Black holes:

on the universality of the Kerr hypothesis

Carlos A. R. Herdeiro
arXiv:2204.05640v1 [gr-qc] 12 Apr 2022

Abstract To what extent are all astrophysical, dark, compact objects both black holes
(BHs) and described by the Kerr geometry? We embark on the exercise of defying
the universality of this remarkable idea, often called the "Kerr hypothesis". After
establishing its rationale and timeliness, we define a minimal set of reasonability
criteria for alternative models of dark compact objects. Then, as proof of principle, we
discuss concrete, dynamically robust non-Kerr BHs and horizonless imitators, that 1)
pass the basic theoretical, and in particular dynamical, tests, 2) match (some of the)
state of the art astrophysical observables and 3) only emerge at some (macroscopic)
scales. These examples illustrate how the universality (at all macroscopic scales) of
the Kerr hypothesis can be challenged.

1 The Kerr hypothesis

The elegant Kerr metric [1, 2],

Δ 2 Σ sin2 𝜃
𝑑𝑠2 = − 𝑑𝑡 − 𝑎 sin2 𝜃𝑑𝜙 + 𝑑𝑟 2 + Σ𝑑𝜃 2 + [𝑎𝑑𝑡 − (𝑟 2 + 𝑎 2 )𝑑𝜙] 2 , (1)
Σ Δ Σ
where Σ ≡ 𝑟 2 + 𝑎 2 cos2 𝜃, Δ ≡ 𝑟 2 − 2𝑀𝑟 + 𝑎 2 , is the currently accepted model
to describe the phenomenology of all astrophysical black hole (BH) candidates.
Since (1) is only described by two macroscopic parameters, the mass 𝑀 and angular
momentum 𝑀𝑎 of the BH, this is clearly an economical scenario. The very same
(almost featureless) theoretical model describes astrophysical objects ranging, at
least, 10 orders of magnitude in mass. There is evidence for BHs in the stellar mass
range, from ∼ 𝑀 to ∼100 𝑀 , obtained from X-ray binaries [3] and gravitational

Carlos A. R. Herdeiro
Departamento de Matemática da Universidade de Aveiro, Campus de Santiago, 3810-193 Aveiro,
Portugal, e-mail: [email protected]

1
2 Carlos A. R. Herdeiro

wave (GW) detections [4, 5, 6] and in the supermassive range, from ∼ 105 to 1010
𝑀 , as radio sources [3], from Very Large Base Line Interferometry [7] and infrared
observations in our own galactic centre [8, 9]. According to the "Kerr hypothesis"
these BHs have exactly the same spacetime structure, simply rescaled by the mass,
and with only one extra macroscopic degree of freedom, their spin. If true, this is
remarkable.
The goal of this article is to discuss, with concrete proof of concept models, the
possibility if (and under which circumstances) astrophysical BH candidates could be
described by something else rather than the Kerr geometry in some range of (macro-
scopic) scales,1 both in view of theoretical consistency, in particular dynamics, as
well as in view of the current observational developments.

2 Non-Kerrness: guiding principles and testing grounds

Non-Kerr models for astrophysical BH candidates must obey theoretical reasonabil-

ity criteria. Here is a possible minimal list [10]:
1) to appear in well motivated and consistent physical theories. In the case of Kerr,
this is vacuum General Relativity (GR), which, albeit an incomplete theory, due 𝑒.𝑔.
to singularities, within its regime of validity fulfills this criterion;
2) to have a dynamical formation mechanism, which for Kerr is gravitational col-
lapse [11], together with accretion and mergers;
3) to be sufficiently stable, meaning it can play a role in astrophysical or cosmological
time scales. For Kerr, mode stability has been established long ago [12].
The two last criteria establish dynamical robustness. This is one of the unifying prin-
ciples of the models discussed below: there should be a route for forming them and
be sufficiently stable against the unavoidable perturbations in any realistic environ-
ment. This is a restrictive criterion. Some models of alternative BHs or horizonless
compact objects, 𝑒.𝑔. wormholes, have no established formation mechanism.
Being a good theoretical model in the sense of the previous paragraph is not,
however, enough to describe Nature. The model must give rise to all the correct phe-
nomenology attributed to astrophysical BHs, both in the electromagnetic and GWs
channels. At the moment, there is not (yet?) clear tension between observations and
the Kerr model. But the limitations of GR (𝑒.𝑔. the unavoidable singularities behind
BH horizons [13]) and of the standard model (SM) of particle physics (which fails
to explain, say, dark matter) leverage us to consider non-Kerr models and how much
we can distinguish them from the paradigm, with state of the art observables. More-
over, we should bear in mind that, at the moment, we are not testing simultaneously
(𝑖.𝑒. for the same mass range) all these observables. Thus, degeneracy in any single
observable (even if not in all), for some model, may be of interest.

1 It is widely accepted that sufficiently microscopic BHs will require quantum gravity effects. This
is not the range of scales we shall be interested in, but rather macroscopic scales for which there is
astrophysical evidence for BHs.
Black holes: on the universality of the Kerr hypothesis 3

The first class of state of the art strong gravity observables, to test against, comes
from GWs. Amongst the many detections, a most interesting (and intriguing) one
is GW190521 [14]. Key novel aspects of this event include: 𝑖) the very massive
progenitors, of about 85 and 66 𝑀 , meaning that, within the uncertainties, at least
one is in the so called "pair instability supernova gap", a gap in the mass spectrum
starting somewhere between 45 and 55 𝑀 wherein BHs cannot form from the
collapse of massive stars according to standard stellar evolution [15]. So, how did
the progenitors of this event come to be? 𝑖𝑖) it is a low frequency event, implying
there is no inspiral in the observed signal, leaving room to speculate about the true
nature of the event. Thus, GW190521 (and similar events) promises to be a fertile
ground for theoretical modelling, in particular for testing the Kerr hypothesis.
A second remarkable observable is the first image of a BH - M87* - resolving
horizon scale structure [7]. Since this observation is probing the strong gravity
region of a few Schwarzschild radii, even though it has only been done for a single
BH and even though there may be a non negligible impact of the astrophysical
environment, which is not fully under control, it is worth understanding how it can
be used to confirm the Kerr hypothesis or constrain deviations thereof, by comparing
the shadow [16, 17] and emission ring in the M87* image with non-Kerr models.

3 Non-Kerrness: two families of examples

The Carter-Robinson uniqueness theorem [18, 19, 20] establishes that physically rea-
sonable BHs in vacuum GR fall into the family (1). Hence they are rather featureless;
they have no-hair [21]. Thus, to find non-Kerr BHs, one must either include matter
in GR or consider modified gravity. This is necessary but not sufficient. For instance,
including minimally coupled matter fields, with standard kinetic terms and obeying
(say) the dominant energy condition is rather restrictive; it prevents non-Kerr BHs
in many models. This conclusion has been established by model-specific no-hair
theorems [22, 23]. Some of these theorems are powerful since they do not require
much, a paradigmatic example being Bekenstein’s theorem ruling out BH “hair" of
a real scalar field minimally coupled to Einstein’s gravity, possibly with a mass term
or even some classes of self interactions [24]. Still, theorems have assumptions, and
dropping some of them one may (sometimes) find reasonable scenarios where new
BHs emerge.
Here, we shall be interested both in examples in GR with minimally coupled scalar
(or massive vector) fields, where the usual theorems are circumvented in a subtle
way, by a property called “symmetry non-inheritance", and in models beyond GR
where the theorems are circumvented by the use of non-minimal couplings. In either
case, the theoretical foundation will invoke new physics which introduces a new
scale. Moreover, in both cases, the non-Kerr BHs co-exist with Kerr as solutions of
the model. At some scales, however, the former emerge dynamically from the latter.
In other words, the Kerr solution becomes unstable due to some new physics and a
new preferred state emerges.
4 Carlos A. R. Herdeiro

The first family is in GR but with matter beyond the SM, in fact ultralight bosonic
particles, that have been proposed as dark matter candidates [25, 26, 27]. The two
basic members of this family are described by the action (𝐺 = 1 = 𝑐)
∫  
(𝑠) 4 √ 𝑅 (𝑠)
𝑆 = 𝑑 𝑥 −𝑔 +L , (2)
16𝜋

where L (0) = −Φ̄, 𝛼 Φ, 𝛼 − 𝜇2 Φ̄Φ and L (1) = −𝐹¯ 𝛼𝛽 𝐹 𝛼𝛽 /4 − 𝜇2 𝐴¯ 𝛼 𝐴 𝛼 /2 for the

scalar (𝑠 = 0) and vector (𝑠 = 1) cases, respectively. Φ and 𝐴 𝜇 are a complex scalar
and vector field, 𝐹 = 𝑑𝐴 and overbar denotes complex conjugate. The field mass
introduces a new scale, 𝜇, which is an inverse length, and which, for the scenario
herein, is taken to be ultralight, with mass between 10−10 to 10−20 eV. Kerr BHs with
a horizon scale comparable to the Compton wavelength of this new particle, become
(efficiently) unstable against a process called superradiance [28], transferring part
of their energy and angular momentum to a scalar cloud around the BH, which
becomes a "BH with synchronised (scalar or Proca) hair” [29, 30], and different
from Kerr. These BHs were first reported in [31] and [32] for the scalar and Proca
case, respectively.
The second family is beyond GR, in modified gravity. A member of this family is
the extended scalar-tensor Gauss-Bonnet (eSTGB) model, described by the action
∫
1 √ 
𝑑 4 𝑥 −𝑔 𝑅 − 2𝜕𝜇 𝜙𝜕 𝜇 𝜙 + 𝜆2 𝑓 (𝜙) (𝑅 𝜇𝜈 𝛼𝛽 𝑅 𝜇𝜈 𝛼𝛽 − 4𝑅 𝜇𝜈 𝑅 𝜇𝜈 + 𝑅 2 ) ,

𝑆=
16𝜋
(3)
where 𝜙 is a real scalar field and 𝑓 (𝜙) an yet unspecified coupling function. The
Gauss-Bonnet (GB) coupling introduces a new scale 𝜆, which has units of length.
Then, BHs with a horizon scale comparable to this new scale, can undergo a strong
gravity phase transition, spontaneous scalarisation [33, 34].2 This leads, dynami-
cally, to new types of BHs, dubbed "scalarised BHs".
Model (2) also accommodates horizonless compact objects known as "bosonic
stars", scalar [36, 37, 38] or vector [39, 40, 41]. The latter are also known as "Proca
stars". Some of these solutions are dynamically robust [42, 43] and have a formation
mechanism, via a process called "gravitational cooling" [44, 45, 46, 43]. When
the new scale, 𝜇 is in the aforementioned range, between 10−10 to 10−20 eV, the
maximal mass of these bosonic stars is in the astrophysical range of 1 − 1010 𝑀 ,
and these objects mimic the mass of astrophysical BHs. Scalar and vector bosonic
stars have some interesting differences which will be emphasised in the examples
below concerning the (BH) imitation game.

2 Spontaneous scalarisation was first proposed for neutron stars, in a different model [35].
Black holes: on the universality of the Kerr hypothesis 5

4 Non-Kerr BHs: the example of synchronisation

4.1 Dynamical considerations

Dynamical synchronisation occurs in many systems, both in biology, 𝑒.𝑔. communi-

ties of fireflies or crickets, and in physics, 𝑒.𝑔. sets of metronomes or pendulums [47].
In these systems, individual cycles converge dynamically to the same phase due to
appropriate interactions, yielding a configuration that would otherwise look fine-
tuned.
In Newtonian gravitational dynamics, synchronisation occurs in binary systems
of extended objects, such as planets or stars [48]. Tidal effects tend to synchronise
and lock orbital and rotational periods. This effect led the moon to always show the
same face towards the Earth, and the Earth is (very slowly) tending to show the same
face towards the moon. This is a ubiquitous behaviour observed in all planets-moons
of the solar system.
In the context of relativistic gravity, the aforementioned BHs with synchronised
bosonic hair can be interpreted as synchronised configurations (hence the name).
The synchronisation condition reads Ω 𝐻 = 𝜔/𝑚 [49, 50, 31], where Ω 𝐻 is the
horizon angular velocity, and 𝜔/𝑚 is the phase angular velocity of the bosonic field,
which has a dependence Ψ ∼ 𝑒 −𝑖 ( 𝜔𝑡−𝑚𝜑) , where Ψ represents either Φ or 𝐴 𝜇 ; 𝑡, 𝜑 are
the time and azimuthal coordinates of the stationary and axisymmetric spacetime;
𝜔 is the frequency of the bosonic field’s harmonic time dependence and 𝑚 is an
integer azimuthal harmonic index. This condition is stating that the phase of the
field is co-rotating in synchrony with the horizon. Are these synchronous hairy BHs
attained dynamically, as in the case of the fireflies or pendulums?
There is indeed one dynamical channel that leads to synchronisation: the process
of superradiance. Fully non-linear numerical simulations have been successfully
performed when the bosonic field is the Proca one [29]. They have shown that the
process of superradiant rotational energy extraction spins down the BH, saturates
and a new equilibrium state is attained, precisely when the BH and the dominant
superradiant mode obey Ω 𝐻 = 𝜔/𝑚. These simulations obtained a maximum of
∼ 9% of energy transfer from the BH into the bosonic cloud. This is close to the
maximum expected in the evolution of the dominant superradiant mode, which is
∼ 10%, also for the scalar case [51]. Using the data from these simulations the new
equilibrium state was identified with a BH with synchronised Proca hair [30].
We thus have a process creating a new sort of BH, which takes some time scale. But
the time scale depends crucially on a resonance between the Compton wavelength of
the fundamental boson 1/𝜇 and the Schwarzschild radius of the BH, ∼ 𝑀. Maximal
efficiency occurs when they are similar, 𝑀 𝜇 ∼ 1 [52]. Otherwise the time scale
quickly grows and becomes larger than the Cosmological time sufficiently far from
the sweet spot. This therefore selects a mass scale of BHs: depending on the sort of
mass of the fundamental boson, BHs in a certain (narrow) mass range become hairy,
but outside this range they (effectively) do not [10]. The punch line is that, under the
assumption a single ultralight boson with some mass 𝜇 exists, non-Kerrness would
6 Carlos A. R. Herdeiro

manifest itself only for BHs in some narrow mass range around 1/𝜇. Observational
evidence for such non-Kerrness would therefore identify 𝜇.
There may be other formation channels for these synchronised BHs, in particular,
from mergers of bosonic stars, both scalar and vector [53]. There is a key difference
in this latter channel: superradiance forms a synchronised configuration by spinning
down the BH. In this new channel, mergers of bosonic stars lead to a synchronised
system by spinning up the BH that results from the merger, and which accretes part
of the field remnant. In the latter case, however, fine tuning seems necessary [53].

4.2 Comparison with observations

Can we constrain BHs with synchronised hair with current observations? Concerning
GWs, both the perturbation theory (for the ringdown) and fully non-linear dynamical
evolutions (for the inspiral and merger) remain essentially unexplored (but see [54]).
Concerning shadows, on the other hand, more progress has been achieved, 𝑒.𝑔. [55,
56].
This family of BHs with synchronised hair interpolates between vacuum Kerr
BHs and bosonic stars. Consequently, close to the Kerr limit phenomenological
differences are as small as desired (compared to Kerr) due to the continuity of the
solutions; but sufficiently far away they are large and can become huge in the solution
space region close to bosonic stars. Fig. 1 shows lensing images obtained by ray
tracing for these BHs and comparing with Kerr [57], using an aesthetically appealing
starry sky as the light source, rather than a realistic astrophysical environment. The
hairy BH shown (righ panel) has the same mass and angular momentum as the Kerr
one (left panel), but 75% of the mass and 85% of the angular momentum are stored
in the scalar field (rather than the horizon). One can see that due to the transfer
of part of the energy and spin of the BH to the scalar “hair” the actual shadow is
considerably smaller than that of a Kerr BH with the same total mass and spin and
comparable observation conditions. Actually it is about 25% smaller in terms of the
average radius, which seems to exclude this particular example of hairy BH as a good
model for M87* with current EHT observations. This illustrates how sufficiently
large departures from Kerr can be falsified even with current observations. However,
how are the differences in the dynamically viable region, assuming formation from
superradiance?
In the dynamically most interesting region of the parameter space, where the hairy
BHs can form from Kerr in an astrophysical time scale and are stable (against higher
superradiant modes) for at least a Hubble time [10], the differences are generically
small. Recall that only about ∼10% of the Kerr BH mass can be transferred to
the bosonic cloud. Yet. . . this is a non-negligible amount of mass for a 6 billion
solar masses BH like M87*. Analysing in this region of the parameter space if the
M87* image could then distinguish the hairy BH from Kerr, one concludes that
all dynamically viable BHs could be mistaken for a Kerr BH within the current
Black holes: on the universality of the Kerr hypothesis 7

Fig. 1 Shadow and lensing of a Kerr BH (left) and a BH with synchronised scalar hair (right)
with the same ADM mass and angular momentum, using a background image of Infant Stars in the
Small Magellanic cloud from the Hubble space telescope. Adapted from [55].

error bars [56]. This conclusion seems quite natural, since roughly, the error in the
measurement of the emission ring diameter is about 10%.
The conclusion is the following. Consider that an ultralight (dark matter) particle
with a mass of ∼ 10−20 eV exists in Nature, so that 𝜇𝑀 𝑀 87∗ ∼ 1. Then, as M87* grew
throughout its history to its current mass, superradiance became efficient around its
current mass, and endowed it with synchronised hair, a process that took a few million
years. The (current) hairy M87* BH is effectively stable against higher superradiant
modes, for a Hubble time. In this scenario, M87* is not a Kerr BH, but rather a
BH with synchronised hair. However, the current Event Horizon Telescope precision
would not be able to tell the difference. Observe, moreover, that all other BHs with
a mass smaller than . 109 𝑀 would not become hairy, as superradiance is not
efficient. They would be Kerr BHs.

5 Non-Kerr BHs: the example of scalarisation

We now consider a different scenario wherein non-Kerr BHs also arise dynamically,
for some scales, but via a different instability of Kerr BHs: spontaneous scalarisation.
In the landscape of modified gravity models there are question marks on the
well-posedness and theoretical consistency of many of them. A fairly well motivated
class of models with higher curvature corrections is the class of eSTGB models (3).
Here, the model is defined by the choice of the coupling between the GB and the
scalar field, 𝑓 (𝜙). Choosing a dilatonic coupling, 𝑓 (𝜙) ∼ 𝑒 𝛽 𝜙 , which is motivated,
say, by string theory or dimensional reduction, Schwarzschild/Kerr are not solutions;
there are new BHs which are stable (in some regime) and which have a qualitatively
new feature, namely a minimal BH size [58, 59]. This can be interpreted as due to a
8 Carlos A. R. Herdeiro

repulsive effect, sourced by the GB term, that destabilises the horizon when the GB
term becomes dominant, 𝑖.𝑒. for sufficiently small BHs.
Changing the scalar-curvature coupling into a more general 𝑓 (𝜙), there are vari-
ous interesting cousin models. For example, one can consider shift-symmetric mod-
els, which are close to the previous dilatonic model but with this additional shift-
symmetry. BHs in this model have been constructed [60, 61] and shown to exhibit
dynamical formation [62]. On the other hand, the condition 𝑑𝑓 /𝑑𝜙(𝜙 = 0) = 0,
yields a class of models admitting both vacuum GR and scalarised BHs [33, 34, 63].
Moreover, depending on the sign of 𝑑 2 𝑓 /𝑑𝜙2 (𝜙 = 0), the GR BHs may be unstable
in some scales, 𝑖.𝑒. when 𝑀/𝜆 is smaller than some threshold, which suggests the
non-GR BHs could emerge dynamically, via spontaneous scalarisation. Entropic
considerations support this possibility in some models, 𝑒.𝑔. [33, 64].
In the case of these models, the fully dynamical process has not been established,
but there are dynamical results showing the exponential growth and saturation of
a self-interacting scalar field in the decoupling limit [65]. Moreover, fully non-
linear numerical evolutions could be performed of the spontaneous scalarisation
of a Reissner-Nordström BH in a cousin model, showing the phenomenon occurs
dynamically and leads to a perturbatively stable scalarised BH [66].
The scalarised Kerr BHs that would be the endpoint of the scalarisation pro-
cess were first constructed in [64] (see also [67]), in model (3) with 𝑓 (𝜙) =
2
(1 − 𝑒 −𝛽 𝜙 ) /(2𝛽). Choosing this coupling is illustrative. Some properties (but not
all) of the scalarised solutions are universal for all couplings allowing scalarisation,
𝑒.𝑔. the threshold between Kerr and scalarised BHs. All solutions (as in the case of
the BHs with synchronised hair) are numerical. Consider first static, spherical BHs,
taking 𝜆 as fixed (new) scale. For sufficiently large BHs as compared to 𝜆 (greater
than ∼ 0.6) there are only Schwarzschild BHs. No scalarised BHs exist. Decreasing
the mass of the BH, Schwarzschild BHs becomes unstable against scalarisation and
the scalarised BHs appear. These are stable, at least against radial perturbations up
to some smaller size, where the scalarised BHs cease to be stable [68]. So, there
is a window of sizes where scalarised BHs are stable and dynamically preferred to
Schwarzschild.
The effect of the angular momentum is interesting: spin suppresses the effects
of scalarisation [64]. This spin suppression of the non-Kerrness can be seen using
the diagnosis of the associated shadows - Fig 2. Using the same background image
as before, one sees that for non-spinning BHs (top panels), the distinction is visible
with a naked eye: the Schwarzschild BH shadow (left panel) and that of a scalarised
BH with the same mass and under similar observation conditions (right panel) are
clearly different. The latter is in the perturbatively stable region. But as the spin
parameter increases (bottom panels) the difference is suppressed and even for still
fairly low spins, say 𝑗 ∼ 0.5, they become negligible.
In [64] a comparison with the Event Horizon Telescope data was performed.
Taking into account that very little is known about the spin of M87*, however, this
comparison is not very informative. Even in the most optimistic scenario (for this
model) where the spin is low, one can only put a rather weak constraint on the new
scale that the model introduces.
Black holes: on the universality of the Kerr hypothesis 9

Fig. 2 Shadow and lensing of scalarised BHs (right panels) and Schwarzschild/Kerr BHs (right
panels) with the same ADM mass and angular momentum, using the same background image as in
Fig. 1. Adapted from [64].

Let us close this discussion on spontaneous scalarisation with a different class

of models that introduces a new twist. In the models we have just discussed,
the Kretschmann scalar of the vacuum BH is providing the instability, endowing
scalar perturbations with a tachyonic mass. This Kretschmann scalar is positive for
Schwarzschild. For Kerr it starts to become negative (around the poles) for the di-
mensionless spin, 𝑗 > 0.5. Thus, if we consider models with the opposite coupling
sign, only sufficiently fast spinning BHs will scalarise. This has been called "spin
induced scalarisation" [69].
In [70] (see also [71]) the domain of existence of the corresponding BHs was
explored, for the same illustrative coupling as before. Whereas in the model before,
spin quenched the Kerr deviations, now spin enhances the Kerr deviations; in fact
it is mandatory. So, this illustrates how there are models in which only some BHs,
either with small or with large spin can differ from Kerr, which moreover only
occurs for some mass scales. In the case of spin induced scalarisation a detailed
phenomenological study of the solutions, namely of the shadows, has not yet been
reported.
10 Carlos A. R. Herdeiro

6 The imitation game: non-BHs mimicking BH observables

Let us now discuss “the imitation game”, or how non-BHs can mimic some (even if
not all) BH observables.
There are several motivations to consider BH mimickers, 𝑒.𝑔., the singularity
problem of BHs. This has led to many models of horizonless compact objects that
could behave as BH imitators. It has been pointed out, however, that an imitator
needs to have a light ring (LR) to mimic the BH ringdown [72]. LRs also play a key
role in determining the edge of the shadow [73]. Thus, it seems that for compact
object to mimic the most strong field current observations they need to possess LRs
(and are then called "ultra-compact").
There may be, however, a generic possible issue with horizonless ultracompact
objects. A theorem established in [74] states that for generic equilibrium ultracompact
objects, resulting from a smooth, incomplete gravitational collapse, (thus, for which
there is plausible formation mechanism) LRs come in pairs and one is stable. A
stable LR has been suggested to trigger a (non-linear) spacetime instability [75],
as massless perturbations can pile up in its vicinity. Not much is known about this
instability or its timescale, but it raises a shadow of doubt about the dynamical
robustness of horizonless, ultracompact objects.
Let us therefore ask the question if real data, like a true GW event or the Event
Horizon Telescope M87* observation could be imitated by a compact object mim-
icker that does not have LRs. Interestingly, as the following case study examples
indicate the answer is yes in both cases.

6.1 Mimicking a GW event

Let us first address the imitation of a GW event. In GR coupled to ultralight bosonic

fields, there has been, for many years, developments in evolving scalar and vector
boson stars dynamically, using numerical relativity techniques [42]. In 2019 an un-
expected difference between spinning scalar and vector bosonic stars was found: the
most fundamental spinning scalar boson stars develop a non-axisymmetric instabil-
ity [43]. When this instability kicks in, these stars collapse into a BH. On the other
hand, this instability is absent in the cousin Proca model. Indeed, even if one perturbs
considerably a spinning Proca star, no instability is seen. So, spinning Proca stars,
without self-interactions, are dynamically robust, unlike their scalar cousins.
Given the dynamically robustness of spinning Proca Stars, recently we considered
simulations of mergers of spinning Proca stars and compared them with real data,
in particular with the intriguing event GW190521 commented on earlier. We have
found, through a Bayesian analysis, that a collision of two spinning Proca stars
actually fits slightly better the data than the vanilla binary BH model that was used
by the LIGO-Virgo collaboration [76].
It follows that if, even just as a proof of concept, one takes seriously the Proca
model, we can use the data to infer the mass of the ultralight bosonic (in this case
Black holes: on the universality of the Kerr hypothesis 11

vector) particle. We have obtained a mass of around 8 × 10−13 eV. More similar
events are needed to confirm this possibility. In the most likely case scenario this is
just a proof of concept showing 1) how there can be degeneracy in real data between
two very different models and 2) how one could extract physical information about
a new fundamental, dark matter particle from GW data (in this case the mass of
the boson). Note that the colliding Proca stars are compact but not ultracompact, so
they do not have LRs. Of course, in a much more exciting possibility, a potential
confirmation of the bosonic star scenario would be a first hint for the long sought
dark matter nature.
Under this rationale, one may ask why could this event be such a Proca star
collision and not the other events detected so far? The point is again mass selection.
The mass of the ultralight bosonic particle determines the maximal mass of the
corresponding stars. For the above quoted ultralight boson, this mass turns out to
be about ∼173 𝑀 . All other events correspond to a smaller final mass. Since we
have observed BH formation, due to the ringdown, then these cannot be Proca star
collisions. For the Proca star collisions to form a BH they have to overshoot this
limit, which requires events as massive as GW190521.

6.2 Mimicking a BH shadow without LRs

Let us now address the imitation of a BH shadow by a mimicker without LRs.

In a generic stationary and axi-symmetric BH spacetime, one can associate the
edge of the BH shadow to a set of photon bound orbits, which we refer to as
"fundamental photon orbits" [73]. In the Schwarzschild case, all of these are planar
LRs. In the Kerr case they are called "spherical photon orbits", since they have a
constant radial Boyer-Lindquist coordinate.
In a real astrophysical environment, however, an effective shadow seen may
depend on the details of the light source. This was nicely illustrated in [77], where GR
magnetohydrodynamic (GRMHD) simulations were performed on the background
of BHs and of some models of static scalar boson stars. None of these stars have
LRs (or a horizon); yet, some could produce an effective shadow, where others did
not. In all boson stars models considered, they admitted stable timelike circular
orbits until their very centre; but in the case where an effective shadow was seen
the angular velocity of the timelike circular orbits attains a maximum at some non-
zero areal radius 𝑅Ω . This new scale is observed to determine the inner edge of the
accretion disk in the simulations, under some assumptions, including that the loss
of angular momentum of the orbiting matter is driven by the magneto-rotational
instability and that the radiation relevant for the BH shadow observations is mostly
due to synchroton emission. There are, however, two caveats for the models in [77]
that produce an effective shadow, to be seriously considered as imitators. First, the
imitation is not perfect, since the efective shadow is considerably smaller than that
of a comparable Scwharzschild BH with the same mass. Secondly, the boson stars
producing an effective shadow are perturbatively unstable.
12 Carlos A. R. Herdeiro

It turns out, as discussed in [78], that this imitation game works better for Proca
stars (rather than scalar boson stars). Within the stable branch of spherical, funda-
mental Proca stars, there are solutions that display the necessary new scale, that is, for
which the timelike circular orbit with the maximal angular velocity is at some radius
𝑅Ω ≠ 0. One can even choose a particular solution for which this new scale equals the
location of the Innermost Stable Circular Orbit (ISCO) of a Schwarzschild BH with
the same mass. So, there is a dynamically robust solution that, under some accretion
models yields an accretion disk morphology mimicking that of a Schwarzschild BH
of the same mass.
To check the potential shadow degeneracy, in [78] ray tracing images were pro-
duced considering a simplified astrophysical setup, wherein the only radiation source
is an opaque and thin accretion disk located on the equatorial plane around the central
compact object. The disk has an inner edge with an areal radius 𝑅Ω = 6𝑀 in both
spacetimes. The most interesting case for degeneracy occurs for an observer close to
the poles (to match the estimated angle at which M87*, was observed from Earth) -
Fig. 3. For this angle, the images of the Schwarzschild (top left panel) and Proca star
(top middle left panel) look similar, although some finer additional lensing features
are still visible in the Schwarzschild case. But these subtle differences are washed
away if one applies a Gaussian blurring filter to the images, to mimic the current
observations limited angular resolution, of the order of the compact object itself.
The images so obtained are shown below the corresponding unblurred image, and
are essentially indistinguishable.

Fig. 3 Shadow and lensing of the Proca star discussed in the text and of a comparable Schwarzschild
BH, both illuminated by a thin, equatorial accretion disk, at two different observation angles. The
bottom images are blurred to mimic current observational limitations. Adapted from [78].

On the other hand, a near equatorial observation leads to fairly different images.
The Schwarzschild one (top middle right panel) resembles the BH shape displayed
in the Hollywood movie “Interstellar”, whereas the Proca star simply looks like a
Black holes: on the universality of the Kerr hypothesis 13

flat accretion disk with a hole in it, as seen from the side (top right panel). This
is because the gravitational potential well of the Proca star is shallow and so the
bending of light it produces is weak. Consequently, the accretion disk has an almost
flat spacetime appearance. Applying the same blurring as before, even with limited
resolution, the two objects could be distinguished. Let us stress that full GRMHD
analysis and ray-tracing in the background of this Proca star are required to fully
settle the question: to what degree can it imitate a BH observation? The case built
herein, nonetheless, clearly confirms a potential degeneracy, but only under some
observation conditions. And these Proca stars do not have LRs and are dynamically
stable.

7 Concluding remarks

Let us close with two following final remarks. Firstly, all these models we have
discussed have caveats; but they illustrate theoretical possibilities of dynamically
robust non-Kerr BHs, or BH imitators, that could manifest themselves only at some
specific scales of mass and spin. Secondly, that producing detailed phenomenology
will constrain the model and the corresponding (admittedly exotic) physics or, in the
best case scenario, provide a smoking gun to this new physics.

Acknowledgements I thank all my collaborators that were fundamental for the work presented
here. Special thanks, by virtue of our long and fruitful collaboration go to P. Cunha, E. Radu
and N. Sanchis-Gual. This work is supported by the Center for Research and Development in
Mathematics and Applications (CIDMA) through the Portuguese Foundation for Science and
Technology (FCT - Fundação para a Ciência e a Tecnologia), references UIDB/04106/2020,
UIDP/04106/2020 and the projects PTDC/FIS-OUT/28407/2017, CERN/FIS-PAR/0027/2019,
PTDC/FIS-AST/3041/2020 and CERN/FIS-PAR/0024/2021. This work has further been supported
by the European Union’s Horizon 2020 research and innovation (RISE) programme H2020-MSCA-
RISE-2017 Grant No. FunFiCO-777740.

References

1. Roy P. Kerr. Gravitational field of a spinning mass as an example of algebraically special

metrics. Phys. Rev. Lett., 11:237–238, 1963.
2. Robert H. Boyer and Richard W. Lindquist. Maximal analytic extension of the Kerr metric. J.
Math. Phys., 8:265, 1967.
3. Ramesh Narayan and Jeffrey E. McClintock. Observational Evidence for Black Holes. 12
2013.
4. B. P. Abbott et al. GWTC-1: A Gravitational-Wave Transient Catalog of Compact Binary
Mergers Observed by LIGO and Virgo during the First and Second Observing Runs. Phys.
Rev. X, 9(3):031040, 2019.
5. R. Abbott et al. GWTC-2: Compact Binary Coalescences Observed by LIGO and Virgo During
the First Half of the Third Observing Run. Phys. Rev. X, 11:021053, 2021.
6. R. Abbott et al. GWTC-3: Compact Binary Coalescences Observed by LIGO and Virgo During
the Second Part of the Third Observing Run. 11 2021.
14 Carlos A. R. Herdeiro

7. Kazunori Akiyama et al. First M87 Event Horizon Telescope Results. I. The Shadow of the
Supermassive Black Hole. Astrophys. J. Lett., 875:L1, 2019.
8. S. Gillessen, F. Eisenhauer, S. Trippe, T. Alexander, R. Genzel, F. Martins, and T. Ott. Mon-
itoring stellar orbits around the Massive Black Hole in the Galactic Center. Astrophys. J.,
692:1075–1109, 2009.
9. A. M. Ghez et al. Measuring Distance and Properties of the Milky Way’s Central Supermassive
Black Hole with Stellar Orbits. Astrophys. J., 689:1044–1062, 2008.
10. Juan Carlos Degollado, Carlos A. R. Herdeiro, and Eugen Radu. Effective stability against
superradiance of Kerr black holes with synchronised hair. Phys. Lett. B, 781:651–655, 2018.
11. T. Nakamura, K. Oohara, and Y. Kojima. General Relativistic Collapse to Black Holes and
Gravitational Waves from Black Holes. Prog. Theor. Phys. Suppl., 90:1–218, 1987.
12. Bernard F. Whiting. Mode Stability of the Kerr Black Hole. J. Math. Phys., 30:1301, 1989.
13. Roger Penrose. Gravitational collapse and space-time singularities. Phys. Rev. Lett., 14:57–59,
1965.
14. R. Abbott et al. GW190521: A Binary Black Hole Merger with a Total Mass of 150𝑀 . Phys.
Rev. Lett., 125(10):101102, 2020.
15. R. Abbott et al. Properties and Astrophysical Implications of the 150 M Binary Black Hole
Merger GW190521. Astrophys. J. Lett., 900(1):L13, 2020.
16. Heino Falcke, Fulvio Melia, and Eric Agol. Viewing the shadow of the black hole at the
galactic center. Astrophys. J. Lett., 528:L13, 2000.
17. Pedro V. P. Cunha and Carlos A. R. Herdeiro. Shadows and strong gravitational lensing: a
brief review. Gen. Rel. Grav., 50(4):42, 2018.
18. B. Carter. Axisymmetric Black Hole Has Only Two Degrees of Freedom. Phys. Rev. Lett.,
26:331–333, 1971.
19. D. C. Robinson. Uniqueness of the Kerr black hole. Phys. Rev. Lett., 34:905–906, 1975.
20. Piotr T. Chrusciel, Joao Lopes Costa, and Markus Heusler. Stationary Black Holes: Uniqueness
and Beyond. Living Rev. Rel., 15:7, 2012.
21. Remo Ruffini and John A. Wheeler. Introducing the black hole. Phys. Today, 24(1):30, 1971.
22. Carlos A. R. Herdeiro and Eugen Radu. Asymptotically flat black holes with scalar hair: a
review. Int. J. Mod. Phys. D, 24(09):1542014, 2015.
23. Mikhail S. Volkov. Hairy black holes in the XX-th and XXI-st centuries. In 14th Marcel Gross-
mann Meeting on Recent Developments in Theoretical and Experimental General Relativity,
Astrophysics, and Relativistic Field Theories, volume 2, pages 1779–1798, 2017.
24. J. D. Bekenstein. Transcendence of the law of baryon-number conservation in black hole
physics. Phys. Rev. Lett., 28:452–455, 1972.
25. Lam Hui, Jeremiah P. Ostriker, Scott Tremaine, and Edward Witten. Ultralight scalars as
cosmological dark matter. Phys. Rev. D, 95(4):043541, 2017.
26. Abril Suárez, Victor H. Robles, and Tonatiuh Matos. A Review on the Scalar Field/Bose-
Einstein Condensate Dark Matter Model. Astrophys. Space Sci. Proc., 38:107–142, 2014.
27. Felipe F. Freitas, Carlos A. R. Herdeiro, António P. Morais, António Onofre, Roman Pasechnik,
Eugen Radu, Nicolas Sanchis-Gual, and Rui Santos. Ultralight bosons for strong gravity
applications from simple Standard Model extensions. JCAP, 12(12):047, 2021.
28. Richard Brito, Vitor Cardoso, and Paolo Pani. Superradiance: New Frontiers in Black Hole
Physics. Lect. Notes Phys., 906:pp.1–237, 2015.
29. William E. East and Frans Pretorius. Superradiant Instability and Backreaction of Massive
Vector Fields around Kerr Black Holes. Phys. Rev. Lett., 119(4):041101, 2017.
30. Carlos A. R. Herdeiro and Eugen Radu. Dynamical Formation of Kerr Black Holes with
Synchronized Hair: An Analytic Model. Phys. Rev. Lett., 119(26):261101, 2017.
31. Carlos A. R. Herdeiro and Eugen Radu. Kerr black holes with scalar hair. Phys. Rev. Lett.,
112:221101, 2014.
32. Carlos Herdeiro, Eugen Radu, and Helgi Rúnarsson. Kerr black holes with Proca hair. Class.
Quant. Grav., 33(15):154001, 2016.
33. Daniela D. Doneva and Stoytcho S. Yazadjiev. New Gauss-Bonnet Black Holes with Curvature-
Induced Scalarization in Extended Scalar-Tensor Theories. Phys. Rev. Lett., 120(13):131103,
2018.
Black holes: on the universality of the Kerr hypothesis 15

34. Hector O. Silva, Jeremy Sakstein, Leonardo Gualtieri, Thomas P. Sotiriou, and Emanuele Berti.
Spontaneous scalarization of black holes and compact stars from a Gauss-Bonnet coupling.
Phys. Rev. Lett., 120(13):131104, 2018.
35. Thibault Damour and Gilles Esposito-Farese. Nonperturbative strong field effects in tensor -
scalar theories of gravitation. Phys. Rev. Lett., 70:2220–2223, 1993.
36. David J. Kaup. Klein-Gordon Geon. Phys. Rev., 172:1331–1342, 1968.
37. Remo Ruffini and Silvano Bonazzola. Systems of selfgravitating particles in general relativity
and the concept of an equation of state. Phys. Rev., 187:1767–1783, 1969.
38. Franz E. Schunck and Eckehard W. Mielke. General relativistic boson stars. Class. Quant.
Grav., 20:R301–R356, 2003.
39. Richard Brito, Vitor Cardoso, Carlos A. R. Herdeiro, and Eugen Radu. Proca stars: Gravitating
Bose–Einstein condensates of massive spin 1 particles. Phys. Lett. B, 752:291–295, 2016.
40. Carlos A. R. Herdeiro, Alexandre M. Pombo, and Eugen Radu. Asymptotically flat scalar,
Dirac and Proca stars: discrete vs. continuous families of solutions. Phys. Lett. B, 773:654–662,
2017.
41. C. Herdeiro, I. Perapechka, E. Radu, and Ya. Shnir. Asymptotically flat spinning scalar, Dirac
and Proca stars. Phys. Lett. B, 797:134845, 2019.
42. Steven L. Liebling and Carlos Palenzuela. Dynamical Boson Stars. Living Rev. Rel., 15:6,
2012.
43. N. Sanchis-Gual, F. Di Giovanni, M. Zilhão, C. Herdeiro, P. Cerdá-Durán, J. A. Font, and
E. Radu. Nonlinear Dynamics of Spinning Bosonic Stars: Formation and Stability. Phys. Rev.
Lett., 123(22):221101, 2019.
44. Edward Seidel and Wai-Mo Suen. Formation of solitonic stars through gravitational cooling.
Phys. Rev. Lett., 72:2516–2519, 1994.
45. F. Siddhartha Guzman and L. Arturo Urena-Lopez. Gravitational cooling of self-gravitating
Bose-Condensates. Astrophys. J., 645:814–819, 2006.
46. Fabrizio Di Giovanni, Nicolas Sanchis-Gual, Carlos A. R. Herdeiro, and José A. Font. Dynam-
ical formation of Proca stars and quasistationary solitonic objects. Phys. Rev. D, 98(6):064044,
2018.
47. S. Strogatz. Sync: How Order Emerges from Chaos in the Universe, Nature, and Daily Life.
Hachette Books, New York, 2004.
48. P. Hut. Tidal evolution in close binary systems. Astron. & Astrophys., 99:126–140, June 1981.
49. Shahar Hod. Stationary Scalar Clouds Around Rotating Black Holes. Phys. Rev. D, 86:104026,
2012. [Erratum: Phys.Rev.D 86, 129902 (2012)].
50. Oscar J. C. Dias, Gary T. Horowitz, and Jorge E. Santos. Black holes with only one Killing
field. JHEP, 07:115, 2011.
51. Carlos A. R. Herdeiro, Eugen Radu, and Nuno M. Santos. A bound on energy extraction (and
hairiness) from superradiance. Phys. Lett. B, 824:136835, 2022.
52. Sam R. Dolan. Superradiant instabilities of rotating black holes in the time domain. Phys. Rev.
D, 87(12):124026, 2013.
53. Nicolas Sanchis-Gual, Miguel Zilhão, Carlos Herdeiro, Fabrizio Di Giovanni, José A. Font,
and Eugen Radu. Synchronized gravitational atoms from mergers of bosonic stars. Phys. Rev.
D, 102(10):101504, 2020.
54. Lucas G. Collodel, Daniela D. Doneva, and Stoytcho S. Yazadjiev. Equatorial EMRIs in
KBHsSH Spacetimes. 8 2021.
55. Pedro V. P. Cunha, Carlos A. R. Herdeiro, Eugen Radu, and Helgi F. Runarsson. Shadows of
Kerr black holes with scalar hair. Phys. Rev. Lett., 115(21):211102, 2015.
56. Pedro V. P. Cunha, Carlos A. R. Herdeiro, and Eugen Radu. EHT constraint on the ultralight
scalar hair of the M87 supermassive black hole. Universe, 5(12):220, 2019.
57. Pedro V. P. Cunha, Carlos A. R. Herdeiro, Eugen Radu, and Helgi F. Runarsson. Shadows of
Kerr black holes with and without scalar hair. Int. J. Mod. Phys. D, 25(09):1641021, 2016.
58. P. Kanti, N. E. Mavromatos, J. Rizos, K. Tamvakis, and E. Winstanley. Dilatonic black holes
in higher curvature string gravity. Phys. Rev. D, 54:5049–5058, 1996.
16 Carlos A. R. Herdeiro

59. Burkhard Kleihaus, Jutta Kunz, Sindy Mojica, and Eugen Radu. Spinning black
holes in Einstein–Gauss-Bonnet–dilaton theory: Nonperturbative solutions. Phys. Rev. D,
93(4):044047, 2016.
60. Thomas P. Sotiriou and Shuang-Yong Zhou. Black hole hair in generalized scalar-tensor
gravity: An explicit example. Phys. Rev. D, 90:124063, 2014.
61. Jorge F. M. Delgado, Carlos A. R. Herdeiro, and Eugen Radu. Spinning black holes in
shift-symmetric Horndeski theory. JHEP, 04:180, 2020.
62. Robert Benkel, Thomas P. Sotiriou, and Helvi Witek. Dynamical scalar hair formation around
a Schwarzschild black hole. Phys. Rev. D, 94(12):121503, 2016.
63. G. Antoniou, A. Bakopoulos, and P. Kanti. Evasion of No-Hair Theorems and Novel Black-
Hole Solutions in Gauss-Bonnet Theories. Phys. Rev. Lett., 120(13):131102, 2018.
64. Pedro V. P. Cunha, Carlos A. R. Herdeiro, and Eugen Radu. Spontaneously Scalar-
ized Kerr Black Holes in Extended Scalar-Tensor–Gauss-Bonnet Gravity. Phys. Rev. Lett.,
123(1):011101, 2019.
65. Daniela D. Doneva and Stoytcho S. Yazadjiev. Dynamics of the nonrotating and rotating black
hole scalarization. Phys. Rev. D, 103(6):064024, 2021.
66. Carlos A. R. Herdeiro, Eugen Radu, Nicolas Sanchis-Gual, and José A. Font. Spontaneous
Scalarization of Charged Black Holes. Phys. Rev. Lett., 121(10):101102, 2018.
67. Lucas G. Collodel, Burkhard Kleihaus, Jutta Kunz, and Emanuele Berti. Spinning and excited
black holes in Einstein-scalar-Gauss–Bonnet theory. Class. Quant. Grav., 37(7):075018, 2020.
68. Jose Luis Blázquez-Salcedo, Daniela D. Doneva, Jutta Kunz, and Stoytcho S. Yazadjiev. Radial
perturbations of the scalarized Einstein-Gauss-Bonnet black holes. Phys. Rev. D, 98(8):084011,
2018.
69. Alexandru Dima, Enrico Barausse, Nicola Franchini, and Thomas P. Sotiriou. Spin-induced
black hole spontaneous scalarization. Phys. Rev. Lett., 125(23):231101, 2020.
70. Carlos A. R. Herdeiro, Eugen Radu, Hector O. Silva, Thomas P. Sotiriou, and Nicolás Yunes.
Spin-induced scalarized black holes. Phys. Rev. Lett., 126(1):011103, 2021.
71. Emanuele Berti, Lucas G. Collodel, Burkhard Kleihaus, and Jutta Kunz. Spin-induced black-
hole scalarization in Einstein-scalar-Gauss-Bonnet theory. Phys. Rev. Lett., 126(1):011104,
2021.
72. Vitor Cardoso, Edgardo Franzin, and Paolo Pani. Is the gravitational-wave ringdown a probe
of the event horizon? Phys. Rev. Lett., 116(17):171101, 2016. [Erratum: Phys.Rev.Lett. 117,
089902 (2016)].
73. Pedro V. P. Cunha, Carlos A. R. Herdeiro, and Eugen Radu. Fundamental photon orbits: black
hole shadows and spacetime instabilities. Phys. Rev. D, 96(2):024039, 2017.
74. Pedro V. P. Cunha, Emanuele Berti, and Carlos A. R. Herdeiro. Light-Ring Stability for
Ultracompact Objects. Phys. Rev. Lett., 119(25):251102, 2017.
75. Joe Keir. Slowly decaying waves on spherically symmetric spacetimes and ultracompact
neutron stars. Class. Quant. Grav., 33(13):135009, 2016.
76. Juan Calderón Bustillo, Nicolas Sanchis-Gual, Alejandro Torres-Forné, José A. Font, Avi
Vajpeyi, Rory Smith, Carlos Herdeiro, Eugen Radu, and Samson H. W. Leong. GW190521 as
a Merger of Proca Stars: A Potential New Vector Boson of 8.7 × 10−13 eV. Phys. Rev. Lett.,
126(8):081101, 2021.
77. Hector Olivares, Ziri Younsi, Christian M. Fromm, Mariafelicia De Laurentis, Oliver Porth,
Yosuke Mizuno, Heino Falcke, Michael Kramer, and Luciano Rezzolla. How to tell an accreting
boson star from a black hole. Mon. Not. Roy. Astron. Soc., 497(1):521–535, 2020.
78. Carlos A. R. Herdeiro, Alexandre M. Pombo, Eugen Radu, Pedro V. P. Cunha, and Nicolas
Sanchis-Gual. The imitation game: Proca stars that can mimic the Schwarzschild shadow.
JCAP, 04:051, 2021.