Quantum Field Theory in Curved Spacetime

Bernard S. Kay
Department of Mathematics, University of York, York YO10 5DD, U.K.

October 3, 2023

Abstract
The 2023 second edition of a 2006 encyclopedia article on mathematical aspects of quantum
field theory in curved spacetimes (QFTCST). Section-titles (in bold, with new section-titles in
bold italics) are: Introduction and preliminaries, Construction of a ∗-algebra for a
arXiv:2308.14517v2 [gr-qc] 29 Sep 2023

real linear scalar field on globally hyperbolic spacetimes and some general theorems,
More about (quasifree) Hadamard states, Particle creation and the limitations of the
particle concept, Theory of the stress-energy tensor, More about the intersection of
QFTCST with AQFT and the Fewster-Verch No-Go Theorem, Hawking and Unruh
effects, More about (classical and) quantum fields on black hole backgrounds, Non-
globally hyperbolic spacetimes and the time-machine question, More about QFT on
non-globally hyperbolic spacetimes, Other related topics and some warnings. The article
contains many references. It also includes a review of, and also compares and contrasts, recent
results on the implications of QFTCST for the question of the instability of three sorts of Cauchy
horizon – first those inside black holes such as especially Reissner-Nordström-de Sitter and Kerr-
de Sitter, second the compactly generated Cauchy horizons of spacetimes in which time-machines
get manufactured, and third the Cauchy horizon of the spacetime which is believed to describe
evaporating black holes and which underlies (one version of) the black hole information-loss puzzle.
—
To the memory of my father, Max Kay (1917-1986) whose love of learning, support, and life-
example continue to sustain me.
And also to Chris Isham, my academic father, together with whom I began the study of this
interesting topic 50 years ago!

Preface to the Second Edition

Since the first edition [1] of this encyclopedia article appeared in 2006, there has continued to be
interesting progress in quantum field theory in curved spacetime (QFTCST) on the mathematical
and conceptual side while the subject has continued to play an important rôle in many areas of
astrophysics and cosmology. Furthermore QFTCST is relevant (or may soon become relevant)
to the analysis of a variety of recent (or recently proposed) laboratory-based (see e.g. [2, 3])
and space-based (see e.g. [4]) experiments to test aspects of quantum behaviour in gravitational
backgrounds and/or in accelerated frames as well as experiments involving condensed matter and
fluid etc. analogues of QFTCST effects such as the Hawking effect [5, 6, 7, 8, 9, 10]. Meanwhile
on the theory side, there has recently been renewed interest (see e.g. [11] and see also the final
paragraph of this preface) in the lessons that QFTCST may teach us towards the quest for a full
quantum theory of gravity.
For all these reasons and more, there continues to be a need for accessible introductions to
the subject and I hope that my 2006 encyclopedia article, re-issued with corrections and updates
here, may usefully contribute to that purpose as far as the mathematical and conceptual basics of
the subject are concerned. For the more practical and computational aspects of the subject see
e.g. (DeWitt, 1975) and (Birrell and Davies, 1982) from the ‘Further Reading’ section and/or the

1
 textbooks (Parker and Toms, 2009) and (Mukhanov and Winitzky, 2007) which we have newly
added to that section.1
The style of the 2006 edition of the encylopedia demanded that no references be given, but
rather a small list of articles and books be appended to each article in a section entitled ‘Further
Reading’. Aside from a few corrections, additions and updates as well as some small rearrange-
ments of the material, we have preserved the 2006 article, with its list of Further Reading (a total
of 18 books and papers to which we have added just two post-2006 textbooks and one collection
of articles) to form the core of this second edition. I believe that this core is still an appropriate
vehicle for introducing the subject.
There were a few places in the first edition where we informally specified a reference for
some specific result which is not in the Further Reading list by specifying the author name(s)
and year of publication, and in those places we have replaced those informal references with full
bibliographical references – included in a new section after the Further Reading section at the
end. These references are indicated in the text by numbers in square brackets. References to the
Further Reading section are instead in the form (Author name(s), year).
Within that core, we have inserted four new sections (Section 6, 3, 8, 10) to the article fully
up to date. One reasonable way to read the article, for someone new to the subject, would be to
skip the new sections at a first reading and then return to them (or just to those that seem of
interest) later.
We have also added to the bibliography quite a big collection of references to the newly added
sections. Still, as for the 2006 core, the choice of additional material and of additional references
inevitably reflects my own interests and tastes and limitations as well as the limitations of space
and the bibliography is in no way to be regarded as comprehensive. Rather, I have attempted to
choose a representative selection of articles on the topics I treat which are either self-contained or
which have a useful list of references themselves, or are themselves much cited. I am sure I must
have unjustly omitted many important and valuable and relevant references. But wherever I have
done so, I hope the authors will understand that I must have done so inadvertently and forgive
me and I would be pleased to be informed about any such omissions. I would, of course, also be
pleased to be informed of any errors.
A feature of the second edition is that it spells out in some detail a number of points which,
while somewhat technical, are useful to understand if one wants to contribute to the subject
and also fairly easy to explain but which, while likely known to experts, are not so well-known
and/or not always easily extractable from the literature. For example, we include a discussion
of the recent correction by Moretti [23] to the definition in (Kay and Wald, 1991) of the notion
of ‘global Hadamard’ and we also include a brief note on the similarities and differences between
the Wald (Wald, 1978) and Décanini-Folacci [24] schemes for the Hadamard renormalization of
the stress-energy tensor. We also provide a simple example that we hope will help the reader to
appreciate at least part of the reason why one would expect the Fewster-Verch no-go theorem
[25] on the impossibility of any locally covariant preferred state construction to hold. For similar
reasons, in Section 11 we have also retained (and added to) our list of warnings. Aside from that
list, Section 11 is intended to at least mention a number of topics which, while important, haven’t
been adequately discussed elsewhere in the article. Note though that some of the material that
was previously in Section 11 has now been moved to, and expanded upon, in earlier sections.
One of the most interesting things about QFTCST is that its study has enabled us to address
questions about quantum, gravitational, and quantum-gravitational physics which one might have
thought would be out of reach of QFTCST because they involve, in an essential way, the quantum
and/or dynamical nature of the gravitational field. Regarding the quantum-gravitational, the
impact and influence that our understanding of the Hawking effect has had on string theory and
AdS/CFT etc. is widely in evidence – see for example the recent lecture notes [26] of Hartman.
Also one version of the black hole information-loss puzzle (see Section 10) – which is of course a
puzzle about quantum gravity – has a lot to do with QFT on a particular class of spacetimes – the
1
Over the years, there have been many other books, lecture notes, review articles etc. with a partly overlapping
scope to that of the present article. See e.g. [12, 13, 14, 15, 16, 17, 18, 19, 20, 21]. These all differ from the present
article and from each other in view of the time at which they were written and in view of their differring emphases on
different aspects of the subject. However all of them contain valuable material and points of view that aren’t easily to
be found elsewhere. See also the article [22] by Hollands in this encyclopedia.

2
 spacetimes of black-hole evaporation – that are conjectured to be solutions to semiclassical gravity
that correspond to Hawking-evaporating black holes. As for gravitational physics, a recurring
theme in QFTCST concerns the rôle quantum effects may have in rendering various sorts of
Cauchy horizon impenetrable (or not). This is reflected in a theme running through the present
article – see our discussion of work in QFTCST related to the question of strong cosmic censorship
for Reissner-Nordström-de Sitter and Kerr-de Sitter Cauchy horizons in Section 8; our discussion
of time machine-like Cauchy horizons in Section 9; and, thirdly, our discussion in Section 10 of
Juárez-Aubry’s recent conjectures [136] about the Cauchy horizon of spacetimes of black-hole
evaporation. In the course of discussing all three of these sorts of spacetime with Cauchy horizon
we shall also point out some interesting similiarities and differences between QFT on them.

1 Introduction and Preliminaries

Quantum Field Theory (QFT) in Curved Spacetime is a hybrid approximate theory in which
quantum matter fields are assumed to propagate in a fixed classical background gravitational
field. Its basic physical prediction is that strong gravitational fields can polarize the vacuum and,
when time-dependent, lead to pair creation just as a strong and/or time-dependent electromagnetic
field can polarize the vacuum and/or give rise to pair-creation of charged particles. One expects
it to be a good approximation to full quantum gravity provided the typical frequencies of the
gravitational background are very much less than the Planck frequency (c5 /Gℏ)1/2 ∼ 1043 s−1 )
and provided, with a suitable measure for energy, the energy of created particles is very much
less than the energy of the background gravitational field or of its matter sources. Undoubtedly
the most important prediction of the theory is the Hawking effect (Hawking, 1975), according to
which a classical black hole will emit thermal radiation at the Hawking temperature κ/2π where
κ is the black hole’s surface gravity. E.g. in the case of an uncharged nonrotating black hole,
T = (8πM )−1 where M is the black hole’s mass. (Here and from now on, we use Planck units
where G, c, ℏ and k [Boltzmann’s constant] are all taken to be 1.)
On the mathematical side, the need to formulate the laws and derive the general properties of
QFT on non-flat spacetimes forces one to state and prove results in local terms and, as a byproduct,
thereby leads to an improved perspective on flat-space-time QFT too. It’s also interesting to
formulate QFT on idealized spacetimes with particular global geometrical features. Thus, QFT
on spacetimes with bifurcate Killing horizons is intimately related to the Hawking effect; QFT on
spacetimes with closed timelike curves is intimately related to the question whether the laws of
physics permit the manufacture of a time-machine.
As is standard in General Relativity, a curved spacetime is modelled mathematically as a
(paracompact, Hausdorff) manifold M equipped with a pseudo-Riemannian metric g of signature
(−, + + +) (we follow the conventions of the standard text “Gravitation” by Misner Thorne and
Wheeler [W.H. Freeman, San Francisco, 1973]). We shall also assume, except where otherwise
stated, our spacetime to be globally hyperbolic, i.e. that M admits a global time coordinate, by
which we mean a global coordinate t such that each constant-t surface is a smooth Cauchy surface
i.e. a smooth spacelike 3-surface cut exactly once by each inextendible causal curve.2 (Without
this default assumption, extra problems arise for QFT which we shall discuss, in connection with
the time-machine question and the spacetime of black hole evaporation in Section 9.) In view of
this definition, globally hyperbolic spacetimes are clearly time-orientable and we shall assume a
choice of time-orientation has been made so we can talk about the “future” and “past” directions.
Modern formulations of the subject take, as the fundamental mathematical structure modelling
the quantum field, a ∗-algebra A (with identity I) together with a family of local sub ∗-algebras
A(O) labelled by bounded open regions O of the spacetime (M, g) and satisfying the isotony or
net condition that O1 ⊂ O2 implies A(O1 ) is a subalgebra of A(O2 ), as well as the condition that
whenever O1 and O2 are spacelike separated, then A(O1 ) and A(O2 ) commute. Observables are
modelled as algebra elements, A for which A∗ = A.
Standard concepts and techniques from algebraic quantum theory (AQFT) 3 are then applica-
2
For the equivalence of this statement with other definitions of global hyperbolicity, see [28].
3
For references on AQFT, see the book (Haag, 2012), the recent review [29] by Fewster and Rejzner and also the
article [30] by Buchholz and Fredenhagen in this encyclopedia.

3
 ble: In particular, states are defined to be positive (this means ω(A∗ A) ≥ 0 ∀A ∈ A) normalized
(this means ω(I) = 1) linear functionals on A. For an observable A, ω(A) is then interpreted as
the expectation value of A in the state ω. One distinguishes between pure states and mixed states,
only the latter being writeable as non-trivial convex combinations of other states. To each state,
ω, the GNS-construction associates a representation, ρω , of A on a Hilbert space Hω together
with a cyclic vector Ω ∈ Hω such that

ω(A) = ⟨Ω|ρω (A)Ω⟩

(and the GNS triple (ρω , Hω , Ω) is unique up to equivalence). There are often technical advantages
in formulating things so that the ∗-algebra is a C∗ -algebra. Then the GNS representation is as
everywhere-defined bounded operators and is irreducible if and only if the state is pure. A useful
concept, due to Haag, is the folium of a given state ω which may be defined to be the set of all
states ωσ which arise in the form Tr(σρω (·)) where σ ranges over the density operators (trace-class
operators with unit trace) on Hω .
Given a state, ω, and an automorphism, α, which preserves the state (i.e. ω ◦α = ω) then there
will be a unitary operator, U , on Hω which implements α in the sense that ρω (α(A)) = U −1 ρω (A)U
and U is chosen uniquely by the condition U Ω = Ω.
On a stationary spacetime, i.e. one which admits a one-parameter group of isometries whose
integral curves are everywhere timelike, the algebra will inherit a one-parameter group (i.e. satis-
fying α(t1 )◦α(t2 ) = α(t1 +t2 )) of time-translation automorphisms, α(t), and, given any stationary
state (i.e. one which satisfies ω ◦ α(t) = ω ∀t ∈ R) these will be implemented by a one-parameter
group of unitaries, U (t), on its GNS Hilbert space satsifying U (t)Ω = Ω. If U (t) is strongly
continuous so that it takes the form e−iHt and if the Hamiltonian, H, is positive, then ω is said
to be a ground state. Typically one expects ground states to exist and often be unique.
Another important class of stationary states for the algebra of a stationary spacetime is the
class of KMS states, ω β , at inverse temperature β; these have the physical interpretation of
thermal equilibrium states. In the GNS representation of one of these, the automorphisms are
also implemented by a strongly-continuous unitary group, e−iHt , which preserves Ω but (in place
of H positive) there is a complex conjugation, J, on Hω such that

e−βH/2 ρω (A)Ω = Jρω (A∗ )Ω (1)

for all A ∈ A.
An attractive aspect of QFTCST is that its main qualitative features are already present
for linear field theories and, unusually in comparison with other questions in QFT, these are
susceptible of a straightforward explicit and rigorous mathematical formulation.4 In fact, as our
principal example, we give in Section (2) a construction for the field algebra for the quantized real
linear Klein Gordon equation
(□g − m2 )ϕ = 0 (2)
of mass m on a globally hyperbolic spacetime (M, g). Here, □g denotes the Laplace Beltrami
operator g ab ∇a ∂b (= (| det(g)|)−1/2 ∂a ((| det(g)|1/2 g ab ∂b )). Note that we may, e.g. replace −m2
by −m2 − V where V is a scalar external background classical field. In case m is zero, taking V to
equal R/6, where R denotes the Riemann scalar, then makes the equation conformally invariant
and the equation is then called the conformally coupled Klein Gordon equation. In the case of a
vacuum solution to Einstein’s equations, R of course vanishes and so the minimally coupled and
conformally coupled equations are the same. However, the stress-energy tensors, Tab , defined as
1 1
−2| det(g)|− 2 δ/δgab S – where the action, S, is the integral with the volume element | det(g)| 2 d4 x
of L where L is, respectively, − 21 (g ab ∂a ϕ∂b ϕ + m2 ϕ2 ) and − 21 (g ab ∂a ϕ∂b ϕ + [m2 + 16 R]ϕ2 ) – will,
of course, differ.
4
What lies behind the tractability of linear QFT is the fact that, unlike for nonlinear QFT, when one formulates the
theory as a dynamical system, the quantum path is the same as the classical path. For Equation (2), this is formalized,
as explained in Section 4, by Equation (5). Moreover, there appears to be a metatheorem: For linear field theories,
every quantum property arises as a corollary to some classical property. For example, the Reeh-Schlieder property (say)
for a wedge (see [31]) may be viewed as inherited from a ‘pre-Reeh-Schlieder’ theorem which refers to the one-particle
sector of the theory and can be understood as a property of the classical equation (2) (See also Footnote 8.) The reader
is invited to supply other illustrations.

4
 The main new feature of quantum field theory in curved spacetime (present already for linear
field theories) is that, in a general (neither flat, nor stationary) spacetime there will not be any
single preferred state but rather a family of preferred states, members of which are best regarded as
on an equal footing with one-another. It is this feature which makes the above algebraic framework
particularly suitable, indeed essential, to a clear formulation of the subject. Conceptually it is
this feature which takes the most getting used to. In particular, one must realize that, as we shall
explain in Section 4, the interpretation of a state as having a particular “particle-content” is in
general problematic because it can only be relative to a particular choice of “vacuum” state and,
depending on the spacetime of interest, there may be one state or several states or, frequently,
no states at all which deserve the name “vacuum” and even when there are states which deserve
this name, they will often only be defined in some approximate or asymptotic or transient sense
or only on some subregion of the spacetime.
Concomitantly, one does not expect global observables such as the “particle number” or the
quantum Hamiltonian of flat-spacetime free field theory to generalize to a general curved space-
time context, and for this reason local observables play a central rôle in the theory. The (quantized
renormalized) stress-energy tensor, Tab , is a particularly natural and important such local observ-
able and the theory of this is central to the whole subject. A brief introduction to it is given in
Section 5.
The main reason for our interest in the quantized stress energy tensor is that its expectation
value, ω(Tab ), in a quantum state, ω, will serve as the right hand side of the semiclassical Einstein
equations Gab = 8πGN ω(Tab ) which, when solved simultaneously with a suitable QFTCST equa-
tion such as (2), is expected, in some situations, to give us some sort of approximate indication
of what the predictions of quantum gravity would be while staying within the framework of a
classical, but now dynamical, spacetime.
There is then a new section, 6, describing some recent developments in AQFT, focusing on
the point of view of locally covariant QFT and discussing the Fewster-Verch no-go theorem on
the non-existence of a locally covariant preferred state. Section 7 is on the Hawking and Unruh
effects. Section 9 is on the problems of extending the theory beyond the “default” setting, to
non-globally hyperbolic spacetimes. Here, we focus on the question of whether it is possible in
principle to manufacture a time-machine, and, in a new section, we also discuss QFT on the (non-
globally hyperbolic) spacetime which is believed to give an approximate description of black-hole
evaporation and which underlies one version of the black-hole information loss puzzle. Finally,
Section 11 briefly mentions a number of other interesting and active areas of the subject as well
as issuing a few warnings to be borne in mind when reading the literature.

2 Construction of a ∗-Algebra for a Real Linear Scalar

Field on Globally Hyperbolic Spacetimes and Some General
Theorems
On a globally hyperbolic spacetime, the classical equation (2) admits well-defined advanced and
retarded Green functions (strictly bidistributions) ∆A and ∆R and the standard covariant quantum
free real (or “Hermitian”) scalar field commutation relations familiar from Minkowski spacetime
free field theory naturally generalize to the (heuristic) equation

[ϕ̂(x), ϕ̂(y)] = i∆(x, y)I

where ∆ is the Lichnerowicz commutator function (again, strictly bidistribution) ∆ = ∆A − ∆R .

Here, the “ ˆ ” on the quantum field ϕ̂ serves to distinguish it from a classical solution ϕ. In
mathematical work, one does not assign a meaning to the field at a point itself, but rather aims to
assign meaning to smeared fields ϕ̂(F ) for all real test functions F ∈ C0∞ (M)5 which are then to
1
be interpreted as standing for M ϕ̂(x)F (x)| det(g)| 2 d4 x. In fact, it is straightforward to define a
R

minimal field algebra (see below) Amin generated by such ϕ̂(F ) which satisfy the suitably smeared
5
Here and elsewhere, when we write C0∞ (X) for some manifold X, we mean, by default, C0∞ (X, R).

5
 version
[ϕ̂(F1 ), ϕ̂(F2 )] = i∆(F1 , F2 )I
of the above commutation relations together with Hermiticity (i.e. ϕ̂(F )∗ = ϕ̂(F )), the property of
being a weak solution of the equation (2) (i.e. ϕ̂((□g − m2 )F ) = 0 ∀F ∈ C0∞ (M)) and linearity
in test functions. There is a technically different alternative formulation of this minimal algebra,
∗
which is known as the Weyl algebra, which is constructed R to be the C algebra generated by
1
operators W (F ) (to be interpreted as standing for exp(i M ϕ̂(x)F (x)| det(g)| 2 d4 x) satisfying

W (F1 )W (F2 ) = exp(−i∆(F1 , F2 )/2)W (F1 + F2 )

together with W (F )∗ = W (−F ) and W ((□g − m2 )F ) = I.6 With either the minimal algebra or
the Weyl algebra one can define, for each bounded open region O, subalgebras A(O) as generated
by the ϕ̂(·) (or the W (·)) smeared with test functions supported in O and verify that they satisfy
the above “net” condition and commutativity at spacelike separation.
Specifying a state, ω, on Amin is tantamount to specifying its collection of n-point distributions
(i.e. smeared n-point functions) ω(ϕ̂(F1 ) . . . ϕ̂(Fn )).7 A particular rôle is played in the theory by
the quasi-free states for which all the truncated n-point distributions except for n = 2 vanish.
Thus all the n-point distributions for odd n vanish while the 4-point distribution is made out of
the 2-point distribution according to

ω(ϕ̂(F1 )ϕ̂(F2 )ϕ̂(F3 )ϕ̂(F4 )) = ω(ϕ̂(F1 )ϕ̂(F2 ))ω(ϕ̂(F3 )ϕ̂(F4 ))

+ω(ϕ̂(F1 )ϕ̂(F3 ))ω(ϕ̂(F2 )ϕ̂(F4 )) + ω(ϕ̂(F1 )ϕ̂(F4 ))ω(ϕ̂(F2 )ϕ̂(F3 ))

etc.8 The anticommutator distribution

G(F1 , F2 ) = ω(ϕ̂(F1 )ϕ̂(F2 )) + ω(ϕ̂(F2 )ϕ̂(F1 )) (3)

of a quasi-free state (or indeed of any state) will satisfy the properties (for all test functions F ,
F1 , F2 etc.):
(a) (symmetry) G(F1 , F2 ) = G(F2 , F1 )
6
If (S, σ) denotes the real vector space, S, of classical real-valued smooth solutions of (2) with compact support on
Cauchy surfaces equipped with the symplectic form σ defined such that σ(ϕ1 , ϕ2 ) is equal to the σ((f01 , p10 ); (f02 , p20 )) of
Equation (4) in Section 4 where (f 1 , p1 ), (f 2 , p2 ) are the Cauchy data of ϕ1 , ϕ2 on any Cauchy surface (which we identify
here with the C0 of Section 4) then what is usually called the Weyl algebra over (S, σ) is the C∗ algebra generated by
operators W̃ (ϕ) (to be interpreted as standing for exp(iσ(ϕ̂, ϕ)) satisfying W̃ (ϕ1 )W̃ (ϕ2 ) = exp(−iσ(ϕ1 , ϕ2 )/2)W̃ (ϕ1 +
ϕ2 ) and the operator W (F ) of the main text may be identified with W̃ (ϕ) where ϕ is equal to the solution ∆F (where
the distribution ∆F is defined by (∆F )(F1 ) = ∆(F1 , F ) for all test functions F1 ∈ C0∞ (M).). See Section 3 of (Kay
and Wald, 1991).
7
In the case of the Weyl algebra, one restricts attention to “regular” states for which the map F → ω(W (F ))
is sufficiently often differentiable on finite dimensional subspaces of C0∞ (M) and defines the n-point distributions in
terms of derivatives with respect to suitable parameters of expectation values of suitable Weyl algebra elements. For
example, one may define the anticommutator function, G(F1 , F2 ), of a state ω, (in place of (3)) to be −2∂ 2 W (t1 F1 +
t2 F2 )/∂t1 ∂t2 |t1 =t2 =0 and the smeared two-point function, ω(ϕ(F1 )ϕ(F2 )), will then be (G(F1 , F2 ) + i∆(F1 , F2 ))/2.
8
Each quasi-free state, ω, can be thought of as arising from a one-particle Hilbert space structure (K, H) over
(S, σ). This consists of a one-particle Hilbert space H and a real-linear map, K from S to H such that the complexified
range of K is dense in H and K is symplectic in the sense that ∀ϕ1 , ϕ2 ∈ S, 2ℑ⟨Kϕ1 |Kϕ2 ⟩ = σ(ϕ1 , ϕ2 ). The
anticommutator function G(F1 , F2 ) of (3) is then equal to 2⟨K∆F1 |K∆F2 ⟩ and in fact, in a Weyl algebra formulation,
we have ω(W (F )) = exp(−∥KF ∥2 /2) where ∥ · ∥ denotes the norm in the one-particle Hilbert space. Also in this Weyl
algebra case, ω is pure if and only if the range of K is dense (without the need for complexification). For the proof of
this and more information about one-particle Hilbert-space structures, see Appendix A in (Kay and Wald, 1991). When
the underlying spacetime is stationary, the space of solutions, (S, σ), will be equipped with a one-parameter group,
T (t) of symplectomorphisms and one can seek a one-particle structure (in the sense introduced in [33]) consisting of
a one-particle Hilbert space structure (K, H) (with ranK dense) for (S, σ) together with a unitary group, e−iHt for a
positive Hamiltonian, H, such that U (t)K = KT . The corresponding state ω is then a ground state. (See [34] for the
uniqueness of such one-particle structures and [35] for constructions for a wide class of stationary spacetimes.) There
is also a notion [36] of ‘KMS one-particle structure’ for which the corresponding state, ω, is a KMS state.)

6
 (b) (weak bisolution property)

G((□g − m2 )F1 , F2 ) = 0 = G(F1 , (□g − m2 )F2 )

(c) (positivity) G(F, F ) ≥ 0 and G(F1 , F1 )1/2 G(F2 , F2 )1/2 ≥ |∆(F1 , F2 )|

and it can be shown that, to every bilinear functional G on C0∞ (M) satisfying (a), (b) and (c),
there is a quasi-free state with two-point distribution (1/2)(G + i∆). One further declares a
quasi-free state to be physically admissible only if (for pairs of points in sufficiently small convex
neighborhoods) it satisfies the
(d) ([local] Hadamard condition) The unsmeared anticommutator function, G(x1 , x2 ) takes the
form “G(x1 , x2 ) = 2π1 2 u(x1 , x2 )P σ1 + v(x1 , x2 ) log |σ| + w(x1 , x2 ) ” where σ denotes the

(signed) geodetic interval between x1 and x2 . u (which satisfies u(x1 , x2 ) = 1 when x1 = x2 )

and v are certain smooth two-point functions determined in terms of the local geometry
(and the local values of any potential term V ) by something called the Hadamard procedure
while the smooth two-point function w depends on the state.9
This last condition expresses the requirement that (locally) the two-point distribution actually
“is” (in the usual sense in which one says that a distribution “is” a function) a smooth function for
pairs of non-null-separated points, and at the same time requires that the two-point distribution
be singular at pairs of null-separated points and locally specifies the nature of the singularity
for such pairs of points with a leading “principal part of 1/σ” type singularity and a subleading
“log |σ|” singularity. The important point is that this local Hadamard condition on the two-
point distribution is believed to be the correct generalization to a curved spacetime of the well-
known universal short-distance behaviour shared by the (truncated) two-point distributions of all
physically relevant states for the special case of our theory when the spacetime is flat (and, in case
one adds a V to m2 ,PV vanishes). In the latter case, u reduces to 1, and v can be expressed as a
∞
simple power series n=0 vn σ n with v0 = m2 /4 etc. (In the general case,P∞ u is the square root of
the Van Vleck-Morette determinant and one again has a power series n=0 vn σ n for v but where
the coefficients, vn , are, in general, symmetric functions of x1 and x2 which are determined by
certain Hadamard recursion relations.)
Actually, it is known (this is the content of “Kay’s Conjecture” which was proved [32] by M.
Radzikowski in 1992) that (a), (b), (c) and (d) together imply that the two-point distribution
is nonsingular at all pairs of (not necessarily close-together) spacelike-separated points. More
important than this result itself is a replacement for the Hadamard condition which Radzikowski
[37] originally introduced as a tool towards its proof:
(d′ ) (Wave Front Set [or Microlocal] Spectrum Condition) WF(G + i∆) = {(x1 , p1 ; x2 , p2 ) ∈
T ∗ (M × M)⧹0| x1 and x2 lie on a single null geodesic, p1 is cotangent to that null geodesic
and future pointing, and p2 when parallel transported along that null geodesic from x2 to
x1 equals −p1 }
Here, the notion of the wave front set of a distribution, D or WF(D) for short, is a notion from
microlocal analysis. For the gist of what it means, it suffices to know that to say that an element
(x, p) of the cotangent bundle of a manifold (excluding the zero section 0) is in the wave front
set, WF, of a given distribution, D, on that manifold may be expressed informally by saying that
that D is singular at the point x in the direction p. (And here the notion is applied in the case
that D is G + i∆, thought of as a distribution on the manifold M × M.)
For many questions, the microlocal spectrum condition, (d′ ) above, is a more convenient and
powerful technical starting point than the original formulation, (d), of the Hadamard condition.
Early applications of (d′ ) were to the question of time-machines (Kay, Radzikowski and Wald,
1997) – see Section 9 – and to certain questions about energy inequalities (see e.g. [38]) and
Wick powers (see [39] and references therein), while it is an important ingredient of pAQFT –
see Section 11. However, it seems that the original formulation, (d), cannot be dispensed with
9
This formula implicitly involves a choice of length scale, say L and the middle term should more strictly be written
v(x1 , x2 ) log(|σ|/L). Changing the length scale from L to, say, L′ will then be accompanied by the change from w to
w + v log(L/L′ ).

7
 altogether. Thus it continues to be needed to define the Hadamard renormalization scheme – see
Section 5 – for the stress-energy tensor, and there are questions, including questions involving the
restriction of two-point functions to (spacelike and lightlike) surfaces, where (cf. the remarks at
the end of Section 4.2.1 in [42]) (d) still seems to be needed. For example, in its ‘global’ version
discussed in Section 3 (d) is essential for the analysis of stationary Hadamard states on spacetimes
with bifurcate Killing horizons in (Kay and Wald, 1991) and of QFT near other types of horizons
– see, for example, [40].
We remark that generically (and, e.g., always if the spatial sections are compact and, in case
one adds a potential term V , if m2 + V (x) is everywhere positive) the Weyl algebra for equation
(2) on a given stationary spacetime will have a unique ground state and unique KMS states at
each temperature and these will be quasi-free. Also [41] (see also [42]) it is known that ground
states and KMS states for (2) and many other linear field theories on stationary spacetimes will
satisfy (d′ ) (or the generalizations thereof to other spins etc.) – i.e. will be Hadamard.
Quasi-free states are important for the theory of Equation (2) also because of a theorem of
Verch [43] (in verification of another conjecture of Kay) that (in the Weyl algebra framework)
on the algebra of any bounded open region, the folia of the quasi-free Hadamard states coincide.
With this result one can extend the notion of physical admissibility to not-necessarily-quasi-free
states by demanding that, to be admissible, a state belong to the resulting common folium when
restricted to the algebra of each bounded open region; equivalently that it be a locally normal
state on the resulting natural extension of the net of local Weyl algebras to a net of local W∗
algebras.

3 More about (Quasifree) Hadamard states

For further reading on quasi-free Hadamard states, see the article [44] in (Brunetti et al., 2005).
Before the work of Radzikowski mentioned above, a notion of global Hadamard was defined
in Section 3 of (Kay and Wald, 1991) for the quantum field theory of (2) on a general globally
hyperbolic spacetime which incorporates both the local Hadamard condition and the absence of
nonlocal spacelike singularities.
An important correction to that definition was made [23] by Moretti in 2021.10 The (thus
corrected) notion of global Hadamard is initially only defined on a causal normal neighborhood,
N , of a Cauchy surface, C – a notion introduced, and shown to always exist, in Section 2 of (Kay
and Wald, 1991) which is defined by the property that every two causally related points in N
belong to a common convex normal neighborhood and thus are connected in N by a unique causal
geodesic. One then argues that this global Hadamard property propagates in the sense that the
set of states which are Hadamard in a causal normal neighborhood of one Cauchy surface, say
C1 , is the same as the set of states which are Hadamard in a causal normal neighborhood of any
other Cauchy surface, C2 .11
10
The error in (Kay and Wald, 1991) which Moretti points out is in the sentence: “The squared geodesic distance,
σ, will be well-defined and smooth in an open neighborhood, O, in M × M , of the set of causally related points (x, y)
such that J + (x) ∩ J − (y) and J − (x) ∩ J + (y) are contained within a convex normal neighborhood.” The point is that
while, for any (x, y) in that set of causally related points, say with y to the future of x, a preferred geodesic connecting
x and y can be picked out by demanding that it stays inside J + (x) ∩ J − (y), for spacelike-related (x, y) in the open
neighborhood O (which may well be connected by more than one geodesic) there is no preferred convex causal set the
staying inside of which could be used as a selection criterion to pick out a preferred exemplar. And the different such
geodesics might well have different lengths. However, it is shown in [23] that as long as M is paracompact, the resulting
different values for the (singular) first two terms of the Hadamard form in (d) above, will differ by a smooth two-point
function.
11
A proof of this was sketched in (Kay and Wald, 1991), a forerunner having been given in [45]. An alternative proof
based on version (d′ ) of the (global) Hadamard condition and the microlocal propagation of singularities theorem was
given in [37]. We give in the rest of this footnote a version of the proof in (Kay and Wald, 1991). Clearly it suffices to
prove it for C2 , to the future of C1 . One can then take an arbitrary compact submanifold, S2 of C2 . The (trapezium-
shaped) region to the past of S2 and to the future of C1 will (if its boundary is included) itself be compact and therefore
coverable with a finite number of overlapping causal normal neighborhoods of a finite number of Cauchy surfaces which
can be taken to be constant-time surfaces for a certain finite set of values of a suitable time-coordinate such that S2
and the causal past of S2 in C1 are themselves constant time surfaces. One may then see that the Hadamard property

8
 The above propagation result may be combined (as explained in [46]) to show, with a smooth
deformation argument, that, once one has a construction of, or proof of the existence of, a (global)
Hadamard state for Equation (2) on some particular globally hyperbolic spacetime with a given
topology then there must be many globally Hadamard states on any globally hyperbolic spacetime
with the same topology. For this deformation argument in the case of a massive scalar field, the
particular spacetime with a given topology was chosen, in [46], to be an ultrastatic spacetime – i.e.
a static spacetime where the metric takes the form −dt2 +(3)gij dxi dxj – with the desired topology
and it was shown in [46] (or follows from the more recent more general argument for stationary
spacetimes in [41]) that ground states on such spacetimes are Hadamard states.
However, this still leaves open many questions as to whether we can construct, and have some
control over the properties of, Hadamard states for certain given spacetimes or classes of space-
times and leaves unanswered questions as to whether certain special states with certain special
properties on certain special spacetimes are (also) Hadamard states. To amplify on this point:
(a) the deformation method of [46] is fairly inexplicit – it is good for showing that there are many
Hadamard states but not for computing with them; (b) the deformation method would not work
for spin-3/2 and spin-2 fields, where the background spacetime has to satisfy the Einstein equa-
tions for the appropriate classical stress-energy tensor (solved simultaneously with the classical
equation(s) of the matter field(s) in the problem), so cannot be smoothly deformed to ultrastatic;
(c) there is a similar problem to show that there are states obeying the analytic wavefront set
condition (cf. e.g. [144]) because one cannot make the deformation in an analytic way.12 One also
wants to generalize to field theories other than that of the scalar equation (2).
Besides ground and KMS states for stationary spacetimes as discussed above, much more work
has been done, in aid of answering the many open questions referred to above, on the construction
of Hadamard states on general (globally hyperbolic) spacetimes as well as on special classes of
spacetimes such as asymptotically flat spacetimes and spacetimes with bifurcate Killing horizons
(see Section 7) using techniques including the theory of pseudo-differential operators [47, 48],
the theory of Calderón projectors [49], scattering theory [50, 51] and the characteristic Cauchy
problem [52, 53] both for the covariant Klein Gordon equation as well as the Dirac field [54, 55]
and the Proca and Maxwell and linearized Yang Mills fields [56, 57], while Gérard and Wrochna
have also explicitly generalized the deformation method of [46] to linearized Yang Mills in [58]
and also given [59] a construction of Hadamard states for linearized quantum gravity – in the
harmonic gauge on analytic backgrounds of bounded geometry.
There is also a way [60] (see also [61]) of obtaining a class of “vacuum-like” Hadamard states,
known as Brum-Fredenhagen (or BM ) states, by making a certain modification of a certain con-
struction of a preferred state for a general spacetime known as the Sorkin-Johnston (or S-J ) state
[62] which however itself fails, in general, to be Hadamard! Let us also remark here that the S-J
state construction also fails to be locally covariant and thus (see [39, 63, 61]) is not in conflict
with the Fewster-Verch no-go theorem which we will discuss in Section 6.
Progress on demonstrating the Hadamard nature of certain states on black hole spacetimes
and other spacetimes with bifurcate Killing horizons will be discussed in Section 7.)
It is natural to wonder whether the Hadamard condition can be replaced with some approxima-
tion or other. E.g. one might contemplate truncating the power series that defines the Hadamard
v function at some finite order. In this connection, it is worth noting that Fewster and Verch have
proven [39] that the Wick squares of all time derivatives of the quantized Klein-Gordon field have
finite fluctuations only if the Wick-ordering is defined with respect to an exactly Hadamard state.
Lastly, let us mention that, in the context of semiclassical gravity, a proposal has recently been
made [64] for a notion of surface Hadamard which is conjectured to be a suitable condition on
gravitational and quantum initial data on an initial surface to ensure that the Cauchy development
of that initial data exists and is a Hadamard solution of semiclassical gravity. Here again, one
propagates from the earlier to the later of any pair of adjacent overlapping such neighborhoods, say Ni and Ni+1 , by
first noting that on their overlap, Ni ∩ Ni+1 , the difference between the two-point function of any global Hadamard
state on Ni+1 and any global Hadamard state on Ni will be a smooth bisolution and then by looking at the Cauchy
evolution of the Cauchy data for this bisolution on a Cauchy surface for the overlap. One can then obviously propagate
from the earliest to the latest causal normal neighborhood in a finite number of such steps. The result follows since the
choice of S was arbitrary.
12
I thank Chris Fewster for making the points (a), (b), (c).

9
might contemplate approximating this notion. But, as is pointed out in [64] (see Footnote 5 there),
any formulation of semiclassical gravity that admits states, ω, which are not exactly Hadamard
would involve metrics that have a lower degree of smoothness than C ∞ .

4 Particle Creation and the Limitations of the Particle

Concept
Global hyperbolicity also entails that the Cauchy problem is well posed for the classical field
equation (2) in the sense that for every Cauchy surface, C, and every pair (f, p) of Cauchy data in
C0∞ (C), there exists a unique solution ϕ in C0∞ (M) such that f = ϕ|C and p = | det(g)|1/2 g 0b ∂b ϕ|C .
Moreover ϕ has compact support on all other Cauchy surfaces. Given a global time coordinate
t, increasing towards the future, foliating M into a family of contant-t Cauchy surfaces, Ct , and
given a choice of global time-like vector field τ a (for example, for suitable t [65] one may take
τ a = g ab ∂b t) enabling one to identify all the Ct , say with C0 , by identifying points cut by the same
integral curve of τ a , a single such classical solution ϕ may be pictured as a family {(ft , pt ) : t ∈ R}
of time-evolving Cauchy data on C0 . Moreover, since (2) implies, for each pair of classical solutions,
ϕ1 , ϕ2 , the conservation (i.e. ∂a j a = 0) of the current j a = | det(g)|1/2 g ab (ϕ1 ∂b ϕ2 − ϕ2 ∂b ϕ1 ), the
symplectic form (on C0∞ (C0 ) × C0∞ (C0 ))
Z
1 1 2 2
σ((ft , pt ); (ft , pt )) = (ft1 p2t − p1t ft2 )d3 x (4)
C0

will be conserved in time. (d3 x denotes dx1 ∧ dx2 ∧ dx3 .)

Corresponding to this picture of classical dynamics, one expects there to be a description of
quantum dynamics in terms of a family of sharp-time quantum fields (φt , πt ) on C0 , satisfying
heuristic canonical commutation relations

[φt (x), φt (y)] = 0, [πt (x), πt (y)] = 0, [φt (x), πt (y)] = iδ 3 (x, y)I

and evolving in time according to the same dynamics as the Cauchy data of a classical solution.
(Both these expectations are correct because the field equation is linear.) An elegant way to
make rigorous mathematical sense of these expectations is in terms of a ∗-algebra with identity
generated by Hermitian objects “σ((φ0 , π0 ); (f, p))” (“symplectically smeared sharp-time fields at
t = 0”) satisfying linearity in f and p together with the commutation relations

[σ((φ0 , π0 ); (f 1 , p1 )), σ((φ0 , π0 ); (f 2 , p2 ))] = iσ((f 1 , p1 ); (f 2 , p2 ))I

and to define (symplectically smeared) time-t sharp-time fields by demanding

σ((φt , πt ); (ft , pt )) = σ((φ0 , π0 ); (f0 , p0 )) (5)

where (ft , pt ) is the classical time-evolute of (f0 , p0 ). This ∗-algebra of sharp-time fields may be
identified with the (minimal) field ∗-algebra of the previous section, the ϕ̂(F ) of the previous
section being identified with σ((φ0 , π0 ); (f, p)) where (f, p) are the Cauchy data at t = 0 of ∆F .
(This identification is of course many-one since ϕ̂(F ) = 0 whenever F arises as (□g − m2 )G for
some test function G ∈ C0∞ (M).)
Specializing momentarily to the case of the free scalar field (□−m2 )ϕ = 0 (m ̸= 0) in Minkowski
space with a flat t = 0 Cauchy surface, the “symplectically smeared” two-point function of the
usual ground state (“Minkowski vacuum state”), ω0 , is given, in this formalism, by
1 1
ω0 (σ((φ, π); (f 1 , p1 ))σ((φ, π); (f 2 , p2 ))) = (⟨f |µf 2 ⟩ + ⟨p1 |µ−1 p2 ⟩ + iσ((f 1 , p1 ); (f 2 , p2 )) (6)
2
where the inner products are in the one-particle Hilbert space H = L2C (R3 ) and µ = (m2 − ∇2 )1/2 .
The GNS representation of this state may be concretely realised on the familiar Fock space F(H)
over H by
ρ0 (σ((φ, π); (f, p))) = −i(↠(a) − (↠(a))∗ )

10
where a denotes the element of H
(µ1/2 f + iµ−1/2 p)
a= √
2
(we note in passing that, if we equip H with the symplectic form 2Im⟨·|·⟩, then K : (f,R p) 7→ a is a
symplectic map – see Footnote 8) and ↠(a) is the usual smeared creation operator (=“ ↠(x)a(x)d3 x”)
on F(H) satisfying
[(↠(a1 ))∗ , ↠(a2 )] = ⟨a1 |a2 ⟩H I.
The usual (smeared) annihilation operator, â(a), is (↠(Ca))∗ where C is the natural complex
conjugation, a 7→ a∗ on H. Both of these operators annihilate the Fock vacuum vector ΩF . In
this representation, the one parameter group of time-translation automorphisms

α(t) : σ((φ0 , π0 ); (f, p)) 7→ σ((φt , πt ); (f, p)) (7)

is implemented by exp(−iHt) where H is the second quantization of µ (i.e. the operator otherwise
known as µ(k)↠(k)â(k)d3 k) on F(H).
R

The most straightforward (albeit physically artificial) situation involving “particle creation” in
a curved spacetime concerns a globally hyperbolic spacetime which, outside of a compact region,
is isometric to Minkowski space with a compact region removed – i.e. to a globally hyperbolic
spacetime which is flat except inside a localized “bump” of curvature. See Figure 1. (One could
also add a potential, V , to m2 in (2) provided it is supported inside the bump.) On the field
algebra (defined as in the previous section) of such a spacetime, there will be an in vacuum
state (which may be identified with the Minkowski vacuum to the past of the bump) and an out
vacuum state (which may be identified with the Minkowski vacuum to the future of the bump)
and one expects e.g. the “in vacuum” to arise as a many particle state in the GNS representation
of the “out vacuum” corresponding to the creation of particles out of the vacuum by the bump of
curvature.

t=T

t=0

Figure 1: A spacetime which is flat outside of a compact bump of curvature.

In the formalism of this section, if we choose our global time coordinate on such a spacetime so
that, say, the t = 0 surface is to the past of the bump and the t = T surface to its future, then the
single automorphism α(T ) (defined as in (7)) encodes the overall effect of the bump of curvature
on the quantum field and one can ask whether α(T ) is implemented by a unitary operator in the
GNS representation of the Minkowski vacuum state (6).
This question may be answered by referring to the real-linear map T : H → H which sends
aT = 2−1/2 (µ1/2 fT + iµ−1/2 pT ) to a0 = 2−1/2 (µ1/2 f0 + iµ−1/2 p0 ). By the conservation in time
of σ and the symplecticity, noted in passing above, of the map K : (f, p) 7→ a, this satisfies the
defining relation
Im⟨T a1 |T a2 ⟩ = Im⟨a1 |a2 ⟩
of a classical Bogoliubov transformation. Splitting T into its complex-linear and complex-antilinear
parts by writing
T = α + βC

11
 where α and β are complex linear operators, this relation may alternatively be expressed in terms
of the pair of relations
α∗ α − β̄ ∗ β̄ = I, ᾱ∗ β̄ = β ∗ α
where ᾱ = CαC, β̄ = CβC.
We remark that there is an easy-to-visualize equivalent way of defining α and β in terms of the
analysis, to the past of the bump, into positive and negative-frequency parts of complex solutions
to (2) which are purely positive-frequency to the future of the bump. In fact, if, for any element
a ∈ H, we identify the positive frequency solution to the Minkowski-space Klein Gordon equation

ϕpos
out (t, x) = ((2µ)
−1/2
exp(−iµt)a)(x)

with a complex solution to (2) to the future of the bump, then (it may easily be seen) to the past
of the bump, this same solution will be identifiable with the (partly positive-frequency, partly
negative-frequency) Minkowski-space Klein-Gordon solution
   
ϕin (t, x) = (2µ)−1/2 exp(−iµt αa)(x) + (2µ)−1/2 exp(iµt)β̄a (x)

and this could be taken to be the defining equation for the operators α and β.
It is known (by a 1962 theorem [66] of Shale) that the automorphism (7) will be unitarily
implemented if and only if β is a Hilbert Schmidt operator on H. Wald [67], in case m ≥ 0
and Dimock [68], in case m ̸= 0 have verified that this condition is satisfied in the case of our
bump-of-curvature situation. In that case, if we denote the unitary implementor by U , we have
the results
(i) The expectation value ⟨U Ω|N (a)U Ω⟩F (H) of the number operator, N (a) = ↠(a)â(a), where
a is a normalized element of H, is equal to ⟨βa|βa⟩H .
= 1 . . . ∞), in H such that
(ii) First note that there exists an orthonormal basis of vectors, ei , (i P
the (Hilbert Schmidt) operator β̄ ∗ ᾱ∗−1 has the canonical form i λi ⟨Cei |·⟩|ei ⟩. We then
have (up to an undertermined phase)
!
1X † †
U Ω = N exp − λi â (ei )â (ei ) Ω.
2 i

where the normalization constant, N , is chosen so that ∥U Ω∥ = 1. This formula makes

manifest that the particles are created in pairs.
We remark that, identifying elements, a, of H with positive-frequency solutions (below, we shall
call them “modes”) as explained above, Result (i) may alternatively be expressed by saying that
the expectation value, ωin (N (a)), in the in-vacuum state of the occupation number, N (a), of a
normalized mode, a, to the future of the bump is given by ⟨βa|βa⟩H .
The formalism and results (i) and (ii) above generalize from spacetimes which are flat outside
of a compact bump of curvature to spacetimes which are only asymptotically flat in suitable senses.
In fact much work has been done on classical scattering theory for linear field equations such as
(2) and on the implications for the quantum theory. At a heuristic level, one will still have notions
of classical solutions which are positive frequency asymptotically towards the future/past, and,
in consequence, one will have well-defined asymptotic notions of in and out vacua and of in and
out particles built on those vacua. At a rigorous algebraic level, say for Equation (2), once one
has established (See e.g. [69] and [70]) the existence of classical wave (or ‘Møller’) operators Ω±
which can be thought of as mapping solutions, ϕ0 , of a comparison Minkowski-space Klein-Gordon
equation to solutions Ω± ϕ0 in the spacetime of interest13 which look like ϕ0 at late/early times
13
Here we assume we can identify the manifold of Minkowski space with the manifold of the spacetime of interest.
When the latter has a different topology as e.g. in the case of exterior Schwarzschild, this is of course not possible
globally but one may overcome this with suitable mollified partial identifications – see e.g. around [70, Equation (4.1)].
[70] also explains that there is considerable leeway in the way these (partial) identifications are made. Note also that,
for the massive Klein-Gordon field, the interaction of the Schwarzschild background with the Klein-Gordon field is
closely analogous to the interaction of a long-range external Coulomb potential with a charged field in Minkowski space
and therefore the wave operators constructed in [70] are suitably Dollard-modified [see [71]] wave operators.

12
 (and once one has also established asymptotic completeness) then one may, e.g., define out and
in fields according to W̃out/in (ϕ0 ) = W̃ (Ω± ϕ0 ) and out and in vacuum states ωout/in according to
ωout/in (W̃ (Ω± ϕ0 ) = ω0 (W̃0 (ϕ0 )) where W̃0 is the Weyl operator and ω0 the usual vacuum state
for the Minkowski space Klein-Gordon equation. We shall say more about classical and quantum
scattering theory on spacetimes related to black holes in Section 7.
Also, in, e.g. cosmological, models where the background spacetime is slowly-varying in time,
one can define approximate adiabatic notions of classical positive frequency solutions, and hence
also of quantum “vacuum” and “particles” at each finite value of the cosmological time. But, at
times where the gravitational field is rapidly varying, one does not expect there to be any sensible
notion of “particles”. And, in a rapidly time-varying background gravitational field which never
settles down one does not expect there to be any sensible particle interpretation of the theory at
all. To understand these statements, it suffices to consider the 1 + 0-dimensional Klein-Gordon
equation with an external potential V :

d2
 
− 2 − m2 − V (t) ϕ = 0
dt

which is of course a system of one degree of freedom, mathematically equivalent to the harmonic
oscillator with a time-varying angular frequency ϖ(t) = (m2 + V (t))1/2 . One could of course
express its quantum theory in terms of a time-evolving Schrödinger wave function Ψ(φ, t) and
attempt to give this a particle interpretation at each time, s, by expanding Ψ(φ, s) in terms of
the harmonic oscillator wave functions for a harmonic oscillator with some particular choice of
angular frequency. But the problem is, as is easy to convince oneself, that there is no such good
choice. For example, one might think that a good choice would be to take, at time s, the set of
harmonic oscillator wave functions with angular frequency ϖ(s). (This is sometimes known as the
method of “instantaneous diagonalization of the Hamiltonian”). But suppose we were to apply
this prescription to the case of a smooth V (·) which is constant in time until time 0 and assume
the initial state is the usual vacuum state. Then at some positive time s, the number of particles
predicted to be present is the same as the number of particles predicted to be present on the same
prescription at all times after s for a V̂ (·) which is equal to V (·) up to time s and then takes
the constant value V (s) for all later times (see Figure 2). But V̂ (·) will generically have a sharp
corner in its graph (i.e. a discontinuity in its time-derivative) at time s and one would expect a
large part of the particle production in the latter situation to be accounted for by the presence of
this sharp corner – and therefore a large part of the predicted particle-production in the case of
V (·) to be spurious.

t
0 s

Figure 2: Plots of ϖ against t for the two potentials V (continuous line) and V̂ (continuous line up to s and
then dashed line) which play a rôle in our critique of “instantaneous diagonalization”.

Back in 1 + 3 dimensions, even where a good notion of particles is possible, it depends on the
choice of time-evolution, as is dramatically illustrated by the Unruh effect (see Section 5).

13
 5 Theory of the Stress-Energy Tensor
To orient ideas, consider first the free (minimally coupled) scalar field, (□−m2 )ϕ = 0, in Minkowski
space. If one quantizes this system in the usual Minkowski-vacuum representation, then the
expectation value of the renormalized stress-energy tensor (which in this case is the same thing
as the normal ordered stress-energy tensor) in a vector state Ψ in the Fock space will be given by
the formal point-splitting expression
 
1
⟨Ψ|Tab (x)Ψ⟩ = lim ∂a1 ∂b2 − ηab (η cd ∂c1 ∂d2 + m2 )
(x1 ,x2 )→(x,x) 2

⟨Ψ|ρ0 (ϕ(x1 )ϕ(x2 ))Ψ⟩ − ⟨ΩF |ρ0 (ϕ(x1 )ϕ(x2 ))ΩF ⟩

(8)
where ηab is the usual Minkowski metric and ΩF the Fock space vacuum vector. A sufficient
condition for the limit here to be finite and well-defined would, e.g., be for Ψ to consist of a (nor-
malised) finite superposition of n-particle vectors of form ↠(a1 ), . . . , ↠(an )ΩF where the smearing
functions a1 , . . . , an are all C ∞ elements of H (i.e. of L2C (R3 ). The reason this works is that the
two-point function in such states shares the same short-distance singularity as the Minkowski-
vacuum two-point function. For exactly the same reason, one obtains a well-defined finite limit if
one defines the expectation value of the stress-energy tensor in any physically admissible quasi-free
state by the expression
 
1 2 1 cd 1 2 2
ω(Tab )(x)) = lim ∂a ∂b − ηab (η ∂c ∂d + m ) (ω(ϕ(x1 )ϕ(x2 )) − ω0 (ϕ(x1 )ϕ(x2 ))) .
(x1 ,x2 )→(x,x) 2
(9)
This latter point-splitting formula generalizes to a definition for the expectation value of the renor-
malized stress-energy tensor for an arbitrary physically admissible quasi-free state (or indeed for
an arbitrary state whose two point function has Hadamard form – i.e. whose anticommutator
function satisfies Condition (d) of Section 2) on the minimal field algebra and to other linear
field theories (including the stress tensor for a conformally coupled linear scalar field) on a gen-
eral globally hyperbolic spacetime (and the result obtained agrees with that obtained by other
methods, including dimensional regularization and zeta-function regularization). However, the
generalization to a curved spacetime involves a number of important new features which we next
briefly list. (See (Wald, 1978) for the original way in which this was done and see Décanini and
Folacci [24] for a technically slightly different and, in some ways simpler, alternative14 .)
First, the subtraction term which replaces ω0 (ϕ(x1 )ϕ(x2 )) is, in general, not the expectation
value of ϕ(x1 )ϕ(x2 ) in any particular state, but rather a particular locally constructed Hadamard
two-point function whose physical interpretation is more subtle; the renormalization is thus in
general not to be regarded as a normal ordering. Second, the immediate result of the resulting
limiting process will not be covariantly conserved and, in order to obtain a covariantly conserved
quantity, one needs to add a particular local geometrical correction term. The upshot of this is
that the resulting expected stress-energy tensor is covariantly conserved but possesses a (state-
independent) anomalous trace. In particular, for a massless conformally coupled linear scalar field,
14
In (Wald, 1978), the locally constructed Hadamard two-point function which isPused as the subtraction piece, say
∞
H(x, x′ ), is defined by demanding that the first term in a Hadamard expansion n=0 wn σ
n
for the w term in the
Hadamard formula (see item (d) in Section 2 but note that Wald’s w corresponds to what in (d) and in [24] would
be U w) takes the same value as in Minkowski space. This results in a subtraction term which, as explained in (Wald,
1978) satisfies (2) in one of its arguments, but (generically) not the other, and which fails to be symmetric. On the
other hand, in [24], Décanini and Folacci (who work with the expectation value in some state of the time-ordered
product of ϕ(x) with ϕ(x′ ) rather than the expectation value of the anticommutator as in (Wald, 1978) and the present
paper, but this makes little difference – other than a factor of 2 – for the stress-energy tensor) take the subtraction
piece, say H ′ (x, x′ ), to be simply the first two terms in the expression in (d) (for some choice of length scale). This is
symmetric but fails (generically) to satisfy (2) in either of its arguments. Nevertheless the procedures in (Wald, 1978)
and [24] both give the same result for the stress-energy tensor (up to the usual ambiguity in local geometrical terms)
and the same formula for the trace anomaly. See also [72] for yet another alternative approach which also spells out the
generalization to other dimenions. Note that Décanini and Folacci [24] also discuss the Hadamard form in dimensions
other than 4. Also see [73] for the generalization – in the case of a charged scalar field – of what is done in [24] to
include an external electromagnetic field.

14
 one has (for all physically admissible quasi-free states, ω) the trace anomaly formula
1
ω(Taa (x)) = (2880π 2 )−1 (Cabcd C abcd + Rab Rab − R2 )
3
– plus an arbitrary multiple of □R. In fact, in general, the thus-defined renormalized stress-energy
tensor operator (see below) is only defined up to a finite renormalization ambiguity which consists
of the addition of arbitrary multiples of the functional derivatives with respect to gab of the four
quantities Z
In = Fn (x)| det(g)|1/2 d4 x,
M

n = 1 . . . 4, where F1 = 1, F2 = R, F3 = R2 , F4 = Rab Rab . In the Minkowski-space case, only

the first of these ambiguities arises and it is implicitly resolved in the formulae (8), (9) inasmuch
as these effectively incorporate the renormalization condition that ω0 (Tab ) = 0. (For the same
reason, the locally-flat example we give below has no ambiguity.)
One expects, in both flat and curved cases, that, for test functions, F ∈ C0∞ (M), there will
exist operators Tab (F ) which are affiliated to the net of local W ∗ -algebras referred to in Section 2
and that it is meaningful to write
Z
ω(Tab (x))F (x)| det(g)|1/2 d4 x = ω(Tab (F ))
M

provided that, by ω on the right-hand side, we understand the extension of ω from the Weyl
algebra to this net. (Tab (F ) is however not expected to belong to the minimal algebra or be
affiliated to the Weyl algebra.)
An interesting simple example of a renormalized stress-energy tensor calculation is the so-called
Casimir effect calculation for a linear scalar field on a (for further simplicity, 1 + 1-dimensional)
timelike cylinder spacetime of radius R (see Figure 3). This spacetime is globally hyperbolic and
stationary and, while locally flat, globally distinct from Minkowski space. As a result, while –
provided the regions O are sufficiently small (such as the diamond region in Figure 3) – elements
A(O) of the minimal net of local algebras on this spacetime will be identifiable, in an obvious
way, with elements of the minimal net of local algebras on Minkowski space, the stationary ground
state ωcylinder will, when restricted to such thus-identified regions, be distinct from the Minkowski
vacuum state ω0 . The resulting renormalized stress-energy tensor (as first pointed out in (Kay,
1979), defineable, once the above identification has been made, exactly as in (9)) turns out to
be non-zero and, interestingly, to have an (in the natural coordinates, constant) negative energy
density T00 . In fact:
1
ωcylinder (Tab ) = ηab .
24πR2

Figure 3: The time-like cylinder spacetime of radius R with a diamond region isometric to a piece of Minkowski
space. See (Kay, 1979).

15
 6 More about the Intersection of QFTCST with AQFT
and the Fewster-Verch No-Go Theorem
We have already indicated in the introduction that there is a close interrelationship between
QFTCST and algebraic quantum field theory (AQFT). In recent years, this connection has got
even closer. See e.g. the collection (Brunetti et al., 2015). This process was greatly accelerated by
two conceptual revolutions, the first of which centred on the introduction in 1992 by Radzikowski
(see Section 2) of the microlocal spectrum condition – which was susceptible [74, 75] to generaliza-
tion to interacting fields. The second conceptual revolution was the adoption of a new perspective
in which, rather than focus on quantum field theory on any single fixed curved spacetime, one
asks instead about the simultaneous specification of a quantum field theory on a whole family of
(at least at first, globally hyperbolic) spacetimes. More precisely one defines a locally covariant
quantum field theory to consist of a covariant functor between the category of globally hyperbolic
spacetimes with isometric embeddings as morphisms and the category of ∗-algebras with unital
injective *-monomorphisms as morphisms. 15 A key paper which might be said to be a manifesto
for this approach is (Brunetti et al., 2003) (see also the review article [20]) although some of the
ingredients in the approach had been apparent in earlier works. In particular, Dimock (Dimock,
1980) had advocated a related shift of perspective while the papers (Kay, 1979) and [128] (see
Footnote 15) had anticipated some aspects of the rôle that the embedding of one spacetime inside
another might play. Also some results had already been obtained in the locally covariant frame-
work before (Brunetti et al., 2003) appeared. See e.g. Verch’s paper [141] on generalizing the spin
statistics theorem to curved spacetimes.
Amongst the several interesting new research directions that flowed from the adoption of
the locally covariant approach to QFTCST, let us mention the work of Fewster and Verch [25]
which shows that a locally covariant QFT will be “the same in all spacetimes”16 in a certain
natural sense provided it satisfies a condition that they introduce and call dynamical locality.
They show that this condition is satisfied by a wide class of quantum field theories, although
they also point out that there are some quantum field theories for which it doesn’t hold – with
interesting consequences. For those theories for which dynamical locality holds (together with
some other mild technical conditions), which includes [76] the quantum theory of the massive
minimally coupled Klein Gordon equation (2), it is proved in [76] (see also [61]) i.a. that it is
impossible to make a locally covariant choice of preferred state in all spacetimes (in the above
category of spacetimes). We shall call this the Fewster-Verch no-go theorem.
To gain some insight into why the Fewster-Verch no-go theorem holds in the case of (for
simplicity, the 1+1-dimensional version of) Equation (2) let us assume that, if there were such
a covariant choice of preferred state, then, in the case of (1+1-dimensional) Minkowski space, it
would coincide with the usual (Poincaré invariant) ground state, ω0 , and that, in the case of the
(1+1-dimensional) timelike cylinder spacetime (see Section 5) it would coincide with the stationary
ground state ωcylinder . (See Section 5.) Then, if there were a covariant choice of preferred state on
all spacetimes, the preferred state on a diamond region of Minkowski space (see Figure 3), viewed
as a globally hyperbolic spacetime in its own right, would have, at the same time, to coincide
with the restriction to it of the Minkowski vacuum when we isometrically embed the diamond
in Minkowski space, and with the restriction to it of ωcylinder when we isometrically embed the
diamond in the timelike cylinder in the way illustrated in Figure 3. But, as we saw in Section 5,
these states cannot coincide since, e.g., they have different expectation values for the renormalized
stress energy tensor at any point in the diamond. A contradiction!

7 Hawking and Unruh effects

The original calculation by Hawking (Hawking, 1975) concerned a model spacetime for a star
which collapses to a black hole. For simplicity, we shall only discuss the spherically-symmetric
15
The morphisms on the spacetime side need to be restricted i.a. so that if one spacetime is embedded as a sub-
spacetime of another another, then their causal structures are compatible. The need for this is illustrated by the helical
strip example of [128].
16
This is what Fewster and Verch call their ‘same physics in all spacetimes’ or ‘SPAS’ property.

16
 case. (See Figure 4.)

... singularity
..✮
.. horizon
.. ✮

interior of star

Figure 4: The spacetime of a star collapsing to a spherical black hole.

Adopting a similar “mode” viewpoint to that mentioned after Results (i) and (ii) in Section 3,
the result of the calculation may be stated as follows: For a real linear scalar field satisfying (2)
with m = 0 (and V = 0) on this spacetime, the expectation value ωin (N (aϖ,ℓ )) of the occupation
number of a one-particle outgoing mode aϖ,ℓ ) localized (as far as a normalized mode can be)
around ϖ in angular-frequency-space and about retarded time v and with angular momentum
“quantum number” ℓ, in the in-vacuum state (i.e. on the minimal algebra for a real scalar field
on this model spacetime) ωin is, at late retarded times, given by the formula

Γ(ϖ, ℓ)
ωin (N (aϖ,ℓ )) =
exp(8πM ϖ) − 1

where M is the mass of the black hole and the absorption factor (alternatively known as grey body
factor) Γ(ϖ, ℓ) is equal to the norm-squared of that part of the one-particle mode, aϖ,ℓ , which,
viewed as a complex positive frequency classical solution propagating backwards in time from late
retarded times, would be absorbed by the black hole. This calculation can be understood as an
application of Result (i) of Section 3 (even though the spacetime is more complicated than one
with a localized “bump of curvature” and even though the relevant overall time-evolution will not
be unitarily implemented, the result still applies when suitably interpreted) and the heart of the
calculation is an asymptotic estimate of the relevant “β” Bogoliubov coefficient which turns out
to be dependent on the geometrical optics of rays which pass through the star just before the
formation of the horizon. This result suggests that the in-vacuum state is indistinguishable at
late retarded times from a state of black-body radiation at the Hawking temperature, THawking =
1/8πM , in Minkowski space from a black-body with the same absorption factor.
The above picture of Hawking radiation for the spacetime of stellar collapse to a black hole
was confirmed by further work by many authors. Much of that work, as well as the original result
of Hawking, was partially heuristic but considerable effort has been made over the nearly 50 years
since Hawking’s work to obtain rigorous results.
A first such rigorous result is that [77] of Fredenhagen and Haag who showed for the (for
simplicity, massless) Klein Gordon equation (2) and for a model of a stationary detector far from
a spherical star collapsing to a black hole, and assuming that the quantum field reflects off the
surface of the collapsing star, that the “counting rate” of the detector will be consistent with
Hawking’s prediction as a consequence of the requirement that the two point function of the state
will have the right scaling limit (a weaker condition than the Hadamard condition) at the sphere
where the surface of the collapsing star crosses the Schwarzschild radius.
It is also, however, desirable to show, in a scattering theory sense, that an in vacuum state
(say viewed as a state on a comparison copy of Minkowski space, or, in the case of (conformally
coupled) massless fields at I −17 evolves into an out state (say viewed as a state in a comparison
17
Here I + and I − – pronounced scri plus and scri minus1– refer to future and past null infinity in the sense of
Penrose’s conformal compactification of asymptotically flat spacetimes (see e.g. [78]). In the case of massive fields, the
notions of I + and I − are of little use since, when multiplied by the appropriate conformal factors, classical solutions
won’t have finite restrictions to them, but one may instead consider the inverse action of (Dollard-modified) in and out

17
 copy of Minkowski space, or, in the case of massless fields, at I + ) which is identical with the state
of thermal radiation predicted by Hawking.
Let us remark here that one expects that the assumption of an in vacuum state in the Hawking
effect is actually not so important and that, e.g. a multiparticle state built on the in vacuum will
lead to similar behaviour at late times. This is because one expects the Hawking radiation from
such an in state will differ from the Hawking radiation from the in vacuum by transients which
disappear completely at “future infinity”. This point was already made by Wald in [79] and is also
implicit in the Fredenhagen-Haag result mentioned above since one would expect multiparticle
in states (with suitable smoothness properties) to have the same scaling limit for their two-point
functions as the in vacuum. Another point (also made in [77] and, earlier by Leahy and Unruh
in [80]) is that for an interacting quantum field theory there is no reason to expect the Hawking
radiation due to stellar collapse to be exactly thermal (i.e. exactly a Gibbs state on modes which
emanate from the black hole). This is to be contrasted with the Hartle-Hawking-Israel state (see
below) on exterior Schwarzschild whose Gibbs-state nature is expected to generalize to interacting
fields.
For the case of a massive or massless Klein Gordon field and spherical collapse, a first step
was taken towards a scattering theoretic understanding of the Hawking effect by Dimock and
Kay [70] who, after developing a suitable classical scattering theory on exterior Schwarzschild,
used that, i.a., to give a rigorous definition of the Unruh state. (In the massive case there was an
assumption of asymptotic completeness which has, however, since been proven [81] by Bachelot. In
the massless case asymptotic completeness was proven by Dimock in [82].) In the case of massless
fields, this is an important state on exterior Schwarzschild introduced in (Unruh, 1976) which is
an in vacuum at the past horizon (with respect to the affine parameter of its null generators18
as well as at I − and which Unruh argued should coincide with the state of Hawking radiation
on the exterior of a collapsing star on I + (when the I + of the spacetime of the collapsing star
is identified with the I + of exterior Schwarzschild). Dimock and Kay proved in both massive
and massless cases that their rigorously defined Unruh state at late times, viewed as a state for
the quantized Equation (2) on a comparison copy of Minkowski space19 is indeed the state of
Hawking radiation that one would expect – namely (see [70] for details) the state of black body
radiation from a black body with an ‘absorption operator’ defined appropriately in terms of the
relevant classical wave operators.20 However, as was written at the end of the introduction of
[70], “a complete treatment of the Hawking effect would [also] require a parallel discussion of the
scattering problem for a generic collapsing star and a discussion of the precise sense in which the
Unruh state approximates the in vacuum state for the spacetime of a collapsing star.”
Complete treatments of the Hawking effect have since been given by proving directly on the
spacetimes of models for a collapsing star (i.e. without consideration of the sense in which the
Unruh state approximates the in vacuum state for the spacetime of a collapsing star) that the in
vacuum state coincides with the expected state of Hawking radiation either on I + (for massless
fields) or on the out auxiliary copy of Minkowski space (for massive or massless fields). The first
such result [83] was obtained by Bachelot for the massive or massless Klein Gordon field on a
spacetime describing a spherical collapse which is stationary in the past in which the field does
not penetrate the surface of the star but rather reflects off it. A similar result [84] for the charged
massive Dirac field was obtained by Melnyk.
More recently, Alford [85] (see also [86]) has given a treatment of the Hawking effect for the
massless Klein-Gordon equation on the spacetime of the Oppenheimer-Snyder solution, treating
the case where the quantum field penetrates the region occupied by the collapsing dust ball as well
wave operators which map full solutions to solutions of a comparison Klein-Gordon equation on Minkowski space. See
around the brief mention of the relation between in and out fields and classical wave operators in Section 4.
18
which is the same thing as a thermal state at the Hawking temperature with respect to Lorentz boosts – see the
discussion below about the Unruh effect and see (Kay and Wald, 1991).
19
In the scattering theory of [70], there are actually two comparison systems. There is the ordinary, 4-dimensional
Klein-Gordon equation on 4-dimensional Minkowski space to capture the behaviour at large distances and early or late
times and there is the 2-dimensional massless wave equation on a 2-dimensional Minkowski space producted with a
2-sphere to capture early and late-time behaviour near the Schwarzschild radius. In the main text here, we are referring
to the former.
20
In showing this, [70] essentially give an appropriate rigorous notion of the ‘grey body’ factors mentioned earlier.

18
 as the case where it reflects off its surface. Even more recently, in [87, 88], Alford has extended
the treatment (still of the chargeless massless Klein-Gordon equation) in the reflective case to
more general collapsing spherical dust clouds which may themselves be electically charged and
which collapse to both subextremal and extremal Reissner Nordström black holes. Amongst other
results, he proves a mathematical result that confirms that, as would be expected, in the extremal
case, the Hawking radiation is transient.
Finally, let us mention that there has been a lot more work on the scattering (and related
properties) of classical linear waves on black hole backgrounds which, while it does not explicitly
explore quantum corollaries that may be derived from the classical results (see Footnote 4) is no
doubt susceptible to interesting such corollaries. See for example [89, 90, 91, 92].
The Hawking black-hole radiation result suggests that there is something fundamentally “ther-
mal” about quantum fields on black-hole backgrounds and our understanding of this is deepened
by consideration of quantum fields, not on the spacetime of stellar collapse, but on the maxi-
mally extended Schwarzschild spacetime (a.k.a. the Kruskal or Kruskal-Szekeres spacetime see
Figure 5). In particular, the theorems in the two papers (Kay and Wald, 1991) and (Kay, 1993),
combined together, tell us that there is a unique state on the Weyl algebra for Kruskal which is
invariant under the Schwarzschild isometry group and whose two-point function has Hadamard
form. Moreover, they tell us that this state, when restricted to a single wedge (i.e. the exterior
Schwarzschild spacetime) is necessarily a KMS state at the Hawking temperature.

future singularity (Schwarzschild case)

exterior Schwarzschild wedge

/Rindler wedge

past singularity (Schwarzschild case)

Figure 5: The geometry of maximally extended Schwarzschild (/or Minkowski) spacetime. In the
Schwarzschild case, every point represents a two-sphere (/in the Minkowski case, a two-plane). The curves
with arrows on them indicate the Schwarzschild time-evolution (/one-parameter family of Lorentz boosts).
These curves include the (straight lines at right angles) event horizons (/Killing horizons).

This unique state is known as the Hartle-Hawking-Israel state (HHI state) These results in
fact apply more generally to a wide class of globally hyperbolic spacetimes with bifurcate Killing
horizons including de Sitter space – where the unique state is sometimes called the Euclidean
and sometimes the Bunch-Davies vacuum state (see e.g. the recent paper [93]) – as well as to
Minkowski space, in which case the unique state is the usual Minkowski vacuum state, the ana-
logue of the exterior Schwarzschild wedge is a so-called Rindler wedge, and the relevant isometry
group is a one-parameter family of wedge-preserving Lorentz boosts. In the latter situation, the
fact that the Minkowski vacuum state is a KMS state (at “temperature” 1/2π) when restricted
to a Rindler wedge and regarded with respect to the time-evolution consisting of the wedge-
preserving one-parameter family of Lorentz boosts is known as the Unruh effect (Unruh, 1976).
See [94] for a review. This latter property of the Minkowski vacuum in fact generalizes to general
Wightman quantum field theories and is in fact, as first pointed out by Sewell [95, 96], an imme-
diate consequence of a combination of the Reeh Schlieder Theorem (applied to a Rindler wedge)

19
and the Bisognano Wichmann Theorem. The latter theorem says that the defining relation (1)
of a KMS state holds if, in (1), we identify the operator J with the complex conjugation which
implements wedge reflection and H with the self-adjoint generator of the unitary implementor of
Lorentz boosts. We remark that the Unruh effect illustrates how the concept of “vacuum” (when
meaningful at all) is dependent on the choice of time-evolution under consideration. Thus the
usual Minkowski vacuum is a ground state with respect to the usual Minkowski time-evolution
but not (when restricted to a Rindler wedge) with respect to a one-parameter family of Lorentz
boosts; with respect to these, it is, instead, a KMS state.
Let us also mention that illuminating heuristic arguments for the existence of the Hartle-
Hawking-Israel state, even in the case of interacting field theories, based on formal Euclidean
path integrals and Wick rotation were given by Jacobson in [97].
A first rigorous construction of the Hartle-Hawking-Israel state on the Kruskal spacetime for
the Klein-Gordon model (2) and the proof of some of its properties was first carried out in [31].
However that construction was not for the full Kruskal spacetime, but rather only for the double
exterior Schwarzschild wedge. In terms of symplectically smeared sharp-time fields on a Cauchy
surface through the bifurcation 2-sphere, the smearing functions were bounded away from the
bifurcation 2-sphere. The removal of this restriction and the thereby the full construction on the
entire Kruskal spacetime, and, more generally, on other spacetimes with static bifurcate Killing
horizons, was first carried out by Sanders [98] based on Euclidean methods and Wick rotation.
Sanders also showed that the thus constructed HHI states are Hadamard. Also using Euclidean
methods and Wick rotation but with a technically different approach based on Calderón projectors,
Gérard [99] has more recently obtained similar results where staticity is relaxed to stationarity.
However, the results in [99] require the bifurcation surface to be compact.
As pointed out in [99], the latter result excludes the Kerr spacetime since the exterior wedge
of that is of course not globally stationary. In fact, it was argued in (Kay and Wald, 1991) that
because of superradiance – or relatedly, because the Killing vector which is null on the horizon
is spacelike well away from the horizon (see also [100]) for the Klein-Gordon equation, (2), There
does not exist any isometry-invariant Hadamard state on (the appropriate globally hyperbolic part
of ) Kerr and there is also a similar no-go result in (Kay and Wald, 1991) for spacetimes with
bifurcate Killing horizons such as Schwarzschild-deSitter which include triple wedges. Another
no-go result of this type was argued for by Kay and Lupo [101], namely that there is no stationary
Hadamard state if one removes from the Kruskal spacetime the part of the right wedge (see Figure
5) to the right of a surface of constant Schwarzschild radius and imposes, say, Dirichlet boundary
conditions on that surface – unless one removes a similar surface and imposes similar boundary
conditions on a ‘mirror surface’ in the left wedge. This raises interesing questions about what
could be the right idealization of a ‘black hole in a (spherical) box’ which plays an important rôle
in black hole thermodynamics [102].

8 More about (Classical and) Quantum Fields on Black

Hole Backgrounds
A construction has been given [103] for a Hartle-Hawking-Israel-like state (as well as several other
states including an Unruh-like state) for the Dirac equation in a Kerr background and several of
its properties obtained. But while some indications were obtained, it seems to still be unknown
whether or not that is everywhere a Hadamard state. (We remark here in relation to this that
the no-go result in (Kay and Wald, 1991) that there is no isometry-invariant Hadamard state
for the Klein-Gordon equation on Kerr may well not generalize to Dirac fields since there is no
superradiance for Dirac fields.)
Concerning the rigorous construction of the Unruh state, as far as I am aware, nothing further
has been done for the massive Klein-Gordon equation on Schwarzschild since the work of Dimock
and Kay – as supplemented by the asymptotic completeness result of Bachelot – as we mentioned
earlier. In particular, while one expects it to be Hadamard, this seems not to have been established.
(The situation is different, however, for Schwarzschild-de Sitter – and its electrically charged and
rotating generalizations – see below.) However, for the massless case, a construction has been
given [104] which extends to the full Kruskal spacetime and the resulting Unruh state is shown

20
 there to be Hadamard. See also [105] for a computation of the two-point function in this state,
especially for pairs of points inside the horizon. A further work [106] obtains a similar result for
the massless Dirac equation on slowly rotating Kerr based on the classical scattering theory of
[91].
A recent development of great interest concerns the behaviour of quantum fields around the
Cauchy horizons of spacetimes such as Reissner Nordström-de Sitter and Kerr-de Sitter. For
ordinary Reissner-Nordström and Kerr, as has been much discussed since Penrose first pointed it
out in [107] in 1968 (see also Simpson and Penrose [108] and Hawking and Ellis [78] as well as [109])
one expects that the classical Einstein equations are such that, due to a blue-shift effect, such
Cauchy horizons will be unstable against small perturbations, say of initial data in the relevant
initially globally hyperbolic region and this leads one to expect/hope that the maximal Cauchy
development of generic initial data close to that of exact Reissner Nordström or Kerr will be
inextendible in a suitable sense – consistently with the strong version [110] of Penrose’s celebrated
cosmic censorship conjecture. (This may be taken to be the [somewhat vague] statement that
the maximal Cauchy development of a generic initial data set should be a [globally hyperbolic]
spacetime which is inextendible in a physically suitable sense. For a recent survey of cosmic
censorship in all its versions, see [111]) For what is known about this rather complicated and subtle
question, we refer to the introduction to the paper [112] of Dafermos and Shlapentokh-Rothman,
the bottom line being, as one may read there, that, satisfactorily, such an inextendibility result
does hold, in a particular sense proposed by Christodoulou [113], which entails that the spacetime
cannot be extended as a weak solution to Einstein’s equations.21 Relatedly, and, say, taking
Equation (2) as a proxy for more general perturbations, if one does QFT for Equation (2) on such
spacetimes, then for a suitable notion of a general initial state, one expects the expectation value
of the stress-energy tensor to diverge at the Cauchy horizon22 sufficiently strongly as to lead one
to expect that any solution to the semiclassical Einstein equations with such initial data would,
again, be suitably inextendible beyond a maximal globally hyperbolic region. Such a QFTCST
divergence result has a long history – let us just mention here the recent papers [114] for Reissner-
Nordström and [115] for Kerr. There is also other evidence that QFTCST effects cause Cauchy
horizons to be unstable from the study of the behaviour of gedanken particle detectors which
move across such horizons. See e.g. [116, 117]. But in view of the classical results, we arguably
don’t need to rely on any sort of QFTCST results to conclude that strong cosmic censorship is
saved for these spacetimes.
However, for Reissner Nordström-de Sitter and Kerr-de Sitter, the black hole blue shift effect
(now for the Einstein equations with cosmological constant) is counteracted by a cosmological
red-shift effect and, classically, it is known, after much work by many authors (see especially
[118]) again for the ‘proxy’ linear Equation (2) on such backgrounds that (at least for many values
for the cosmological constant, the black hole mass and the scalar field mass) the singularity at the
Cauchy horizon is too weak for the spacetime to be inextendible in the sense of Christodoulou –
thus seemingly insufficient to save strong cosmic censorship.
In the same paper [112] by Dafermos and Shlapentokh-Rotman recommended above, a sug-
gestion is made for a possible way out of this seeming impasse, staying within classical general
relativity, by taking a suitable view as to which class of initial data is the right physical class – in
particular by taking a class of initial data which includes data that is suitably rough. However,
it is of interest to explore whether quantum theory might save strong cosmic censorship, without
having to get involved with debates about the right notion of roughness of classical initial data.
To this end, Hollands, Wald and Zahn [119] and Hollands, Klein and Zahn [120] have studied the
expectation value of the quantum stress energy tensor again for the ‘proxy’ quantum Equation (2)
near the Cauchy horizon for all initially Hadamard states on Reissner Nordström-de Sitter space-
time and found, indeed that (generically in parameters – see Footnote 23) this will blow-up as one
approaches the Cauchy horizon strongly enough23 to lead one to expect that a solution to semi-
21
Christodoulou’s proposal is to declare a spacetime to be extendible if it is extendible as a manifold and a chart may
be found around each point of the extended manifold such that the metric components have continuous extensions and
the Christoffel symbols are locally square integrable.
22
This is stronger than just a failure of the two-point function of the initial state to extend as a Hadamard two-point
1
function beyond the Cauchy horizon and corresponds to failure to be extendible in Hloc .
23
In suitable double-null Kruskal coordinates, (U, V, θ, ϕ), adapated to the Cauchy horizon (see [119]) they show

21
 classical gravity would convert the Cauchy horizon into a singularity through which the spacetime
could not be extended as a (weak) solution of the semiclassical Einstein equation and moreover
be such that a physical object attempting to cross the horizon would be infinitely stretched or
crushed.
We remark that an important step in the work in [119, 120] involves the construction of an
Unruh-like state on a suitable region of Reissner-Nordström-de Sitter for massive or massless Klein
Gordon. (The rôle this plays is that then the two-point function in the generic initial Hadamard
state they consider will differ from the two-point function in the Unruh state by a smooth two-
point function.) This has antecedents both in the work [121] of Marković and Unruh and also
in the earlier work of Kay [122] and Hollands [123] towards a semi-local vacuum notion for the
neighbourhood of a point p based on the imposition of a suitable vacuum-like condition on the
boundary of the future light-cone of p. In [119], this semi-local vacuum construction is developed
further and adapted to the region of Reissner-Nordström-de Sitter bounded in the past by the
past black hole horizon and the past cosmological horizon. Recently, Klein [124] has carried out
an analogous construction of an Unruh-like state for Kerr-de Sitter – an important step towards
establishing the (expected) instability of the Kerr-de Sitter Cauchy horizon.
In conclusion, there seems now to be increasing evidence that, for de Sitter black holes,
QFTCST has an important rôle to play in arguing that strong cosmic censorship is not violated.

9 Non-Globally Hyperbolic Spacetimes and the “Time Ma-

chine” Question
In (Hawking, 1992) it is argued that a spacetime in which a time-machine gets manufactured
should be modelled (see Figure 6) by a spacetime with an initial globally hyperbolic region with
a region containing closed timelike curves to its future and such that the future boundary of the
globally hyperbolic region is a compactly generated Cauchy horizon (see (Hawking, 1992) for the
definition). For such a spacetime, (Kay, Radzikowski and Wald, 1997) use microlocal methods
and in particular certain propagation of singularity theorems to prove that it is impossible for
any distributional bisolution which satisfies (even a certain weakened version of) the Hadamard
condition on the initial globally hyperbolic region to continue to satisfy that condition on the full
spacetime – the (weakened) Hadamard condition being necessarily violated at at least one point of
the Cauchy horizon. This result implies that, however one extends a state, satisfying our conditions
(a), (b), (c) and (d), on the minimal algebra for Equation (2) on the initial globally hyperbolic
region, the expectation value of its stress-energy tensor must necessarily become singular, or rather
undefinable, on the Cauchy horizon. This result, together with many heuristic results and specific
examples considered by many other authors was argued in (Kay, Radzikowski and Wald, 1997) to
support the validity of the (Hawking, 1992) chronology protection conjecture according to which
it is impossible in principle to manufacture a time machine. For example, in (for simplicity, two-
dimensional) Misner space24 one finds (see e.g. [125]) that, generically25 the expectation value
of TV V , where V (see again [125]) is a coordinate26 that is regular across the Cauchy horizon,
will behave like C/V 2 for a nonzero constant C. However, the value of C depends on the state,
and, as pointed out by Krasnikov [126] there are states for which the value of C is zero! This is
qualitatively different from the behaviour, discussed in Section 8, of the expectation value of the
stress-energy tensor in initially Hadamard states near the Cauchy horizon of (1+3-dimensional)

that, in any of their states and generically in parameters by which we mean for almost all values of the cosmological
constant, the black hole mass and charge and the scalar field mass, ω(TV V ) blows up like C/V 2 .
24
Misner space is a two-dimensional locally flat spacetime with the topology of a cylinder which is the disjoint union
of a lower half-cylinder, which is an initially globally hyperbolic region, an upper half-cylinder which has closed timelike
curves and a compactly generated Cauchy horizon in between these two halves which consists of a null circle – see e.g.
Hawking and Ellis [78].
25
Note that here the genericity is in the state, in contrast to the situation for Reissner Nordström-de Sitter Cauchy
horizons, where – see Footnote 23 in Section 8 – it was in the parameters – i.e. the cosmological constant, the black
hole mass and charge and the field mass.
26
The V coordinate can locally can be identified with the V of standard double null coordinates, (U, V ), in Minkowski
space such that the metric is dU dV .

22
Reissner-Nordström-de Sitter where (see Footnote 25) as was shown in [119], the constant, C, in
the C/V 2 behaviour near the horizon (which holds except for certain discrete combinations of the
parameters of that problem – i.e. the cosmological constant, black hole mass and charge and field
mass) is the same for all initial Hadamard states!
The above-mentioned difference between the Reissner-Nordström-de Sitter Cauchy horizon
story and the Misner-space Cauchy horizon story seems to help us to appreciate an important
loophole, first argued for by Visser in [127] (see also (Visser, 2003)), in the physical interpretation
of the Cauchy horizon quantum stress-energy result in the latter case. Visser argues that the
semiclassical approximation to quantum gravity might well be expected to be a bad approximation
within a Planck distance or so (he gives a suitable definition for this notion in the above-cited
papers) of the Cauchy horizon and yet, say in the Krasnikov state, the expectation value of the
stress-energy tensor is only predicted to be singular/undefinable on the Cauchy horizon, while it
is bounded (in fact zero!) everywhere else (to the past of the Cauchy horizon). Instead, Visser
suggests that, before the Cauchy horizon is reached, there may well be big quantum fluctuations
in the geometry which would render any semiclassical approximation inapplicable and could only
be understood in full quantum gravity and, for all we know, quantum gravity will effectively allow
the Cauchy horizon to be traversed and thereby allow time travel to be possible and Hawking’s
Chronology Protection Conjecture to be violated.
To summarize: While the singularity/ill-definedness of the expectation value of the stress-
energy tensor on the Cauchy horizon indicates a breakdown of semiclassical gravity there, that
doesn’t necessarily mean that physics breaks down there because semiclassical gravity may cease to
be a good approximation to physics before one reaches the Cauchy horizon. Therefore the question
of whether or not chronology is protected is still open.

... Region with closed timelike curves ...

.......... .
... ...............
.... .
........ . ....
.... ..
..... ......... .
Cauchy horizon ! ......
...... .......
.
........ ..............
.
Initial globally hyperbolic region

Figure 6: The schematic geometry of a spacetime in which a time-machine gets manufactured.

10 More about QFT on Non-Globally Hyperbolic Space-

times
In 1992, Kay [128] proposed the F-locality condition as a possible condition one might wish to
impose on any net of local ∗-algebras on any (non-globally hyperbolic) spacetime. Namely that
every neighbourhood of every point should contain a globally hyperbolic subneighbourhood of
that point for which the algebra coincides with the algebra one would obtain were one to regard
the subneighbourhood as a spacetime in its own right and quantize – with some choice of time-
orientation – according to the standard rules for quantum field theory on globally hyperbolic
spacetimes.
Recently Janssen [129] has introduced an interesting general class of (not necessarily globally
hyperbolic) spacetimes which he calls semi-globally hyperbolic spacetimes and for some of these, he
has proposed a scheme for extending the usual construction of Equation (2) on globally hyperbolic
spacetimes to spacetimes in this class with a resulting (minimal-like) net of local ∗-algebras which,

23
 inter alia, satisfies (a stronger condition than) the F-locality property.27
Moreover, Janssen discusses a generalization of his construction to the (semi-globally hyper-
bolic) spacetime – whose Penrose diagram first appeared in (Hawking, 1975) and is reproduced
here, in the diagram 7 – which is assumed to be the best spacetime picture we can have, at a
semiclassical level of description, for a (spherical) evaporating black hole. (A relevant recent paper
with many references is [130].)

r=0 I+
PE
r=0
n

r=0
I−

Figure 7: The schematic geometry of a (spherically symmetric) spacetime of black-hole evaporation.

Even though there are no closed timelike curves in this spacetime, it is nevertheless a non-
globally hyperbolic spacetime since the point marked ‘n’ on Figure 7 is a naked singularity (strictly
this point is not in the spacetime manifold) and the dashed line to its future is a Cauchy horizon.
We shall call the region labelled ‘PE’ in Figure 7 the post evaporation region. The complement of
PE in this spacetime is easily seen to have Cauchy surfaces and therefore be globally hyperbolic;
we shall call it the initial globally hyperbolic region.
Janssen obtains a net of local ∗-algebras for the QFT of Equation (2) on this spacetime.
However he leaves as an open question whether it admits any states. As explained in his paper,
there is a corresponding “Weyl-like” algebra which will admit states, but it seems to be unknown
at present whether any of these states will satisfy the necessary regularity [see Section 2] for any
of their n-point functions to exist. Thus one cannot even begin to talk about Hadamard states
since of course the existence of states with 2-point functions is a prerequisite for even saying what
a Hadamard state would be!
The spacetime of Figure 7 is of importance since at least one version28 of the black hole
information loss puzzle [132, 133, 134, 135] is based on the seeming implication of this diagram
that some of the information contained in an initial state on the initial globally hyperbolic region
will be lost on the future singularity (the thick horizontal black line labelled r = 0 in Figure 7)
in the sense that the state on the initial globally hyperbolic region will not be retrodictable from
knowledge of the state on the post-evaporation region in Figure 7. And it is claimed to be difficult
to imagine how this state of affairs could fail to persist in a suitable sense in full quantum gravity.
In a separate development, Juárez-Aubry [136] has very recently put forward the conjecture
that, for any linear field theory (or interacting field theory when treated perturbatively) any
extension of any pure Hadamard state on the initial globally hyperbolic region of the spacetime
of Figure 7) to any F-local net of ∗-algebras for the full spacetime will fail to be Hadamard on the
27
Interestingly, Janssen arrives at this ∗-algebra construction by first constructing a patchwork of one-particle Hilbert
space structures (see Footnote 8). At first sight, this may seem strange since one traditionally thinks of a one-particle
Hilbert space structure on a symplectic space of solutions as an auxiliary structure that help to define certain (quasi-
free) states on the ∗-algebra associated with that symplectic space. Thus, in a sense, Janssen’s construction gives
primacy to states over field operators, in a reversal of the familiar story for globally hyperbolic spacetimes where the
construction of a ∗-algebra has primacy over the definition of states.
28
The expression ‘black hole information loss puzzle’ is also sometimes taken to refer to the distinct but closely related
puzzle called in the most recent version of [131] ‘the second law puzzle in the case of black holes’.

24
 full spacetime.29 (As Juárez-Aubry also points out, this conjecture, if true, is reminiscent of the
result of Fewster and Verch in [39] that a pure Hadamard state on a double cone in Minkowski
space must fail to have a Hadamard extension beyond the double cone.) He further conjectures
that, in fact, the state will fail to be Hadamard on the Cauchy horizon sufficiently severely that
the stress energy tensor will not be definable there – rather as happens for the time-machine
Cauchy horizons we discussed in Section 10.
Juárez-Aubry then points out that, if these conjectures hold, then one could conclude that
semiclassical gravity would break down on the Cauchy horizon. Thus it would not be true that
the entire Figure 7 depicts a solution to semiclassical gravity. To have a semiclassical solution,
one would e.g. have to remove the PE region from Figure 7. (See Figure 2 in [136].) But there
is no information loss puzzle associated with the resulting truncated figure! So, assuming his
conjectures are true, we could say that it had actually never been true that any sort of information
loss puzzle was ever suggested by any solution to semiclassical gravity appropriate to black hole
evaporation.
This is of course not to say that such a semiclassical solution, with the PE region excised,
would be physically correct. If it were, it would seem to mean that spacetime comes to an end
at the (retarded) final moment of black hole evaporation! Rather, and somewhat similarly to
Visser’s conclusions in the case of the time-machine Cauchy horizons discussed in Section 10, it
seems reasonable to assume that quantum gravity would cause both the black hole singularity
and the Cauchy horizon to be resolved and, insofar as it is representable by a classical spacetime
picture, the solution to quantum gravity to again include a post-evaporation region. But there is
no reason to assume that this spacetime would resemble Figure 7. Instead it is argued in [136]
that it is more reasonable to assume it would be globally hyperbolic (as depicted in Figure 3 in
[136]). It is also pointed out in [136] that the conjectured ill-definedness of the stress-energy tensor
on the Cauchy horizon corresponds to the ‘thunderbolt’ singularity argued for by Hawking and
Stewart in [137]. Hawking and Stewart suggest in that paper that, rather than spacetime coming
to an end, quantum gravitational effects “would soften the thunderbolt singularity into a burst
of high energy particles” at the endpoint of black hole evaporation. Such a physical expectation
would, as is pointed out in [136] seem to be consistent with the globally hyperbolic spacetime
depicted in Figure 3 in [136].

11 Other Related Topics and Some Warnings

There is a vast computational literature, calculating the expectation values of stress-energy tensors
in states of interest for scalar and higher spin linear fields (and also some work for interacting
fields) on interesting cosmological and black-hole backgrounds. For an early survey, see (Birrell
and Davies, 1982). For a few recent papers on computational and approximation methods, see
e.g. [138, 139, 140] as well, e.g. as recent papers such as [114, 119, 120] where (see Section 8) a
computation of an expectation value of a renormalized stress-energy tensor forms forms part of a
specific investigation.
Quantum field theory on de Sitter space is a big subject area in its own right with many
interesting features, while quantum field theory on anti-de Sitter space is of sustained interest
because of its relevance to string theory and holography. Also important on black hole backgrounds
is the calculation of grey-body factors, again with renewed interest because of relevance to string
theory and to brane-world scenarios.
Further to the topics discussed in Section 6, there are many further mathematically rigorous
results on algebraic and axiomatic quantum field theory in a curved spacetime setting, including
versions of the PCT (see [143]), Spin-Statistics (see Verch’s paper, [141], which we mentioned
above and also [142]) Reeh-Schlieder theorems (see [144, 145] and also the very recent paper
[146] which also generalizes certain theorems of Araki and Borchers) as well as the split property
[147, 148].
29
A variation of this conjecture, which doesn’t make any assumptions about the nature of any field algebra beyond
the Cauchy horizon, and in particular therefore does not need to mention the F-locality condition, would be that, say
for Equation (2), the two point function of any Hadamard state in the initial region will fail to have any Hadamard
extension to beyond the Cauchy horizon.

25
 There are also rigorous energy inequalities bounding the extent to which expected energy
densities etc. can be negative. For more about this extensive topic see [149, 150]. See also [151].
Perturbative renormalization theory of interacting QFT (including gauge theories – see e.g.
[152]) in curved spacetime is also now a highly developed subject, which, in the context of AQFT,
is known as pAQFT. For more information about pAQFT, see the article [153] by Fredenhagen and
Rejzner and the book [154] by Rejzner as well as the article [155] by Rejzner in this encyclopedia.
pAQFT not only generalises what was previously understood about perturbative renormalization
in Minkowski space to curved spacetimes but has also achieved a new understanding (see e.g. the
book [156] by Dütsch) as to how pAQFT can be understood as arising from a deformation of
a perturbative treatment of classical field theory. Moreover (see [157]) it has been shown how
pAQFT when combined with ideas from locally covariant QFT (see Section 6) leads to a new
understanding of how perturbative QFT can be consistent with background independence and
this, in turn, has led (see [158, 159]) to a new background-independent perturbative approach to
quantum gravity30 , which is also known as a relational (and gauge-invariant) version of pertur-
bation theory. For an illustration of its power, and fuller references to earlier work, see e.g. the
recent papers [160, 161] by Lima and by Fröb, Rein and Verch. Treating perturbative quantum
gravity as an effective field theory, the former computes quantum-gravitational one-loop correc-
tions to a gauge-invariant observable measuring the space-time local expansion rate in slow-roll
inflation thereby contributing to the resolution of some earlier controversies and ambiguities; the
latter computes quantum gravitational corrections to the Newtonian potential, giving, for the first
time, unambiguous answers for the appropriate generalization to non-trivial backgrounds. (See
also [162].)
Let us also mention here the construction of scalar quantum fields with polynomial interactions
in 2-dimensional de Sitter space by Barata, Jäkel and Mund – see [163] and references therein.
As one may read there, this work makes interesting and novel use of the notion of one-particle
structures (see Footnote 8) and of Tomita-Takesaki modular theory and also introduces and uses
a new version of the Osterwalder-Schrader reconstruction theorem.
Beyond quantum field theory in a fixed curved spacetime is semiclassical gravity which takes
into account the back reaction of the expectation value of the stress-energy tensor on the classical
gravitational background and which we have referred to in several places in this article. It is
beyond the scope of this article to review the subject. But we would mention the review article,
[164] of Ford and the book [165] of Hu and Verdaguer. Aside from [64] which we referred to in
Section 3, an incomplete, but we hope useful, list of recent references is [166, 167, 168, 169, 170,
171, 172, 173, 174, 175].
Readers exploring the wider literature, or doing further research on the subject should be
aware that the word “vacuum” is sometimes used to mean “ground state” and sometimes just
to mean (pure) “quasi-free state”. Furthermore they should be cautious of attempts to define
particles on Cauchy surfaces in instantaneous diagonalization schemes (cf. the remarks at the
end of Section 4). When studying (or performing) calculations of the “expectation value of
the stress-energy tensor” it is always important to ask oneself with respect to which state the
expectation value is being taken. It is also important to remember to check that candidate two-
point (anticommutator) functions satisfy the positivity condition (c) of Section 2. Typically two-
point distributions obtained via mode sums automatically satisfy Condition (c) (and Condition
(d)), but those obtained via image methods don’t always satisfy it. (When they don’t, the
presence of non-local spacelike singularities is often a tell-tale sign as can be inferred from Kay’s
Conjecture/Radzikowski’s Theorem discussed in Section 2.) There are a number of apparent
implicit assertions in the literature that some such two-point functions arise from “states” when
of course they can’t. Some of these concern proposed analogues to the Hartle-Hawking-Israel
state for the (appropriate maximal globally hyperbolic portion of the maximally extended) Kerr
spacetime. That they can’t belong to states is clear from the theorem in (Kay and Wald, 1991) (see
Section 7) which states that there is no isometry-invariant Hadamard state on this spacetime at all.
Others of them concern claimed “states” on spacetimes such as those discussed in Section 9 which,
if they really were states would seem to be in conflict with the chronology protection conjecture.
30
As pointed out in [159], this work in some respects supercedes other approaches to (classical and quantum, gauge
invariant, nonlinear) cosmological perturbation theory.

26
(In the absence of a full understanding of quantum gravity, it is of course still unknown whether
or not that conjecture holds, but, as far as I am aware, there are still no known violations of it
which are semiclassically describable.) Finally, beware states (such as the so-called α-vacua of de
Sitter spacetime) whose two-point distributions violate the “Hadamard” Condition (d) of Section
2 and which therefore do not have a well-defined finite expectation value for the renormalized
stress-energy tensor. As we discussed in Section 3 the same is true of the (unmodified) S-J states.

See also
Black hole thermodynamics. ***Is there still an article on that topic?*** Algebraic ap-
proach to quantum field theory. The operator product expansion in quantum field
theory in curved spacetimes. Measurement in quantum field theory. Microlocal anal-
ysis in quantum field theory. Perturbative algebraic quantum field theory. Scattering,
asymptotic completeness and bound states. Tomita–Takesaki modular theory.

Acknowledgments
I thank Chris Fewster, Atsushi Higuchi and Benito Juárez-Aubry for many suggestions and Klaas
Landsman for commenting on an earlier version.

Further Reading
DeWitt BS (1975) Quantum field theory in curved space-time. Physics Reports 19, No. 6,
295-357.
Birrell ND and Davies PCW (1982) Quantum Fields in Curved Space. Cambridge: Cambridge
University Press.
Haag R (2012) Local Quantum Physics. Berlin: Springer.
Dimock J (1980) Algebras of local observables on a manifold. Communications in Mathematical
Physics 77: 219-228.
Brunetti R, Fredenhagen K and Verch R (2003) The generally covariant locality principle – A new
paradigm for local quantum physics. Communications in Mathematical Physics 237:31-68.
Kay, BS (2000) Application of linear hyperbolic PDE to linear quantum fields in curved space-
times: especially black holes, time machines and a new semi-local vacuum concept. Journées
Équations aux Dérivées Partielles, Nantes, 5-9 juin 2000 GDR 1151 (CNRS): IX1-IX19.
(Also available at [Link] or as gr-qc/0103056.)
Wald RM (1978) Trace anomaly of a conformally invariant quantum field in a curved spacetime.
Physical Review D17: 1477-1484.
Kay BS (1979) Casimir effect in quantum field theory. (Original title: The Casimir effect without
magic.) Physical Review D20: 3052-3062.
Hawking SW (1975) Particle Creation by Black Holes. Communications in Mathematical Physics
43: 199-220.
Kay BS and Wald RM (1991) Theorems on the uniqueness and thermal properties of stationary,
nonsingular, quasifree states on spacetimes with a bifurcate Killing horizon. Physics Reports
207, No. 2, 49-136.
Kay BS (1993) Sufficient conditions for quasifree states and an improved uniqueness theorem for
quantum fields on space-times with horizons. Journal of Mathematical Physics 34: 4519-
4539.
Wald RM (1994) Quantum Field Theory in Curved Spacetime and Black Hole Thermodynamics.
Chicago: University of Chicago Press.
Hartle JB and Hawking SW (1976) Path-integral derivation of black-hole radiance. Physical
Review D13: 2188-2203.

27
Israel W (1976) Thermo-field dynamics of black holes. Physics Letters A 57: 107-110.
Unruh W (1976) Notes on black hole evaporation. Physical Review D14: 870-892.
Hawking SW (1992) The Chronology Protection Conjecture. Physical Review D46: 603-611.
Kay BS, Radzikowski MJ and Wald RM (1997) Quantum field theory on spacetimes with a
compactly generated Cauchy horizon. Communications in Mathematical Physics 183: 533-
556.
Visser M (2003) The quantum physics of chronology protection. In: Gibbons GW, Shellard EPS
and Rankin SJ The Future of Theoretical Physics and Cosmology. Cambridge: Cambridge
University Press.
Parker L and Toms D (2009) Quantum Field Theory In Curved Spacetime Cambridge: Cambridge
University Press.
Mukhanov V and Winitzki S (2007) Introduction to Quantum Effects in Gravity. Cambridge:
Cambridge University Press.
Brunetti R, Dappiaggi C, Fredenhagen K and Yngvason J eds. (2015) Advances in Algebraic
Quantum Field Theory. Cham: Springer International Publishing. [Link]
com/book/10.1007/978-3-319-21353-8

References
[1] Kay BS (2006) Quantum field theory in curved spacetime. In: Françoise J-P, Naber G
and Tsou TS (eds) Encyclopedia of Mathematical Physics. Amsterdam, New York, London:
Academic (Elsevier) 4: 202-212. [arXiv:gr-qc/0601008]
[2] Barbado LC, Báez-Camargo and Fuentes I (2021): Evolution of confined quantum scalar
fields in curved spacetime. Part II: Spacetimes with moving boundaries in any synchronous
gauge. European Physics Journal C 81: 953. [arXiv:1811.10507]
[3] Barbado LC, Báez-Camargo and Fuentes I (2020) Evolution of confined quantum scalar
fields in curved spacetime. Part I: Spacetimes without boundaries or with static boundaries
in a synchronous gauge. European Physical Journal C 80: 796. [arXiv:2106.14923]
[4] Kohlrus J, Louko J, Fuentes I and Bruschi DE (2018): Wigner phase of photonic helicity
states in the spacetime of the Earth. arXiv:1810.10502.
[5] Unruh WG (2008) Dumb holes: analogues for black holes. Philosophical Transactions of the
Royal Society A: Mathematical, Physical and Engineering Sciences 366: 2905-2913.
[6] Barcelo C, Liberati S and Visser M (2011) Analogue Gravity. Living Reviews in Relativity
14: 3. [Link]
[7] Weinfurtner S, Tedford EW, Penrice MCJ, Unruh WG, Lawrence GA (2011) Measure-
ment of stimulated Hawking emission in an analogue system. Phys. Rev. Lett. 106: 021302.
[arXiv:1008.1911]
[8] Drori J, Rosenberg Y, Bermudez D, Silberberg Y and Leonhardt U (2019) Observation of
Stimulated Hawking Radiation in an Optical Analogue. Phys. Rev. Lett. 122: 010404.
[9] Mertens L, Moghaddam AG, Chernyavsky D, Morice C, van den Brink J and van
Wezel J (2002) Thermalization by a synthetic horizon. Phys. Rev. Research 4: 043084.
[arXiv:2206.08041]
[10] Braunstein SL, Faizal M, Krauss LM, Marino F and Shah NA (2023) Analogue simula-
tions of quantum gravity with fluids. Nat. Rev. Phys. (2023) [Link]
s42254-023-00630-y
[11] Witten E (2022) Why does quantum field theory in curved spacetime make sense? And what
happens to the algebra of observables in the thermodynamic limit? In: Dialogues Between
Physics and Mathematics: CN Yang at 100. (Cham: Springer International Publishing)
pp. 241-284. [arXiv:2112.11614]

28
[12] Gibbons GW (1979) Quantum field theory in curved spacetime. In: Hawking SW, Israel
W (eds) (1979) General Relativity, An Einstein Centenary Survey Cambridge, Cambridge
University Press.
[13] Ford LH (1997) Quantum field theory in curved spacetime. arXiv: gr-qc/9707062
[14] Jacobson T (2005) Introduction to quantum fields in curved spacetime and the Hawking
effect. arXiv:gr-qc/0308048
[15] Wald RM (2012) The history and present status of quantum field theory in curved spacetime
In: Lehner C, Renn J and Schemmel M (eds) Einstein and the Changing Worldviews of
Physics. Einstein Studies, vol 12 Birkhäuser Boston. pp. 317-331. [arXiv:gr-qc/0608018]
[16] Hollands S and Wald RM (2015) Quantum fields in curved spacetime. Physics Reports 574:
1-35. [arXiv:1401.2026]
[17] Wald RM (2018) The formulation of quantum field theory in curved spacetime. In: Rowe
DE, Sauer T, Walter SA (eds) Beyond Einstein: Perspectives on Geometry, Gravitation, and
Cosmology in the Twentieth Century New York, Birkhäuser. pp. 439-449. [arXiv:0907.0416]
[18] Brunetti R and Fredenhagen K (2009) Quantum field theory on curved backgrounds. In: Bär
C and Fredenhagen K (eds) (2009) Quantum Field theory on Curved Spacetimes: Concepts
and Mathematical Foundations Berlin: Springer [arXiv:0901.2063]
[19] Bär C, Ginoux N and Pfäffle F (2007) Wave Equations on Lorentzian Manifolds and Quan-
tization European Mathematical Society. [arXiv:0806.1036]
[20] Fewster CJ and Verch R (2015) Algebraic quantum field theory in curved spacetimes. In:
(Brunetti et al., 2015) (see Further Reading) 125-189. [arXiv:1504.00586]
[21] Hack TP (2015) Cosmological applications of algebraic quantum field theory in curved space-
times Cham: Springer. [arXiv:1506.01869]
[22] Hollands S (2023) *** The operator product expansion in quantum field theory in curved
spacetimes *** to appear, this encyclopedia ***
[23] Moretti V (2021) On the global Hadamard parametrix in QFT and the signed squared
geodesic distance defined in domains larger than convex normal neighbourhoods. Letters in
Mathematical Physics 111: 130. [arXiv:2107.04903]
[24] Décanini Y and Folacci A (2008) Hadamard renormalization of the stress-energy tensor for
a quantized scalar field in a general spacetime of arbitrary dimension. Physical Review D
78: 044025. [arXiv:gr-qc/0512118]
[25] Fewster CJ and Verch R (2012) Dynamical locality and covariance: What makes a physical
theory the same in all spacetimes? Annales Henri Poincaré 13: 1613-1674. [arXiv:1106.4785]
[26] Hartman T (2015) Lectures on Quantum Gravity and Black Holes. (Cornell University)
[Link]
[27] Juárez-Aubry BA (2023) Quantum strong cosmic censorship and black hole evaporation.
arXiv:2305.01617.
[28] Bernal AN and Sánchez M (2003) On smooth Cauchy hypersurfaces and Geroch’s splitting
theorem. Commun. Math. Phys. 243: 461-470.
[29] Fewster CJ and Rejzner K (2020) Algebraic quantum field theory: An introduction.
In: Finster F, Giulini D, Kleiner J and Tolksdorf J (eds) Progress and Visions in
Quantum Theory in View of Gravity: Bridging Foundations of Physics and Mathe-
matics. Cham, Birkhäuser. pp. 1-61. [Link]
978-3-030-38941-3_1 [arXiv:1904.04051]
[30] Buchholz D and Fredenhagen K (2023) Algebraic quantum field theory: objectives, methods,
and results ***to appear, this encyclopedia*** [arXiv:2305.12923]
[31] Kay BS (1985) The double-wedge algebra for quantum fields on Schwarzschild and
Minkowski spacetimes. Communications in Mathematical Physics 100: 57-81. And Erra-
tum to appear.

29
[32] Radzikowski MJ (1996): A Local-to-Global Singularity Theorem for Quantum Field Theory
on Curved Space-Time. Commun. Math. Phys. 180: 1-22. (Includes an Appendix by Verch
R.)
[33] Kay BS (1977) Quantum fields in time-dependent backgrounds and in curved spacetimes.
PhD thesis, University of London.
[34] Kay BS (1979) A uniqueness result in the Segal-Weinless approach to linear Bose fields.
Journal of Mathematical Physics 20:1712-1713.
[35] Kay BS (1978) Linear spin-zero quantum fields in external gravitational and scalar fields.
Commun. Math. Phys 62: 55-70.
[36] Kay BS (1985) A uniqueness result for quasi-free KMS states. (1985) Helv. Phys. Acta
58:1017-1029.
[37] Radzikowski MJ (1996): Micro-local approach to the Hadamard condition in quantum field
theory on curved space-time. Commun. Math. Phys. 179: 529-553.
[38] Fewster CJ (2000) A general worldline quantum inequality. Class. Quantum Grav. 17:1897.
[arXiv:gr-qc/9910060]
[39] Fewster CJ and Verch R (2013) The Necessity of the Hadamard Condition. Class. Quant.
Grav. 30:235027. [arXiv:1307.5242]
[40] Kurpicz F, Pinamonti N and Verch R (2021) Temperature and entropy-area relation of
quantum matter near spherically symmetric outer trapping horizons. Letters in Mathemat-
ical Physics: 111: 110. [arXiv:2102.11547]
[41] Sahlmann H and Verch R (2000) Passivity and microlocal spectrum condition. Com-
mun. Math. Phys. 214: 705. [arXiv:math-ph/0002021]
[42] Sanders K (2013) Thermal equilibrium states of a linear scalar quantum field in stationary
spacetimes. International Journal of Modern Physics A 28: 1330010. [arXiv:1209.6068]
[43] Verch R (2004) Local definiteness, primarity and quasiequivalence of quasifree Hadamard
quantum states in curved spacetime. Commun. Math. Phys. 160: 507-536.
[44] Khavkine I and Moretti V (2015) Algebraic QFT in curved spacetime and quasifree
Hadamard states: an introduction. In: (Brunetti et al., 2015) (see Further Reading) 191-251.
[arXiv:1412.5945]
[45] Fulling SA, Sweeny M and Wald RM (1978) Singularity structure of the two-point function
in quantum field theory in curved spacetime. Communications in Mathematical Physics 63:
257-264.
[46] Fulling SA, Narcowich FJ and Wald RM (1981) Singularity structure of the two-point func-
tion in quantum field theory in curved spacetime, II. Annals of Physics 136: 243-272.
[47] Junker W (1996) Hadamard States, Adiabatic vacua and the construction of physical states
for scalar quantum fields on curved spacetime. Reviews in Mathematical Physics 08: 1091-
1159. And Erratum (2002) Reviews in Mathematical Physics 14: 511-51.
[48] Gérard C and Wrochna M (2014) Construction of Hadamard states by pseudo-differential
calculus. Commun. Math. Phys. 325: 713-755. [arXiv:1209.2604]
[49] Gérard C and Wrochna M (2019) Analytic Hadamard States, Calderón Projectors and
Wick Rotation Near Analytic Cauchy Surfaces. Commun. Math. Phys. 366: 29-65.
[arXiv:1706.08942]
[50] Gérard C and Wrochna M (2017) Hadamard property of the in and out states for Klein-
Gordon fields on asymptotically static spacetimes. Annales Henri Poincaré 18: 2715-2756.
[51] Moretti V, Murro S and Volpe D (2022) Paracausal deformations of Lorentzian metrics and
Møller isomorphisms in algebraic quantum field theory. To appear in Selecta Mathematica,
new series [arXiv:2109.06685]
[52] Dappiaggi C, Moretti V, Pinamonti N (2017) Hadamard States From Light-like Hypersur-
faces SpringerBriefs in Mathematical Physics 2017 [arXiv:1706.09666]

30
[53] Gérard C and Wrochna M (2016) Construction of Hadamard states by characteristic Cauchy
problem. Analysis and PDE 9: 111-149. [arXiv:1409.6691]
[54] Hollands S (2001) The Hadamard condition for Dirac fields and adiabatic states on
Robertson-Walker spacetimes. Communications in Mathematical Physics 216: 635-661.
[arXiv:gr-qc/9906076]
[55] Gérard C and Stoskopf T (2022) Hadamard states for quantized Dirac fields on
Lorentzian manifolds of bounded geometry. Reviews in Mathematical Physics 34: 2250008.
[arXiv:2108.11630]
[56] Fewster CJ, Pfenning MJ (2003). A quantum weak energy inequality for spin-one fields in
curved space-time. Journal of Mathematical Physics 44: 4480-4513. [arXiv:gr-qc/0303106]
[57] Wrochna M and Zahn J (2017) Classical phase space and Hadamard states in the BRST
formalism for gauge field theories on curved spacetime. Reviews in Mathematical Physics
29: 1750014. [arXiv:1407.8079]
[58] Gérard C and Wrochna M (2015) Hadamard states for the linearized Yang-Mills equa-
tion on curved spacetime. Communications in Mathematical Physics 337: 253-320.
[arXiv:1403.7153]
[59] Gérard C, Murro S and Wrochna M (2022) Quantization of linearized gravity by Wick
rotation in Gaussian time. arXiv:2204.01094.
[60] Brum M and Fredenhagen K (2013) ‘Vacuum-like’ Hadamard states for quantum fields on
curved spacetimes. Classical and Quantum Gravity 31: 025024. [arXiv:1307.0482]
[61] Fewster CJ (2018) The art of the state. International Journal of Modern Physics D 27:
1843007. [arXiv:1803.06836]
[62] Afshordi N, Aslanbeigi S and Sorkin RD (2012) A distinguished vacuum state for a quantum
field in a curved spacetime: formalism, features, and cosmology. Journal of High Energy
Physics 2012: 1-29. [arXiv:1205.1296]
[63] Fewster CJ and Verch R (2012) On a recent construction of ‘vacuum-like’ quantum field
states in curved spacetime. Class. Quantum Grav. 29: 205017. [arXiv:1206.1562]
[64] Juárez-Aubry BA, Kay BS, Miramontes T, Sudarsky D (2023) On the initial value problem
for semiclassical gravity without and with quantum state collapses. Journal of Cosmology
and Astroparticle Physics 01: 040. [arXiv:2205.11671]
[65] Bernal AN and Sánchez M (2005) Smoothness of time functions and the metric splitting of
globally hyperbolic spacetimes. Commun. Math. Phys. 257: 43-50. [arXiv:gr-qc/0401112]
[66] Shale D (1962) Linear symmetries of free Boson fields. Transactions of the American Math-
ematical Society 103: 149-167.
[67] Wald RM (1979) Existence of the S-matrix in quantum field theory in curved space-time.
Annals of Physics 118: 490-510.
[68] Dimock J (1979) Scalar quantum field in an external gravitational field. Journal of Mathe-
matical Physics 20: 2549-2555.
[69] Kay BS (1982) Quantum fields in curved space-times and scattering theory. In: Doebner
HD, Andersson SI, Petry HR (eds) (1980) Differential Geometric Methods in Mathematical
Physics: Proceedings, Clausthal Springer Lecture Notes in Mathematics 905 pp. 272-295.
Berlin-Heidelberg: Springer.
[70] Dimock J and Kay BS (1987) Classical and quantum scattering theory for linear scalar fields
on the Schwarzschild metric [Link] of Physics 175: 366-426.
[71] Duch P and Dybalski W (2023) Infrared problem in quantum electrodynamics. ***to appear,
this Encyclopedia*** [arXiv:2307.06114]
[72] Moretti V (2003) Comments on the stress-energy tensor operator in curved spacetime. Com-
mun. Math. Phys. 232:189-221. [arXiv:gr-qc/0109048]
[73] Balakumar V and Winstanley E (2020) Hadamard renormalization for a charged scalar field.
Classical and Quantum Gravity 37: 065004.

31
[74] Verch R (1999) Wavefront sets in algebraic quantum field theory. Commun. Math. Phys.
205: 337-367. [arXiv:math-ph/9807022]
[75] Brunetti R and Fredenhagen K (2000) Microlocal analysis and interacting quantum field the-
ories: Renormalization on physical backgrounds. Communications in Mathematical Physics
208: 623-661. [arXiv:math-ph/9903028]
[76] Fewster CJ and Verch R (2012) Dynamical locality of the free scalar field. Annales Henri
Poincaré 13: 1675-1709. [arXiv:1109.6732]
[77] Fredenhagen K and Haag R (1990) On the derivation of Hawking radiation associated with
the formation of a black hole. Communications in mathematical physics 127: 273-284.
[78] Hawking SW and Ellis GFR (1973) The large scale structure of space-time. Cambridge:
Cambridge University Press.
[79] Wald RM (1976) Stimulated-emission effects in particle creation near black holes. Physical
Review D 13: 3176.
[80] Leahy DA and Unruh WG (1983). Effects of a λΦ4 interaction on black-hole evaporation in
two dimensions. Physical Review D 28: 694.
[81] Bachelot A. (1994) Asymptotic completeness for the Klein-Gordon equation on the
Schwarzschild metric. Annales de l’Institut Henri Poincaré section A 61: 411-441.
[82] Dimock J (1985) Scattering for the Wave Equation on the Schwarzschild Metric.
Gen. Rel. Grav. 17: 353-369.
[83] Bachelot A (1999) The Hawking effect. Annales de l’Institut Henri Poincaré, Physique
théorique 70: 41-99.
[84] Melnyk F (2004) The Hawking Effect for Spin 1/2 Fields. Commun. Math. Phys. 244: 483-
525
[85] Alford F (2022) A Mathematical Study of Hawking Radiation on Collapsing,
Spherically Symmetric Spacetimes. (Doctoral dissertation, University of Cam-
bridge). [Link]
8e4ec008-0f5d-4711-b78e-67bf01aff0d3/content
[86] Alford, F (2020) The scattering map on Oppenheimer-Snyder space-time. Annales Henri
Poincaré 21: 2031-2092. [arXiv:1910.02752]
[87] Alford, F (2023) The scattering map on collapsing charged spherically symmetric spacetimes.
arXiv:2309.03022
[88] Alford, F (2023) A rigorous study of Hawking radiation on collapsing charged spherically
symmetric spacetimes. arXiv:2309.03022
[89] Dafermos M and Rodnianski I (2013) Lectures on black holes and linear waves.
Clay Math. Proc 17: 97-205. [arXiv:0811.0354]
[90] Dafermos M, Rodnianski I and Shlapentokh-Rothman Y (2018) A scattering theory for the
wave equation on Kerr black hole exteriors. Annales Scientifiques de l’ENS 51: 371-486
[arXiv:1412.8379]
[91] Häfner D and Nicolas JP (2004) Scattering of massless Dirac fields by a Kerr black hole.
Reviews in Mathematical Physics 16: 29-123.
[92] Batic D (2007) Scattering for massive Dirac fields on the Kerr metric. Journal of mathe-
matical physics 48: 022502. [arXiv: gr-qc/0606051]
[93] Higuchi A and Yamamoto (2018) Vacuum state in de Sitter spacetime with static charts.
Phys. Rev. D98: 065014. [arXiv:1808.02147]
[94] Crispino LCB, Higuchi A and Matsas GEA (2008) The Unruh effect and its applications.
Rev. Mod. Phys. 80: 787. [arXiv:0710.5373]
[95] Sewell GL (1980) Relativity of temperature and the Hawking effect. Physics Letters A 79:
23-24.
[96] Sewell GL (1982) Quantum fields on manifolds: PCT and gravitationally induced thermal
states. Annals of Physics 141: 201-224.

32
 [97] Jacobson T (1994) A note on Hartle-Hawking vacua Phys. Rev. D50:6031-6032. [arXiv:
gr-qc/9407022]
[98] Sanders K (2015) On the construction of Hartle-Hawking-Israel states across a static bifur-
cate Killing horizon. Lett. Math. Phys. 105: 575-640. [arXiv:1310.5537]
[99] Gérard C (2021) The Hartle-Hawking-Israel state on spacetimes with stationary bifurcate
Killing horizons. Reviews in Mathematical Physics 33: 2150028. [arXiv:1806.07645]
[100] Pinamonti N, Sanders K and Verch R (2019) Local incompatibility of the microlocal spec-
trum condition with the KMS property along spacelike directions in quantum field theory
on curved spacetime. Lett. Math. Phys. 109: 1735-1745. [arXiv:1806.02124]
[101] Kay BS and Lupo U (2016) Non-existence of isometry-invariant Hadamard states for a
Kruskal black hole in a box and for massless fields on 1+ 1 Minkowski spacetime with a uni-
formly accelerating mirror. Classical and Quantum Gravity 33: 215001. [arXiv: 1502.06582]
[102] Hawking SW (1976) Black holes and thermodynamics. Phys. Rev. D 13: 191-197.
Quantization of fermions on Kerr space-time
[103] Casals M, Dolan SR, Nolan BC, Ottewill AC and Winstanley E Phys. Rev. D 87: 064027.
[arXiv:1207.7089]
[104] Dappiaggi C, Moretti V and Pinamonti N. (2011) Rigorous construction and Hadamard
property of the Unruh state in Schwarzschild spacetime. Advances in Theoretical and Math-
ematical Physics 15: 355-447. [arXiv:0907.1034 ]
[105] (2022) Zilberman N, Casals M, Ori A and Ottewill AC (2022) Two-point function of a
quantum scalar field in the interior region of a Kerr black hole. Phys. Rev. D 106:125011.
[arXiv:2203.07780]
[106] Gérard C, Häfner D and Wrochna M. (2020) The Unruh state for massless fermions on Kerr
spacetime and its Hadamard property. arXiv:2008.10995.
[107] Penrose R (1968) Structure of space-time. Battelle Rencontres. 1967 Lectures in Mathemat-
ics and Physics deWitt C and Wheeler J (eds) (New York: Benjamin)
[108] Simpson M and Penrose R (1973) Internal instability in a Reissner-Nordström black hole
International Journal of Theoretical Physics 7: 183-187.
[109] Misner CW and Taub AH (1969) A singularity-free empty universe. Soviet Physics JETP
28: 122-133.
[110] Penrose R (1974) Gravitational collapse. In: Dewitt-Morette C ed. Gravitational radiation
and gravitational collapse. International Astronomical Union Symposium No. 64. pp. 82-91.
(Dordrecht: Reidel)
[111] Landsman K (2022) Penrose’s 1965 singularity theorem: from geodesic incompleteness to
cosmic censorship. General Relativity and Gravitation 54: 115. [arXiv:2205.01680]
[112] Dafermos M and Shlapentokh-Rothman Y (2018) Rough initial data and the strength of
the blue-shift instability on cosmological black holes with Λ > 0. Classical and Quantum
Gravity 35: 195010. [arXiv:1805.08764]
[113] Christodoulou D (2009) The Formation of Black Holes in General Relativity (EMS Mono-
graphs in Mathematics) (Zürich: European Mathematical Society)
[114] Lanir A, Ori A, Zilberman N, Sela O, Maline A and Levi A (2019) Analysis of quantum ef-
fects inside spherical charged black holes. Physical Review D 99: 061502. [arXiv:1811.03672]
[115] Zilberman N, Casals M, Ori A and Ottewill AC (2022) Quantum Fluxes at the Inner Horizon
of a Spinning Black Hole. Phys. Rev. Lett. 129: 261102 [arXiv:2203.08502].
[116] Juárez-Aubry BA (2015) Can a particle detector cross a Cauchy horizon? International
Journal of Modern Physics D 24:1542005. [arXiv:1502.02533]
[117] Juárez-Aubry BA and Louko J (2022). Quantum kicks near a Cauchy horizon. AVS Quantum
Science 4: 013201. [arXiv:2109.14601]
[118] Hintz P and Vasy A (2017) Analysis of linear waves near the Cauchy horizon of cosmological
black holes. Journal of Mathematical Physics 58: 081509. [arXiv:1512.08004]

33
[119] Hollands S, Wald RM and Zahn J (2020) Quantum instability of the Cauchy horizon
in Reissner-Nordström-deSitter spacetime. Classical and Quantum Gravity 37: 115009.
[arXiv:1912.06047]
[120] Hollands S, Klein C and Zahn J (2020) Quantum stress tensor at the Cauchy horizon of the
Reissner-Nordström-de Sitter spacetime. Physical Review D 102: 085004. [arXiv:2006.10991]
[121] Marković D and Unruh WG (1991) Vacuum for a massless scalar field outside a collapsing
body in de Sitter space-time Phys. Rev. D 43:332.
[122] Kay BS (2001) Application of linear hyperbolic PDE to linear quantum fields in curved
space-times: Especially black holes, time machines and a new semilocal vacuum concept.
Journées Équations aux dérivées partielles, Nantes, 5 au 9 juin 2000, GDR 1151 (CNRS)
pp. IX-1 to IX-19 (Also available at [Link]
2000/html) [arXiv:gr-qc/0103056]
[123] Hollands S (2000) Aspects of Quantum Field Theory on Curved Spacetime PhD Thesis
University of York.
[124] Klein CKM (2023) Construction of the Unruh state for a real scalar field on the Kerr-de
Sitter spacetime. Annales Henri Poincaré 24: 2401-2442. [arXiv:2206.05073]
[125] Cramer CR and Kay BS (1998) Thermal and two-particle stress-energy must be ill defined on
the two-dimensional Misner space chronology horizon. Physical Review D 57:1052. [arXiv:gr-
qc/9708028]
[126] Krasnikov SV (1996) Quantum stability of the time machine. Physical Review D 54: 7322.
[arXiv:gr-qc/9508038]
[127] Visser M (1997) The reliability horizon for semi-classical quantum gravity: Metric fluctua-
tions are often more important than back-reaction. Physics Letters B 415: 8-14. [arXiv:gr-
qc/9702041]
[128] Kay BS (1992) The principle of locality and quantum field theory on (non globally hyper-
bolic) curved spacetimes. Reviews in Mathematical Physics 4: 167-195. (Available at https:
//[Link]/publication/234525000_The_Principle_of_Locality_and_
Quantum_Field_Theory_on_non_Globally_Hyperbolic_Curved_Spacetimes)
[129] Janssen DW (2022) Quantum fields on semi-globally hyperbolic space-times. Communica-
tions in Mathematical Physics 391: 669-705. [arXiv:2111.01643]
[130] Meda P, Pinamonti N, Roncallo S and Zanghı̀ N (2021) Evaporation of four-dimensional
dynamical black holes sourced by the quantum trace anomaly. Classical and Quantum Grav-
ity 38: 195022. (See also Erratum (2022) Classical and Quantum Gravity 39: 059501.)
[arXiv:2103.02057]
[131] Kay BS (2022) The black hole information loss puzzle, matter-gravity entanglement entropy
and the second law. arXiv:2206.07445v4. (A shorter version of this appears will appear as:
Kay BS (2023) Matter-gravity entanglement entropy and the second law for black holes.
International Journal of Modern Physics D (to appear) [arXiv:2305.11723])
[132] Hawking SW (1976) Breakdown of predictability in gravitational collapse. Phys. Rev. D 14:
2460.
[133] Unruh WG and Wald RM (2017) Information loss. Reports on Progress in Physics 80:
092002. [arXiv:1703.02140]
[134] Maudlin T (2017) (Information) paradox lost. arXiv:1705.03541.
[135] Okon E and Sudarsky D (2015) The Black Hole Information Paradox and the Collapse of
the Wave Function Found. Phys. 45:461-470 [arXiv:1710.01451]
[136] Juárez-Aubry BA (2023) Quantum strong cosmic censorship and black hole evaporation.
arXiv:2305.01617v1.
[137] Hawking SW and Stewart JM (1993). Naked and thunderbolt singularities in black hole
evaporation. Nuclear Physics B 400: 393-415. [arXiv:hep-th/9207105]

34
[138] Anderson PR and Eaker W (1999) Analytic approximation and an improved method for
computing the stress-energy of quantized scalar fields in Robertson-Walker spacetimes. Phys-
ical Review D 61: 024003. [arXiv:gr-qc/9906055]
[139] Levi A and Ori A (2016) Versatile method for renormalized stress-energy computation in
black-hole spacetimes. Physical Review Letters 117: 231101. [arXiv:1608.03806]
[140] Taylor P, Breen C and Ottewill A (2022) Mode-sum prescription for the renormalized stress
energy tensor on black hole spacetimes. Physical Review D 106: 065023. [arXiv: 2201.05174]
[141] Verch R (2001) A spin-statistics theorem for quantum fields on curved spacetime manifolds
in a generally covariant framework. Commun. Math. Phys. 223: 261-288.
[142] Fewster CJ (2016) On the spin-statistics connection in curved spacetimes. In: Finster
F, Kleiner J, Röken C and Tolksdorf J (eds) Quantum Mathematical Physics. Cham:
Birkhäuser. [arXiv:1503.05797]
[143] Hollands S (2004) A General PCT theorem for the operator product expansion in curved
space-time. Commun. Math. Phys. 244: 209-244. [arXiv:gr-qc/0212028]
[144] Strohmaier A, Verch R and Wollenberg M (2002) Microlocal analysis of quantum fields on
curved space-times: Analytic wave front sets and Reeh-Schlieder theorems. J. Math. Phys.
43: 5514. [arXiv:math-ph/0202003]
[145] Sanders, K (2009) On the Reeh-Schlieder Property in Curved Spacetime. Commun. Math.
Phys. 288: 271-285. [arXiv: 0801.4676]
[146] Strohmaier A and Witten E (2023) Analytic states in quantum field theory on curved
spacetimes. arXiv:2302.02709.
[147] Fewster CJ (2015) The split property for locally covariant quantum field theories in curved
cpacetime. Lett. Math. Phys. 105: 1633-1661 [arXiv:1501.02682]
[148] Fewster CJ (2016) The split property for quantum field theories in flat and curved space-
times. Abh. Math. Semin. Univ. Hambg. 86: 153-175 [arXiv:1601.06936]
[149] Fewster CJ (2012) Lectures on energy inequalities. arXiv:1208.5399.
[150] Kontou E-A, Sanders, K (2020) Energy conditions in general relativity and quantum field
theory. [arXiv: 2003.01815]
[151] Witten E (2020) Remarks on energy inequalities. YouTube video of conference talk https:
//[Link]/watch?v=0Oh-Kmy-mx0
[152] Hollands S (2008) Renormalized quantum Yang-Mills fields in curved spacetime. Reviews in
Mathematical Physics 20: 1033-1172. [arXiv:0705.3340]
[153] Fredenhagen K and Rejzner K (2015) Perturbative construction of models of algebraic quan-
tum field theory. In (Brunetti et al., 2015) (see ‘Further Reading’).
[154] Rejzner K. (2016) Perturbative Algebraic Quantum Field Theory. Math. Phys. Stud Springer
International Publishing.
[155] Rejzner K (2023) *** Perturbative algebraic quantum field theory. *** to appear, this
encyclopedia ***
[156] Dütsch M (2019) From Classical Field Theory to Perturbative Quantum Field Theory. Cham:
Birkhäuser.
[157] Fredenhagen K and Rejzner K (2012). Local covariance and background independence.
Quantum Field Theory and Gravity: In: Finster F, Müller O, Nardmann M, Tolksdorf J and
Zeidler E (eds) (2012) Quantum Field Theory and Gravity, Conceptual and Mathematical
Advances in the Search for a Unified Framework (Birkhäuser, Springer-Basel) pp. 15-23.
[arXiv:math-ph/1102.2376]
[158] Brunetti R, Fredenhagen K and Rejzner K (2016) Quantum gravity from the point of view
of locally covariant quantum field theory. Communications in Mathematical Physics 345:
741-779. [arXiv:1306.1058]
[159] Brunetti R, Fredenhagen K, Hack TP, Pinamonti N and Rejzner K (2016) Cosmologi-
cal perturbation theory and quantum gravity. Journal of High Energy Physics, 2016: 32.
[arXiv:1605.02573]

35
[160] Lima WCC (2021) Graviton backreaction on the local cosmological expansion in slow-roll
inflation. Class. Quantum Grav. 38: 135015 [arXiv:2007.04995]
[161] Fröb MB, Rein C and Verch R (2022) Graviton corrections to the Newtonian potential using
invariant observables. Journal of High Energy Physics 2022: 180. [arXiv:2109.09753]
[162] Brunetti R, Fredenhagen K and Rejzner K (2022) Locally covariant approach to effective
quantum gravity. arXiv:2212.07800.
[163] Barata JCA, Jäkel CD and Mund J (2023) The P (ϕ)2 model on de Sitter space. Memoirs
of the American Mathematical Society 281: 1389.
[164] L.H. Ford (2005) Spacetime in semiclassical gravity. In Ashtekar A (ed) 100 Years of Rela-
tivity – Space-time Structure: Einstein and Beyond (Singapore, World Scientific) [arXiv:gr-
qc/0504096]
[165] Hu BLB and Verdaguer E (2020) Semiclassical and Stochastic Gravity: Quantum Field
Effects on Curved Spacetime. Cambridge, Cambridge University Press.
[166] Juárez-Aubry BA (2019) Semi-classical gravity in de Sitter spacetime and the cosmological
constant. Physics Letters B 797: 134912. [arXiv:1903.03924]
[167] Juárez-Aubry BA, Miramontes T, Sudarsky D (2020) Semiclassical theories as initial value
problems. J. Math. Phys. 61: 032301 [arXiv:1907.09960]
[168] Gottschalk H and Siemssen D. (2021) The cosmological semiclassical Einstein equation
as an infinite-dimensional dynamical system. Annales Henri Poincaré 22: 3915-3964.
[arXiv:1809.03812]
[169] Meda P, Pinamonti N and Siemssen D (2021) Existence and uniqueness of solutions of the
semiclassical Einstein equation in cosmological models. Annales Henri Poincaré 22:3965-
4015. [arXiv:2007.14665]
[170] Gottschalk H, Rothe NR, Siemssen D (2022) Special cosmological models derived from the
semiclassical Einstein equation on flat FLRW space-times. Classical and Quantum Gravity
39: 125004 [arXiv:2112.15050]
[171] Sanders K (2022) Static symmetric solutions of the semi-classical Einstein-Klein-Gordon
system. Annales Henri Poincaré 23: 1321-1358. [arXiv:2007.14311]
[172] Juárez-Aubry BA (2022) Semiclassical gravity in static spacetimes as a constrained initial
value problem. Annales Henri Poincaré 23: 1451-1487 [arXiv:2011.05947]
[173] Juárez-Aubry BA and Modak SK (2022) Semiclassical gravity with a conformally covari-
ant field in globally hyperbolic spacetimes. Journal of Mathematical Physics 63: 092303.
[arXiv:2110.01719]
[174] Meda P, Pinamonti N (2023) Linear stability of semiclassical theories of gravity. Annales
Henri Poincaré 24: 1211-1243. [arXiv:2201.10288]
[175] Gottschalk H, Rothe NR, Siemssen D (2023) Cosmological de Sitter Solutions of the Semi-
classical Einstein Equation. Annales Henri Poincaré 24: 2949–3029. [arXiv:2206.07774]

36