Arxiv:1805.00122V2 [Astro-Ph.CO] 6 Nov 2018 ∼ ÷ 26]

CERN-TH-2018-102, INR-TH-2018-008, KCL-PH-TH/2018-16

Galactic Rotation Curves vs. Ultra-Light Dark Matter: Implications of the Soliton – Host Halo Relation

Nitsan Bar,1 Diego Blas,2, 3 Kfir Blum,1, 2 and Sergey Sibiryakov2, 4, 5 1Weizmann Institute of Science, Rehovot, Israel 2Theory department, CERN, CH-1211 Geneve 23, Switzerland 3Theoretical Particle Physics and Cosmology Group, Department of Physics, King’s College London, Strand, London WC2R 2LS, UK 4LPPC, IPHYS, Ecole Politechnique Fédérale de Lausanne, CH-1015 Lausanne, Switzerland 5Institute for Nuclear Research of the Russian Academy of Sciences, 60th October Anniversary Prospect, 7a, 117312 Moscow, Russia (Dated: November 7, 2018) Bosonic ultra-light dark matter (ULDM) would form cored density distributions at the center of galaxies. These cores, seen in numerical simulations, admit analytic description as the lowest energy bound state solution (“soliton”) of the Schroedinger-Poisson equations. Numerical simulations of ULDM galactic halos found empirical scaling relations between the mass of the large-scale host halo and the mass of the central soliton. We discuss how the simulation results of different groups can be understood in terms of the basic properties of the soliton. Importantly, simulations imply that the energy per unit mass in the soliton and in the virialised host halo should be approximately equal. This relation lends itself to observational tests, because it predicts that the peak circular velocity, measured for the host halo in the outskirts of the galaxy, should approximately repeat itself in the central region. Contrasting this prediction to the measured rotation curves of well-resolved near-by galaxies, we show that ULDM in the mass range m ∼ (10−22 ÷ 10−21) eV, which has been invoked as a possible solution to the small-scale puzzles of ΛCDM, is in tension with the data. We suggest that a dedicated analysis of the Milky Way inner gravitational potential could probe ULDM up to −19 m . 10 eV.

I. INTRODUCTION galactic halos, roughly within the de Broglie wavelength,

1 1 v − m − xdB 190 22 pc. (1) Two forms of matter could provide the cosmological ≈ 100 km/s 10− eV equation of state, required for dark matter (DM): non- relativistic massive particles (e.g. WIMPs), or the classi- The core, referred to as “soliton” in the literature [4, cal background of an ultra-light bosonic field oscillating 5, 8, 9], corresponds to a coherent quasi-stationary solution of the ULDM equations of motion and its den- around a minimum of its potential (e.g. axions or axion- 1 like particles). On cosmological scales, WIMPs and ultra- sity profile can be derived analytically . This match be- light dark matter (ULDM) behave similarly. However, on tween analytic and numerical simulation results offers a scales comparable to its de Broglie wavelength, ULDM unique opportunity to understand and extend the simu- behaves markedly different from WIMPs [1–13]. lations, making the ULDM model potentially more pre- dictive than the WIMP paradigm. For near-by dwarf Axion-like particles with exponentially suppressed galaxies (v 10 km/s) or Milky Way (MW)-like galaxies masses arise in the context of string theory [12, 14–16]. (v 100 km/s),∼ the soliton core could be resolved with They are produced by misalignment and for mass in the current∼ observational tools, offering a test of the model. 22 18 range m (10− 10− ) eV their cosmic abundance Attempts to detect, constrain, or fit ULDM soliton could naturally∼ match÷ the observed DM density. cores using galactic rotation velocity and velocity disper-

22 21 sion data were discussed in the literature [2, 3, 6–8, 23– ULDM with mass m (10− 10− ) eV is partic- arXiv:1805.00122v2 [astro-ph.CO] 6 Nov 2018 ∼ ÷ 26]. Some of these analyses fitted a soliton profile to ularly motivated due to several puzzles, facing the stan- describe entire galaxies [2, 3, 23]. This exercise leads to dard WIMP paradigm on galactic scales (see, e.g. [12, 17] 23 small ULDM particle mass, m . 10− eV, in strong for recent reviews). This range of m is in some tension tension with cosmological bounds [18–22]. Furthermore, with the matter power spectrum, inferred from Ly-α for- 21 numerical simulations [6, 7, 10, 11, 13] have shown that est analyses [18–22], which yields a bound m & 10− eV. galactic halos above a certain mass would exhibit a more Nevertheless, since the strongest Ly-α bound constraints come from the smallest and most non-linear scales, where systematic effects are challenging, we believe it prudent to seek additional methods to constrain the model for 1 Strictly speaking, the soliton solution is only known numerically. 22 m & 10− eV. With some abuse of the word, we refer to it here as analytical to emphasize that the solution can be found by integrating a Numerical simulations [6, 7, 10, 11, 13] show that simple differential equation in a procedure that takes < 1 sec on ULDM forms cored density profiles in the inner region of a standard laptop. 2 complex structure, with the central soliton transiting into results are confirmed, solitons formed by ULDM with 19 a large-scale host halo composed of an incoherent su- masses m . 10− eV will produce order-one contribu- perposition of multiple scalar field wavepackets. Conse- tion into the mass budget of the inner MW. A dedicated quently, other works [8, 24–26] analysed galactic profiles analysis of inner MW kinematics – including simultane- using a soliton+host halo description. ous modeling of the baryonic mass and the soliton – could 19 A key question in comparing the soliton+host halo potentially test ULDM up to m . 10− eV. model to kinematical data, is how to model the transi- The outline of this paper is as follows. In Sec. II we tion between the central soliton and the host halo. Pre- review some basic properties of the soliton. In Sec. III vious phenomenological analyses [8, 24–26] defined sepa- we discuss the soliton–host halo relations found in sim- rate free parameters for the host halo and for the soliton, ulations, and show that these relations can be under- for each individual galaxy in the sample. On the numer- stood in terms of fundamental properties of the soliton ical simulation side, several studies focused on finding a and the halo. The simulations of [6, 7] are considered soliton–host halo relation [6, 10, 11, 13]. The point in in Sec. III A; they can be summarised by the statement finding a soliton–host halo relation is, of course, that it (E/M) soliton = (E/M) halo, where E is the total energy | | could tie together the behaviour of ULDM in the large- (kinetic+gravitational, within the virial radius, for the scale host halo to predict the central core with fewer free halo) and M is the mass (again within the virial radius, parameters. for the halo). The simulations of [13] are considered in In this work we consider the soliton–host halo relations, Sec. III B; their soliton–halo result can be summarised by the statement E = E . We argue that this reported by different numerical simulation groups [6, 7, |soliton |halo 10, 11, 13]. Our first observation is that properties of the result is unlikely to represent realistic galaxies above a analytic soliton solution can provide important insight certain size. on the numerical results. We show that: (i) the soliton– In Sec. IV, adopting the soliton–host halo relation host halo relation, reported in [13], essentially attributes of [6, 7], we work out its observational consequences. the total energy (kinetic+gravitational) of the halo to We show that in ULDM galaxies satisfying this relation, the dominant soliton. This energetic dominance of the the rotation velocity in the inner core should be approxi- soliton is unlikely to hold for realistic galaxies above a mately as high as the peak rotation velocity in the outer certain size; (ii) the soliton–host halo relation, reported part of the galaxy. In Sec. IV A we compare this analysis in [7], essentially equates the energy per unit mass in the to numerical profiles taken directly from the published soliton to the energy per unit mass in the virialised halo. simulation results, finding good agreement. In Sec. IV B This relation can apply to real galaxies. we compare this prediction to real galaxies from [27, 28], 22 21 finding tension for m (10− 10− ) eV. We show Assuming the soliton–host halo relation of [7], we show ∼ ÷ that it leads to a prediction: the peak circular velocity that the intrinsic scatter in the soliton–host halo relation characterising the host halo on large scales (few kpc for does not resolve the discrepancy. typical (109 1010)M galaxies) should repeat itself in In Sec. V we calculate how baryonic effects could mod- ÷ ify the soliton solution. In Sec. V A we consider a smooth the core on small scales (. 1 kpc), insensitive to the details of the host halo density profile. This implies an fixed distribution of baryonic mass, deferring the case of observational constraint that can be tested without free a super-massive black hole to App. B. These results are parameters. Applying this test to high resolution rota- applied to the MW in Sec. V B. tion curves of late-type galaxies from [27, 28], we find In Sec. VI we compare our results to previous litera- 22 21 ture and discuss some caveats to our analysis. Sec. VII that ULDM in the mass range m (10− 10− ) eV is disfavoured by the data. Baryonic∼ physics÷ is unlikely contains a summary of our results and open questions. to cure the tension: the discrepancy between the predictions of ULDM (with the soliton–host halo relation) and II. SOLITON PROPERTIES: ANALYTIC the data is too large. In many galaxies in the sample CONSIDERATIONS we analyse, baryonic feedback would need to overcome a dark matter-to-baryon mass ratio of 10 : 1, to destroy & In this section we review the relevant properties of the the soliton. As a result, if the soliton–host halo relation soliton that help to understand results from numerical of [6, 7] holds for real systems, ULDM is disfavoured be- simulations. low m 10 21 eV and is unlikely to play a role in solving − We consider a real, massive, free2 scalar field φ, satisfy- the small∼ scale puzzles of ΛCDM. ing the Klein-Gordon equation of motion and minimally We also discuss observational imprints of the soliton in coupled to gravity. In the non-relativistic regime it is big galaxies, such as the MW. In this case the shape of convenient to decompose φ as the soliton is modified due to the gravitational potential 1 imt of baryonic matter and supermassive black hole (SMBH). φ(x, t) = e− ψ(x, t) + c.c., (2) Preliminary numerical simulations of ULDM halos with √2m stars [29] indicate that in the presence of baryons, the soliton mass stays the same or increases compared to 2 the soliton–host halo prediction of pure ULDM. If these Analyses of interacting fields can be found in, e.g. [4, 5, 30, 31]. 3 with complex field ψ that varies slowly in space and time, 1.5 such that ψ m ψ and ψ˙ m ψ . The field ψ |∇ | | | | | | | χ1 satisfies the Schroedinger-Poisson (SP) equations [32] 1.0 Vcirc,1 1 2 . i∂tψ = ψ + mΦψ, (3) 0.5 −2m∇ r 2Φ = 4πG ψ 2. (4) 0.0 ∇ | | rc1 2 4 6 8 10 We look for a quasi-stationary phase-coherent solution, -0.5 3 described by the ansatz Φ1 -1.0 mMpl iγmt ψ(x, t) = e− χ(x). (5) √4π -1.5 The ULDM mass density is FIG. 1. Profile of the “standard” χ1 soliton with λ = 1 (blue (mM )2 solid). We also show the corresponding gravitational potential ρ = pl χ2 (6) 4π (orange dashed) and circular velocity of a test particle (dotted 2 green). 14 m 2 3 4.1 10 22 χ M /pc . ≈ × 10− eV The parameter γ is proportional to the ULDM energy per Other solutions of Eqs. (7-8) can be obtained from unit mass. Validity of the non-relativistic regime requires χ1(r), Φ1(r) by a scale transformation. That is, the func- γ 1. Since we are looking for gravitationally bound tions χλ(r), Φλ(r), together with the eigenvalue γλ, given configurations,| | γ < 0. by

Assuming spherical symmetry and defining r = mx, 2 the SP equations for χ and Φ are given by χλ(r) = λ χ1(λr), (12) 2 Φλ(r) = λ Φ1(λr), (13) ∂2 (rχ) = 2r (Φ γ) χ, (7) 2 r γλ = λ γ1, (14) 2 2 − ∂r (rΦ) = rχ . (8) also satisfy Eqs. (7-8) with correct boundary conditions Finding the ground state solution amounts to solving for any λ > 0. The soliton mass and core radius for χλ Eqs. (7-8) subject to χ(r 0) = const, χ(r ) = 0, are with no nodes. Given the→ boundary value of χ→at ∞r 0, the solution is found for a unique value of γ. → Mλ = λM1, (15) 1 It is convenient to first solve Eqs. (7-8) with the bound- xcλ = λ− xc1. (16) ary condition χ(0) = 1. Let us call this auxiliary solution A mnemonic for the numerical value of λ is given by χ1(r), with γ1. A numerical calculation gives [4, 5, 8] 4 m Mλ γ1 0.69, (9) λ = 3.6 10− 22 9 . (17) ≈ − × 10− eV 10 M and the solution is plotted in Fig. 1. The mass of the χ 1 The product of the soliton mass and core radius is inde- soliton is pendent of λ, M 2 Z pl ∞ 2 2 2 M1 = drr χ (r) (10) 8 m − 1 Mλxcλ 2.27 10 kpc M . (18) m 0 22 ≈ × 10− eV 1 12 m − 2.79 10 22 M . Formally, solutions exist for any positive value of λ and ≈ × 10− eV hence for any soliton mass. However, if we select λ & 1 Its core radius, defined as the radius where the mass den- we reach γλ > 1, outside of the regime of validity of sity drops by a factor of 2 from its value at the origin, the non-relativistic| | approximation. Thus, self-consistent is solutions are limited to λ 1 and their eigenvalue γ = λ2 γ 1, consistent with the non-relativistic 1 λ 1 m − |approximation.| | | xc1 0.082 22 pc. (11) ≈ 10− eV The energy in an arbitrary non-relativistic ULDM configuration is ! Z ψ 2 Φ ψ 2 √ 3 3 E = d x |∇ 2| + | | = Ek + Ep, (19) Mpl = 1/ G. 2m 2 4

4 with kinetic (potential) energy Ek (Ep). For the ansatz halo in the numerical simulations . In Fig. 2 we col- (5), integrating by parts and using Eqs. (3-5) we have lect representative density profiles from Ref. [6] (blue), Ref. [13] (orange), and Ref. [10] (green). We refer to 1 E = M γ. (20) those papers for more details on the specific set-ups in 3 each simulation. To make Fig. 2, in each case, we find Note that spherical symmetry is not needed for Eq. (20) the λ parameter that takes the numerical result into the to hold. χ1 soliton, rescale the numerical result accordingly and 2 Considering the χλ solitons, we find Ep,λ = 2Ek,λ = present it in comparison with the analytic χ1(r) profile. ■ − ■ ■ ■ ●▲ ●▲■●●▲▲●▲■●▲●▲ 2E with 1 ●▲■▲●▲▲●■▲ λ ▲●▲▲●■ ▲●▲■ 2 ▲●▲ (r) ▲●■▲ 2 ■ ■ ■ ▲● ■ ●▲ ●▲■●●▲▲●▲■●▲●■■ ■ M 0.100 ■ ▲●▲■▲●▲ ●▲ ●▲■▲■● ●▲▲● pl 1 1 ▲●■▲▲● ●▲ ▲■●▲●▲■ 3 ▲▲●■ ●▲▲●▲ R ▲●▲■ ▲ ▲●■▲ ● Eλ 0.476 λ , (21) ▲●▲ ●▲■ ▲●▲ ▲●■ ▲ ▲●■ �� ▲ ●■ ▲● ● RSchive et al (2014) ≈ − m ▲● ▲ ▲ ■ ▲ 0.1000.010 ▲■ ●■ ▲● ●▲ ▲ ● ▲ 2 ▲ ▲ ■ ● ▲●■ RMocz et al (2017) M ●▲ ■ ● ▲ ●■ pl ■ ▲ ■ ● ▲ �� 0.100 ▲● ▲ ● ● R ▲■ ■ ▲ Schwabe et al (2016) Mλ 2.06 λ . (22) 0.0100.001 ●▲ ● ■ ▲ ▲■ ▲ ■ ● ▲● ≈ m ■●▲ R ▲ ▲● ● ■ ● ▲ ■ ▲ ▲■ ■ ● ● ●▲ ▲ ● ■ ■ ▲ ■ ● 0.001 ▲ ▲ This leads to a relation for an isolated soliton [4, 5], ▲ ■ ● ● ■ 0.10.0100.5 1 5 ▲10 ■ 50● 100▲ ▲ ● ● ▲ ■ 1 ■ ● ■ 3 ▲ ● 0.1 0.5 1 5 10 50100 ■ ● ▲ Mλ Eλ ■ ● 2.64 . (23) ▲ ● 0.001 ■ 2 2 ▲ ● (M /m) ≈ (M /m) ▲ pl pl ● ▲ ■ ▲ ■ ● Another useful relation gives the energy per unit mass ▲ ● ■ 0.5 1 5 10 50 100 from the scaling parameter λ, r E | λ| 0.23 λ2, (24) Mλ ≈ FIG. 2. Review of results from numerical simulations by different groups. Markers show density profiles of simulated which can also be written as halos from Schive et al. [6] (blue circles), Mocz et al. [13] 1 2 2 (orange squares), and Schwabe et al. [10] (green triangles). Eλ Mpl Mλ 4.3 | | . (25) The central regions of the halos are described by the soliton ≈ Mλ m (solid line). The circular velocity curve for a test particle in the soliton gravitational potential is given by While different groups agree that solitons form in the centers of halos, they do not appear to agree on the 2 Vcirc,λ(r) = r∂rΦλ(r). (26) matching between the inner soliton profile and the host halo. Refs. [6, 7] and Ref. [13] reported scaling relations The circular velocity rises as Vcirc,λ r at small r and 1 ∝ between the central soliton and the host halo. As we show decreases as V r− 2 at large r, see Fig. 1. The circ,λ ∝ below, the scaling relations found by both groups are con- peak of Vcirc is obtained at nected to properties of a single, isolated, self-gravitating 1 1 m − soliton (part of these observations were made in [10, 11]). xpeak,λ 0.16 λ− 22 pc (27) ≈ 10− eV 2 1 m − Mλ − A. Soliton vs. host halo: the simulations of 460 22 9 pc, ≈ 10− eV 10 M Ref. [6, 7] and the peak velocity is At cosmological redshift z = 0, the numerical simula- 5 maxVcirc,λ 2.3 10 λ km/s (28) tions of [6, 7] yield approximately NFW-like halos which ≈ × m Mλ transit, in the central region, into a core with core radius 83 22 9 km/s. and mass density ≈ 10− eV 10 M 1 3 1 Mh − m − xc 160 12 22 pc, (29) ≈ 10 M 10− eV III. MAKING CONTACT WITH NUMERICAL SIMULATIONS 2 4 m − xc − 190 10−22 eV 100 pc ρ(x) M pc 3, (30) 8 − We now discuss results from the numerical simula- ≈ 2 1 + 0.091 x tions of three different groups, Refs. [6, 7], Ref. [13], and xc Refs. [10, 11]. The first point to note is that soliton configurations, in a form close to the idealised form discussed in Sec. II, 4 The first simulations of cosmological ULDM galaxies [33] did not actually occur dynamically in the central region of the have sufficient resolution to resolve the central core. 5 where Mh is the virial mass of the host halo. As noted However, this is just Eq. (23), if we replace the halo en- in [6–8], Eqs. (29-30) are an excellent numerical fit for a ergy Eh by the energy of the soliton. Because the central soliton χλ. The mass of this soliton is density profile found in [13] was a χλ soliton, to a good

1 approximation, it must be the case that the total energy 1 3 9 m − Mh of the halo in the simulations of [13] was dominated by M 1.4 10 22 12 M , (31) ≈ × 10− eV 10 M the central soliton contribution. This situation is unlikely to hold for realistic cosmological host halos with so its λ parameter is Mh significantly above Mh,min. 1 3 How could this have happened? The initial conditions 4 Mh λ 4.9 10− . (32) in the simulations of [13] were a collection of N solitons, ≈ × 1012 M which were then allowed to merge. It appears that these Note that Eq. (31) is applicable only as long as the halo initial conditions were constructed such that one initial exceeds a minimal mass, state soliton – the soliton of initially largest mass – grew to absorb the entire energy of the system. Differently 3/2 from Ref. [7] that considered initial conditions of N iden- 7 m − Mh,min 5.2 10 M . (33) tical initial solitons, the simulations of [13] initiated their ∼ × 10 22eV − N solitons with a random flat distribution in soliton ra- Smaller mass halos would be dominated by the soliton. dius. Such distribution would be skewed towards large 3 Ref. [7] showed that Eq. (31) is consistent with the soliton energy because Eλ xcλ− . Considering the ini- relation, tial condition set-up as explained∝ in [13], we find that

1 the most massive initial state soliton typically needed to 2 2 Eh Mpl grow in mass by only a factor of 1.5 2, to absorb the Mc α | | , (34) ÷ ≈ Mh m entire energy of the halo. Note that energy dominance of the central soliton over where Mc is the core mass (mass within x < xc); Mh,Eh the host halo, implied by Eq. (36), is not the same, of are the virial mass and energy of the simulated halo; and course, as equating the energy per unit mass of the soliton α = 1 provides a good fit to the data. Ref. [7] gave a and the halo, implied by Eq. (35). Halos in [6, 7] attained heuristic argument, pointing out that Eq. (34) relates the masses up to two orders of magnitude larger than the soliton scale radius (chosen as the core radius xc in [7]) central soliton mass, meaning their halo energy was two with the velocity dispersion of particles in the host halo, orders of magnitude larger than the energy of the soliton. in qualitative agreement with a wave-like “uncertainty principle”. However, there is another way to express Eq. (34). The C. Comments core mass of a χλ soliton is related to its total mass via Mcλ 0.236Mλ. Thus, using Eq. (25) we have an an- ≈ 1 2 2 As far as we can currently determine, Eq. (35) may Eλ Mpl alytic relation Mcλ 1.02 | | . This allows us indeed reflect a realistic soliton–host halo relation for ≈ Mλ m to rephrase the empirical Eq. (34) by a more intuitive large enough cosmological halos. In the following sec- (though equally empirical) expression: tions, we take a leap of faith and assume that the simulations of [6, 7] produced the correct scaling relation. E E We stress that Eq. (35) is an empirical result, and was . (35) M soliton ≈ M halo only tested in [6, 7] for host halo masses ranging from 108 M to 1011 M . Our key numerical analysis Therefore, the soliton–host halo relation in the simula- will∼ concern systems∼ in this range of mass. tions of Ref. [6, 7] can be summarised by the statement We defer a theoretical study of the origin of Eq. (35) that the energy per unit mass of the soliton matches the to future work. Here we give only a few comments. We energy per unit mass of the host halo. stress that the discussion in the rest of this section does not affect any of our results. B. Soliton vs. host halo: the simulations of For a soliton, E/M = γ/3. On the other hand, γm can Ref. [13] be associated with the chemical potential of ULDM particles in the soliton (see e.g. [34] and references therein). This may appear to hint that Eq. (35) corresponds to The simulations of Ref. [13] pointed to an empirical thermodynamic equilibrium between the ULDM parti- scaling relation between the soliton mass M and the total cles in the host halo and in the soliton. However, there energy of the ULDM distribution in the simulation box, is some evidence to the contrary from simulations. E , h Ref. [29] simulated ULDM, adding collisionless point 1 3 particles (“stars”). The stars aggregated dynamically in

M Eh a cuspy profile, resulting in a more massive soliton com- 2 2.6 2 . (36) (Mpl/m) ≈ (Mpl/m) pared to the pure ULDM simulations [6, 7] with a given 6 host halo mass. Testing the reversibility of the system, We can estimate the energy per unit mass of the viri- Ref. [29] adiabatically “turned off” the stars after the alised halo by initial system virialised. When eliminating the stars, the R R200 2 soliton+halo system did not relax back to Eq. (35). In- E dxx ρNFW (x)ΦNFW (x) π 0 . (41) stead, the excess ULDM mass that was contained in the M halo ≈ M200 soliton in the presence of stars remained captured in the soliton, and did not return to the host halo. The final This gives state of the system was not described by Eq. (35): the E c˜ soliton ended up containing larger (negative) E/M than Φh, (42) M 4 the halo, and larger mass compared with Eq. (31). halo ≈ where IV. SOLITON-HOST HALO RELATION AND c ln(1 + c) c˜ = − . (43) GALACTIC ROTATION CURVES (1 + c) ln(1 + c) c − Typical values of the concentration parameter are in the As we have seen, the soliton–host halo relation found in range c 5 30 [35]. In this range,c ˜ varies between the simulations of [6, 7] can be summarised by Eq. (35), c˜ 0.55∼ 0.÷35, respectively. (For reference, fits of the equating the energy per unit mass of the virialised host MW∼ outer÷ rotation curve give c 10 20 [36].) halo to that in the soliton component. For a virialised Plugging Eq. (42) into the soliton–host∼ ÷ halo relation system, the energy per unit mass maps to kinetic energy Eq. (35), the scaling parameter λ is fixed as density: in particular, the characteristic circular velocity

(or, up to an (1) geometrical factor, the velocity dis- 2 Eλ c˜ persion) of testO particles in the halo and in the soliton 0.23 λ Φh, (44) − ≈ Mλ ≈ 4 should match. The peak circular velocity of the soliton, given by Eqs. (27-28), occurs deep in the inner part, which implies5 x < 1 kpc, of the galaxy; while the peak circular velocity M 2 of an NFW-like halo occurs far out at x 2 R , with p pl ∼ s Mλ 2.1 c˜Φh Rs the NFW characteristic radius, of order 10 kpc for a ≈ − m 1 MW-like galaxy. Thus, if the scaling derived from the 9 m − 2.4 10 22 simulations of [6, 7] is correct, ULDM predicts that the ≈ × 10− eV 1 1 peak rotation velocity in the outskirts of a halo should 3 3 H(z) M200 approximately repeat itself in the deep inner region. We 12 f(c)M , (45) × H0 10 M now discuss this result quantitatively. Consider a halo with an NFW density profile with v ρcδc u c c ln(1 + c) ρNFW (x) = 2 , (37) u x x f(c) = 0.54 − 2 . 1 + t 1 + c c Rs Rs ln(1 + c) − 1+c where Eq. (45) depends weakly on the NFW concentration pa- 2 3 3H (z) 200 c rameter, via the factor f(c) that varies in the range ρc(z) = , δc = . (38) 8πG 3 ln(1 + c) c 0.9 1.1 for c = 5 30. It agrees parametrically with − 1+c the÷ simulation result,÷ Eq. (31) (including the redshift de- The profile has two parameters: the radius R and the s pendence, which we have suppressed in Eq. (31)). It also concentration parameter c = R /R , where R is the 200 s 200 agrees quantitatively to about 20%; to see this, we need radius where the average density of the halo equals 200 to account for the slightly different definition of the halo times the cosmological critical density, roughly indicating mass M , used in [7], and our M . We do this compar- the virial radius of the halo. The gravitational potential h 200 ison in App. A. of the halo is Consider the rotation velocity curve of an ULDM 3 4πGρcδcRs x galaxy satisfying Eq. (35). The NFW rotation curve is ΦNFW (x) = ln 1 + . (39) − x Rs given by

Near the origin, x Rs,ΦNFW is approximately con- V 2 (x) circ,h 2(1 + ξ) ln(1 + ξ) 2ξ x stant, ΦNFW (x Rs) Φh, and is related to the mass 2 = − , ξ .(46) ≈ 4π 3 3 Vcirc,h(Rs) ξ(1 + ξ)(ln(4) 1) ≡ Rs of the halo, M200 = 200ρc 3 c Rs, via − 1   3 1 2 4πδc 3 3 Φ = G   ρc M . (40) 5 h −  2  200 In the numerical estimates we use H0 = 70 km/s/Mpc for the ln(1 + c) c present-day Hubble constant. − 1+c 7

22 This halo rotation curve peaks at x 2.16 Rs with a bottom panels, respectively. For larger m > 10− eV, peak value ≈ the soliton bump in the rotation curve would shift to smaller x according to Eq. (50), but would maintain its 5 1 maxVcirc,h 1.37 10 ( Φh) 2 km/s. (47) height. ≈ × − On the other hand, in the inner galaxy x Rs, the circu- V [km/s] 12 lar velocity due to the soliton peaks to a local maximum circ M200 = 10 M of 300

1 250 2 c = 25 5 c˜ 1 c = 15 maxV 1.51 10 ( Φ ) 2 km/s, (48) 200 circ,λ ≈ × 0.4 − h c = 10 150 where we used Eq. (44) to ﬁx λ and Eq. (28) to relate it to maxVcirc,λ. As anticipated in the beginning of this section, Eq. (35) 100 predicts approximately equal peak circular velocities for the inner soliton component and for the host halo,

1 maxV c˜ 2 circ,λ 1.1 , (49) 50 maxV 0.4 10 100 1000 104 105 106 circ,h ≈ x [pc] independent of the particle mass m, independent of the V [km/s] 10 circ M200 =5 10 M ⇥ halo mass M200, and only weakly dependent on the de- 1 150 tails of the halo via the factor (˜c/0.4) 2 . Eq. (49) is plot- 125 ted in Fig. 3 as function of the concentration parameter. 100 c = 25 75 c = 15 maxVcirc, c = 10 maxVcirc,h 50 1.30

1.25

1.20 25

1.15 10 100 1000 104 105 x [pc] 1.10 FIG. 4. Rotation curves for the ULDM soliton+halo system, 1.05 obtained for a DM-only NFW halo using the soliton–halo relation Eq. (35) with m = 10−22 eV. Solid black, dot-dashed 5 10 15 20 25 30 c orange, and dashed blue show Vcirc due to the total soliton+halo system, the soliton only, and the halo only. Results are shown for NFW concentration parameter c = 10, 15, 25, FIG. 3. Ratio between halo and soliton peak circular veloci- 12 10 with M200 = 10 M and 5 × 10 M on the upper and ties as a function of the halo concentration. lower panels, respectively.

While maxV and the approximate equality circ,λ In Fig. 4, to define the rotation velocity for the total Eq. (49) are m-independent, the soliton peak velocity system, we set the ULDM mass density for the total sys- occurs in an m-dependent location, tem to be ρ(x) = max ρλ(x), ρNFW (x) , calculate the 1 { } 10 22 eV maxV − resulting mass profile M(x), and use spherical symme- x 184 − circ,λ pc. (50) p peak,λ ≈ m 200 km/s try to find Vcirc(x) = GM(x)/x. This prescription for matching between the soliton and NFW parts is ad-hoc Fig. 4 shows the circular velocity curve for the NFW and only roughly consistent with the simulations of [6, 7]. halo+soliton system, following from Eq. (35), with The true transition region between the NFW part and the 22 ULDM particle mass m = 10− eV. The solid black, dot- soliton part probably deviates from the pure NFW form. dashed orange, and dashed blue lines show the contribu- Ref. [37] considered this transition region and concluded tions to Vcirc due to the total system, the soliton only, and that the density profile in this region should follow ap- 5 the halo only. Results are shown for three different val- proximately ρ x− 3 , steeper than the usual inner NFW 1 ∼ ues of the NFW concentration parameter, c = 10, 15, 25, form ρ x− . This would affect the detailed shape of 12 10 ∼ with M200 = 10 M and 5 10 M on the top and the rotation curve in the intermediate region between the × 8 two peaks, but not our general results. B. Comparison to real galaxies Eq. (49) was derived for an NFW host halo, but it is the manifestation of Eq. (35) that is not tied to a particular We now consider some observational consequences of parametrisation of the halo profile. Building on Eq. (35), our analysis. We choose to do so by examining the ro- we expect in general that for DM-dominated galaxies, tation curves of nearby disc galaxies with halo masses in the soliton peak circular velocity should roughly equal the range covered by the simulations of [6, 7], and above the peak circular velocity in the host halo. The NFW the minimal mass of an ULDM halo, see Eq. (33). We example demonstrates that details of the host halo profile divide our analysis into two parts. First, in Sec. IV B 1, affect this result at the 10% level or so. we concentrate on a small set of galaxies for which high- In the rest of this paper, when we refer to Eq. (49), resolution kinematical and photometric data is available. we set the RHS to unity. Approximating the RHS of For this small set of galaxies, we present detailed rotation Eq. (49) by unity, and replacing maxVcirc,h instead of curve data and compare them to the ULDM prediction. maxVcirc,λ in Eq. (28), the peak circular velocity of a We find clear tension between the data and the ULDM host halo allows to predict the scale parameter λ and prediction. Next, in Sec. IV B 2, we extend the analysis thus the soliton relevant for that host halo. to include a large sample of galaxies, confirming and re- inforcing the tension between the ULDM prediction and the rotation curve data. A. Comparison to numerical simulations

In Fig. 5 we compare our results to two soliton+halo 1. A few specific examples configurations from the simulations of [6] and [29] (for [6], we take the largest halo, and for [29] we take the ini- In our first analysis of the data, we study a set of four tial state of Case C). To calculate the soliton, we read representative rotation curves from Ref. [27] (see Ref. [28] maxVcirc,h from the large-scale peak (at x 20 kpc) for a recent rendering of these and many other rota- of the numerically extracted halo rotation curves∼ (solid tion curves), for which high-resolution kinematical data lines). Following Eq. (49), we use maxVcirc,h instead of is available. Our goal is to check if the ULDM prediction, maxVcirc,λ in Eq. (28), and read off the value of λ. The that we demonstrated in the previous section using sim- predicted soliton bump is shown in dashed lines. It gives ulated galaxies, actually holds in real observed galaxies. the correct soliton peak rotation velocity to 20% accu- The key prediction we check is summarised by Eq. (49), racy in both cases. ∼ and was demonstrated for simulated galaxies in Figs. 5 and 6: this prediction states that given a measurement

Vcirc [km/s] of the large-radius halo rotation curve, ULDM prescribes a soliton-induced peak in the inner part of the halo, with 100 Schive et al (2014) 80 height speciﬁed by Eq. (49), at a location speciﬁed by Eq. (50). 60 The four measured rotation curves are shown in 40

Vcirc Chan et al (2017) maxVcirc, 2.0 20 1.5

1.0 0.5 1 5 10 50 100 x [kpc]

FIG. 5. Comparison of the prediction of Eq. (49) (dashed 0.5 lines) to the numerical simulation results (solid lines) of Schive et al (2014) [6] and Chan et al (2017) [29].

0.1 0.5 1 5 10 50 In Fig. 6 we show the velocity proﬁles for 11 simu- x (rescaled) lated halos, calculated for 6 halos from [7] (solid lines) and 5 halos from [6] (dashed lines). The rotation curves FIG. 6. Circular velocity proﬁles for halos, calculated for sim- are scaled and normalised such that xpeak,λ = 1 and ulation results published in [7] (solid lines) and [6] (dashed maxVcirc,λ = 1 in each case. All of the halos satisfy lines), scaled and normalised such that xpeak,λ = 1 and maxV 0.65 < circ,λ < 1.4; for later reference, the shaded maxVcirc,λ = 1. Shaded band highlights the range 0.5 < maxVcirc,h maxVcirc,λ maxVcirc,λ < 1.5. band highlights the range 0.5 < < 1.5. maxVcirc,h maxVcirc,h 9

Figs. 7-10, as blue markers. For each of these rota- corrections from the detailed halo shape. The baryonic tion curves, we use the measured circular velocity at (stellar and gas) contribution to the gravitational poten- the farthest radius, to serve as an input for maxVcirc,h tial affects the halo velocity at a similar level: in Ref. [28], in Eq. (49). In turn, Eq. (49) gives as output the pre- the baryonic contribution to the halo peak rotation ve- dicted soliton-induced peak rotation velocity, maxVcirc,λ. locity of UGC 1281, UGC 4325, and NGC 100 was esti- Given the soliton-induced peak rotation velocity, the soli- mated from photometric data to be well bellow the DM ton λ parameter is fixed by Eq. (28) and with it, the full 6 (bar) (obs) contribution , Vcirc,h /Vcirc,h < 0.5. This means that the soliton-induced rotation curve. The result is plotted as observed velocity V (obs) is equal to the DM-induced ve- dashed line in Figs. 7-10; the upper panels show the re- circ,h 22 (DM) sult for m = 10− eV and the lower panels show it for locity Vcirc,h to better than 15%. On the other hand, 21 m = 10− eV. this simple procedure relieves us from the need to fit for We find that the ULDM soliton–host halo relation sig- the virial mass or other details of the host halo. All that nificantly overestimates the rotation velocity in the inner is needed is the peak halo rotation velocity, a directly part of all of the galaxies in Figs. 7-10. This puts the pre- observable quantity7. For completeness, we will return 22 21 dictions of ULDM in the mass range m (10− 10− ) to the issue of baryonic effects and deal with it more sys- eV in tension with the data. ∼ ÷ tematically in the second part of this section, considering SPARC data [28].

Vcirc [km/s] Eq. (49) (represented by the dashed line in Figs. 7-10) 80 corresponds to the central value of the soliton–host halo UGC 1281

60 Vcirc [km/s] 140 UGC 4325 40 120

100

20 80 22 m = 10 eV 60 0 0 1 2 3 4 5 6 40 x [kpc]

20 22 Vcirc [km/s] m = 10 eV 80 0 UGC 1281 0 1 2 3 4 5 x [kpc]

60 Vcirc [km/s] 140 UGC 4325 40 120

100 20 21 80 m = 10 eV 60 0 0 1 2 3 4 5 6 40 x [kpc] 21 m = 10 eV 20

FIG. 7. Measured rotation curve of UGC 1281 superimposed 0 on the prediction from Eq. (49) following from the soliton– 0 1 2 3 4 5 host halo relation. The ULDM mass is m = 10−22 eV (upper x [kpc] panel) and m = 10−21 eV (lower panel). The shaded band accounts for the intrinsic scatter of the soliton–host halo re- FIG. 8. Same as Fig. 7 for UGC 4325. lation.

We emphasize that in using Eq. (49) to predict the soli- 6 ton, we set the RHS of that equation to unity, and thus For this estimate we use 3.6µm mass-to-light ratio Υ∗ = 0.5 M /L . we ignore any details of the shape of the host halo. We 7 The rotation curves in Figs. 7-10 do not show a clear peak within also neglect corrections due to baryons. On the one hand, the range of the measurement; this means that our soliton bump, as we have learned from the NFW analysis, this prescrip- derived from the maximal velocity seen in the data, underesti- tion for deriving the soliton profile would suffer (10%) mates the true predicted soliton and is thus conservative. O 10 relation. Ref. [6, 7] showed a scatter of about a factor limit ourselves to galaxy masses that are comfortably of two around Eq. (34) between simulated halos. This above the minimal halo mass (33), and not above the translates to a factor of two scatter in the soliton λ pa- range simulated in [6, 7]. Our results are not sensitive to rameter (we have illustrated this scatter for a sub-sample the details of this mass cut. of simulated halos in Fig. 6). In Figs. 7-10 we represent Next, for each galaxy we determine the observed max- this scatter by a shaded band, showing the results when imal halo rotation velocity maxVcirc,h, and use it to com- the λ parameter inferred from Eq. (49) is changed by a pute the soliton prediction from Eq. (49). To avoid con- factor of 2. fusion between halo peak velocity and soliton peak velocity, we search for the halo peak velocity restricting to ra- 22 1 dial distance x > 3 m/10− eV − kpc. Galaxies with 2. A large sample of galaxies: SPARC 22 1 no data above x = 3 m/10− eV − kpc are discarded. Our results are not sensitive to this criterion; defining the It is important to check if scatter between different 1 halo cut anywhere at 1 m/10 22 eV − kpc guaran- galaxies could explain the discrepancy, with the four & − tees that such confusion is avoided. galaxies in Figs. 7-10 being accidental outliers. To ad- dress this question, we analyse the 175 rotation curves SPARC galaxies come with photometric data, allow- contained in the SPARC data base [28]. This sample in- ing to model the baryonic contribution to the gravita- cludes, in particular, the galaxies UGC 1281, UGC 4325, tional potential [28]. We use this information to limit and NGC 100, shown in Figs. 7-10. baryonic effects on our analysis, and to explore the sen- Our SPARC analysis is as follows. For each galaxy, we sitivity of our results to baryonic corrections. For each make a crude estimate of the halo mass contained within galaxy, we estimate the baryonic contribution to the ob- 2 served rotation velocity using the mass models of [28] the observed rotation curve profile by Mgal RV /G, where R is the radial distance of the last∼ data point with 3.6µm mass-to-light ratio Υ = 0.5 M /L . Set- (bar),2 (DM),2 (obs),2 ∗ in the rotation curve, and V the corresponding veloc- ting Vcirc,h + Vcirc,h = Vcirc,h , we calculate the ratio 11 (bar) (DM) ity. We keep only galaxies with 5 10 M > Mgal > f = V /V . We present results when cut- × bar2DM circ,h circ,h 8 22 3/2 5 10 m/10− eV − M . We do this in order to ting on different values of fbar2DM < 1, 0.5, 0.33. ×

Vcirc [km/s] Vcirc [km/s] 100 120 NGC 1560 NGC 100 80 100

80 60 60 40 40

20 20 22 22 m = 10 eV m = 10 eV 0 0 0 2 4 6 8 0 2 4 6 8 x [kpc] x [kpc]

Vcirc [km/s] Vcirc [km/s] 100 120 NGC 1560 NGC 100 100 80 80 60 60 40 40 21 21 m = 10 eV m = 10 eV 20 20

0 0 0 2 4 6 8 0 2 4 6 8 x [kpc] x [kpc]

FIG. 9. Same as Fig. 7 for NGC 1560. FIG. 10. Same as Fig. 7 for NGC 100. 11

Our first pass on the data includes only galaxies for We have limited our attention to the range m = 22 21 which the predicted soliton is resolved, namely, xpeak,λ (10− 10− ) eV, for which we believe the results are (DM) clear. We÷ leave a detailed study of the precise exclusion from Eq. (50), with maxVcirc,λ = maxVcirc,h , lies within the rotation curve data. For these galaxies, we compute range to future work. We note that for lower particle 23 from data the ratio mass, m . 10− eV, the soliton contribution extends over much of the velocity profile of many of the SPARC Vcirc, obs(xpeak,λ) galaxies, leaving little room for a host halo. This limit, (DM) . (51) maxVcirc,h where the galaxies are essentially composed of a single giant soliton, was considered in other works. We do not Here, Vcirc, obs(xpeak,λ) is the measured velocity at the pursue it further, one reason being that this range of expected soliton peak position. small m is in significant tension with Ly-α data [18, 19]. The results of this first pass on the data are shown in 21 For higher particle mass, m & 10− eV, the soliton Fig. 11. Red, blue, and green histograms show the result peak is pushed deep into the inner 100 pc of the rotation when imposing fbar2DM < 1, 0.5, 0.33, respectively. For 22 curve. Although high resolution data (e.g. NGC 1560, m = 10− eV, we find 45, 26, and 5 galaxies that pass Fig. 9) is sensitive to and disfavours this situation, a more the resolved soliton cut for fbar2DM < 1, 0.5, 0.33. For 21 m = 10− eV, only 4 galaxies pass the resolved soliton cut for fbar2DM < 1, and none for fbar2DM < 0.5, 0.33. 15 Including only galaxies with a resolved soliton causes us to loose many rotation curves with discriminatory -22 power. For example, UGC 4325 drops out of the anal- m=10 eV ysis for m = 10 21 eV, though it clearly constrains the soliton resolved − 10 model, as seen from the lower panel of Fig. 8. To overcome this without complicating the analysis, we perform a second pass on the data. Here, we allow galaxies with unresolved soliton, as long as the innermost data point is

# of galaxies 5 located not farther than 3 xpeak,λ. We need to correct for the fact that the soliton× peak velocity is outside of the measurement resolution. To do this, we modify our observable as r 0 Vcirc, obs(xpeak,λ) Vcirc, obs(xmin,data) xmin,data 0 0.5 1 1.5 2 (DM) (DM) , → × xpeak,λ (velocityV atcirc soliton, obs(x peakpeak)/(max,)/max Vhalocirc,h velocity) maxVcirc,h maxVcirc,h (52) 5 where xmin,data is the radius of the ﬁrst data point. This 4 correction is conservative, because it takes the fall-oﬀ of the soliton gravitational potential at at x > xpeak,λ to -21 be the same as for a point mass. In reality, the potential 3 m=10 eV decays slower and the soliton-induced velocity decreases soliton resolved slower. Keeping this caveat in mind, Fig. 12 presents our 22 2

results including unresolved solitons. For m = 10− eV, # of galaxies with fbar2DM < 1, 0.5, we find 48 and 16 galaxies with unresolved soliton, that can be added to the sample of 1 Fig. 11. No galaxy is added for fbar2DM < 0.33. For m = 21 10− eV, 16 and 5 galaxies are added with fbar2DM < 0 1, 0.5, and none for fbar2DM < 0.33. 0 0.5 1 1.5 2 In Figs. 11-12, vertical dashed line indicates the (velocityV atcirc soliton, obs(x peakpeak)/(max,)/max Vhalocirc,h velocity) soliton–host halo prediction. The shaded region shows the range of the prediction, modifying the RHS of FIG. 11. Distribution of SPARC galaxies [28] with respect Eq. (49) between 0.5 1.5, consistent with the scatter − to the ratio of observed circular velocity at the soliton peak seen in the simulations. to the maximal circular velocity of the halo. The vertical We conclude that the four galaxies in Figs. 7-10 are not dashed line shows the prediction for the mean implied by the outliers: they are representative of a systematic discrep- soliton-host halo relation and the shaded region accounts for ancy, that would be difficult to attribute to the scatter the intrinsic scatter in this relation. The ULDM mass is m = seen in the simulations. If the soliton-host halo rela- 10−22 eV (upper panel) and m = 10−21 eV (lower panel). tion of [6, 7] is correct, then ULDM in the mass range Red, blue, green histograms correspond to the cuts fbar2DM < 22 21 1, 0.5, 0.33, respectively. Only rotation curves with resolved m 10− eV to m 10− eV is in tension with the data.∼ ∼ solitons are included (see the main text for details). 12 careful analysis would be needed to draw a definitive con- observed velocities by a factor of 3-5, implying that the clusion. Note that in this range of m, ULDM ceases to soliton over-predicts the mass in the central region of the offer a solution to the small-scale puzzles of ΛCDM (see, galaxy by about an order of magnitude. This leaves lit- e.g., review in [12]). tle room for baryons (and, we think, baryonic feedback) to significantly affect the dynamics of the inner ULDM halo. V. BARYONIC EFFECTS Proceeding to the analysis, we denote the gravitational potential due to baryons by Φb(r). We assume that the 8 A. A fixed distribution of baryons distribution of baryonic mass is spherically symmetric and dies off at infinity sufficiently fast, so that Φ (r b → ) = GMb/r, where Mb is the total baryonic mass. We now consider the soliton solution, and some aspects ∞ − Adding Φb(r) changes Eqs. (7-8) into [2] of the soliton-host halo relation, with a coexisting distribution of baryonic mass. In this section we consider a ∂2 (rχ) = 2r (Φ + Φ (r) γ) χ, (53) r b − smooth mass distribution, deferring the analysis of the 2 2 ∂r (rΦ) = rχ . (54) effect of a super-massive black hole (SMBH) to App. B. The analysis of baryonic effects is qualitatively impor- It remains convenient to solve the problem using bound- tant, and quantitatively relevant to ULDM in, e.g., MW- ary conditions with χ(0) = 1. Let us denote this so- like galaxies. We will see that in order to make a signif- lution (satisfying χ(0) = 1) by χ1(r;Φb), accompanied icant impact, baryons need to constitute an (1) frac- by the soliton potential Φ1(r;Φb). The general solution O tion of the total mass in the soliton region. For many of χλ(r;Φb) satisfying Eqs. (53-54) with boundary condi- 2 the galaxies that we analysed in Sec. IV B, the predicted tion χλ(0; Φb) = λ is given by soliton contribution to the rotation velocity exceeds the 2 2 1 χλ(r;Φb) = λ χ1 λr; λ− Φb λ− r , (55) 2 2 1 Φλ(r;Φb) = λ Φ1 λr; λ− Φb λ− r . (56) 20 Defining the soliton mass and energy in the presence of the baryons by Mλ(Φb),Eλ(Φb), we have 15 2 1 Mλ(Φb) = λM1 λ− Φb λ− r , (57) 3 2 1 m=10-22 eV Eλ(Φb) = λ E1 λ− Φb λ− r , (58) 10 allow unresolved where the ULDM energy is soliton Z 2 ! # of galaxies ψ Φ E(Φ ) = d3x |∇ | + + Φ (x) ψ 2 . 5 b 2m2 2 b | | (59) 0 Given the function Φ , and plotting M (Φ ) and E (Φ ) 0 0.5 1 1.5 2 b λ b λ b vs. λ, we can find the value of λ and hence the profile (velocityV atcirc soliton, obs(x peakpeak)/(max,)/max Vhalocirc,h velocity) for a soliton solution of any desired mass or energy. 10 Solutions of Eqs. (55-56) can be compared with the results of numerical simulations. Ref. [29] added a dis- 8 tribution of mass in the form of point particles (“stars”) -21 22 m=10 eV to ULDM simulations with m = 0.8 10− eV. The allow unresolved first toy model they studied included an× isolated soliton 6 (without an ULDM host halo) of mass M = 3.3 108 M , to which a stellar distribution was added and evolved× to soliton 4 a virialised state. From Fig. 1 of Ref. [29] we derive the

# of galaxies baryonic contribution to the gravitational potential, and solve for the distorted soliton at the stated ULDM mass. 2 In the top panel of Fig. 13 we show Mλ vs. λ, derived

0 0 0.5 1 1.5 2 (velocityV atcirc soliton, obs(x peakpeak)/(max,)/max Vhalocirc,h velocity) 8 The assumption of spherical symmetry is not realistic; some of the galaxies considered in Sec. IV B may in fact exhibit maximal discs. Nevertheless, we expect our simpliﬁed analysis to give us FIG. 12. Same as Fig. 11, including galaxies with unresolved correct order of magnitude estimates, and defer the analysis of solitons (see the main text). non-spherical baryonic+soliton+halo systems to future work. 107

106 13 105 �� from Eq. (57). Blue solid (green dashed) lines show Mλ fixed by the SP equations up to the ambiguity of λ. We with (without) the stellar potential. Black horizontal line checked10 that4 in all of the virialised soliton+halo+stellar denotes the value of M that was fixed in the simulation: mass simulations, presented in Ref. [29], the soliton pro- the intersection of the black and the blue lines gives the file matches that of the analytic distorted soliton, once parameter λ describing the distorted soliton. In the bot- accounting1000 for the stellar potential. In the top panel of tom panel we show the density profiles. Red lines are Fig. 14 we illustrate this point for the t = tH snapshot taken from [29], while green lines show the analytic soli- of Case100 C, described in Fig. 5 of Ref. [29] ton solutions. Clearly, the numerical profile from [29] 0.1 0.5 1 5 10 closely matches the analytically-derived distorted soliton ⇢/⇢ ρstars ρDM (sim) of Eq. (55). bkg ρsol (analytic) ρsol (no stars) 107

M [M ] 6 1× 1010 10 5× 109 105

1× 109 104 5× 108 1000 1× 108 5× 107 100 0.1 0.5 1 5 10 x [kpc]

1× 107 50 -5 -4 -4 -3 V [km/s] 5.× 10 1.× 10 5.× 10 10 circ 50 20 ⇢/10⇢bkg7 �� 107 106 ((tt==0) 0)DMDM sim 20 10 6 soliton �� 10105 ((tt==tHt)HDM) DM sim sim soliton w w// stars stars 4 10 10 0.5 1 5 10 105 ULDM+stars, t = tH ULDM, t =0 0.2 0.5 1 2 5 ULDM, t = tH 104 0.5 1 5 10 x [kpc]

FIG. 14. Top: Soliton proﬁles with and without stars, de- 0.2 0.5 1 2 5 x [kpc] scribing the inner part of the halo obtained in the numerical simulation (Case C in Fig. 5 of Ref. [29]). Green and red solid FIG. 13. Top: Soliton mass vs. λ parameter, determined lines show the stellar and ULDM densities, respectively, taken using Eq. (57) with the stellar gravitational potential as in from the simulation. Blue dashed show a distorted soliton so- the numerical simulation shown in Fig. 1 of Ref. [29]. Blue lution, computed including the stellar potential. Dotted black solid (green dashed) lines show the Mλ vs. λ with (without) shows an un-distorted soliton with the same mass. Bottom: the stellar potential. Black horizontal (dot-dashed) line de- Rotation curve for the total ULDM+stellar system (red solid) notes the mass M chosen in the simulation. Bottom: Soliton and including only the ULDM contribution (green dashed), proﬁles with and without stars, describing the results of nu- at t = tH . We also show the initial ULDM only distribution merical simulation shown in Fig. 1 of Ref. [29] (ρbkg is the at t = 0 (blue dotted). cosmological background matter density).

With stars included, the total t = tH mass distribu- Ref. [29] also simulated the soliton in an ULDM halo, tion, given by summing the red (DM) and green (stars) the output of a DM-only cosmological simulation. Then, solid lines in the upper panel of Fig. 14, leads to the a stellar mass distribution was introduced as before and rotation curve shown by the red solid line in the lower the system allowed to evolve. panel of the same figure. This rotation curve is peaked When embedded in a halo, the mass of the soliton is at small radius due to the distorted soliton, that is more not constant anymore but can grow by absorbing mass massive than the soliton-halo prediction of the DM-only from the halo. However, regardless of the large-scale halo, simulations. The inner velocity exceeds the prediction in the core region the perturbed soliton profile is still of Eq. (49) by a factor of two. We also show, by green 14 dashed line, the rotation velocity at t = tH due to ULDM the soliton profile, are included and discussed below. alone. To compare, the blue dotted curve shows the rotation curve of the initial ULDM system, extracted from 1012 the cosmological simulations. This is the same curve we m=10-19 eV m=10-20 eV showed in Fig. 5; as we discussed, it satisfies Eq. (49) to -21 1011 m=10 eV 20%. m=10-22 eV Ghez 2003 These numerical results suggest that the presence of McGinn 1989 10 Fritz 2016 10 Nuclear Bulge (disc+star cluster) NFW fit, Lindqvist 1992 ] from photometry, Launhardt (2002) Piffl (2015) baryonic matter tends to make the soliton somewhat Schodel 2014 Sofue 2009 more compact and massive than predicted by the pure 9 Sofue 2012 10 Sofue 2013 ULDM soliton–host halo relation (31). This, in turn, in- Chatzopoulos 2015 Deguchi 2004 creases the peak in the rotation curve due to the soliton, 8 Oh 2009 10 Trippe 2008 suggesting that our constraints on ULDM obtained in Gilessen 2008 enclosed mass [M Sec. IV B are robust with respect to inclusion of baryons. 107 We now consider an example, where baryonic effects are expected to be important: the Milky Way. 106

105 10-4 10-2 100 102 104 B. The Milky Way: Nuclear Bulge vs. Soliton r [pc]

Refs. [6, 7] commented that for a MW-like galaxy, FIG. 15. Spherically-averaged mass profile in the Milky Way, 22 Eq. (31) predicts that ULDM with m 10− eV should vs. ULDM soliton contributions. See text for details. produce a bump in the rotation curve∼ at a radial distance 200 pc, consistent at face value with a feature We stress that the purpose of Fig. 15 is to illustrate the observed∼ in the MW. Our results in Sec. IV B show that possible signature of ULDM in the inner MW, and not for this value of m is in tension with observations of small statistical analyses of the MW mass distribution. Mod- disc galaxies. Nevertheless, the MW provides an inter- eling the inner MW is a complicated task. The measure- esting example for studying the impact of baryons and of ment of inner kinematics of the galaxy, below a few kpc, a SMBH on an ULDM soliton, and developing intuition is subject to large systematic uncertainties due, among as to what extent this effect can be important. Our goal other issues, to the effects of the Galactic bar and spiral in this section is to provide a preliminary study along arm structures [51], which impact tangent-point velocity these lines, using photometric baryonic mass estimates. measurements like those utilised in [52, 53]. Our simpli- As an interesting outcome, we find that precision kine- fied derivation of the spherically-averaged mass profile in matical studies of the MW inner bulge could in principle Fig. 15 combines many tracers with different systematics, 21 19 be sensitive to ULDM with 10− eV . m . 10− eV, and accounts for none of these subtleties. where the analysis along the lines of Sec. IV B may be- Ref. [54] analysed the MW central gravitational po- come challenging. tential using a large set of observational constraints. In Fig. 15 shows a spherically-averaged enclosed mass addition to the classical bulge and disc, Ref. [54] found profile, derived for the MW via various dynamical trac- dynamical evidence for the presence of a mass compo- ers [38–51]. Flattening of the enclosed mass at small radii nent of 2 109 M extending to 250 pc. This mass reflects the contribution of the SMBH. From the large ra- component∼ is× visible as a mass bump∼ in Fig. 15 (see, e.g. dius (r & 10 kpc) part of the rotation curve, the param- green data points extracted from [52]). Consistent with eters of the large-scale DM host halo can be extracted comments in [6, 7], the bump is in tantalising agreement 22 by standard analysis. We demonstrate this in Fig. 15 with the soliton prediction of Eq. (31) for m = 10− eV by a black dashed line, showing an NFW profile fitted (blue shaded band). in Ref. [36] to r & 10 kpc SDSS data. The large-scale Unfortunately, there are about a billion stars in there, analysis yields for the MW an host halo mass roughly in too: the bump in the mass profile at r 200 pc has been 12 ∼ the range Mh = (0.8 2) 10 M . associated in the literature with the nuclear bulge (NB). ÷ × Given the host halo mass Mh, the soliton–host halo Ref. [55] fitted the NB mass and light by a dense disc of relation, Eq. (31), predicts the soliton mass M. This, in stars, with mass density ρ 200 M /pc3, scale height ∗ ∼ turn, allows us to extract the soliton scaling parameter . 45 pc and scale radius 230 pc. In all, the NB is ∼ 9 λ (from Eq. (17), with Mλ = M) and with it the entire thought to contain (1.4 0.6) 10 M in stars, roughly predicted soliton contribution to the enclosed mass. In enough to match the dynamically± × inferred mass. Sub- Fig. 15, shaded bands show the resulting soliton-induced sequent kinematic detection supporting the stellar mass 22 19 rotation curves for m in the range (10− 10− ) eV. and disc-like morphology of this component was given The width of the shaded bands comes from÷ varying the in [56]. Microlensing analyses [57] lend further support 12 host halo mass Mh in the range (0.8 2) 10 M . The to the results of [54–56] down to r & 220 pc. approximate effects of baryon mass÷ contributions,× which The photometrically-derived NB mass model of [55] is cannot be neglected in the inner MW and which distort superimposed as purple line in Fig. 15. We stress that 15

19 the photometric derivation is subject to large uncertain- tons all the way up to m 10− eV. For larger m, the ties due to the need to correct for very strong extinction expected soliton becomes∼ subdominant with respect to and due to unknown stellar mass-to-light ratios. What the SMBH. we learn from this photometric mass model, therefore, is Another limitation comes from absorption of the that stars could plausibly account for all of the kinemat- ULDM from the soliton by SMBH. As discusses in ically inferred mass in this region. App. B 2, the accretion rate is negligible as long as 20 Assuming that the NB is due to stars, we now use a m . 5 10− eV. However, for higher ULDM masses the toy model of this mass distribution to see its effect on time scale× of accretion can become shorter than the age an ULDM soliton. We replace the disc-like morphology of the universe and our steady-state Newtonian analysis of the NB in [55] by a spherical model with the same in this section may not apply. radially averaged mass. The nominal model, containing the NB and additional subleading components described in [55], contains 1.7 109 M in stars inside of r = VI. DISCUSSION ∼ × 6 300 pc. Adding a SMBH of MBH = 4.3 10 M [38], × we calculate soliton solutions in this baryonic potential. A. Comparison to previous work Fig. 16 shows the soliton mass as function of the λ parameter, for m = 10 22 eV. Green dashed line shows − An earlier analysis of ULDM halos including the ef- the unperturbed M vs. λ relation. Solid, dashed, and λ fect of baryons was done in [2] (see also [3]). This work dotted black lines show the relation for the nominal NB attempted to fit the rotation curves of spiral galaxies model and for two other models, obtained by scaling the from [58], assuming that the entire ULDM halo is con- NB mass density by an over-all factor of 0.5 and 2, re- tained in a giant soliton (see also [59] for a similar ap- spectively. For orientation, shaded blue band shows the proach in modeling low surface brightness galaxies). This soliton mass predicted by Eq. (31) for a host halo with 23 12 exercise led [2] to report m < 10− eV, in tension with mass Mh = (0.8 2) 10 M . − × Ly-α data. This exercise is different from the current work. As suggested by numerical simulations, we expect M [M ] 22 heavier ULDM particles with m & 10− eV to produce galaxies with a large scale host halo following roughly 1× 1010 the usual NFW profile, in which the soliton affects the rotation curve only in the inner part of the halo. 9 5× 10 Refs. [6, 8] performed a Jeans analysis, fitting a soliton+halo configuration to a small sample of dispersion-dominated MW-satellite dwarf spheroidal galaxies (dSph). They found that the modelling of stellar kinematics in the dSph is consistent with a cored inner re- 1× 109 mNB(300) = mNB(300) = gion, as expected in ULDM, with good fits to the data for 8.5 108 M 3.4 109 M 22 ⇥ ⇥ m . 10− eV. The fitted core found in [6] was consistent 5× 108 5.× 10-4 0.001 0.002 0.005 with expectations from the soliton–host halo relation. In [8], the transition between an assumed host halo NFW profile and the soliton was modelled using correspond- FIG. 16. Soliton mass vs. λ parameter, accounting for the ing free parameters. To conclude, these analyses con- spherically-averaged gravitational potential due to stars [55]. sidered a small sample of dispersion-dominated galaxies, The ULDM particle mass is m = 10−22 eV. See text for more as opposed to our larger sample of rotation-dominated 22 details. galaxies. Their best-fit ULDM mass m . 10− eV is disfavoured by our findings. 9 For Mλ & 3 MNB 5 10 M , the NB makes a Ref. [26] fitted ULDM soliton+halo profiles to negligible impact on the∼ soliton.× For larger ratio of the SPARC [28] galaxies. Without a soliton–host halo re- stellar to ULDM mass, Mλ . MNB, the NB becomes lation, Ref. [26] assigned separate free parameters to the important, contracting the soliton profile. For the MW, NFW halo and to the soliton, for each galaxy. As a re- this is the parametric region predicted by Eq. (31), im- sult of this freedom, it is difficult to understand from that plying that the solitons would receive significant distor- analysis if ULDM is ruled out or not in the mass range 22 21 tion. In Fig. 15 we illustrated this effect by presenting, m = (10− 10− ) eV. Nevertheless, Ref. [26] did state ÷ 23 in shaded bands, the soliton mass profiles computed ac- a best fit of m 0.5 10− eV, and noted the tension counting for the nominal NB model. We observe that with Ly-α data,≈ in agreement× with our findings. Our 22 19 the solitons for m in the range (10− 10− ) eV are work here shows that the soliton–host halo relation re- expected to affect the potential at an order÷ unity level. duces the modelling freedom, and leads to disagreement Thus, a dedicated analysis using our formalism to calcu- with rotation curve data. late steady state soliton profiles consistent with a given Recently, Ref. [60] addressed the problem from a differ- baryonic mass model could be sensitive to ULDM soli- ent angle. It builds on the results of Ref. [61] that fitted 16 rotations curves of a sample of galaxies assuming that one would conclude that the relaxation takes longer than 22 their halos are described by the Burkert profile. Ref. [60] the age of the universe if m & 3 10− eV. points out a correlation between the parameters of the However, it can be incorrect to× use Eq. (61) for the esti- fitted profiles — the central dark matter density ρc and mate of the soliton formation time inside ULDM halos in the size of the core Rc, the cosmological context. Indeed, this equation has been derived and tested in the kinetic regime corresponding β ρ R− (60) c ∝ c to large velocity dispersion of ULDM particles, such that their de Broglie wavelength is much shorter than the size with β 1. This is different from β = 4 predicted by of the system. Starting from initial conditions of a gas ULDM∼ if the cores are identified with the solitons, see of wavepackets with large velocity dispersion in a box, Eq. (30). Thus, Ref. [60] concludes that ULDM cannot Ref. [63] observed a fast formation of a virialised halo, explain the origin of the halo cores, and cannot solve the followed by a long period of kinetic evolution populat- core–cusp problem of ΛCDM. Our results are consistent ing the low-lying energy levels, until the system was able with these findings and strengthen them by showing that to form the soliton through Bose–Einstein condensation. the soliton–host halo relation implies a tension between The time scale (61) refers to the long second stage of the predictions of ULDM and the data for m (10 22 − process. On the other hand, the cosmological initial con- 10 21) eV, disfavouring this mass range altogether.∼ ÷ − ditions prior to virialisation are characterised by very low velocities, where the kinetic description does not apply. In this situation a long relaxation seems unnecessary and B. Caveats the formation of the soliton can proceed much faster, on the scale of the halo free-fall time, see discussion in [63]. 1. Soliton formation time This is supported by the simulations with cosmological initial conditions which appear to imply formation of the Our analysis has been based on the assumption that soliton in every halo at the moment of virialisation [6, 62]. ULDM solitons satisfying the soliton–host halo relation We believe that the issue of the soliton formation time exist in the centres of all (or most) galactic halos. For- requires further investigation, including possible effects mation of such solitons within the lifetime of the universe of baryonic matter. has been demonstrated in numerical simulations [6, 62] 22 for ULDM masses m 10− eV and halos with mass (109 1011) M . Here∼ we discuss whether such solitons 2. Non-gravitational interactions have÷ enough time to form for higher values of m and larger halos. In our analysis we have neglected any non-gravitational Recently, Ref. [63] considered the scenario, where soli- interactions of the ULDM particles. Let us discuss under tons form as a result of gravitational relaxation of the which conditions this approximation is justified. As an inner part of the halo. In this case the relaxation time example we consider quartic self-interaction of the scalar can be estimated as, field φ with coupling constant κ, 4 b√2 m3v6 κφ τ = , (61) δV (φ) = . (63) relax 12π3G2ρ2Λ 4 The self-interaction is negligible as long as the corre- where v and ρ are the mean velocity and density in the sponding potential energy is much smaller than the gra- inner part of the halo, Λ = log(mvR) is the Coulomb dient energy of the field, logarithm depending on the size R of the relaxation region, and b is a numerical coefficient close to one. The κ φ4 ( φ)2 | | ∇ . (64) formula has been confirmed in [63] by numerical simula- 4 2 tions that found b to range between 0.7 0.9 depending ÷ For a soliton with the scale parameter λ, this translates on the precise form of the initial conditions. Substitut- into a bound, ing into Eq. (61) the values for a typical halo and taking 2 Λ 4, we obtain, 2m ∼ κ < 2 . (65) | | x ρcλ 3 cλ 8 m τrelax 7 10 22 Using the soliton–host halo relation in the form of ∼ × 10− eV 6 2 (62) Eqs. (29), (30) we obtain the condition, v ρ − yr . 2 3 2 3 × 100 km/s 0.1M /pc m Mh − 93 κ < 4 10− 22 12 . (66) | | × 10− eV 10 M We observe that the relaxation time has strong depen- dence on the velocity and the density in the central re- If this condition is satisfied, our analysis is applicable. gion of the halo, that have large uncertainties. Fixing Otherwise, the effects of the self-interaction on the soliton the fiducial values of v = 100 km/s and ρ = 0.1M /pc3 properties must be taken into account.

While the condition (66) appears stringent, it is natu- less than a factor of two, around this relation. This E/M rally fulfilled in a broad class of theories containing axion- relation could apply to real galaxies. like particles. These typically have a periodic potential The E/M soliton–host halo relation implies that the of the form, peak circular velocity of the large-scale halo should approximately reproduce itself in the inner soliton region. V (φ) = m2f 21 cos(φ/f) . (67) − This can be tested against observations without free pa- 22 18 rameters. Contrasting this prediction with high resolu- For the masses of interest m (10− 10− ) eV, the ∼ ÷ 16 tion disc galaxy data, we showed that ULDM in the mass axion periodicity should be in the range f (10 22 21 17 ∼ ÷ range m (10− 10− ) eV is disfavoured by observa- 10 ) GeV to produce the right dark matter abundance tions. Given∼ that smaller÷ m is in tension with cosmolog- via the misalignment mechanism (see e.g. [12]). Expand- ical measurements, this disfavours ULDM as a solution ing the potential (67) in powers of φ we obtain the value to the small scale puzzles of ΛCDM. of the coupling, We also analysed the effect of a fixed background m2 distribution of baryons and of a super-massive black κ = hole (SMBH) on the soliton solution. We showed that − 6f 2 (68) this analysis would be important towards using high- 2 2 97 m f − resolution kinematical data in the Milky Way or Milky 1.7 10− 22 17 . ≈ − × 10− eV 10 GeV Way-like galaxies. Mapping the Milky Way gravitational 2 potential down to r 10− pc may allow to probe ULDM ∼ 19 We see that (66) is indeed satisfied for all galactic halos with mass up to m 10− eV. considered in this paper, independently of the ULDM It should be stressed∼ that the E/M soliton–host halo mass. relation (35) is an empirical result, deduced from numer- Similar analysis can be performed in the case of non- ical simulations without baryons and tested in a limited gravitational interactions between ULDM and baryonic range of ULDM and halo masses. The importance of its matter, though in this case it will be strongly model- phenomenological implications motivates further investi- dependent. gation to better understand the physics underlying this relation and map out its domain of validity. Throughout the paper we have assumed that ULDM VII. SUMMARY comprises the total amount of dark matter. We leave for future the study of scenarios where only a fraction of Bosonic ultra-light dark matter (ULDM) is an inter- dark matter is in the form of ULDM. esting paradigm for dark matter. For particle mass 22 m 10− eV, ULDM would form solitonic cores in the ∼ center of galaxies, and it has been suggested that this ACKNOWLEDGMENTS could solve several puzzles of ΛCDM on small scales. We analysed the results of numerical simulations of We thank James Hung-Hsu Chan, Josh Eby, Lam Hui, ULDM, which have found scaling relations between the Dmitry Levkov, Moti Milgrom, Eran Ofek, Alexander mass of the central soliton and the mass or energy of Panin, and Igor Tkachev for discussions, and Yossi Nir, the host halo. Simulations by different groups converge Scott Tremaine, and Eli Waxman for comments on an on a soliton profile in good agreement with the analyt- early version of the draft. KB thanks the participants of ical solution of the Schroedinger-Poisson (SP) equation, the VBSM Polynesia workshop, where part of this work admitting important analytic insight into the numerical was conducted, for useful discussions. KB is incumbent results. of the Dewey David Stone and Harry Levine career devel- We showed that the simulations of Ref. [13] contain opment chair. The work of KB and NB is supported by a central soliton that dominates the total energy of the grant 1937/12 from the I-CORE program of the Plan- entire soliton+halo system. This situation is unlikely to 9 ning and Budgeting Committee and the Israel Science describe realistic galactic halos with mass above 5 Foundation and by grant 1507/16 from the Israel Sci- 7 22 3/2 ∼ × 10 m/10− eV − M . ence Foundation. SS is grateful to Columbia University

We have demonstrated that the soliton–host halo re- for hospitality during completion of this work. lation found in the simulations of Refs. [6, 7] can be summarised by the statement, that (E/M) soliton = (E/M) . The simulations of [7] show a small| spread, |halo Appendix A: Soliton mass–host halo mass relation, Mh vs. M200

Ref. [7] deﬁned the halo virial mass Mh and radius Rv, 9 This does not invalidate the beautiful simulations of [13]. We only think that their setup for initial conditions does not match 4π real galaxies. M = R3 ζ(z) ρ , (A1) h 3 v m0 18

where ρm0 is the present matter density and ζ(z) = this solution (satisfying χ(0) = 1) by χ1(r; A), accom- (18π2 82(1 Ω (z)) 39(1 Ω (z))2)/Ω (z). The panied by the potential Φ (r; A). Having found χ (r; A) − − m − − m m 1 1 halo energy per unit mass was estimated as Eh/Mh = and Φ1(r; A) for any value of A, the physical solution 1 1 2 3 4π 3 3 3 χλ(r; A) satisfying Eqs. (B1-B2) with boundary condi- 3GMh/(10Rv) = G 10 3 ζ Ωm0 ρc Mh . In 2 − − tion χλ(0; A) = λ is given by Sec. IV, we prefer to work with M200 and R200 and have calculated the halo energy per unit mass, using Eq. (42), 2 1 χλ(r; A) = λ χ1(λr; A/λ), (B4) 3 1 2 c˜ 4πδc 3 3 2 as E200/M200 = G 2 ρc M . Φλ(r; A) = λ Φ1(λr; A/λ). (B5) 4 ln(1+c) c 200 − ( − 1+c ) To express the soliton mass–host halo mass relation, Defining Mλ(A), with obvious notation, the soliton mass Eq. (45), in terms of Mh of Ref. [7], we need to do so at is fixed E /M = E /M E/M. This gives: 200 200 h h ≡ 1 Mλ(A) = λM1(A/λ). (B6)   2 1 1 3 4π 3 2 3  ζ Ω  Fig. 17 shows the density profile χ (r; A) for different M200  10 3 m0  1 = 1 .(A2) 2   values of A. For A . 10− the solution converges to Mh E/M 3  c˜ 4πδc  2 the unperturbed χ1. For A & 1 the solution approaches 4 (ln(1+c) c ) Ar − 1+c χ1(r; A ) e− which is nothing, but the wave- function→ of ∞ the→ ground state in the Coulomb potential. For c in the range 5 30, the RHS of Eq. (A2) varies This means that for A/λ 1, solitons with different val- between 0.58 and 0.47.÷ Plugging this result into Eq. (45) & ues of λ have the same profile up to over-all normaliza- shows that it agrees with Eq. (31) of [7] to 20%. 2 Ar tion: χ (r; A λ) λ e− . In this limit the soli- λ & → ton core radius is rcλ(A) ln(2)/(2A), and the mass is 4 2 → λ Mpl Appendix B: Adding a super-massive black hole Mλ(A) 4A3 m . More generally, for any A the mass- radius relation→ is Most galaxies, if not all, host a SMBH, and we should check how it affects the ULDM soliton. This is an im- Mλ(A) rcλ(A) = M1(A/λ) rc1(A/λ). (B7) portant point to verify and we do this analysis here in The mass-radius product is shown by the solid blue some detail. The upshot of our discussion is that a curve in Fig. 18 (top panel). The large-A analytic so- SMBH would not affect our results for small disc galax- 2 ln(2) Mpl ies, whereas in the MW it becomes relevant at the upper lution, M1(A) rc1(A) A 8A4 m (recall r = m x), 20 | →∞ → end of considered ULDM masses, m & 5 10− eV. is shown in dotted black, normalised to the A = 0 result. We start with a Newtonian analysis, dealing× with the ULDM configuration far from the BH Schwarzschild ra-1.0 dius. Then we will consider the limitations of this analy- 2(r; A) 0.8 1 sis imposed by absorption of ULDM on SMBH that leads 1.0 to the decay of the soliton+SMBH configuration. A=0=0 0.6 A=0.1=0.1 �� 0.8 A=1=1 0.4 A=2=2 1. Soliton shape in the presence of SMBH 0.6 0.2 Adding a SMBH, coincident with the soliton center of 0.4 0.0 mass, changes Eqs. (7-8) into 0 1 2 3 4 5 0.2 A ∂2 (rχ) = 2r Φ γ χ, (B1) r − r − 0.0 2 2 0 1 2 3 4 5 ∂r (rΦ) = rχ , (B2) r

where 2 FIG. 17. χ1(A) proﬁles, with A = 0, 0.1, 1, 2 shown in blue, orange, green and red, respectively. For the A = 1 and A = 2 A = GMBH m (B3) proﬁles, the large-A analytic solution is shown in dotted black. 6 MBH m 3 10− 6 22 . ≈ × 4 10 M 10− eV × As discussed in Sec. III A, Eqs. (31) and (35) are equiv- We chose the reference value for MBH to represent the alent when the ULDM soliton is self-gravitating, but this case of the MW. equivalence is broken by the SMBH gravitational poten- It remains convenient to solve the problem using tial. It is unlikely that Eq. (31) remains realistic when boundary conditions with χ(0) = 1. Let us denote the SMBH becomes important. Instead, Eq. (35) may 19

still be valid. In computing the ULDM energy we need perturbed soliton mass using Eq. (15), to account for the SMBH contribution, M A/λ 2.06 BH (B10) Z 2 ! ≈ Mλ(A = 0) 3 ψ Φ A 2 E(A) = d x |∇ | + ψ , 1 2m2 2 − m x | | MBH Mλ(A = 0) − 0.8 . ≈ 4 106 M 107 M (B8) × Next, we see from Eqs. (B4), (B5) that A/λ gives an which means that effective A-parameter that must be substituted in the χ1-soliton to obtain the actual solution. Thus, whether 3 Eλ(A) = λ E1(A/λ). (B9) the effects of SMBH are important or not can be read from the ratio of the SMBH mass to the mass of the undistorted soliton. The large A limit (attained for a χλ(r; A) soliton at A & 4 2 λ Mpl To analyse the rotation curve for the soliton+SMBH λ) is Eλ(A ) . In this limit Eλ/Mλ → ∞ → − 8A m system, it is useful to view Vcirc for the χ1(A) soliton, tends to a λ-independent constant, Eλ/Mλ(A ) 2 → ∞ → obtained for different values of A. This is shown in A /2, because the soliton is not held together by self Fig. 20. Because χ (A) solitons are a scale transforma- gravity,− but by the SMBH potential. The E/M relation λ tion of χ1(A/λ), plotting χ1(A) with different values of for the χ1(A) soliton is shown in the bottom panel of A is equivalent to plotting the rescaled velocity curve of Fig. 18. solitons at different values of MBH /Mλ. For example, us- In Fig. 19, solid lines show E/M for the soliton with ing Eq. (B10), the A = 0.2 curve in Fig. 20 corresponds 2 different values of particle mass m, in the presence of a to a soliton with Mλ 10 MBH ; while the A = 10− 6 ≈ SMBH with MBH = 4 10 M . The dashed lines show corresponds to Mλ 200 MBH . the undistorted result,× obtained from Eq. (25). The large ≈ For Mλ & 10 MBH , the soliton’s circular velocity peak A limit with constant Eλ/Mλ is attained when the soliton and the peak location are affected very little by the mass becomes smaller than the SMBH mass. SMBH. However, at small x approaching the SMBH, To further clarify this point we relate A/λ to the un- the velocity curve turns back up reflecting the gravitational potential of the SMBH itself. This implies that the rotation curve of a galaxy hosting a SMBH M1rc1(A) may not decrease appreciably below xpeak,λ, even when M1rc1(0) MBH is much smaller than Mλ. For example, even for 1 MBH = 0.05 Mλ (A = 0.1 in Fig. 20), the circular veloc- 0.500 ity decreases by only 25% at x < xpeak,λ before it goes back up due to the BH.∼ For the same parameters, Fig. 18 0.100 shows that Eλ/Mλ is essentially unperturbed, meaning 0.050 that the soliton-host halo relation of Eq. (35) and the rotation curve properties discussed in Sec. IV remain un- 0.010

0.005 aﬀected. Such a galaxy would have an approximately

10 10 10 10

0.05 0.10 0.50 1 E(A) 10 10 9 8 7 A 6 M(A) -6 E1/M1(A) 10

E1/M1(0)

10-7 10

- 100 11

50 10-8

- 10-9 10

eV 10

10-10 20 20

5 10 eV 10

eV 10

- 9

�� - 21 21

10-11 10 eV eV 10

- 22 22

1 10 eV

6 8 9 10 10 - 0.5 10 107 10 10 10 8

0.05 0.10 0.50 1 5 10

A M(A)[M ]

- FIG. 18. Mass-core radius relation (top panel) and E/M FIG. 19. Solid lines: energy per unit mass vs. mass for 7

(bottom panel) for the χ1(A) soliton, as function of A. The a soliton, including the eﬀect of a SMBH with MBH = 4 ×

large-A analytic solution is shown in dashed black. 10 M . Dashed lines: switching oﬀ the SMBH. 10 - 6 20

Vcirc,1 scalar field configuration. 5 The growth of BH mass due to an infalling flux of particles with density ρ is, 5 1 dM 32π2(GM )3m2ρ BH = BH . (B13) dt k(1 e ζ ) 0.50 − −

1 AA=0=0 In the case of the soliton we can estimate the absorption 4 AA=10=-4 10 2 rate by substituting into this expression the central den- 0.50 AA=10=-2 10 �� 0.10 AA=0.1=0.1 sity of the soliton and the characteristic momentum of AA=0.2=0.2 particles in the soliton that can be read from the soliton 0.05 AA=0.5=0.5 profile (30), 0.10 0.05 0.10 0.50 1 5 10 r 1 k 0.3 x− . (B14) 0.05 ∼ c FIG. 20. Rotation curve for the χ (A) soliton. 0.05 0.10 0.50 1 5 10 1 This approximation is justified as long as the mass of the soliton is bigger than the BH mass, so that the effect of BH on the soliton shape can be neglected. flat rotation curve all the way in to the region of SMBH The expression for the accretion rate simplifies in the dominance. limiting cases when the parameter ζ is small or large. The impact of a SMBH on the analysis of Sec. IV can Using Eq. (29) (which is consistent with the soliton–host be concluded as follows. For M 4M (correspond- λ & BH halo relation) we obtain, ing, by using Eq. (B10), to effective A/λ 0.5), Fig. 18 ≈ shows that Eλ/Mλ is corrected by . 25% compared to 1 3 its unperturbed value. Eqs. (28) and (44) then imply m MBH Mh − ζ = 0.16 22 6 12 . that λ and maxVcirc,λ are corrected by an insignificant 10− eV 4 10 M 10 M × 12%. For a large SMBH with MBH & Mλ, on the ∼other hand, the SMBH gravitational potential itself must If ζ 1, the characteristic accretion time, during which produce high rotation velocity at xpeak,λ, comparable or the BH mass grows by a factor 2 is, higher than what the soliton itself would provide. As a 2 result, besides from the fact that the rotation curve may 17 m − τζ 1 2.4 10 22 (B15) not decrease towards lower x below the soliton peak, our ∼ × 10− eV analysis in and observational limits from Sec. IV should 1 4 M − M − 3 be robust against the effect of a SMBH10. BH h yr . × 4 106M 1012M × In the opposite regime, ζ 1, the characteristic time is, 2. Absorption of soliton by SMBH 3 18 m − We now discuss accretion of ULDM from the soli- τζ 1 1.5 10 22 (B16) ∼ × 10− eV ton onto SMBH. To estimate the lifetime of the soli- 2 1 M − M − ton+SMBH configuration we follow the approach of [12] BH h yr . (see also [65, 66] for the study of scalar bound states × 4 106M 1012M × in external Schwarzschild metric and [67] for a numerical analysis of the lifetime of soliton+BH system). Un- In deriving these expressions we made use of Eqs. (29,30). ruh [68] has derived the cross section for absorption Both expressions give the accretion time longer than of a scalar particle with mass m and momentum k by the age of the universe for MW’s SMBH with MBH 6 20 ≈ a Schwarzschild BH, whose size is much smaller than 4.3 10 M and m . 5 10− eV. Our Newtonian × × the Compton wavelength of the particle. In the non- analysis, addressing stationary ULDM soliton solutions, relativistic limit, k m, it has the form, is adequate in this case. For galaxies with a more massive 8 SMBH, e.g. M31 with MBH 10 M [69], Eq. (B16) 32π2(GM )3m3 ∼ σ = BH , (B11) implies that absorption of the ULDM field into the k2(1 e ζ ) SMBH becomes important already for m 6 10 21 eV, where − − & − and our stationary Newtonian analysis is not× expected to 2 ζ = 2πGMBH m /k . (B12) capture the physics correctly for such values of m. Fi- 22 nally, for the reference value of m = 10− eV, our anal- As the cross section is s-wave dominated, it is appropriate 10 ysis should be valid as long as MBH . 10 M . for calculation of accretion from a spherically symmetric 21

[1] W. Hu, R. Barkana, and A. Gruzinov, “Cold and fuzzy [14] P. Svrcek and E. Witten, “Axions In String Theory,” dark matter,” Phys. Rev. Lett. 85 (2000) 1158–1161, JHEP 06 (2006) 051, arXiv:hep-th/0605206 [hep-th]. arXiv:astro-ph/0003365 [astro-ph]. [15] A. Arvanitaki, S. Dimopoulos, S. Dubovsky, N. Kaloper, [2] A. Arbey, J. Lesgourgues, and P. Salati, “Quintessential and J. March-Russell, “String Axiverse,” Phys. Rev. haloes around galaxies,” Phys. Rev. D64 (2001) D81 (2010) 123530, arXiv:0905.4720 [hep-th]. 123528, arXiv:astro-ph/0105564 [astro-ph]. [16] D. J. E. Marsh, “Axion Cosmology,” Phys. Rept. 643 [3] J. Lesgourgues, A. Arbey, and P. Salati, “A light scalar (2016) 1–79, arXiv:1510.07633 [astro-ph.CO]. field at the origin of galaxy rotation curves,” New [17] A. Del Popolo and M. Le Delliou, “Small scale problems Astron. Rev. 46 (2002) 791–799. of the ΛCDM model: a short review,” Galaxies 5 no. 1, [4] P.-H. Chavanis, “Mass-radius relation of Newtonian (2017) 17, arXiv:1606.07790 [astro-ph.CO]. self-gravitating Bose-Einstein condensates with [18] B. Bozek, D. J. E. Marsh, J. Silk, and R. F. G. Wyse, short-range interactions: I. Analytical results,” Phys. “Galaxy UV-luminosity function and reionization Rev. D84 (2011) 043531, arXiv:1103.2050 constraints on axion dark matter,” Mon. Not. Roy. [astro-ph.CO]. Astron. Soc. 450 no. 1, (2015) 209–222, [5] P. H. Chavanis and L. Delfini, “Mass-radius relation of arXiv:1409.3544 [astro-ph.CO]. Newtonian self-gravitating Bose-Einstein condensates [19] E. Armengaud, N. Palanque-Delabrouille, D. J. E. with short-range interactions: II. Numerical results,” Marsh, J. Baur, and C. Yche, “Constraining the mass of Phys. Rev. D84 (2011) 043532, arXiv:1103.2054 light bosonic dark matter using SDSS Lyman-α forest,” [astro-ph.CO]. Mon. Not. Roy. Astron. Soc. 471 no. 4, (2017) [6] H.-Y. Schive, T. Chiueh, and T. Broadhurst, “Cosmic 4606–4614, arXiv:1703.09126 [astro-ph.CO]. Structure as the Quantum Interference of a Coherent [20] V. Irˇsiˇc,M. Viel, M. G. Haehnelt, J. S. Bolton, and Dark Wave,” Nature Phys. 10 (2014) 496–499, G. D. Becker, “First constraints on fuzzy dark matter arXiv:1406.6586 [astro-ph.GA]. from Lyman-α forest data and hydrodynamical [7] H.-Y. Schive, M.-H. Liao, T.-P. Woo, S.-K. Wong, simulations,” Phys. Rev. Lett. 119 no. 3, (2017) T. Chiueh, T. Broadhurst, and W. Y. P. Hwang, 031302, arXiv:1703.04683 [astro-ph.CO]. “Understanding the Core-Halo Relation of Quantum [21] J. Zhang, J.-L. Kuo, H. Liu, Y.-L. S. Tsai, K. Cheung, Wave Dark Matter from 3D Simulations,” Phys. Rev. and M.-C. Chu, “Is Fuzzy Dark Matter in tension with Lett. 113 no. 26, (2014) 261302, arXiv:1407.7762 Lyman-alpha forest?,” arXiv:1708.04389 [astro-ph.GA]. [astro-ph.CO]. [8] D. J. E. Marsh and A.-R. Pop, “Axion dark matter, [22] T. Kobayashi, R. Murgia, A. De Simone, V. Irˇsiˇc,and solitons and the cusp-core problem,” Mon. Not. Roy. M. Viel, “Lyman-alpha Constraints on Ultralight Scalar Astron. Soc. 451 no. 3, (2015) 2479–2492, Dark Matter: Implications for the Early and Late arXiv:1502.03456 [astro-ph.CO]. Universe,” arXiv:1708.00015 [astro-ph.CO]. [9] S.-R. Chen, H.-Y. Schive, and T. Chiueh, “Jeans [23] E. Calabrese and D. N. Spergel, “Ultra-Light Dark Analysis for Dwarf Spheroidal Galaxies in Wave Dark Matter in Ultra-Faint Dwarf Galaxies,” Mon. Not. Roy. Matter,” Mon. Not. Roy. Astron. Soc. 468 no. 2, (2017) Astron. Soc. 460 no. 4, (2016) 4397–4402, 1338–1348, arXiv:1606.09030 [astro-ph.GA]. arXiv:1603.07321 [astro-ph.CO]. [10] B. Schwabe, J. C. Niemeyer, and J. F. Engels, [24] A. X. Gonzáles-Morales, D. J. E. Marsh, J. Peñarrubia, “Simulations of solitonic core mergers in ultralight and L. UreñaLópez, “Unbiased constraints on axion dark matter cosmologies,” Phys. Rev. D94 no. 4, ultralight axion mass from dwarf spheroidal galaxies,” (2016) 043513, arXiv:1606.05151 [astro-ph.CO]. Mon. Not. Roy. Astron. Soc. 472 (2017) 1346, [11] J. Veltmaat and J. C. Niemeyer, “Cosmological arXiv:1609.05856 [astro-ph.CO]. particle-in-cell simulations with ultralight axion dark [25] V. H. Robles and T. Matos, “Flat Central Density matter,” Phys. Rev. D94 no. 12, (2016) 123523, Profile and Constant DM Surface Density in Galaxies arXiv:1608.00802 [astro-ph.CO]. from Scalar Field Dark Matter,” Mon. Not. Roy. [12] L. Hui, J. P. Ostriker, S. Tremaine, and E. Witten, Astron. Soc. 422 (2012) 282–289, arXiv:1201.3032 “Ultralight scalars as cosmological dark matter,” Phys. [astro-ph.CO]. Rev. D95 no. 4, (2017) 043541, arXiv:1610.08297 [26] T. Bernal, L. M. Fernández-Hernández,T. Matos, and [astro-ph.CO]. M. A. Rodrguez-Meza, “Rotation Curves of [13] P. Mocz, M. Vogelsberger, V. H. Robles, J. Zavala, High-Resolution LSB and SPARC Galaxies in Wave M. Boylan-Kolchin, A. Fialkov, and L. Hernquist, (Fuzzy) and Multistate (Ultra-light Boson) Scalar Field “Galaxy formation with BECDM: I. Turbulence and Dark Matter,” arXiv:1701.00912 [astro-ph.GA]. relaxation of idealized haloes,” Mon. Not. Roy. Astron. [27] W. J. G. de Blok and A. Bosma, “High-resolution Soc. 471 no. 4, (2017) 4559–4570, arXiv:1705.05845 rotation curves of low surface brightness galaxies,” [astro-ph.CO]. Astron. Astrophys. 385 (2002) 816, arXiv:astro-ph/0201276 [astro-ph]. [28] F. Lelli, S. S. McGaugh, and J. M. Schombert, “SPARC: 10 This conclusion seems in contradiction with the claim of Ref. [64], Mass Models for 175 Disk Galaxies with Spitzer who proposed that ULDM solitons distorted by SMBHs could Photometry and Accurate Rotation Curves,” Astron. J. explain the MBH − σ correlation, observed between SMBHs and 152 (2016) 157, arXiv:1606.09251 [astro-ph.GA]. the stellar bulge of the galaxies hosting them. 22

[29] J. H. H. Chan, H.-Y. Schive, T.-P. Woo, and T. Chiueh, [43] S. Trippe, S. Gillessen, O. E. Gerhard, H. Bartko, T. K. “How do stars affect ψDM halos?,” ArXiv e-prints Fritz, H. L. Maness, F. Eisenhauer, F. Martins, T. Ott, (Dec., 2017) , arXiv:1712.01947. K. Dodds-Eden, and R. Genzel, “Kinematics of the old [30] T. Rindler-Daller, T. Rindler-Daller, P. R. Shapiro, and stellar population at the galactic centre,” AAP 492 P. R. Shapiro, “Finding New Signature Effects on (Dec., 2008) 419–439, arXiv:0810.1040. Galactic Dynamics to Constrain [44] S. Oh, S. S. Kim, and D. F. Figer, “Mass Distribution Bose-Einstein-Condensed Cold Dark Matter,” in 6th in the Central Few Parsecs of Our Galaxy,” J. Korean International Meeting on Gravitation and Cosmology Astron. Soc. 42 (2009) 17, arXiv:0906.0765 Guadalajara, Jalisco, Mexico, May 21-25, 2012, [astro-ph.GA]. pp. 163–182. 2012. arXiv:1209.1835 [astro-ph.CO]. [45] R. Schdel, A. Feldmeier, D. Kunneriath, S. Stolovy, https://inspirehep.net/record/1184893/files/ N. Neumayer, P. Amaro-Seoane, and S. Nishiyama, arXiv:1209.1835.pdf. “Surface brightness profile of the Milky Ways nuclear [31] V. Desjacques, A. Kehagias, and A. Riotto, “Impact of star cluster,” Astron. Astrophys. 566 (2014) A47, ultralight axion self-interactions on the large scale arXiv:1403.6657 [astro-ph.GA]. structure of the Universe,” Phys. Rev. D97 no. 2, [46] S. Chatzopoulos, T. Fritz, O. Gerhard, S. Gillessen, (2018) 023529, arXiv:1709.07946 [astro-ph.CO]. C. Wegg, R. Genzel, and O. Pfuhl, “The old nuclear star [32] R. Ruffini and S. Bonazzola, “Systems of selfgravitating cluster in the milky way: dynamics, mass, statistical particles in general relativity and the concept of an parallax, and black hole mass,” Monthly Notices of the equation of state,” Phys. Rev. 187 (1969) 1767–1783. Royal Astronomical Society 447 no. 1, (2014) 948–968. [33] T.-P. Woo and T. Chiueh, “High-Resolution Simulation [47] T. Fritz, S. Chatzopoulos, O. Gerhard, S. Gillessen, on Structure Formation with Extremely Light Bosonic R. Genzel, O. Pfuhl, S. Tacchella, F. Eisenhauer, and Dark Matter,” Astrophys. J. 697 (2009) 850–861, T. Ott, “The nuclear cluster of the milky way: Total arXiv:0806.0232 [astro-ph]. mass and luminosity,” The Astrophysical Journal 821 [34] J. A. Eby, Phenomenology and Astrophysics of no. 1, (2016) 44. Gravitationally-Bound Condensates of Axion-Like [48] Y. Sofue, M. Honma, and T. Omodaka, “Unified Particles. PhD thesis, Cincinnati U., 2017. rotation curve of the galaxydecomposition into de http://lss.fnal.gov/archive/thesis/2000/ vaucouleurs bulge, disk, dark halo, and the 9-kpc fermilab-thesis-2017-19.pdf. rotation dip,” Publications of the Astronomical Society [35] M. Boylan-Kolchin, V. Springel, S. D. M. White, and of Japan 61 no. 2, (2009) 227–236. A. Jenkins, “There’s no place like home? Statistics of [49] Y. Sofue, “A Grand Rotation Curve and Dark Matter Milky Way-mass dark matter halos,” Mon. Not. Roy. Halo in the Milky Way Galaxy,” Publ. Astron. Soc. Astron. Soc. 406 (2010) 896, arXiv:0911.4484 Jap. 64 (2012) 75, arXiv:1110.4431 [astro-ph.GA]. [astro-ph.CO]. [50] Y. Sofue, “Rotation Curve and Mass Distribution in the [36] T. Piffl et al., “Constraining the Galaxy’s dark halo Galactic Center — From Black Hole to Entire Galaxy with RAVE stars,” Mon. Not. Roy. Astron. Soc. 445 —,” Publ. Astron. Soc. Jap. 65 (2013) 118, no. 3, (2014) 3133–3151, arXiv:1406.4130 arXiv:1307.8241 [astro-ph.GA]. [astro-ph.GA]. [51] L. Chemin, F. Renaud, and C. Soubiran, “The incorrect [37] J. Vicens, J. Salvado, and J. Miralda-Escudé,“Bosonic rotation curve of the milky way,” 1504.01507v2. dark matter halos: excited states and relaxation in the [52] Y. Sofue, “Rotation Curve and Mass Distribution in the potential of the ground state,” arXiv:1802.10513 Galactic Center — From Black Hole to Entire Galaxy [astro-ph.CO]. —,” Publ. Astron. Soc. Jap. 65 (2013) 118, [38] A. M. Ghez, E. Becklin, G. Duchene, S. D. Hornstein, arXiv:1307.8241 [astro-ph.GA]. M. Morris, S. Salim, and A. Tanner, “Full three [53] Y. Sofue, “Dark halos of M-31 and the Milky Way,” dimensional orbits for multiple stars on close approaches Publ. Astron. Soc. Jap. 67 no. 4, (2015) 75, to the central supermassive black hole,” Astron. Nachr. arXiv:1504.05368 [astro-ph.GA]. 324 (2003) S1, arXiv:astro-ph/0303151 [astro-ph]. [54] M. Portail, O. Gerhard, C. Wegg, and M. Ness, [39] A. Feldmeier-Krause, L. Zhu, N. Neumayer, G. van de “Dynamical modelling of the galactic bulge and bar: Ven, P. T. de Zeeuw, and R. Schdel, “Triaxial the Milky Way’s pattern speed, stellar and dark matter orbit-based modelling of the milky way nuclear star mass distribution,” Mon. Not. Roy. Astron. Soc. 465 cluster,” 1701.01583v1. no. 2, (2017) 1621–1644, arXiv:1608.07954 [40] M. T. McGinn, K. Sellgren, E. E. Becklin, and D. N. B. [astro-ph.GA]. Hall, “Stellar kinematics in the galactic center,” [55] R. Launhardt, R. Zylka, and P. G. Mezger, “The Astrophysics Journal 338 (March, 1989) 824–840. nuclear bulge of the galaxy. 3. Large scale physical [41] M. Lindqvist, H. J. Habing, and A. Winnberg, “Oh/ir characteristics of stars and interstellar matter,” Astron. stars close to the galactic centre. ii - their spatial and Astrophys. 384 (2002) 112–139, kinematic properties and the mass distribution within arXiv:astro-ph/0201294 [astro-ph]. 5-100 pc from the galactic centre,” AAP 259 (June, [56] R. Schnrich, M. Aumer, and S. E. Sale, “Kinematic 1992) 118–127. detection of the galactic nuclear disk,” The [42] S. Deguchi, H. Imai, T. Fujii, I. S. Glass, Y. Ita, Astrophysical Journal Letters 812 no. 2, (2015) L21. H. Izumiura, O. Kameya, A. Miyazaki, Y. Nakada, and http://stacks.iop.org/2041-8205/812/i=2/a=L21. J.-I. Nakashima, “Sio maser survey of the [57] C. Wegg, O. Gerhard, and M. Portail, “The Initial Mass large-amplitude variables in the galactic center,” PASJ Function of the Inner Galaxy Measured from OGLE-III 56 (Apr., 2004) 261–294, astro-ph/0403113. Microlensing Timescales,” The Astrophysical Journal Letters 843 (July, 2017) L5, 23

arXiv:1706.04193. Field Dark Matter,” arXiv:1512.02351 [58] M. Persic, P. Salucci, and F. Stel, “The Universal [astro-ph.CO]. rotation curve of spiral galaxies: 1. The Dark matter [65] J. Barranco, A. Bernal, J. C. Degollado, connection,” Mon. Not. Roy. Astron. Soc. 281 (1996) A. Diez-Tejedor, M. Megevand, M. Alcubierre, 27, arXiv:astro-ph/9506004 [astro-ph]. D. Nunez, and O. Sarbach, “Are black holes a serious [59] L. M. Fernandez-Hernandez, M. A. Rodriguez-Meza, threat to scalar ﬁeld dark matter models?,” Phys. Rev. and T. Matos, “Comparison between two scalar ﬁeld D84 (2011) 083008, arXiv:1108.0931 [gr-qc]. models using rotation curves of spiral galaxies,” [66] J. Barranco, A. Bernal, J. C. Degollado, arXiv:1708.06681 [astro-ph.GA]. A. Diez-Tejedor, M. Megevand, M. Alcubierre, [60] H. Deng, M. P. Hertzberg, M. H. Namjoo, and D. Nunez, and O. Sarbach, “Schwarzschild black holes A. Masoumi, “Can Light Dark Matter Solve the can wear scalar wigs,” Phys. Rev. Lett. 109 (2012) Core-Cusp Problem?,” arXiv:1804.05921 081102, arXiv:1207.2153 [gr-qc]. [astro-ph.CO]. [67] J. Barranco, A. Bernal, J. C. Degollado, [61] D. C. Rodrigues, A. del Popolo, V. Marra, and P. L. C. A. Diez-Tejedor, M. Megevand, D. Nunez, and de Oliveira, “Evidence against cuspy dark matter haloes O. Sarbach, “Self-gravitating black hole scalar wigs,” in large galaxies,” Mon. Not. Roy. Astron. Soc. 470 Phys. Rev. D96 no. 2, (2017) 024049, no. 2, (2017) 2410–2426, arXiv:1701.02698 arXiv:1704.03450 [gr-qc]. [astro-ph.GA]. [68] W. G. Unruh, “Absorption Cross-Section of Small [62] J. Veltmaat, J. C. Niemeyer, and B. Schwabe, Black Holes,” Phys. Rev. D14 (1976) 3251–3259. “Formation and structure of ultralight bosonic dark [69] R. Bender et al., “Hst stis spectroscopy of the triple matter halos,” arXiv:1804.09647 [astro-ph.CO]. nucleus of m31: two nested disks in keplerian rotation [63] D. G. Levkov, A. G. Panin, and I. I. Tkachev, “Bose around a supermassive black hole,” Astrophys. J. 631 Condensation by Gravitational Interactions,” (2005) 280, arXiv:astro-ph/0509839 [astro-ph]. arXiv:1804.05857 [astro-ph.CO]. [64] J.-W. Lee, J. Lee, and H.-C. Kim, “The M-sigma Relation of Super Massive Black Holes from the Scalar