [33] V. Takhistov, G. M. Fuller, and A. Kusenko, Phys. [61] H. Bondi, Mon. Not. Roy. Astron. Soc. 112, 195 (1952). Rev. Lett. 126, 071101 (2021), arXiv:2008.12780 [astro- [62] F. Yuan and R. Narayan, Ann. Rev. Astron. Astrophys. ph.HE]. 52, 529 (2014), arXiv:1401.0586 [astro-ph.HE]. [34] B. P. Abbott et al. (Virgo, LIGO Scientific), Phys. Rev. [63] N. Shakura and R. Sunyaev, Astron. Astrophys. 24, 337 Lett. 116, 061102 (2016), arXiv:1602.03837 [gr-qc]. (1973). [35] B. P. Abbott et al. (Virgo, LIGO Scientific), Phys. Rev. [64] J. Pringle, Ann. Rev. Astron. Astrophys. 19, 137 (1981). Lett. 116, 241103 (2016), arXiv:1606.04855 [gr-qc]. [65] R. Narayan and I.-s. Yi, Astrophys. J. Lett. 428, L13 [36] B. P. Abbott et al. (VIRGO, LIGO Scientific), Phys. Rev. (1994), arXiv:astro-ph/9403052. Lett. 118, 221101 (2017), arXiv:1706.01812 [gr-qc]. [66] R. Mahadevan, Astrophys. J. 477, 585 (1997), [37] R. Bean and J. Magueijo, Phys. Rev. D66, 063505 arXiv:astro-ph/9609107 [astro-ph]. (2002), arXiv:astro-ph/0204486 [astro-ph]. [67] H. A. Bethe and E. E. Salpeter, “Quantum mechanics [38] M. Kawasaki, A. Kusenko, and T. T. Yanagida, Phys. of one- and two-electron systems,” in Atoms I / Atome Lett. B711, 1 (2012), arXiv:1202.3848 [astro-ph.CO]. I (Springer Berlin Heidelberg, Berlin, Heidelberg, 1957) [39] S. Clesse and J. Garcia-Bellido, Phys. Rev. D92, 023524 pp. 88–436. (2015), arXiv:1501.07565 [astro-ph.CO]. [68] I. M. Band, M. B. Trzhaskovskaia, D. A. Verner, and [40] T. Nakamura, M. Sasaki, T. Tanaka, and K. S. Thorne, D. G. Iakovlev, Astron. Astrophys. 237, 267 (1990). Astrophys. J. Lett. 487, L139 (1997), arXiv:astro- [69] K. Olive, Chinese Physics C 38, 090001 (2014). ph/9708060 [astro-ph]. [70] S. R. Furlanetto and S. J. Stoever, Mon. Not. R. Astron. [41] S. Bird et al., Phys. Rev. Lett. 116, 201301 (2016), Soc 404, 1869 (2010), arXiv:0910.4410 [astro-ph.CO]. arXiv:1603.00464 [astro-ph.CO]. [71] J. Binney and S. Tremaine, Galactic Dynamics: Second [42] M. Raidal, V. Vaskonen, and H. Veermäe, JCAP 09, 037 Edition (2008). (2017), arXiv:1707.01480 [astro-ph.CO]. [72] E. C. Ostriker, Astrophys. J. 513, 252 (1999), [43] Y. N. Eroshenko, J. Phys. Conf. Ser. 1051, 012010 arXiv:astro-ph/9810324 [astro-ph]. (2018), arXiv:1604.04932 [astro-ph.CO]. [73] R. D. Blandford and M. C. Begelman, Mon. Not. Roy. [44] M. Sasaki, T. Suyama, T. Tanaka, and S. Yokoyama, Astron. Soc. 303, L1 (1999), arXiv:astro-ph/9809083. Phys. Rev. Lett. 117, 061101 (2016), arXiv:1603.08338 [74] F. Yuan, D. Bu, and M. Wu, Astrophys. J. 761, 130 [astro-ph.CO]. (2012), arXiv:1206.4173 [astro-ph.HE]. [45] S. Clesse and J. Garc´ıa-Bellido, Phys. Dark Univ. 18, 105 [75] F.-G. Xie and F. Yuan, Mon. Not. Roy. Astron. Soc. 427, (2017), arXiv:1610.08479 [astro-ph.CO]. 1580 (2012), arXiv:1207.3113 [astro-ph.HE]. [46] R. Abbott et al. (LIGO Scientific, Virgo), Phys. Rev. [76] B. Pang, U.-L. Pen, C. D. Matzner, S. R. Green, and Lett. 125, 101102 (2020), arXiv:2009.01075 [gr-qc]. M. Liebendörfer, Mon. Not. Roy. Astron. Soc. 415, 1228 [47] B. Carr, K. Kohri, Y. Sendouda, and J. Yokoyama, (2011), arXiv:1011.5498 [astro-ph.GA]. (2020), arXiv:2002.12778 [astro-ph.CO]. [77] J. Li, J. Ostriker, and R. Sunyaev, Astrophys. J. 767, [48] B. J. Carr and M. Sakellariadou, Astrophys. J. 516, 195 105 (2013), arXiv:1206.4059 [astro-ph.GA]. (1999). [78] F. Yuan, D. Bu, and M. Wu, in Feeding Compact Ob- [49] C. G. Lacey and J. P. Ostriker, Astrophys. J. 299, 633 jects: Accretion on All Scales, IAU Symposium, Vol. 290, (1985). edited by C. M. Zhang, T. Belloni, M. Méndez,and S. N. [50] T. Totani, PASJ 62, L1 (2010), arXiv:0908.3295 [astro- Zhang (2013) pp. 86–89, arXiv:1210.6093 [astro-ph.HE]. ph.CO]. [79] J. J. Bailey, I. B. Abdurakhmanov, A. S. Kadyrov, [51] D. Gaggero, G. Bertone, F. Calore, R. M. T. Connors, I. Bray, and A. M. Mukhamedzhanov, Phys. Rev. A M. Lovell, S. Markoff, and E. Storm, Phys. Rev. Lett. 99, 042701 (2019). 118, 241101 (2017), arXiv:1612.00457 [astro-ph.HE]. [80] M. Oguri, J. M. Diego, N. Kaiser, P. L. Kelly, [52] Y. Inoue and A. Kusenko, JCAP 1710, 034 (2017), and T. Broadhurst, Phys. Rev. D 97, 023518 (2018), arXiv:1705.00791 [astro-ph.CO]. arXiv:1710.00148 [astro-ph.CO]. [53] A. Bhoonah, J. Bramante, F. Elahi, and S. Schon, Phys. [81] Y. Ali-Ha¨ımoudand M. Kamionkowski, Phys. Rev. D95, Rev. Lett. 121, 131101 (2018), arXiv:1806.06857 [hep- 043534 (2017), arXiv:1612.05644 [astro-ph.CO]. ph]. [82] P. D. Serpico, V. Poulin, D. Inman, and K. Kohri, Phys. [54] G. R. Farrar, F. J. Lockman, N. McClure-Griffiths, Rev. Res. 2, 023204 (2020), arXiv:2002.10771 [astro- and D. Wadekar, Phys. Rev. Lett. 124, 029001 (2020), ph.CO]. arXiv:1903.12191 [hep-ph]. [83] R. Murgia, G. Scelfo, M. Viel, and A. Raccanelli, Phys. [55] D. Wadekar and G. R. Farrar, (2019), arXiv:1903.12190 Rev. Lett. 123, 071102 (2019), arXiv:1903.10509 [astro- [hep-ph]. ph.CO]. 7

[84] T. D. Brandt, Astrophys. J. Lett. 824, L31 (2016), [88] Y. Faerman, A. Sternberg, and C. F. McKee, Astrophys. arXiv:1605.03665 [astro-ph.GA]. J. 777, 119 (2013), arXiv:1309.0815 [astro-ph.CO]. [85] S. M. Koushiappas and A. Loeb, Phys. Rev. Lett. 119, [89] E. V. Ryan-Weber, A. Begum, T. Oosterloo, S. Pal, 041102 (2017), arXiv:1704.01668 [astro-ph.GA]. M. J. Irwin, V. Belokurov, N. W. Evans, and D. B. [86] M. A. Monroy-Rodr´ıguez and C. Allen, Astrophys. J. Zucker, Mon. Not. Roy. Astron. Soc. 384, 535 (2008), 790, 159 (2014), arXiv:1406.5169 [astro-ph.GA]. arXiv:0711.2979 [astro-ph]. [87] B. D. Smith et al., Mon. Not. Roy. Astron. Soc. 466, [90] E. N. Kirby, J. D. Simon, M. Geha, P. Guhathakurta, 2217 (2017), arXiv:1610.09591 [astro-ph.CO]. and A. Frebel, Astrophys. J. Lett. 685, L43 (2008), arXiv:0807.1925 [astro-ph].