PHYSICAL REVIEW LETTERS 134, 101501 (2025)

Black Hole Singularity Resolution in Unimodular Gravity from Unitarity

*
Steffen Gielen
School of Mathematical and Physical Sciences, University of Sheffield,
Hicks Building, Hounsfield Road, Sheffield S3 7RH, United Kingdom
†
Lucía Menéndez-Pidal
Departamento de Física Teórica, Universidad Complutense de Madrid, Parque de Ciencias 1, 28040 Madrid, Spain

(Received 13 September 2024; revised 15 October 2024; accepted 23 January 2025; published 11 March 2025)

We study the quantum dynamics of an interior planar anti–de Sitter black hole, requiring unitarity in the
natural time coordinate conjugate to the cosmological “constant of motion” appearing in unimodular
gravity. Both the classical singularity and the horizon are replaced by a nonsingular highly quantum region;
semiclassical notions of spacetime evolution are only valid in an intermediate region. For the singularity,
our results should be applicable to general black holes: unitarity in unimodular time always implies
singularity resolution.

DOI: 10.1103/PhysRevLett.134.101501

Introduction—The fate of classical singularities is one of result means that the property of singularity resolution
the most important questions for any theory of quantum depends on the choice of clock [4], and signifies a clash
gravity; indeed, the incompleteness of classical relativity between general covariance and unitarity [5], itself a
formalized in the Penrose-Hawking singularity theorems [1] somewhat controversial topic in quantum gravity.
is often cited as a main motivation for why gravity must be Here, we note that the results of Ref. [4] extend
quantum. The most important singularities of direct rel- straightforwardly to the study of black hole singularities,
evance to our Universe are at the big bang and at the center in particular for the planar anti–de Sitter (AdS) black holes
of black holes. In a first approximation, these situations can studied in Ref. [6] and related previous work in the context
be represented through idealized, spatially homogeneous of AdS/CFT [7]. The interior metric studied in these works
geometries whose high degree of symmetry allows for a is of Kasner type, with a single anisotropy variable, and
quantum description at least at an effective level. One can dynamically equivalent to a flat homogeneous, isotropic
then ask in such simple models what happens to the classical cosmology with a massless scalar field. The cosmological
singularity. constant is taken to be negative and fixed, but reinterpreting
In the context of homogeneous cosmology, the question it as a constant of motion as suggested by unimodular
of whether singularities are resolved through quantum gravity [8] turns it into another global degree of freedom,
effects (in what is usually called quantum cosmology) conjugate to unimodular time. The black hole interior is
does not have a clear answer, since it depends on the criteria then classically identical to the cosmology studied in [4],
for singularity resolution and on the precise definition of and its canonical quantization can be studied along the
the model including the choice of quantum state [2]. same lines. If the Schrödinger-like unimodular time is used
Nevertheless, one can make general statements if the to define evolution and one requires the theory to be
quantum theory is required to be unitary with respect to unitary, all singularities are resolved [9].
a given choice of time [3]: one would expect singularities to Unimodular gravity refers to modifications of general
be resolved if they are only a finite “time” away, since relativity in which only the trace-free Einstein equations are
singular evolution seems incompatible with requiring a imposed. The cosmological constant becomes an integra-
global time translation operator. Inevitably, such a general tion constant, and there is one additional global degree of
freedom. Classically unimodular gravity and general rel-
ativity are indistinguishable; there may be differences in
*
Contact author: [email protected] the quantum regime [10]. While we focus on unimodular
†
Contact author: [email protected] gravity, our mechanism for singularity resolution could be
applied to other extensions of general relativity that include
Published by the American Physical Society under the terms of clocks to define unitarity, such as the vacuum sequester [11]
the Creative Commons Attribution 4.0 International license.
Further distribution of this work must maintain attribution to in its local version [12], models with additional matter
the author(s) and the published article’s title, journal citation, fields [13], or with constants of nature becoming constants
and DOI. Funded by SCOAP3. of motion [14].

0031-9007=25=134(10)=101501(6) 101501-1 Published by the American Physical Society

 PHYSICAL REVIEW LETTERS 134, 101501 (2025)

The notion of “singularity” in this context does not only where 0 denotes derivative with respect to r. We have
apply to curvature singularities; a singularity is any point at implicitly assumed that the overall numerical factor arising
which the classical evolution terminates, and where non- from performing the integration over t, x, and y has been set
classical behavior is required for a unitary quantum theory. to one by a choice of coordinates.
These can be coordinate singularities [5]. In our black hole The Legendre transform leads to a Hamiltonian,
model, at the horizon the spatial volume goes to zero and pffiffiffi  2 
the classical solution cannot continue. Quantum unitarity 3 πk 2
H¼ Nv 2 − π v þ Λ ; ð4Þ
with respect to the preferred clock (unimodular time) would 2 v
then require the horizon to be similarly replaced by a highly
pffiffiffi
quantum region in which classical evolution is not valid. where we have made the unit choice κ ¼ 2= 3 to obtain a
This specific conclusion is sensitive to having a foliation simpler form. The resulting Hamiltonian constraint [6]
that becomes singular at the horizon. However, the con-
clusion regarding black hole singularities is more generally π 2k
valid since the singularity is only a finite time away for many C¼ − π 2v þ Λ ≈ 0 ð5Þ
v2
observers. The Belinski-Khalatnikov-Lifshitz conjecture
states that approach to a generic singularity is described is exactly the one studied in Ref. [4], where it was
by Kasner-like dynamics, like the example studied here; this interpreted as describing a flat homogeneous, isotropic
suggests that for many clocks the classical singularity would cosmology with a free massless scalar field and a perfect
need to be replaced by well-defined quantum evolution fluid (playing the role of dark energy). In this interpretation,
leading to the emergence of a white hole. Our results Λ is no longer a parameter but a (conserved) momentum
demonstrate that black hole singularities either lead to a conjugate to a time variable T. This assumption can be
failure of quantum unitarity (in unimodular time), or to a new justified by promoting Λ in (3) to a dynamical variable and
scenario for a quantum transition of a black hole to a adding a term ΛT 0 to the action. The resulting action is then
white hole. the reduction of the Henneaux-Teitelboim action for
Quantum theory of black hole interior—The classical unimodular gravity [16],
action for general relativity with cosmological constant Λ,
Z    
including the Gibbons-Hawking-York boundary term, is 1 pffiffiffiffiffiffi 1
SHT ¼ d4 x −g R − Λ þ Λ∂μ T μ ; ð6Þ
Z   Z κ 2
1 4 pffiffiffiffiffiffi 1 1 pffiffiffi
S¼ d x −g R − Λ − d3 x qK; ð1Þ
κ 2 κ with suitable boundary terms to a spatially homogeneous
geometry. Similar constructions would be possible for other
where gμν is the spacetime metric, R the Ricci scalar, qij the theories as mentioned above.
boundary metric, and K the extrinsic curvature; κ ¼ 8πG The classical solutions in the time T are found to be
where G is Newton’s constant.
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
The interior planar black hole (Kasner) metric studied in π2
Ref. [6] and previous papers [7] is given by vðTÞ ¼ − k þ 4ΛðT − T 0 Þ2 ; ð7Þ
Λ
 
ds2 ¼ −N 2 dr2 þ v2=3 e4k=3 dt2 þ e−2k=3 ðdx2 þ dy2 Þ ; ð2Þ 1 π − 2ΛðT − T 0 Þ
kðTÞ ¼ log k þ k0 ; ð8Þ
2 π k þ 2ΛðT − T 0 Þ
where N, v, and k are functions of r only. Thought of as a
radial coordinate outside the horizon, in the interior r is where T 0 and k0 are integration constants. The metric (2)
timelike and hence this metric is spatially homogeneous. It has singularities (v → 0) for T − ¼ T 0 − ðπ k =2ΛÞ and
corresponds to a locally rotationally symmetric Bianchi I T þ ¼ T 0 þ ðπ k =2ΛÞ. The Kretschmann scalar
model with metric written in the Misner parametrization
(see, e.g., Ref. [15]). One important feature of this para-  
μνξη 8Λ2 2π 2k
metrization is that the anisotropy variables (here a single Rμνξη R ¼ 1þ ð9Þ
3 ð2ΛðT − T 0 Þ þ π k Þ2
one, k) behave as free massless scalar fields in a flat
isotropic geometry, as we will see explicitly.
diverges for T → T − (black hole singularity, k → þ∞) but
Substituting the metric ansatz (2) into (1) reduces the
is finite for T ¼ T þ (black hole horizon, k → −∞).
action to
The constraint (5) can be written as C ¼ gAB π A π B þ Λ,
Z  0 2 2  making the dynamical problem equivalent to relativistic
1 ∞ ðk Þ v − ðv0 Þ2
S¼ dr − ΛNv ; ð3Þ particle motion in a configuration space (minisuperspace)
κ −∞ 3Nv parametrized by v and k and with a flat metric,

101501-2
 PHYSICAL REVIEW LETTERS 134, 101501 (2025)
 
v2 0 thus obtaining a unitary quantum theory. Such a boundary
gAB ¼ : ð10Þ condition can be seen as reflection from the singularity,
0 −1
similar to what is proposed in Ref. [18].
This minisuperspace is equivalent to the Rindler wedge, a We are interested in Λ < 0 solutions, which are the
portion of full (1 þ 1) dimensional Minkowski spacetime analog of bound states. Normalized solutions to the
with boundary at v ¼ 0. This viewpoint suggests a natural Wheeler-DeWitt equation (11) and the boundary condition
operator ordering in canonical quantization [4,6,17,18], (14) can be expressed as
leading to the Wheeler-DeWitt equation Z sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
dp ipk X iΛpn T −2Λpn sinhðjpjπÞ
∞
  Ψðv; k; TÞ ¼ e e
1 2 2 1 −∞ 2π jpjπ
ð□ þ ΛÞΨ ≔ − 2 ∂k þ ∂v þ ∂v þ Λ Ψ ¼ 0; ð11Þ n∈Z
v v pffiffiffiffiffiffiffiffiffi
× αðp; Λpn ÞK ijpj −Λpn v ; ð15Þ
where □ is the Laplace-Beltrami operator for gAB. R P
Equation (11) is covariant with respect to variable trans- where ðdp=2πÞ n ∈ Z jαðp; Λpn Þj2 ¼ 1 and
formations of v and k and, because gAB is flat, with respect
ð2nþ1Þπ
to lapse redefinitions that act as conformal transformations Λpn ¼ −e− jpj
þθðpÞ
ð16Þ
on gAB [19].
The general solution to Eq. (11) can be straightforwardly is a discrete set of allowed negative Λ values. Here, the free
given as function θðpÞ parametrizes different self-adjoint extensions
Z of □; in a sense, each choice of θðpÞ defines a different
∞
dp ipk pffiffiffiffi 
Ψðv; k; ΛÞ ¼ e αðp; ΛÞJip Λv theory. Physically, θðpÞ encodes a phase shift as the modes
−∞ 2π are reflected off v ¼ 0; an extensive discussion can be found
pffiffiffiffi 
þ βðp; ΛÞJ−ip Λv ; ð12Þ in Ref. [9]. We are interested in qualitative features of a
theory that replaces the black hole singularity and horizon by
where αðp; ΛÞ and βðp; ΛÞ are so far arbitrary and J ν ðxÞ is nonsingular quantum regions, which do not depend on the
a Bessel function of the first kind. Equation (12) gives the detailed choice. This insensitivity was also studied in detail
general solution as a function of Λ, a dynamical variable in in the same model (with a different clock) in Ref. [22];
our psetup; for Λ < 0 it is less ambiguous to pass from different self-adjoint extensions show only small quantita-
ffiffiffiffi pffiffiffiffiffiffiffi tive differences. In the following we choose θðpÞ ¼ π=jpj.
Jip ð ΛvÞ to the modified Bessel functions K ip ð −ΛvÞ
pffiffiffiffiffiffiffi Equation (15) represents the quantum states of a planar
and I ip ð −ΛvÞ [4]. The appearance of Bessel function AdS black hole interior as superpositions of different values
solutions for the black hole interior is familiar from other of momentum p and cosmological constant Λpn . Crucially,
contexts [20]. Fourier transform converts the wave function because of the reflecting boundary condition the allowed
given as a function of Λ into a time-dependent wave bound states are modified Bessel functions of the second
function dependent on T, the conjugate to Λ. kind, behaving near v ¼ 0 as
Asking whether the resulting quantum theory is unitary
pffiffiffiffiffipffi
with respect to evolution in unimodular time T [21] is pffiffiffiffiffiffiffiffiffi −Λn v

equivalent to asking whether □ is self-adjoint in a suitable Γð−ipÞ ip log

−Λpn v ∼
2
K ijpj e þ c:c:; ð17Þ
inner product. The most natural choice of inner product is 2
induced by gAB,
i.e., as superpositions of positive and negative p modes
Z ∞
Z ∞
with equal magnitude. Since changing the sign of p is
hΨjΦi ¼ dv v dk Ψ Φ: ð13Þ equivalent to time reversal, swapping the roles of horizon
0 −∞ and singularity, or switching between classical black-hole
and white-hole solutions [see Eq. (9) and below; this
It is easy to see □ is then not self-adjoint, as expected for a corresponds to the parameter π k in the classical solution],
space with boundaries that can be reached by classical none of these bound states can correspond to a single
solutions in a finite amount of time. Here, this corresponds semiclassical trajectory that ends in a singularity. The
to both the black hole horizon and the singularity being necessary superpositions of black-hole and white-hole
only a finite (unimodular) time away. Self-adjoint exten- solutions then lead to singularity resolution in this theory.
sions can be defined by restricting wave functions to a Singularity resolution—To see this explicitly, we
subspace satisfying the boundary condition [4,9] numerically study the evolution of a semiclassical state
Z  
∞
 ∂Φ ∂Ψ p
pffiffiffiffiffi −ðp−p20 Þ2 −ðΛn −Λ2 0 Þ2
lim dk v Ψ −Φ ¼ 0; ð14Þ
v→0 −∞ ∂v ∂v αðp; Λpn Þ ¼ N e 2σp 2σ
Λ ð18Þ

101501-3
 PHYSICAL REVIEW LETTERS 134, 101501 (2025)

with free parameters p0 , Λ0 , σ 2p , and σ 2Λ and a normalization

factor N . For a semiclassical interpretation we need
σ p ≪ p0 , σ Λ ≪ jΛ0 j, and p0 ≫ 1. The latter condition
then also guarantees that the allowed discrete Λpn values are
reasonably close together.
Our main observable is the volume vðTÞ. Expectation
values and moments of v in our state take the form
qffiffiffiffiffiffiffiffiffiffiffiffi
Z ∞
Z ∞ dp X 2 Λpn Λpn0 sinhðjpjπÞ
hvα ðTÞi ¼ N dv vαþ1
0 −∞ 2π n;n0 jpjπ
p p 2
ðp−p0 Þ2 ðΛ −Λ Þ2 ðΛ 0 −Λ0 Þ
− − n 2 0 − n 2 þiðΛpn −Λp0 ÞT
σ 2p
×e e 2σ
Λ
2σ
Λ
n

pffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffi
× K ijpj −Λpn v K ijpj −Λpn0 v : ð19Þ

The v integral in Eq. (19) can be done analytically, leaving

the sums and the p integral for numerical evaluation.
Because of the Gaussians inside the integral, the contri-
butions from jp − p0 j ≫ σ p and jΛpn − Λ0 j ≫ σ Λ are very
small and we can replace the infinite p integral and sums
over n and n0 by finite sums by introducing cutoffs. We
chose these so that the integrals and sums include the
regions jp − p0 j ≤ 5σ p and jΛnp − Λ0 j ≤ 5σ Λ ; increasing
the cutoff does not lead to any noticeable improvement.
The expectation value hvðTÞi can be compared with FIG. 1. Quantum expectation values hvðTÞi (400 black data
classical solutions given in Eq. (7), where π k and Λ are points) with fluctuations (in gray) for a state with p0 ¼ 70,
replaced by the average of p and Λ in our chosen states; due σ p ¼ 0.1, Λ0 ¼ −1, and σ Λ ¼ 0.15 (top) and one with p0 ¼ 30,
to the discrete spacing of possible Λ values these averages σ p ¼ 0.05, Λ0 ¼ −0.01, and σ Λ ¼ 0.0019 (bottom), compared
are not exactly equal (but close) to p0 and Λ0 . with the corresponding classical solution (7) (in blue). The points
In Fig. 1 we show the quantum expectationffi value hvðTÞi where classically v → 0 correspond to the horizon and the
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi singularity.
and fluctuations Δv ¼ hvðTÞ i − hvðTÞi2 for such a
2

state [23]. We can see that, as expected, for small jTj

the expectation value stays close to its corresponding showing disappearance of the classical horizon and singu-
classical solution (7), but it departs strongly near the larity seems universal, resulting from the reflecting boun-
horizon or singularity where interference between an dary condition (14).
ingoing (black-hole) and outgoing (white-hole) solution Discussion—It has long been stated that a quantum
becomes important. There is a finite minimal value for v theory of black hole dynamics that is required to be unitary
and in this sense, both singularity and horizon are replaced must deviate strongly from semiclassical expectations.
by quantum “bounces.” Where the expectation value Usually this is discussed in the context of unitarity of
closely follows the classical curve the variance is small, black hole formation and evaporation, leading to the
but at the bounces the variance grows, indicating strong famous issue of information loss [24]. Preserving unitarity
quantum fluctuations where the state is reflected. As together with a few other “reasonable” assumptions can
required by unitarity, all expectation values and higher lead to disasters such as a firewall at the horizon [25], or
moments are globally defined, not just for the finite T more generally the conclusion that there is no simple
interval in which the classical solution is valid. Taken at semiclassical resolution of the paradox. What we are
face value the resulting quantum solution describes cycles discussing here is different; we studied a simple quantum
of local expansion and contraction, corresponding to a model of the black hole interior, in which the gravitational
sequence of black-hole and white-hole interiors passing degrees of freedom (truncated to a Kasner metric) are
from horizon to singularity and back. Over longer time- quantised using the Wheeler-DeWitt equation. In this
scales the variances grow, suggesting a spreading in the setting too, there is a clash between unitarity and consis-
state and eventual breakdown of the semiclassical picture. tency with semiclassical physics: unitarity means globally
While all the specific features displayed here depend on the well-defined time evolution that is incompatible with
chosen parameters in the state, the qualitative behavior singularities, even a coordinate singularity at the horizon.

101501-4
 PHYSICAL REVIEW LETTERS 134, 101501 (2025)

It is a tricky issue to define unitarity in a fundamentally stays regular at the horizon, so that collapse from an
timeless setting such as canonical quantum gravity. The key asymptotically flat (or asymptotically AdS or de Sitter)
ingredient in our discussion was to use unimodular gravity region could be described as a unitary process. While these
and its preferred choice of time, implemented through alternative constructions will change the interpretation of
auxiliary fields as in Ref. [16]. With respect to this clock, the horizon, in the theory we have defined unitary quantum
both the horizon and the singularity are only a finite time dynamics will always necessarily replace the universal
away, as they would be for an infalling observer. Unitarity singularity by nonsingular evolution into a white hole:
forces us to replace the horizon and singularity with highly either unitarity fails, or there is no black hole singularity.
quantum bounce regions. Unitarity with respect to a
different clock would generally lead to different conclu- Acknowledgments—The work of S. G. is funded by the
sions [5]. Our general conclusions should be valid for any Royal Society through the University Research Fellowship
standard of time such that the singularity or horizon are Renewal URFnRn221005. L. M. P. is supported by the
only a finite time away; it would be interesting to construct Leverhulme Trust.
other analytically accessible examples. Again we point out
that this could be done in other theories that extend general Data availability—The data that support the findings of
relativity by including local clock variables [11–14]. this Letter are openly available [23]; embargo periods
The work of Ref. [6] studied a number of clocks in the may apply.
same interior black hole spacetime (see also Ref. [26] for an
extension to charged black holes). For instance, one can use
the classically monotonic anisotropy parameter k. With [1] S. W. Hawking and G. F. R. Ellis, The Large Scale Structure
of Space-Time (Cambridge University Press, Cambridge,
respect to k or other clocks such as v and π v , no deviation
England, 1973).
from semiclassical physics was found in Ref. [6]. These [2] V. Husain and O. Winkler, Singularity resolution in quan-
observations are fully consistent with the results of Ref. [4] tum gravity, Phys. Rev. D 69, 084016 (2004); A. Ashtekar,
for k and v, and with the general conjecture of Ref. [3], Singularity resolution in loop quantum cosmology: A brief
where different clocks were classed as “slow” and “fast.” overview, J. Phys. Conf. Ser. 189, 012003 (2009); C. Kiefer,
Unitarity with respect to a fast clock, which runs to ∞ at a On the avoidance of classical singularities in quantum
singularity, does not require any deviation from semi- cosmology, J. Phys. Conf. Ser. 222, 012049 (2010).
classical physics. In our model k, v, and π v are all fast. [3] M. J. Gotay and J. Demaret, Quantum cosmological singu-
Similar behavior is also found for a massless scalar field in larities, Phys. Rev. D 28, 2402 (1983).
homogeneous Wheeler-DeWitt cosmology [27]. However, [4] S. Gielen and L. Menéndez-Pidal, Singularity resolution
depends on the clock, Classical Quantum Gravity 37,
such clocks do not describe the experience of local
205018 (2020); Unitarity, clock dependence and quantum
observers; classical singularities are troublesome exactly recollapse in quantum cosmology, Classical Quantum
because one can reach them in finite time. When such a Gravity 39, 075011 (2022); L. Menéndez-Pidal, The prob-
slow clock is studied, on approach to the singularity one lem of time in quantum cosmology, PhD thesis, University
must either give up unitarity or find a generic resolution of of Nottingham, 2022; arXiv:2211.09173.
singularities and possibly horizons. [5] S. Gielen and L. Menéndez-Pidal, Unitarity and quantum
The metric form (2) corresponds to a simple model resolution of gravitational singularities, Int. J. Mod. Phys. D
of a black hole with planar symmetry, but only a slight 31, 2241005 (2022).
extension—adding a second possible anisotropy variable [6] S. A. Hartnoll, Wheeler–DeWitt states of the AdS–
Schwarzschild interior, J. High Energy Phys. 01 (2023) 066.
in the Misner parametrization—turns it into the general
[7] J. P. S. Lemos, Two-dimensional black holes and planar
Kasner form describing, according to the Belinski- general relativity, Classical Quantum Gravity 12, 1081
Khalatnikov-Lifshitz conjecture, successive periods during (1995); E. Witten, Anti-de Sitter space, thermal phase
the generic approach to a spacelike singularity, even for transition, and confinement in gauge theories, Adv. Theor.
more realistic black holes (see also the related Ref. [18]). Math. Phys. 2, 505 (1998); A. Frenkel, S. A. Hartnoll, J.
Since this second anisotropy variable again acts as a Kruthoff, and Z. D. Shi, Holographic flows from CFT to the
massless scalar field in an isotropic Universe, the results Kasner universe, J. High Energy Phys. 08 (2020) 003; N.
illustrated here would be expected to hold more generically Bilic and J. C. Fabris, Thermodynamics of AdS planar black
for singularities. For horizons, the general picture is less holes and holography, J. High Energy Phys. 11 (2022) 013.
clear since the model studied here sees the horizon as a [8] J. L. Anderson and D. Finkelstein, Cosmological constant
and fundamental length, Am. J. Phys. 39, 901 (1971); J. J.
coordinate singularity, and the black hole metric at a
van der Bij, H. van Dam, and Y. J. Ng, The exchange of
horizon is in general more complicated. Already for the massless spin two particles, Physica (Amsterdam) 116A,
usual AdS-Schwarzschild black hole, the positive curvature 307 (1982); W. G. Unruh, Unimodular theory of canonical
of constant time slices would contribute at the horizon and quantum gravity, Phys. Rev. D 40, 1048 (1989); W.
potentially change the conclusions. In general, it is in Buchmüller and N. Dragon, Einstein gravity from restricted
principle always possible to construct a notion of time that coordinate invariance, Phys. Lett. B 207, 292 (1988); Gauge

101501-5
 PHYSICAL REVIEW LETTERS 134, 101501 (2025)

fixing and the cosmological constant, Phys. Lett. B 223, 313 [15] M. Bojowald, Canonical Gravity and Applications: Cos-
(1989). mology, Black Holes, and Quantum Gravity (Cambridge
[9] S. Gryb and K. P. Y. Thébault, Superpositions of the University Press, Cambridge, England 2010).
cosmological constant allow for singularity resolution and [16] M. Henneaux and C. Teitelboim, The cosmological
unitary evolution in quantum cosmology, Phys. Lett. B constant and general covariance, Phys. Lett. B 222, 195
784, 324 (2018); Bouncing unitary cosmology I: Mini- (1989).
superspace general solution, Classical Quantum Gravity 36, [17] S. W. Hawking and D. N. Page, Operator ordering and the
035009 (2019); Bouncing unitary cosmology II: Mini- flatness of the universe, Nucl. Phys. B264, 185 (1986).
superspace phenomenologyClassical Quantum Gravity 36, [18] M. J. Perry, No future in black holes, arXiv:2106.03715.
035010 (2019). [19] J. J. Halliwell, Derivation of the Wheeler–DeWitt equation
[10] E. Álvarez, S. González-Martín, M. Herrero-Valea, and C. P. from a path integral for minisuperspace models, Phys. Rev.
Martín, Quantum corrections to unimodular gravity, J. High D 38, 2468 (1988).
Energy Phys. 08 (2015) 078; R. de León Ardón, N. Ohta, [20] M. Bouhmadi-López, S. Brahma, C. Y. Chen, P. Chen, and
and R. Percacci, Path integral of unimodular gravity, Phys. D. h. Yeom, Annihilation-to-nothing: A quantum gravita-
Rev. D 97, 026007 (2018); R. Carballo-Rubio, L. J. Garay, tional boundary condition for the Schwarzschild black hole,
and G. García-Moreno, Unimodular gravity vs general J. Cosmol. Astropart. Phys. 11 (2020) 002.
[21] To see why T is unimodular time, note the global factor
relativity: A status report, Classical Quantum Gravity 39, pffiffiffiffiffiffi
Nv ¼ −g in the Hamiltonian H. For H to be given just by
243001 (2022); E. Álvarez and E. Velasco-Aja, A primer on pffiffiffiffiffiffi
unimodular gravity, Phys. Part. Nucl. 54, 908 (2023). the constraint C, we must choose N so that −g is constant.
[11] N. Kaloper and A. Padilla, Sequestering the standard model [22] T. Pawłowski and A. Ashtekar, Positive cosmological
vacuum energy, Phys. Rev. Lett. 112, 091304 (2014); constant in loop quantum cosmology, Phys. Rev. D 85,
Vacuum energy sequestering: The framework and its cos- 064001 (2012).
mological consequences, Phys. Rev. D 90, 084023 (2014); [23] The supporting data for this article (in the form of a
90, 109901(E) (2014); 90, 084023 (2014). Mathematica notebook) are openly available from the
[12] N. Kaloper, A. Padilla, D. Stefanyszyn, and G. Zahariade, arXiv version of the Letter.
Manifestly local theory of vacuum energy sequestering, [24] S. W. Hawking, Breakdown of predictability in gravitational
Phys. Rev. Lett. 116, 051302 (2016). collapse, Phys. Rev. D 14, 2460 (1976).
[13] J. D. Brown and K. V. Kuchaŕ, Dust as a standard of space [25] A. Almheiri, D. Marolf, J. Polchinski, and J. Sully, Black
and time in canonical quantum gravity, Phys. Rev. D 51, holes: Complementarity or firewalls?, J. High Energy Phys.
5600 (1995); V. Husain and T. Pawłowski, Time and a 02 (2013) 062.
physical Hamiltonian for quantum gravity, Phys. Rev. Lett. [26] M. J. Blacker and S. Ning, Wheeler DeWitt states of a charged
AdS4 black hole, J. High Energy Phys. 12 (2023) 002.
108, 141301 (2012).
[27] A. Ashtekar and P. Singh, Loop quantum cosmology: A
[14] J. Magueijo, Cosmological time and the constants of nature,
status report, Classical Quantum Gravity 28, 213001 (2011).
Phys. Lett. B 820, 136487 (2021).

101501-6