Algebraic Quantum Field Theory on Spacetimes with Timelike Boundary

Algebraic Quantum Field Theory on Spacetimes with Timelike Boundary Ann. Henri Poincar´ e Online First c 2018 The Author(s) Annales Henri Poincar´ e https://doi.org/10.1007/s00023-018-0687-1 Algebraic Quantum Field Theory on Spacetimes with Timelike Boundary Marco Benini , Claudio Dappiaggi and Alexander Schenkel Dedicated to Klaus Fredenhagen on the occasion of his 70th birthday. Abstract. We analyze quantum field theories on spacetimes M with time- like boundary from a model-independent perspective. We construct an adjunction which describes a universal extension to the whole spacetime M of theories defined only on the interior intM . The unit of this adjunc- tion is a natural isomorphism, which implies that our universal extension satisfies Kay’s F-locality property. Our main result is the following char- acterization theorem: Every quantum field theory on M that is additive from the interior (i.e., generated by observables localized in the interior) admits a presentation by a quantum field theory on the interior intM and an ideal of its universal extension that is trivial on the interior. We shall illustrate our constructions by applying them to the free Klein–Gordon field. 1. Introduction and Summary Algebraic quantum field theory is a powerful and far developed framework to address model-independent aspects of quantum field theories on Minkowski spacetime [18] and more generally on globally hyperbolic spacetimes [7]. In addition to establishing the axiomatic foundations for quantum field theory, the algebraic approach has provided a variety of mathematically rigorous con- structions of non-interacting models, see e.g., the reviews [1, 3, 4], and more in- terestingly also perturbatively interacting quantum field theories, see e.g., the recent monograph [26]. It is worth emphasizing that many of the techniques involved in such constructions, e.g., existence and uniqueness of Green’s opera- tors and the singular structure of propagators, crucially rely on the hypothesis that the spacetime is globally hyperbolic and has empty boundary. M. Benini et al. Ann. Henri Poincar´ e Even though globally hyperbolic spacetimes have plenty of applications to physics, there exist also important and interesting situations which require non-globally hyperbolic spacetimes, possibly with a non-trivial boundary. On the one hand, recent developments in high energy physics and string theory are strongly focused on anti-de Sitter spacetime, which is not globally hyperbolic and has a (conformal) timelike boundary. On the other hand, experimental setups for studying the Casimir effect confine quantum field theories between several metal plates (or other shapes), which may be modeled theoretically by introducing timelike boundaries to the system. This immediately prompts the question whether the rigorous framework of algebraic quantum field theory admits a generalization to cover such scenarios. Most existing works on algebraic quantum field theory on spacetimes with a timelike boundary focus on the construction of concrete examples, such as the free Klein–Gordon field on simple classes of spacetimes. The basic strategy em- ployed in such constructions is to analyze the initial value problem on a given spacetime with timelike boundary, which has to be supplemented by suitable boundary conditions. Different choices of boundary conditions lead to different Green’s operators for the equation of motion, which is in sharp contrast to the well-known existence and uniqueness results on globally hyperbolic spacetimes with empty boundary. Recent works addressing this problem are [19, 20, 33], the latter extending the analysis of [31]. For specific choices of boundary con- ditions, there exist successful constructions of algebraic quantum field theories on spacetimes with timelike boundary, see e.g., [8, 10–12]. The main message of these works is that the algebraic approach is versatile enough to account also for these models, although some key structures, such as for example the notion of Hadamard states [11, 32], should be modified accordingly. Unfortunately, model-independent results on algebraic quantum field the- ory on spacetimes with timelike boundary are more scarce. There are, how- ever, some notable and very interesting works in this direction: On the one hand, Rehren’s proposal for algebraic holography [25] initiated the rigorous study of quantum field theories on the anti-de Sitter spacetime. This has been further elaborated in [13] and extended to asymptotically AdS spacetimes in [27]. On the other hand, inspired by Fredenhagen’s universal algebra [15–17], a very interesting construction and analysis of global algebras of observables on spacetimes with timelike boundaries has been performed in [30]. The most notable outcome is the existence of a relationship between maximal ideals of this algebra and boundary conditions, a result which has been of inspiration for this work. In the present paper, we shall analyze quantum field theories on space- times with timelike boundary from a model-independent perspective. We are mainly interested in understanding and proving structural results for whole categories of quantum field theories, in contrast to focusing on particular the- ories. Such questions can be naturally addressed by using techniques from the recently developed operadic approach to algebraic quantum field theory [5]. Let us describe rather informally the basic idea of our construction and its implications: Given a spacetime M with timelike boundary, an algebraic AQFT on Spacetimes with Timelike Boundary quantum field theory on M is a functor B : R → Alg assigning algebras of observables to suitable regions U ⊆ M (possibly intersecting the boundary), which satisfies the causality and time-slice axioms. We denote by QFT(M ) the category of algebraic quantum field theories on M . Denoting the full sub- category of regions in the interior of M by R ⊆R , we may restrict int M M any theory B ∈ QFT(M)toatheory res B ∈ QFT(int M ) defined only on the interior regions. Notice that it is in practice much easier to analyze and construct theories on int M as opposed to theories on the whole spacetime M . This is because the former are postulated to be insensitive to the boundary by Kay’s F-locality principle [22]. As a first result, we shall construct a left adjoint of the restriction functor res : QFT(M ) → QFT(int M ), which we call the universal extension functor ext : QFT(int M ) → QFT(M ). This means that given any theory A ∈ QFT(int M ) that is defined only on the interior regions in M , we obtain a universal extension ext A ∈ QFT(M ) to all regions in M , including those that intersect the boundary. It is worth to emphasize that the adjective universal above refers to the categorical concept of universal properties. Below we explain in which sense ext is also “universal” in a more physical meaning of the word. It is crucial to emphasize that our universal extension ext A ∈ QFT(M ) is always a bona fide algebraic quantum field theory in the sense that it satisfies the causality and time-slice axioms. This is granted by the operadic approach to algebraic quantum field theory of [5]. In particular, the ext  res adjunction investigated in the present paper is one concrete instance of a whole fam- ily of adjunctions between categories of algebraic quantum field theories that naturally arise within the theory of colored operads and algebras over them. A far reaching implication of the above mentioned ext  res adjunction is a characterization theorem that we shall establish for quantum field theories on spacetimes with timelike boundary. Given any theory B ∈ QFT(M)ona spacetime M with timelike boundary, we can restrict and universally extend to obtain another such theory ext res B ∈ QFT(M ). The adjunction also provides us with a natural comparison map between these theories, namely the counit  :extres B → B of the adjunction. Our result in Theorem 5.6 and Corollary 5.7 is that  induces an isomorphism ext res B/ ker  B B B of quantum field theories if and only if B is additive from the interior as formalized in Definition 5.5. The latter condition axiomatises the heuristic idea that the theory B has no degrees of freedom that are localized on the boundary of M , i.e., all its observables may be generated by observables supported in the interior of M . Notice that the results in Theorem 5.6 and Corollary 5.7 give the adjective universal also a physical meaning in the sense that the extensions are sufficiently large such that any additive theory can be recovered by a quotient. We strengthen this result in Theorem 5.10 by constructing an equivalence between the category of additive quantum field theories on M and a category of pairs (A, I) consisting of a theory A ∈ QFT(int M)onthe interior and an ideal I ⊆ ext A of the universal extension that is trivial on the interior. More concretely, this means that every additive theory B ∈ QFT(M ) may be naturally decomposed into two distinct pieces of data: (1) A theory M. Benini et al. Ann. Henri Poincar´ e A ∈ QFT(int M ) on the interior, which is insensitive to the boundary as postulated by F-locality, and (2) an ideal I ⊆ ext A of its universal extension that is trivial on the interior, i.e., that is only sensitive to the boundary. Specific examples of such ideals arise from imposing boundary conditions. We shall illustrate this fact by using the free Klein–Gordon theory as an example. Thus, our results also provide a bridge between the ideas of [30] and the concrete constructions in [8, 10–12]. The remainder of this paper is structured as follows: In Sect. 2, we recall some basic definitions and results about the causal structure of spacetimes with timelike boundaries, see also [9, 29]. In Sect. 3, we provide a precise definition of the categories QFT(M)and QFT(int M ) by using the ideas of [5]. Our universal boundary extension is developed in Sect. 4, where we also provide an explicit model in terms of left Kan extension. Our main results on the characterization of additive quantum field theories on M are proven in Sect. 5. Section 6 illustrates our construction by focusing on the simple example of the free Klein–Gordon theory, where more explicit formulas can be developed. It is in this context that we provide examples of ideals implementing boundary conditions and relate to analytic results, e.g., [12]. We included “Appendix A” to state some basic definitions and results of category theory which will be used in our work. 2. Spacetimes with Timelike Boundary We collect some basic facts about spacetimes with timelike boundary, following [29, Section 3.1] and [9, Section 2.2]. For a general introduction to Lorentzian geometry, we refer to [2, 24], see also [1, Sections 1.3 and A.5] for a concise presentation. We use the term manifold with boundary to refer to a Hausdorff, second countable, m-dimensional smooth manifold M with boundary, see e.g., [23]. This definition subsumes ordinary manifolds as manifolds with empty bound- ary ∂M = ∅. We denote by int M ⊆ M the submanifold without the boundary. Every open subset U ⊆ M carries the structure of a manifold with (possibly empty) boundary and one has int U = U ∩ int M . Definition 2.1. A Lorentzian manifold with boundary is a manifold with bound- ary that is equipped with a Lorentzian metric. Definition 2.2. Let M be a time-oriented Lorentzian manifold with boundary. The Cauchy development D(S) ⊆ M of a subset S ⊆ M is the set of points p ∈ M such that every inextensible (piecewise smooth) future directed causal curve stemming from p meets S. The following properties follow easily from the definition of Cauchy de- velopment. Proposition 2.3. Let S, S ⊆ M be subsets of a time-oriented Lorentzian man- ifold M with boundary. Then, the following holds true: (a) S ⊆ S implies D(S) ⊆ D(S ); AQFT on Spacetimes with Timelike Boundary (b) S ⊆ D(S)= D(D(S)); (c) D(D(S) ∩ D(S )) = D(S) ∩ D(S ). We denote by J (S) ⊆ M the causal future/past of a subset S ⊆ M , i.e., the set of points that can be reached by a future/past directed causal curve stemming from S. Furthermore, we denote by I (S) ⊆ M the chronological future/past of a subset S ⊆ M , i.e., the set of points that can be reached by a future/past directed timelike curve stemming from S. Definition 2.4. Let M be a time-oriented Lorentzian manifold with boundary. + − We say that a subset S ⊆ M is causally convex in M if J (S) ∩ J (S) ⊆ S. M M We say that two subsets S, S ⊆ M are causally disjoint in M if (J (S) ∪ J (S)) ∩ S = ∅. The following properties are simple consequences of these definitions. Proposition 2.5. Let S, S ⊆ M be two subsets of a time-oriented Lorentzian manifold M with boundary. Then, the following holds true: (a) D(S) and D(S ) are causally disjoint if and only if S and S are causally disjoint; (b) Suppose S and S are causally disjoint. Then, the disjoint union S S ⊆ M is causally convex if and only if both S and S are causally convex. The following two definitions play an essential role in our work. Definition 2.6. A spacetime with timelike boundary is an oriented and time- oriented Lorentzian manifold M with boundary, such that the pullback of the Lorentzian metric along the boundary inclusion ∂M → M defines a Lorentzian metric on the boundary ∂M . Definition 2.7. Let M be a spacetime with timelike boundary. (i) R denotes the category whose objects are causally convex open subsets U ⊆ M and whose morphisms i : U → U are inclusions U ⊆ U ⊆ M . We call it the category of regions in M . (ii) C ⊆ Mor R is the subset of Cauchy morphisms in R , i.e., inclusions M M M i : U → U such that D(U)= D(U ). (iii) R ⊆R is the full subcategory whose objects are contained in the int M M interior int M . We denote by C ⊆C the Cauchy morphisms between int M M objects of R . int M Proposition 2.8. Let M be a spacetime with timelike boundary. For each subset S ⊆ M and each object U ∈R , i.e., a causally convex open subset U ⊆ M, the following holds true: ± ± (a) I (S) is the largest open subset of J (S); M M ± ± ± ± ± (b) J (I (S)) = I (S)= I (J (S)); M M M M M (c) S ⊆ int M implies D(S) ⊆ int M; (d) D(U ) ⊆ M is causally convex and open, i.e., D(U ) ∈R . M M. Benini et al. Ann. Henri Poincar´ e Proof. (a) and (b): These are standard results in the case of empty bound- ary, see e.g., [1, 2, 24]. The extension to spacetimes with non-empty timelike boundary can be found in [29, Section 3.1.1]. (c): We show that if D(S) contains a boundary point, so does S: Suppose p ∈ D(S) belongs to the boundary of M . By Definition 2.6, the boundary ∂M of M can be regarded as a time-oriented Lorentzian manifold with empty boundary; hence, we can consider a future directed inextensible causal curve γ in ∂M stemming from p. Since ∂M is a closed subset of M , γ must be inextensible also as a causal curve in M , hence γ meets S because it stems from p ∈ D(S). Since γ lies in ∂M by construction, we conclude that S contains a boundary point of M . (d): D(U ) ⊆ M is causally convex by the definition of Cauchy develop- ment and by the causal convexity of U ⊆ M . To check that D(U ) ⊆ M is open, we use quasi-limits as in [24, Definition 14.7 and Proposition 14.8]: First, ob- serve that I (U ) ⊆ M is open by (a). Hence, it is the same to check whether + − a subset of I := I (U ) ∪ I (U)isopenin M or in I (with the induced topol- M M ogy). Indeed, U ⊆ D(U ) ⊆ I because U is open in M . From now on, in place of M , let us therefore consider I, equipped with the induced metric, orientation and time-orientation. By contradiction, assume that there exists p ∈ D(U )\U such that all of its neighborhoods intersect the complement of D(U ). Then, there exists a sequence {α } of inextensible causal curves in I never meeting U such that {α (0)} converges to p. We fix a convex cover of I refining the open + − cover {I (U ),I (U )}. Relative to the fixed convex cover, the construction M M of quasi-limits allows us to obtain from {α } an inextensible causal curve λ through p ∈ D(U ). Hence, λ meets U,say in q. By the construction of a quasi- limit, q lies on a causal geodesic segment between p and p , two successive k k+1 limit points for {α } contained in some element of the fixed convex cover. It + − follows that either p or p belongs to J (U ) ∩ J (U ), which is contained k k+1 I I in U by causal convexity. Hence, we found a subsequence {α } of {α } and n n a sequence of parameters {s } such that {α (s )} converges to a point of U j n j (either p or p ). By construction the sequence {α (s )} is contained in k k+1 n j I\U ; however, its limit lies in U . This contradicts the hypothesis that U is open in I. The causal structure of a spacetime M with timelike boundary can be affected by several pathologies, such as the presence of closed future directed causal curves. It is crucial to avoid these issues in order to obtain concrete examples of our constructions in Sect. 6. The following definition is due to [9, Section 2.2] and [29, Section 3.1.2]. Definition 2.9. A spacetime M with timelike boundary is called globally hy- perbolic if the following two properties hold true: (i) Strong causality Every open neighborhood of each point p ∈ M contains a causally convex open neighborhood of p. + − (ii) Compact double-cones J (p) ∩ J (q) is compact for all p, q ∈ M . M M AQFT on Spacetimes with Timelike Boundary Remark 2.10. In the case of empty boundary, this definition agrees with the usual one in [1, 2, 24]. Simple examples of globally hyperbolic spacetimes with m−1 m non-empty timelike boundary are the half space {x ≥ 0}⊆ R , the spatial m−1 m 1 2 m−1 2 m slab {0 ≤ x ≤ 1}⊆ R and the cylinder {(x ) +···+(x ) ≤ 1}⊆ R in Minkowski spacetime R ,for m ≥ 2, as well as all causally convex open subsets thereof. The following results follow immediately from Definition 2.9 and Propo- sition 2.8. Proposition 2.11. Let M be a globally hyperbolic spacetime with timelike bound- ary. (a) M admits a cover by causally convex open subsets. (b) For each U ∈R , i.e., a causally convex open subset U ⊆ M, both U and D(U ) are globally hyperbolic spacetimes with (possibly empty) boundary when equipped with the metric, orientation and time-orientation induced by M.Ifmoreover U ⊆ int M is contained in the interior, then both U and D(U ) have empty boundary. 3. Categories of Algebraic Quantum Field Theories Let M be a spacetime with timelike boundary. (In this section we do not have to assume that M is globally hyperbolic in the sense of Definition 2.9). Recall the category R of open and causally convex regions in M and the subset C M M of Cauchy morphisms (cf. Definition 2.7). Together with our notion of causal disjointness from Definition 2.4, these data provide the geometrical input for the traditional definition of algebraic quantum field theories on M . Definition 3.1. An algebraic quantum field theory on M is a functor A : R −→ Alg (3.1) with values in the category Alg of associative and unital ∗-algebras over C, which satisfies the following properties: (i) Causality axiom For all causally disjoint inclusions i : U → U ← U : i , 1 1 2 2 the induced commutator A(i )(−), A(i )(−) : A(U ) ⊗ A(U ) −→ A(U ) (3.2) 1 2 1 2 A(U ) is zero. (ii) Time-slice axiom For all Cauchy morphisms (i : U → U ) ∈C , the map A(i): A(U ) −→ A(U ) (3.3) is an Alg-isomorphism. We denote by qft(M ) ⊆ Alg the full subcategory of the category of func- tors from R to Alg whose objects are all algebraic quantum field theories on M , i.e., functors fulfilling the causality and time-slice axioms. (Morphisms in this category are all natural transformations). M. Benini et al. Ann. Henri Poincar´ e We shall now show that there exists an alternative, but equivalent, de- scription of the category qft(M ) which will be more convenient for the tech- nical constructions in our paper. Following [5, Section 4.1], we observe that the time-slice axiom in Definition 3.1(ii) is equivalent to considering functors −1 B : R C → Alg that are defined on the localization of the category R at the set of Cauchy morphisms C . See Definition A.4 for the defini- M M tion of localizations of categories. By abstract arguments as in [5, Section 4.6], one observes that the universal property of localizations implies that the cat- egory qft(M ) is equivalent to the full subcategory of the functor category −1 R C −1 M [ ] Alg whose objects are all functors B : R C → Alg that satisfy −1 the causality axiom for the pushforward orthogonality relation on R C . Loosely speaking, this means that the time-slice axiom in Definition 3.1(ii) −1 can be hard-coded by working on the localized category R C instead of using the usual category R of regions in M . The aim of the remainder of this section is to provide an explicit model −1 for the localization functor R →R C . With this model, it will be- M M come particularly easy to verify the equivalence between the two alternative descriptions of the category qft(M ). Let us denote by −1 R C ⊆R (3.4a) M M the full subcategory of R whose objects V ⊆ M are stable under Cauchy de- velopment, i.e., D(V )= V where D(V ) ⊆ M denotes the Cauchy development (cf. Definition 2.2). In the following, we shall always use letters like U ⊆ M for generic regions in R and V ⊆ M for regions that are stable under Cauchy −1 development, i.e., objects in R C . Recall from Definition 2.7 that each object U ∈R is a causally convex open subset U ⊆ M , hence the Cauchy de- velopment D(U ) ⊆ M is a causally convex open subset by Proposition 2.8(d), which is stable under Cauchy development by Proposition 2.3(b). This shows −1 that D(U ) is an object of R C . Furthermore, a morphism i : U → U in R is an inclusion U ⊆ U , which induces an inclusion D(U ) ⊆ D(U )by −1 Proposition 2.3(a) and hence a morphism D(i): D(U ) → D(U )in R C . We define a functor −1 D : R −→ R C (3.4b) M M by setting on objects and morphisms U −→ D(U ), (i : U → U ) −→ D(i): D(U ) → D(U ) . (3.4c) Furthermore, let us write −1 I : R C −→ R (3.5) M M for the full subcategory embedding. Lemma 3.2. D and I form an adjunction (cf. Definition A.1) −1 D : R  R C : I (3.6) M M M AQFT on Spacetimes with Timelike Boundary −1 whose counit is a natural isomorphism (in fact, the identity), hence R C is a full reflective subcategory of R . Furthermore, the components of the unit are Cauchy morphisms. Proof. For U ∈R ,the U -component of the unit η :id −→ ID (3.7) is given by the inclusion U ⊆ D(U)of U into its Cauchy development, which is a Cauchy morphism, see Proposition 2.3(b) and Definition 2.7(ii). For V ∈ −1 R C ,the V -component of the counit : DI −→ id −1 (3.8) R C M [ ] is given by the identity of the object D(V )= V . The triangle identities hold trivially. −1 −1 Proposition 3.3. The category R C and the functor D : R →R C M M M M M defined in (3.4) provide a model for the localization of R at C . M M Proof. We have to check all the requirements listed in Definition A.4. (a) By Definition 2.7, for each Cauchy morphism i : U → U one has D(U)= −1 D(U ) and hence D(i)=id is an isomorphism in R C . D(U ) M (b) Let F : R → D be any functor to a category D that sends morphisms in C to D-isomorphisms. Using Lemma 3.2, we define F := FI : M W −1 R C → D and consider the natural transformation Fη : F → FI D = F D obtained by the unit of the adjunction D  I. Because all components of η are Cauchy morphisms (cf. Lemma 3.2), Fη is a natural isomorphism. −1 (c) Let G, H : R C → D be two functors. We have to show that the map Hom G, H −→ Hom R GD, HD (3.9) −1 R C D [ ] is a bijection. Let us first prove injectivity: Let ξ, ξ : G → H be two natural transformations such that ξD = ξD. Using Lemma 3.2, we obtain commutative diagrams ξDI ξDI GDI HDI GDI HDI (3.10) G H G H G H G H where the vertical arrows are natural isomorphisms because the counit is an isomorphism. Recalling that by hypothesis ξD = ξD, it follows that ξ = ξ. Hence, the map (3.9) is injective. M. Benini et al. Ann. Henri Poincar´ e It remains to prove that (3.9) is also surjective. Let χ : GD → HD be any natural transformation. Using Lemma 3.2, we obtain a commutative diagram GD HD (3.11) GDη HDη GDID HDID χID where the vertical arrows are natural isomorphisms because the components of the unit η are Cauchy morphisms and D assigns isomorphisms to them. Let us define a natural transformation ξ : G → H by the commutative diagram G H (3.12) −1 GDI HDI χI where we use that  is a natural isomorphism (cf. Lemma 3.2). Combining the last two diagrams, one easily computes that ξD = χ by using also the triangle identities of the adjunction D  I. Hence, the map (3.9) is surjective. We note that there exist two (a priori different) options to define an −1 orthogonality relation on the localized category R C , both of which are provided by [5, Lemma 4.29]: (1) The pullback orthogonality relation along −1 the full subcategory embedding I : R C →R and (2) the pushforward M M −1 orthogonality relation along the localization functor D : R →R C . M M In our present scenario, both constructions coincide and one concludes that −1 two R C -morphisms are orthogonal precisely when they are orthogonal in R . Summing up, we obtain −1 Lemma 3.4. We say that two morphisms in the full subcategory R C ⊆ R are orthogonal precisely when they are orthogonal in R (i.e., causally M M −1 disjoint, cf. Definition 2.4). Then, both functors D : R →R C and M M −1 I : R C →R preserve (and also detect) the orthogonality relations. M M Proof. For I this holds trivially, while for D see Proposition 2.5(a). With these preparations, we may now define our alternative description of the category of algebraic quantum field theories. −1 R [C ] Definition 3.5. We denote by QFT(M ) ⊆ Alg the full subcategory −1 whose objects are all functors B : R C → Alg that satisfy the following version of the causality axiom: For all causally disjoint inclusions i : V → 1 1 −1 V ← V : i in R C , i.e., V , V and V are stable under Cauchy develop- 2 2 M 1 2 ment, the induced commutator B(i )(−), B(i )(−) : B(V ) ⊗ B(V ) −→ B(V ) (3.13) 1 2 1 2 B(V ) is zero. AQFT on Spacetimes with Timelike Boundary Theorem 3.6. By pullback, the adjunction D  I of Lemma 3.2 induces an adjoint equivalence (cf. Definition A.2) ∗ ∗ I : qft(M )  ∼ QFT(M): D . (3.14) In particular, the two categories qft(M ) of Definition 3.1 and QFT(M ) of Definition 3.5 are equivalent. −1 Proof. It is trivial to check that the adjunction D : R  R C : I M M induces an adjunction −1 ∗ R R [C ] ∗ M M I : Alg  Alg : D (3.15) ∗ ∗ between functor categories. Explicitly, the unit η  :id R → D I has Alg M components ∗ ∗ η  := Aη : A −→ D (I (A)) = AID, (3.16) where A : R → Alg is any functor and η :id → ID denotes the unit of M R ∗ ∗ D  I. The counit   : I D → id has components −1 R C [ ] Alg ∗ ∗ := B : I (D (B)) = BDI −→ B, (3.17) −1 where B : R C → Alg is any functor and  : DI → id −1 denotes M R C M [ ] ∗ ∗ the counit of D  I. The triangle identities for I  D follow directly from those of D  I. Next, we have to prove that this adjunction restricts to the claimed source and target categories in (3.14). Given A ∈ qft(M ) ⊆ Alg , the functor ∗ −1 I (A)= A I : R C → Alg satisfies the causality axiom of Definition 3.5 because of Lemma 3.4. Hence, I (A) ∈ QFT(M ). Vice versa, given B ∈ −1 R C M [ ] ∗ QFT(M ) ⊆ Alg , the functor D (B)= B D : R → Alg satisfies the causality axiom of Definition 3.5 because of Lemma 3.4 and the time-slice axiom of Definition 3.5 because D sends by construction morphisms in C to isomorphisms. Hence, D (B) ∈ qft(M ). Using Lemma 3.2, we obtain that the counit   of the restricted adjunc- tion (3.14) is an isomorphism. Furthermore, all components of η are Cauchy morphisms, hence η  = Aη is an isomorphism for all A ∈ qft(M ), i.e., the unit η  is an isomorphism. This completes the proof that (3.14) is an adjoint equivalence. Remark 3.7. Theorem 3.6 provides us with a constructive prescription of how to change between the two equivalent formulations of algebraic quan- tum field theories given in Definitions 3.1 and 3.5. Concretely, given any A ∈ qft(M ), i.e., a functor A : R → Alg satisfying the causality and time-slice axioms as in Definition 3.1, the corresponding quantum field the- ory I (A) ∈ QFT(M ) in the sense of Definition 3.5 reads as follows: It ∗ −1 is the functor I (A)= A I : R C → Alg on the category of regions V ⊆ M that are stable under Cauchy development, which assigns to V ⊆ M the algebra A(V ) ∈ Alg and to an inclusion i : V → V the algebra map M. Benini et al. Ann. Henri Poincar´ e A(i): A(V ) → A(V ). More interestingly, given B ∈ QFT(M ), i.e., a func- −1 tor B : R C → Alg satisfying the causality axiom as in Definition 3.5, the corresponding quantum field theory D (B) ∈ qft(M ) in the sense of Def- inition 3.1 reads as follows: It is the functor D (B)= B D : R → Alg defined on the category of (not necessarily Cauchy development stable) re- gions U ⊆ M , which assigns to U ⊆ M the algebra B(D(U )) corresponding to the Cauchy development of U and to an inclusion i : U → U the al- gebra map B(D(i)) : B(D(U )) → B(D(U )) associated with the inclusion D(i): D(U ) → D(U ) of Cauchy developments. Remark 3.8. It is straightforward to check that the results of this section still hold true when one replaces R with its full subcategory R of regions M int M contained in the interior of M and C with C (cf. Definition 2.7). This M int M follows from the observation that the Cauchy development of a subset of the interior of M is also contained in int M , as shown in Proposition 2.8(c). We denote by −1 R C int M [ ] int M QFT(int M ) ⊆ Alg (3.18) the category of algebraic quantum field theories in the sense of Definition 3.5 on the interior regions of M . Concretely, an object A ∈ QFT(int M ) is a functor −1 A : R [C ] → Alg that satisfies the causality axiom of Definition 3.5 int M int M for causally disjoint interior regions. 4. Universal Boundary Extension The goal of this section is to develop a universal construction to extend quan- tum field theories from the interior of a spacetime M with timelike boundary to the whole spacetime. (Again, we do not have to assume that M is globally hyperbolic in the sense of Definition 2.9). Loosely speaking, our extended quan- tum field theory will have the following pleasant properties: (1) It describes precisely those observables that are generated from the original theory on the interior, (2) it does not require a choice of boundary conditions, (3) specific choices of boundary conditions correspond to ideals of our extended quantum field theory. We also refer to Sect. 5 for more details on the properties (1) and (3). The starting point for this construction is the full subcategory inclusion R ⊆R defined by selecting only the regions of R that lie in the int M M M interior of M (cf. Definition 2.7). We denote the corresponding embedding functor by j : R −→ R (4.1) int M M and notice that j preserves (and also detects) causally disjoint inclusions, i.e., j is a full orthogonal subcategory embedding in the terminology of [5]. Making use of Proposition 3.3, Lemma 3.2 and Remark 3.8, we define a functor J : −1 −1 R [C ] →R C on the localized categories via the commutative int M M int M M AQFT on Spacetimes with Timelike Boundary diagram −1 −1 R [C ] R [C ] (4.2a) int M M int M M I D R R int M M Notice that J is simply an embedding functor, which acts on objects and morphisms as V ⊆ int M −→ V ⊆ M, (i : V → V ) −→ (i : V → V ). (4.2b) From this explicit description, it is clear that J preserves (and also detects) causally disjoint inclusions, i.e., it is a full orthogonal subcategory embedding. The constructions in [5, Section 5.3] (see also [6] for details how to treat ∗- algebras) then imply that J induces an adjunction ext : QFT(int M )  QFT(M):res (4.3) between the category of quantum field theories on the interior int M (cf. Re- mark 3.8) and the category of quantum field theories on the whole spacetime M . The right adjoint res := J : QFT(M ) → QFT(int M ) is the pullback −1 −1 along J : R [C ] →R C , i.e., it restricts quantum field theories int M M int M M defined on M to the interior int M . The left adjoint ext : QFT(int M ) → QFT(M ) should be regarded as a universal extension functor which extends quantum field theories on the interior int M to the whole spacetime M.The goal of this section is to analyze the properties of this extension functor and to develop an explicit model that allows us to do computations in the sections below. An important structural result, whose physical relevance is explained in Remark 4.2 below, is the following proposition. Proposition 4.1. The unit η :id −→ res ext (4.4) QFT(int M ) of the adjunction (4.3) is a natural isomorphism. Proof. This is a direct consequence of the fact that the functor J given in (4.2) is a full orthogonal subcategory embedding and the general result in [5, Proposition 5.6]. Remark 4.2. The physical interpretation of this result is as follows: Let A ∈ QFT(int M ) be a quantum field theory defined only on the interior int M of M and let B := ext A ∈ QFT(M ) denote its universal extension to the whole spacetime M.The A-component η : A −→ res ext A (4.5) of the unit of the adjunction (4.3) allows us to compare A with the restriction res B of its extension B =ext A. Since η is an isomorphism by Propo- sition 4.1, restricting the extension B recovers our original theory A up to M. Benini et al. Ann. Henri Poincar´ e isomorphism. This allows us to interpret the left adjoint ext : QFT(int M ) → QFT(M ) as a genuine extension prescription. Notice that this also proves that the universal extension ext A ∈ QFT(M ) of any theory A ∈ QFT(int M)on the interior satisfies F-locality [22]. We next address the question how to compute the extension functor ext : QFT(int M ) → QFT(M ) explicitly. A crucial step toward reaching this goal is to notice that ext may be computed by a left Kan extension. Proposition 4.3. Consider the adjunction −1 −1 R [C ]  R [C ] int M M ∗ int M M Lan : Alg Alg : J , (4.6) −1 corresponding to left Kan extension along the functor J : R [C ] → int M int M −1 R C . Then, the restriction of Lan to QFT(int M ) induces a functor M J Lan : QFT(int M ) −→ QFT(M ) (4.7) that is left adjoint to the restriction functor res : QFT(M ) → QFT(int M ) in (4.3). Due to uniqueness (up to unique natural isomorphism) of adjoint functors (cf. Proposition A.3), it follows that ext Lan , i.e., (4.7) is a model for the extension functor ext in (4.3). Proof. A general version of this problem has been addressed in [5, Section 6]. Using in particular [5, Corollary 6.5], we observe that we can prove this propo- −1 sition by showing that every object V ∈R C is J -closed in the sense of [5, Definition 6.3]. In our present scenario, this amounts to proving that for all −1 causally disjoint inclusions i : V → V ← V : i with V ,V ∈R [C ] 1 1 2 2 1 2 int M int M −1 in the interior and V ∈R C not necessarily in the interior, there ex- ists a factorization of both i and i through a common interior region. Let 1 2 us consider the Cauchy development D(V V ) of the disjoint union and 1 2 the canonical inclusions j : V → D(V V ) ← V : j . As we explain be- 1 1 1 2 2 2 −1 low, D(V V ) ∈R [C ] is an interior region and j ,j provide the 1 2 int M 1 2 int M desired factorization: Since the open set V V ⊆ int M is causally con- 1 2 vex by Proposition 2.5(b), D(V V ) is causally convex, open and contained 1 2 in the interior int M by Proposition 2.8(c–d). It is, moreover, stable under Cauchy development by Proposition 2.3(b), which also provides the inclusion V ⊆ V V ⊆ D(V V ) inducing j ,for k =1, 2. Consider now the chain k 1 2 1 2 k of inclusions V ⊆ V V ⊆ V corresponding to i ,for k =1, 2. From the k 1 2 k stability under Cauchy development of V , V and V , we obtain also the chain 1 2 of inclusions V ⊆ D(V V ) ⊆ V ,for k =1, 2, that exhibits the desired k 1 2 factorization V (4.8) i i 1  2 D(i i ) 1 2 V D(V V ) V 1 1 2 2 j j 1 2 which completes the proof.  AQFT on Spacetimes with Timelike Boundary We shall now briefly review a concrete model for left Kan extension along full subcategory embeddings that was developed in [5, Section 6]. This model is obtained by means of abstract operadic techniques, but it admits an intuitive graphical interpretation that we explain in Remark 4.4 below. It allows us to compute quite explicitly the extension ext A =Lan A ∈ QFT(M)ofa quantum field theory A ∈ QFT(int M ) defined on the interior int M to the −1 whole spacetime M . The functor ext A : R C → Alg describing the −1 extended quantum field theory reads as follows: To V ∈R C it assigns a quotient algebra ext A(V )= A(V ) ∼ (4.9) i:V →V that we will describe now in detail. The direct sum (of vector spaces) in (4.9) −1 runs over all tuples i : V → V of morphisms in R C , i.e., i =(i : V → M 1 1 V,...,i : V → V ) for some n ∈ Z , with the requirement that all sources n n ≥0 −1 V ∈R [C ] are interior regions. (Notice that the regions V are not k int M k int M assumed to be causally disjoint and that the empty tuple, i.e., n = 0, is also allowed). The vector space A(V ) is defined by the tensor product of vector spaces |V | A(V ):= A(V ), (4.10) k=1 where |V | is the length of the tuple. (For the empty tuple, we set A(∅)= C). This means that the (homogeneous) elements (i,a) ∈ A(V ) (4.11) i:V →V are given by pairs consisting of a tuple of morphisms i : V → V (with all V in the interior) and an element a ∈ A(V ) of the corresponding tensor product vector space (4.10). The product on (4.11) is given on homogeneous elements by (i,a)(i ,a ):= (i,i ),a ⊗ a , (4.12) where (i,i )=(i ,...,i ,i ,...,i ) is the concatenation of tuples. The unit 1 n 1 m element in (4.11)is 1 := (∅, 1), where ∅ is the empty tuple and 1 ∈ C, and the ∗-involution is defined by ∗ ∗ (i ,...,i ),a ⊗ ··· ⊗ a := (i ,...,i ),a ⊗ ··· ⊗ a (4.13) 1 n 1 n n 1 n 1 and C-antilinear extension. Finally, the quotient in (4.9) is by the two-sided ∗-ideal of the algebra (4.11) that is generated by i i ,...,i ,a ⊗ ··· ⊗ a − i, A(i ) a ⊗ ··· ⊗ A(i ) a ∈ A(V ) , 1 n 1 n 1 1 n n i:V →V (4.14) for all tuples i : V → V of length |V | = n ≥ 1 (with all V in the interior), all −1 tuples i : V → V of R [C ]-morphisms (possibly of length zero), for k int M k k int M M. Benini et al. Ann. Henri Poincar´ e k =1,...,n, and all a ∈ A(V ), for k =1,...,n. The tuple in the first term k k of (4.14) is defined by composition i i ,...,i := i i ,...,i i ,...,i i ,...,i i (4.15a) 1 11 1 1|V | n n1 n n|V | 1 n 1 n and the expressions A(i) a in the second term are determined by A(i): A(V ) −→ A(V ),a ⊗ ··· ⊗ a −→ A(i ) a ··· A(i ) a , (4.15b) 1 n 1 1 n n where multiplication in A(V ) is denoted by juxtaposition. To a morphism −1 −1 i : V → V in R C , the functor ext A : R C → Alg assigns the M M M M algebra map ext A(i ) : ext A(V ) −→ ext A(V ), [i,a] −→ i (i),a , (4.16) where we used square brackets to indicate equivalence classes in (4.9). Remark 4.4. The construction of the algebra ext A(V ) above admits an intu- itive graphical interpretation: We shall visualize the (homogeneous) elements (i,a)in(4.11) by decorated trees ··· a a 1 n (4.17) where a ∈ A(V ) is an element of the algebra A(V ) associated to the interior k k k region V ⊆ V , for all k =1,...,n. We interpret such a decorated tree as a formal product of the formal pushforward along i : V → V of the family of observables a ∈ A(V ). The product (4.12) is given by concatenation of the k k inputs of the individual decorated trees, i.e., V V V ··· ··· ··· a a a 1 n 1 a a a 1 m m (4.18) where the decorated tree on the right-hand side has n + m inputs. The ∗- involution (4.13) may be visualized by reversing the input profile and applying ∗ to each element a ∈ A(V ). Finally, the ∗-ideal in (4.14) implements the fol- k k lowing relations: Assume that (i,a) is such that the sub-family of embeddings (i ,i ,...,i ):(V ,V ,...,V ) → V factorizes through some common in- k k+1 l k k+1 l terior region, say V ⊆ V . Using the original functor A ∈ QFT(int M ), we may form the product A(i )(a ) ··· A(i )(a ) in the algebra A(V ), which we k k l l AQFT on Spacetimes with Timelike Boundary denote for simplicity by a ··· a ∈ A(V ). We then have the relation k l V V ··· · ·· ··· ··· · ·· a a a a a a ··· a a 1 k l n 1 k l n (4.19) which we interpret as follows: Whenever (i ,i ,...,i ):(V ,V ,...,V ) → k k+1 l k k+1 l V is a sub-family of embeddings that factorizes through a common interior re- gion V ⊆ V , then the formal product of the formal pushforward of observables is identified with the formal pushforward of the actual product of observables on V . 5. Characterization of Boundary Quantum Field Theories In the previous section, we established a universal construction that allows us to extend quantum field theories A ∈ QFT(int M ) that are defined only on the interior int M of a spacetime M with timelike boundary to the whole spacetime. The extension ext A ∈ QFT(M ) is characterized abstractly by the ext  res adjunction in (4.3). We now shall reverse the question and ask which quantum field theories B ∈ QFT(M)on M admit a description in terms of (quotients of) our universal extensions. Given any quantum field theory B ∈ QFT(M ) on the whole spacetime M , we can use the right adjoint in (4.3) in order to restrict it to a theory res B ∈ QFT(int M ) on the interior of M . Applying now the extension func- tor, we obtain another quantum field theory ext res B ∈ QFT(M)onthe whole spacetime M , which we would like to compare to our original theory B ∈ QFT(M ). A natural comparison map is given by the B-component of the counit  :extres → id of the adjunction (4.3), i.e., the canonical QFT(M ) QFT(M )-morphism :extres B −→ B. (5.1a) Using our model for the extension functor given in (4.9) (and the formulas −1 following this equation), the R C  V -component of  explicitly reads M B as ( ) : B(V ) ∼−→ B(V ), [i,b] −→ B(i) b . (5.1b) B V i:V →V In order to establish positive comparison results, we have to introduce the concept of ideals of quantum field theories. Definition 5.1. An ideal I ⊆ B of a quantum field theory B ∈ QFT(M)isa −1 functor I : R C → Vec to the category of complex vector spaces, which satisfies the following properties: M. Benini et al. Ann. Henri Poincar´ e −1 (i) For all V ∈R C , I(V ) ⊆ B(V ) is a two-sided ∗-ideal of the unital ∗-algebra B(V ). −1 (ii) For all R C -morphisms i : V → V , the linear map I(i): I(V ) → I(V ) is the restriction of B(i): B(V ) → B(V ) to the two-sided ∗-ideals () () I(V ) ⊆ B(V ). Lemma 5.2. Let B ∈ QFT(M ) and I ⊆ B any ideal. Let us define B/I(V ):= −1 B(V )/I(V ) to be the quotient algebra, for all V ∈R C ,and B/I(i): B/I(V ) → B/I(V ) to be the Alg-morphism induced by B(i): B(V ) → −1 B(V ), for all R C -morphisms i : V → V .Then B/I ∈ QFT(M ) is a quantum field theory on M which we call the quotient of B by I. −1 Proof. The requirements listed in Definition 5.1 ensure that B/I : R C → Alg is an Alg-valued functor. It satisfies the causality axiom of Definition 3.5 because this property is inherited from B ∈ QFT(M ) by taking quotients. Lemma 5.3. Let κ : B → B be any QFT(M )-morphism. Define the vec- tor space ker κ(V):=ker κ : B(V ) → B (V ) ⊆ B(V ), for all V ∈ −1 R C ,and ker κ(i):ker κ(V ) → ker κ(V ) to be the linear map induced −1 by B(i): B(V ) → B(V ), for all R C -morphisms i : V → V .Then −1 ker κ : R C → Vec is an ideal of B ∈ QFT(M ), which we call the kernel of κ. Proof. The fact that ker κ defines a functor follows from naturality of κ.Prop- erty (ii) in Definition 5.1 holds true by construction. Property (i) is a conse- quence of the fact that kernels of unital ∗-algebra morphisms κ : B(V ) → B(V ) are two-sided ∗-ideals. Remark 5.4. Using the concept of ideals, we may canonically factorize (5.1) according to the diagram ext res B B (5.2) ext res B ker where both the projection π and the inclusion λ are QFT(M )-morphi- B B sms. As a last ingredient for our comparison result, we have to introduce a suitable notion of additivity for quantum field theories on spacetimes with timelike boundary. We refer to [14, Definition 2.3] for a notion of additivity on globally hyperbolic spacetimes. Definition 5.5. A quantum field theory B ∈ QFT(M ) on a spacetime M with timelike boundary is called additive (from the interior) at the object −1 V ∈R C if the algebra B(V ) is generated by the images of the Alg- −1 morphisms B(i ): B(V ) → B(V ), for all R C -morphism i : V → int int M int int −1 V whose source V ∈R [C ] is in the interior int M of M . We call int int M int M B ∈ QFT(M ) additive (from the interior) if it is additive at every object AQFT on Spacetimes with Timelike Boundary −1 V ∈R C . The full subcategory of additive quantum field theories on M add is denoted by QFT (M ) ⊆ QFT(M ). We can now prove our first characterization theorem for boundary quan- tum field theories. Theorem 5.6. Let B ∈ QFT(M ) be any quantum field theory on a (not necessarily globally hyperbolic) spacetime M with timelike boundary and let −1 V ∈R C . Then the following are equivalent: (1) The V -component (λ ) :extres B(V ) ker  (V ) −→ B(V ) (5.3) B V B of the canonical inclusion in (5.2) is an Alg-isomorphism. −1 (2) B is additive at the object V ∈R C . −1 Proof. Let i : V → V be any R C -morphism whose source V ∈ int int M int −1 R [C ] is in the interior int M of M . Using our model for the extension int M int M functor given in (4.9) (and the formulas following this equation), we obtain an Alg-morphism [i , −]: B(V ) −→ ext res B(V ),b −→ [i ,b]. (5.4) int int int Composing this morphism with the V -component of  given in (5.1), we obtain a commutative diagram ( ) B V ext res B(V ) B(V ) (5.5) [i ,−] B(i ) int  int B(V ) int for all i : V → V with V in the interior. int int int Next we observe that the images of the Alg-morphisms (5.4), for all i : V → V with V in the interior, generate ext res B(V ). Combining int int int this property with (5.5), we conclude that ( ) is a surjective map if and B V only if B is additive at V . Hence, (λ ) given by (5.2), which is injective by B V construction, is an Alg-isomorphism if and only if B is additive at V . Corollary 5.7. λ :extres B ker  → B given by (5.2) is a QFT(M )- B B add isomorphism if and only if B ∈ QFT (M ) ⊆ QFT(M ) is additive in the sense of Definition 5.5. We shall now refine this characterization theorem by showing that add QFT (M ) is equivalent, as a category, to a category describing quantum field theories on the interior of M together with suitable ideals of their uni- versal extensions. The precise definitions are as follows. Definition 5.8. Let B ∈ QFT(M ). An ideal I ⊆ B is called trivial on the −1 interior if its restriction to R [C ] is the functor assigning zero vector int M int M −1 spaces, i.e., I(V ) = 0 for all V ∈R [C ] in the interior int M of M . int int int M int M M. Benini et al. Ann. Henri Poincar´ e Definition 5.9. Let M be a spacetime with timelike boundary. We define the category IQFT(M ) as follows: • Objects are pairs (A, I) consisting of a quantum field theory A ∈ QFT(int M ) on the interior int M of M and an ideal I ⊆ ext A of its universal extension ext A ∈ QFT(M ) that is trivial on the interior. • Morphisms κ :(A, I) → (A , I )are QFT(M )-morphisms κ :ext A → ext A between the universal extensions that preserve the ideals, i.e., κ restricts to a natural transformation from I ⊆ ext A to I ⊆ ext A . There exists an obvious functor add Q : IQFT(M ) −→ QFT (M ), (5.6a) which assigns to (A, I) ∈ IQFT(M ) the quotient add Q(A, I):=ext A I ∈ QFT (M ). (5.6b) Notice that additivity of ext A I follows from that of the universal exten- sion ext A (cf. the arguments in the proof of Theorem 5.6) and the fact that quotients preserve the additivity property. There exists also a functor add S : QFT (M ) −→ IQFT(M ), (5.7a) which ‘extracts’ from a quantum field theory on M the relevant ideal. Explic- add itly, it assigns to B ∈ QFT (M ) the pair SB := res B, ker  . (5.7b) Notice that the ideal ker  ⊆ ext res B is trivial on the interior: Applying the restriction functor (4.3)to  , we obtain a QFT(int M )-morphism res  : res ext res B −→ res B. (5.8) Proposition 4.1 together with the triangle identities for the adjunction ext res in (4.3) then imply that (5.8) is an isomorphism with inverse given by η :res B → res ext res B. In particular, res  has a trivial kernel and res B B hence ker  ⊆ ext res B is trivial on the interior (cf. Definition 5.8). Our refined characterization theorem for boundary quantum field theories is as follows. Theorem 5.10. The functors Q and S defined in (5.6) and (5.7) exhibit an equivalence of categories add QFT (M ) IQFT(M ). (5.9) add Proof. We first consider the composition of functors QS : QFT (M ) → add add QFT (M ). To B ∈ QFT (M ), it assigns QSB =ext res B ker  . (5.10) The QFT(M )-morphisms λ :ext res B/ ker  → B given by (5.2) define a B B natural transformation λ : QS → id add , which is a natural isomorphism QFT (M ) due to Corollary 5.7. AQFT on Spacetimes with Timelike Boundary Let us now consider the composition of functors SQ : IQFT(M ) → IQFT(M ). To (A, I) ∈ IQFT(M ), it assigns SQ(A, I)= res ext A I , ker  = res ext A, ker  , (5.11) ext A/I ext A/I where we also used that res ext A I = res ext A because I is by hypothesis trivial on the interior. Using further the QFT(int M )-isomorphism η : A → res ext A from Proposition 4.1, we define a QFT(M )-morphism q via the (A,I) diagram (A,I) ext A ext A I (5.12) ext η A = ext A/I ext res ext A ext res ext A I Using the explicit expression for  given in (5.1) and the explicit expres- ext A/I sion for η given by (η ) : A(V ) −→ res ext A(V )= A(V ) ∼, int int A V int i:V →V int a −→ [id ,a], (5.13) int −1 for all V ∈R [C ], one computes from the diagram (5.12) that q is int int M (A,I) int M the canonical projection π :ext A → ext A/I. Hence, the QFT(M )- isomorphisms ext η :ext A → ext res ext A induce IQFT(M )-isomorphisms ext η :(A, I) −→ SQ(A, I), (5.14) which are natural in (A, I), i.e., they define a natural isomorphism ext η : id → SQ. IQFT(M ) Remark 5.11. The physical interpretation of this result is as follows: Every add additive quantum field theory B ∈ QFT (M ) on a (not necessarily glob- ally hyperbolic) spacetime M with timelike boundary admits an equivalent description in terms of a pair (A, I) ∈ IQFT(M ). Notice that the roles of A and I are completely different: On the one hand, A ∈ QFT(int M)isa quantum field theory on the interior int M of M and as such it is independent of the detailed aspects of the boundary. On the other hand, I ⊆ ext A is an ideal of the universal extension of A that is trivial on the interior, i.e., it only captures the physics that happens directly at the boundary. Examples of such ideals I arise by imposing specific boundary conditions on the universal ex- tension ext A ∈ QFT(M ), i.e., the quotient ext A/I describes a quantum field theory on M that satisfies specific boundary conditions encoded in I. We shall illustrate this assertion in Sect. 6 below using the explicit example given by the free Klein–Gordon field. Let us also note that there is a reason why our universal extension cap- tures only the class of additive quantum field theories on M . Recall that ext A ∈ QFT(M ) takes as an input a quantum field theory A ∈ QFT(int M ) on the interior int M of M . As a consequence, the extension ext A can only M. Benini et al. Ann. Henri Poincar´ e have knowledge of the ‘degrees of freedom’ that are generated in some way out of the interior regions. Additive theories in the sense of Definition 5.5 are precisely the theories whose ‘degrees of freedom’ are generated out of those localized in the interior regions. 6. Example: Free Klein–Gordon Theory In order to illustrate and make more explicit our abstract constructions de- veloped in the previous sections, we shall consider the simple example given by the free Klein–Gordon field. From now on M will be a globally hyperbolic spacetime with timelike boundary, see Definition 2.9. This assumption implies that all interior regions R are globally hyperbolic spacetimes with empty int M boundary, see Proposition 2.11. This allows us to use the standard techniques of [1, Section 3] on such regions. −1 6.1. Definition on R [C ]: int M int M Let M be a globally hyperbolic spacetime with timelike boundary, see Def- −1 inition 2.9. The free Klein–Gordon theory on R [C ] is given by the int M int M following standard construction, see e.g., [3, 4] for expository reviews. On the interior int M , we consider the Klein–Gordon operator 2 ∞ ∞ P :=  + m : C (int M ) −→ C (int M ) , (6.1) where  is the d’Alembert operator and m ≥ 0 is a mass parameter. When −1 restricting P to regions V ∈R [C ], we shall write int M int M ∞ ∞ P : C (V ) −→ C (V ). (6.2) It follows from [1] that there exists a unique retarded/advanced Green’s oper- ator ± ∞ ∞ G : C (V ) −→ C (V ) (6.3) V c −1 for P because every V ∈R [C ] is a globally hyperbolic spacetime V int M int M with empty boundary, cf. Proposition 2.11. The Klein–Gordon theory K ∈ QFT(int M ) is the functor −1 K : R [C ] → Alg given by the following assignment: To any V ∈ int M int M −1 R [C ] it assigns the associative and unital ∗-algebra K(V ) that is freely int M int M generated by Φ (f ), for all f ∈ C (V ), modulo the two-sided ∗-ideal gener- ated by the following relations: • Linearity Φ (αf + βg)= α Φ (f)+ β Φ (g), for all α, β ∈ R and V V V f, g ∈ C (V ); ∗ ∞ • Hermiticity Φ (f ) =Φ (f ), for all f ∈ C (V ); V V • Equation of motion Φ (P f ) = 0, for all f ∈ C (V ); V V • Canonical commutation relations (CCR) Φ (f)Φ (g) − Φ (g)Φ (f ) V V V V =i τ (f, g) 1, for all f, g ∈ C (V ), where τ (f, g):= fG (g)vol (6.4) V V V V AQFT on Spacetimes with Timelike Boundary + − with G := G − G the causal propagator and vol the canonical vol- V V V V ume form on V . (Note that τ is antisymmetric, see e.g., [1, Lemma 4.3.5]). −1 −1 To a morphism i : V → V in R [C ], the functor K : R [C ] → int M int M int M int M Alg assigns the algebra map that is specified on the generators by pushforward along i (which we shall suppress) K(i): K(V ) −→ K(V ), Φ (f ) −→ Φ  (f ). (6.5) V V The naturality of τ (i.e., naturality of the causal propagator, cf. e.g., [1, Sec- tion 4.3]) entails that the assignment K defines a quantum field theory in the sense of Definition 3.5. 6.2. Universal Extension Using the techniques developed in Sect. 4, we may now extend the Klein– Gordon theory K ∈ QFT(int M ) from the interior int M to the whole space- time M . In particular, using (4.9) (and the formulas following this equa- tion), one could directly compute the universal extension ext K ∈ QFT(M ). The resulting expressions, however, can be considerably simplified. We there- fore prefer to provide a more convenient model for the universal extension ext K ∈ QFT(M ) by adopting the following strategy: We first make an ‘ed- ext ucated guess’ for a theory K ∈ QFT(M ) which we expect to be the uni- versal extension of K ∈ QFT(int M ). (This was inspired by partially simpli- fying the direct computation of the universal extension). After this, we shall ext prove that K ∈ QFT(M ) satisfies the universal property that characterizes ext ext K ∈ QFT(M ). Hence, there exists a (unique) isomorphism ext K = K in ext QFT(M ), which means that our K ∈ QFT(M ) is a model for the universal extension ext K. −1 ext Let us define the functor K : R C → Alg by the following assign- −1 ment: To any region V ∈R C , which may intersect the boundary, we ext assign the associative and unital ∗-algebra K (V ) that is freely generated by Φ (f ), for all f ∈ C (int V ) in the interior int V of V , modulo the two-sided ideal generated by the following relations: • Linearity Φ (αf + βg)= α Φ (f)+ β Φ (g), for all α, β ∈ R and V V V f, g ∈ C (int V ); ∗ ∞ • Hermiticity Φ (f ) =Φ (f ), for all f ∈ C (int V ); V V • Equation of motion Φ (P f ) = 0, for all f ∈ C (int V ); V int V • Partially-defined CCR Φ (f)Φ (g) − Φ (g)Φ (f)=i τ (f, g) 1,for V V V V V int −1 all interior regions V ∈R [C ] with V ⊆ int V and f, g ∈ int int M int int M C (int V ) with supp(f ) ∪ supp(g) ⊆ V . int Remark 6.1. We note that our partially-defined CCR are consistent in the () −1 following sense: Consider V ,V ∈R [C ] with V ⊆ int V and int int M int int M int () f, g ∈ C (int V ) with the property that supp(f ) ∪ supp(g) ⊆ V .Us- c int ing the partially-defined CCR for both V and V , we obtain the equality int int ext ext i τ (f, g) 1 =i τ  (f, g) 1 in K (V ). To ensure that K (V ) is not the V V int int zero-algebra, we have to show that τ (f, g)= τ (f, g). This holds true V V int int due to the following argument: Consider the subset V ∩ V ⊆ int M . This int int M. Benini et al. Ann. Henri Poincar´ e is open, causally convex and by Proposition 2.3(c) also stable under Cauchy −1 development, hence V ∩ V ∈R [C ]. Furthermore, the inclusions int int M int int M () −1 V ∩ V → V are morphisms in R [C ]. It follows by construction int int M int int int M that supp(f ) ∪ supp(g) ⊆ V ∩ V and hence due to naturality of the τ ’s we int int obtain τ (f, g)= τ (f, g)= τ (f, g). (6.6) V V ∩V V int int int int Hence, for any fixed pair f, g ∈ C (int V ), the partially-defined CCR are independent of the choice of V (if one exists). int −1 −1 ext To a morphism i : V → V in R [C ], the functor K : R C → M M M M Alg assigns the algebra map that is specified on the generators by the push- forward along i (which we shall suppress) ext ext ext K (i): K (V ) −→ K (V ), Φ (f ) −→ Φ (f ). (6.7) V V ext Compatibility of the map (6.7) with the relations in K is a straightforward check. −1 −1 Recalling the embedding functor J : R [C ] →R C given in int M M int M M (4.2), we observe that the diagram of functors −1 R [C ] Alg (6.8) int M int M ext J K −1 R [C ] ext commutes via the natural transformation γ : K → K J with components specified on the generators by the identity maps ext γ : K(V ) −→ K (V ), Φ (f ) −→ Φ (f ), (6.9) int int V V V int int int −1 for all V ∈R [C ]. Notice that γ is a natural isomorphism because int int M int M int V = V and the partially-defined CCR on any interior region V coin- int int int cides with the CCR. −1 Theorem 6.2. (6.8) is a left Kan extension of K : R [C ] → Alg along int M int M −1 −1 J : R [C ] →R C . As a consequence of uniqueness (up to unique int M M int M M natural isomorphism) of left Kan extensions and Proposition 4.3,itfollows ext ext that K ext K, i.e., K ∈ QFT(M ) is a model for our universal extension ext K ∈ QFT(M ) of the Klein–Gordon theory K ∈ QFT(int M ). Proof. We have to prove that (6.8) satisfies the universal property of left Kan −1 extensions: Given any functor B : R C → Alg and natural transfor- mation ρ : K → B J , we have to show that there exists a unique natural ext transformation ζ : K → B such that the diagram K B J (6.10) ζJ ext K J AQFT on Spacetimes with Timelike Boundary commutes. Because γ is a natural isomorphism, it immediately follows that ζJ is uniquely fixed by this diagram. Concretely, this means that the components −1 ζ corresponding to interior regions V ∈R [C ] are uniquely fixed V int int M int int M by −1 ext ζ := ρ γ : K (V ) −→ B(V ). (6.11) int int V V V int int int It remains to determine the components ext ζ : K (V ) −→ B(V ) (6.12) −1 ext for generic regions V ∈R C . Consider any generator Φ (f)of K (V ), M V where f ∈ C (int V ), and choose a finite cover {V ⊆ int V } of supp(f)by −1 interior regions V ∈R [C ], together with a partition of unity {χ } α int M α int M subordinate to this cover. (The existence of such a cover is guaranteed by the assumption that M is a globally hyperbolic spacetime with timelike boundary, see Proposition 2.11). We define ζ Φ (f ) := B(i ) ζ Φ (χ f ) , (6.13) V V α V V α α α where i : V → V is the inclusion. Our definition (6.13) is independent of the α α choice of cover and partition of unity: For any other {V ⊆ int V } and {χ }, β β we obtain B(i ) ζ Φ (χ f ) = B(i ) ζ Φ (χ χ f ) β V V αβ V ∩V V ∩V α β α α β β β β β β α,β = B(i ) ζ Φ (χ f ) , (6.14) α V V α α α where i : V → V and i : V ∩ V → V are the inclusions. In particular, this β αβ α β β implies that (6.13) coincides with (6.11) on interior regions V = V .(Hint: int Choose the cover given by the single region V together with its partition of int unity). ext We have to check that (6.13) preserves the relations in K (V ). Preserva- tion of linearity and Hermiticity is obvious. The equation of motion relations are preserved because ζ Φ (P f ) = B(i ) ζ Φ (χ P f ) V V α V V α int V α α V = B(i ) ζ Φ (χ P χ f ) αβ V ∩V V ∩V α β α β α β V ∩V α β α,β = B(i ) ζ Φ (P χ f ) =0. (6.15) β V V β β β V −1 Regarding the partially-defined CCR, let V ∈R [C ] with V ⊆ int int M int int M int V and f, g ∈ C (int V ) with supp(f ) ∪ supp(g) ⊆ V .Wemay choose int the cover given by the single region i : V → V together with its partition of int M. Benini et al. Ann. Henri Poincar´ e unity. We obtain for the commutator ζ Φ (f ) ,ζ Φ (g) = B(i) ζ Φ (f ) , B(i) ζ Φ (g) V V V V V V V V int int int int =i τ (f, g) 1, (6.16) int which implies that the partially-defined CCR are preserved. Naturality of the components (6.13) is easily verified. Uniqueness of the ext resulting natural transformation ζ : K → B is a consequence of uniqueness of ext ext ζJ and of the fact that the Alg-morphisms K (i ): K (V ) K(V ) → int int int ext ext K (V ), for all interior regions i : V → V , generate K (V ), for all V ∈ int int −1 R C . This completes the proof. 6.3. Ideals from Green’s Operator Extensions The Klein–Gordon theory K ∈ QFT(int M ) on the interior int M of the glob- ally hyperbolic spacetime M with timelike boundary and its universal exten- ext sion K ∈ QFT(M ) depend on the local retarded and advanced Green’s oper- ± −1 ∞ ∞ ators G : C (V ) → C (V ) on all interior regions V ∈R [C ] int int int M c int V int M int in M . For constructing concrete examples of quantum field theories on globally hyperbolic spacetimes with timelike boundary as in [12], one typically imposes suitable boundary conditions for the field equation in order to obtain also global retarded and advanced Green’s operators on M . Inspired by such examples, we + − shall now show that any choice of an adjoint-related pair (G ,G ) consisting of a retarded and an advanced Green’s operator for P on M (see Definition 6.3 ext below) defines an ideal I ± ⊆ K ∈ QFT(M ) that is trivial on the interior ext (cf. Definition 5.8). The corresponding quotient K I ± ∈ QFT(M ) then may be interpreted as the Klein–Gordon theory on M , subject to a specific choice of boundary conditions that is encoded in G . Definition 6.3. A retarded/advanced Green’s operator for the Klein–Gordon ± ∞ ∞ operator P on M is a linear map G : C (int M ) → C (int M ) which satisfies the following properties, for all f ∈ C (int M ): (i) PG (f)= f , (ii) G (Pf)= f,and (iii) supp(G (f )) ⊆ J (supp(f )). + − A pair (G ,G ) consisting of a retarded and an advanced Green’s operator + − for P on M is called adjoint-related if G is the formal adjoint of G , i.e., + − G (f ) g vol = fG (g)vol , (6.17) M M M M for all f, g ∈ C (int M ). Remark 6.4. In contrast to the situation where M is a globally hyperbolic spacetime with empty boundary [1], existence, uniqueness and adjoint- relatedness of retarded/advanced Green’s operators for the Klein–Gordon op- erator P is in general not to be expected on spacetimes with timelike boundary. Positive results seem to be more likely on globally hyperbolic spacetimes with AQFT on Spacetimes with Timelike Boundary non-empty timelike boundary, although the general theory has not been de- veloped yet to the best of our knowledge. Simple examples of adjoint-related pairs of Green’s operators were constructed, e.g., in [12]. −1 Given any region V ∈R C in M , which may intersect the boundary, we use the canonical inclusion i : V → M to define local retarded/advanced Green’s operators ∞ ∞ C (int V ) C (int V ) (6.18) ∗ i ∞ ∞ C (int M ) C (int M ) where i denotes the pushforward of compactly supported functions (i.e., ex- tension by zero) and i the pullback of functions (i.e., restriction). Since V ⊆ M ± ± is causally convex, it follows that J (p) ∩ V = J (p) for all p ∈ V . Therefore M V G satisfies the axioms of a retarded/advanced Green’s operator for P on V . V V (Here, we regard V as a globally hyperbolic spacetime with timelike bound- ary, see Proposition 2.11. J (p) denotes the causal future/past of p in the −1 spacetime V ). In particular, for all interior regions V ∈R [C ]in M , int int M int M by combining Proposition 2.11 and [1, Corollary 3.4.3] we obtain that G int as defined in (6.18) is the unique retarded/advanced Green’s operator for the restricted Klein–Gordon operator P on the globally hyperbolic spacetime int V with empty boundary. int + − Consider any adjoint-related pair (G ,G ) of Green’s operator for P on −1 ext M . For all V ∈R C ,weset I ± (V ) ⊆ K (V ) to be the two-sided ∗-ideal generated by the following relations: ± ∞ • G -CCR Φ (f)Φ (g) − Φ (g)Φ (f)=i τ (f, g) 1, for all f, g ∈ C V V V V V (int V ), where τ (f, g):= fG (g)vol (6.19) V V V + − with G := G − G the causal propagator and vol the canonical V V V V volume form on V . + − The fact that the pair (G ,G ) is adjoint-related (cf. Definition 6.3) implies −1 that for all V ∈R C the causal propagator G is formally skew-adjoint, M V hence τ is antisymmetric. ext Proposition 6.5. I ± ⊆ K is an ideal that is trivial on the interior (cf. Definition 5.8). −1 Proof. Functoriality of I ± : R C → Vec is a consequence of (6.18), G M ext hence I ± ⊆ K is an ideal in the sense of Definition 5.1. It is trivial on the −1 interior because for all interior regions V ∈R [C ], the Green’s op- int M int int M erators defined by (6.18) are the unique retarded/advanced Green’s operators for P and hence the relations imposed by I ± (V ) automatically hold true G int int ext in K (V ) on account of the (partially-defined) CCR. int M. Benini et al. Ann. Henri Poincar´ e Remark 6.6. We note that the results of this section still hold true if we slightly weaken the hypotheses of Definition 2.9 by assuming the strong causality and the compact double-cones property only for points in the interior int M of M . In fact, int M can still be covered by causally convex open subsets and any causally convex open subset U ⊆ int M becomes a globally hyperbolic space- time with empty boundary once equipped with the induced metric, orientation and time-orientation. m−1 m Example 6.7. Consider the sub-spacetime M := R × [0,π] ⊆ R of m- dimensional Minkowski spacetime, which has a timelike boundary ∂M = m−1 + − R ×{0,π}. The constructions in [12] define an adjoint-related pair (G ,G ) of Green’s operators for P on M that corresponds to Dirichlet boundary conditions. Using this as an input for our construction above, we obtain a ext quantum field theory K I ± ∈ QFT(M ) that may be interpreted as the Klein–Gordon theory on M with Dirichlet boundary conditions. It is worth to emphasize that our theory in general does not coincide with the one con- structed in [12]. To provide a simple argument, let us focus on the case of m = 2 dimensions, i.e., M = R × [0,π], and compare our global algebra BDS ext DNP A (M):= K I ± (M ) with the global algebra A (M ) constructed in [12]. Both algebras are CCR-algebras, however the underlying symplec- tic vector spaces differ: The symplectic vector space underlying our global BDS ∞ ∞ algebra A (M)is C (int M ) PC (int M ) with the symplectic structure c c (6.19). Using that the spatial slices of M = R × [0,π] are compact, we observe DNP that the symplectic vector space underlying A (M ) is given by the space Sol (M)of all solutions with Dirichlet boundary condition on M (equipped Dir with the usual symplectic structure). The causal propagator defines a sym- ∞ ∞ plectic map G : C (int M ) PC (int M ) → Sol (M ), which however is not Dir c c surjective for the following reason: Any ϕ ∈ C (int M ) has by definition com- pact support in the interior of M , hence the support of Gϕ ∈ Sol (M)is Dir schematically as follows supp Gϕ supp ϕ x = π x =0 (6.20) 2 2 The usual mode functions Φ (t, x)=cos( k + m t)sin(kx) ∈ Sol (M ), k Dir ∞ ∞ for k ≥ 1, are clearly not of this form, hence G : C (int M ) PC (int M ) → c c Sol (M ) cannot be surjective. As a consequence, the models constructed in Dir ext [12] are in general not additive from the interior and our construction K I ± should be interpreted as the maximal additive subtheory of these examples. AQFT on Spacetimes with Timelike Boundary It is interesting to note that there exists a case where both constructions m−1 m coincide: Consider the sub-spacetime M := R × [0, ∞) ⊆ R of Minkowski spacetime with m ≥ 4 even and take a massless real scalar field with Dirichlet boundary conditions. Using Huygens’ principle and the support properties of BDS the Green’s operators, one may show that our algebra A (M ) is isomorphic to the construction in [12]. Acknowledgements We would like to thank Jorma Louko for useful discussions about boundaries in quantum field theory and also Didier A. Solis for sending us a copy of his PhD thesis [29]. The work of M. B. is supported by a research grant funded by the Deutsche Forschungsgemeinschaft (DFG, Germany). Furthermore, M. B. would like to thank Istituto Nazionale di Alta Matematica (INdAM, Italy) for the financial support provided during his visit to the School of Mathematical Sciences of the University of Nottingham, whose kind hospitality is gratefully acknowledged. The work of C. D. is supported by the University of Pavia. C. D. gratefully acknowledges the kind of hospitality of the School of Mathematical Sciences of the University of Nottingham, where part of this work has been done. A. S. gratefully acknowledges the financial support of the Royal Society (UK) through a Royal Society University Research Fellowship, a Research Grant and an Enhancement Award. He also would like to thank the Research Training Group 1670 “Mathematics inspired by String Theory and Quantum Field Theory” for the invitation to Hamburg, where part of this work has been done. Open Access. This article is distributed under the terms of the Creative Com- mons Attribution 4.0 International License (http://creativecommons.org/licenses/ by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Appendix A: Some Concepts from Category Theory Adjunctions This is a standard concept, which is treated in any category theory textbook, e.g., [28]. Definition A.1. An adjunction consists of a pair of functors F : C D : G (A.1) M. Benini et al. Ann. Henri Poincar´ e together with natural transformations η :id → GF (called unit)and  : FG → id (called counit ) that satisfy the triangle identities Fη ηG F FGF G GF G (A.2) F G id id F G F G We call F the left adjoint of G and G the right adjoint of F , and write F  G. Definition A.2. An adjoint equivalence is an adjunction F : C  ∼ D : G (A.3) for which both the unit η and the counit  are natural isomorphisms. Existence of an adjoint equivalence in particular implies that C = D are equivalent as categories. Proposition A.3. If a functor G : D → C admits a left adjoint F : C → D, then F is unique up to a unique natural isomorphism. Vice versa, if a functor F : C → D admits a right adjoint G : D → C, then G is unique up to a unique natural isomorphism. Localizations Localizations of categories are treated for example in [21, Section 7.1]. In our paper we restrict ourselves to small categories. Definition A.4. Let C be a small category and W ⊆ Mor C a subset of the set −1 of morphisms. A localization of C at W is a small category C[W ] together −1 with a functor L : C → C[W ] satisfying the following properties: (a) For all (f : c → c ) ∈ W , L(f): L(c) → L(c ) is an isomorphism in −1 C[W ]. (b) For any category D and any functor F : C → D that sends morphisms −1 in W to isomorphisms in D, there exists a functor F : C[W ] → D and a natural isomorphism F F L. −1 C[W ] (c) For all objects G, H ∈ D in the functor category, the map Hom −1 G, H −→ Hom C GL,H L (A.4) C[W ] is a bijection of Hom-sets. −1 Proposition A.5. If C[W ] exists, it is unique up to equivalence of categories. References [1] B¨ ar, C., Ginoux, N., Pf¨ affle, F.: Wave Equations on Lorentzian Manifolds and Quantization. European Mathematical Society, Zuric ¨ h (2007) [2] Beem, J.K., Ehrlich, P.E., Easley, K.L.: Global Lorentzian geometry. Marcel Dekker, New York (1996) AQFT on Spacetimes with Timelike Boundary [3] Benini, M., Dappiaggi, C.: Models of free quantum field theories on curved back- grounds. In: Brunetti, R., Dappiaggi, C., Fredenhagen, K., Yngvason, J. (eds.) Advances in Algebraic Quantum Field Theory, pp. 75–124. Springer, Heidelberg (2015). arXiv:1505.04298 [math-ph] [4] Benini, M., Dappiaggi, C., Hack, T.-P.: Quantum field theory on curved backgrounds—a primer. Int. J. Mod. Phys. A 28, 1330023 (2013). arXiv:1306.0527 [gr-qc] [5] Benini, M., Schenkel, A., Woike, L.: Operads for algebraic quantum field theory. arXiv:1709.08657 [math-ph] [6] Benini, M., Schenkel, A., Woike, L.: Involutive categories, colored ∗-operads and quantum field theory. arXiv:1802.09555 [math.CT] [7] Brunetti, R., Fredenhagen, K., Verch, R.: The generally covariant locality prin- ciple: A new paradigm for local quantum field theory. Commun. Math. Phys. 237, 31 (2003). arxiv:math-ph/0112041 [8] Bussola, F., Dappiaggi, C., Ferreira, H.R.C., Khavkine, I.: Ground state for a massive scalar field in the BTZ spacetime with Robin boundary conditions. Phys. Rev. D 96(10), 105016 (2017). arXiv:1708.00271 [gr-qc] [9] Chru´ sciel, P.T., Galloway, G.J., Solis, D.: Topological censorship for Kaluza– Klein space-times. Ann. Henri Poincar´ e 10, 893 (2009). arXiv:0808.3233 [gr-qc] [10] Dappiaggi, C., Ferreira, H.R.C.: Hadamard states for a scalar field in anti-de Sit- ter spacetime with arbitrary boundary conditions. Phys. Rev. D 94(12), 125016 (2016). arXiv:1610.01049 [gr-qc] [11] Dappiaggi, C., Ferreira, H.R.C.: On the algebraic quantization of a massive scalar field in anti-de-Sitter spacetime. Rev. Math. Phys. 30, 1850004 (2018). arXiv:1701.07215 [math-ph] [12] Dappiaggi, C., Nosari, G., Pinamonti, N.: The Casimir effect from the point of view of algebraic quantum field theory. Math. Phys. Anal. Geom. 19, 12 (2016). arXiv:1412.1409 [math-ph] [13] Dutsc ¨ h, M., Rehren, K.-H.: Generalized free fields and the AdS-CFT correspon- dence. Ann. Henri Poincar´ e 4, 613 (2003). arXiv:math-ph/0209035 [14] Fewster, C.J.: Endomorphisms and automorphisms of locally covariant quantum field theories. Rev. Math. Phys. 25, 1350008 (2013). arXiv:1201.3295 [math-ph] [15] Fredenhagen, K.: Generalizations of the theory of superselection sectors. In: Kastler, D. (ed.) The Algebraic theory of Superselection Sectors: Introduction and Recent Results, p. 379. World Scientific Publishing, Singapore (1990) [16] Fredenhagen, K.: Global observables in local quantum physics. In: Araki, H., Ito, K.R., Kishimoto, A., Ojima, I. (eds.) Quantum and Non-commutative Analysis: Past, Present and Future Perspectives, pp. 41–51. Kluwer Academic Publishers, Dordrecht (1993) [17] Fredenhagen, K., Rehren, K.-H., Schroer, B.: Superselection sectors with braid group statistics and exchange algebras II: geometric aspects and conformal co- variance. Rev. Math. Phys. 4, 113 (1992) [18] Haag, R., Kastler, D.: An algebraic approach to quantum field theory. J. Math. Phys. 5, 848 (1964) [19] Ishibashi, A., Wald, R.M.: Dynamics in non-globally-hyperbolic static space- times II: general analysis of prescriptions for dynamics. Class. Quantum Gravity. 20, 3815 (2003). arXiv:gr-qc/0305012 M. Benini et al. Ann. Henri Poincar´ e [20] Ishibashi, A., Wald, R.M.: Dynamics in non-globally-hyperbolic static space- times III: anti-de Sitter space-time. Class. Quantum Gravity 21, 2981 (2004). arXiv:hep-th/0402184 [21] Kashiwara, M., Schapira, P.: Categories and Sheaves. Springer, Berlin (2006) [22] Kay, B.S.: The principle of locality and quantum field theory on (non globally hyperbolic) curved spacetimes. Rev. Math. Phys. 4, 167 (1992) [23] Lee, J.M.: Introduction to Smooth Manifolds. Springer, New York (2013) [24] O’Neill, B.: Semi-Riemannian Geometry. Academic Press, New York (1983) [25] Rehren, K.-H.: Algebraic holography. Ann. Henri Poincar´ e 1, 607 (2000). arXiv:hep-th/9905179 [26] Rejzner, K.: Perturbative Algebraic Quantum Field Theory: An Introduction for Mathematicians. Springer, Heidelberg (2016) [27] Ribeiro, P.L.: Structural and Dynamical Aspects of the AdS/CFT Correspon- dence: A Rigorous Approach. Ph.D. Thesis, University of Sao Paulo (2007). arXiv:0712.0401 [math-ph] [28] Riehl, E.: Category Theory in Context. Dover Publications Inc, New York (2016) [29] Solis, D.A.: Global Properties of Asymptotically de Sitter and Anti de Sitter Spacetimes. Ph.D. Thesis, University of Miami (2006). https:// scholarlyrepository.miami.edu/dissertations/2414/ [30] Sommer, C.: Algebraische Charakterisierung von Randbedingungen in der Quan- tenfeldtheorie. Diploma Thesis in German, University of Hamburg (2006). http://www.desy.de/uni-th/lqp/psfiles/dipl-sommer.ps.gz [31] Wald, R.M.: Dynamics in nonglobally hyperbolic, static space-times. J. Math. Phys. 21, 2802 (1980) [32] Wrochna, M.: The holographic Hadamard condition on asymptotically anti-de Sitter spacetimes. Lett. Math. Phys. 107, 2291 (2017). arXiv:1612.01203 [math- ph] [33] Zahn, J.: Generalized Wentzell boundary conditions and quantum field theory. Ann. Henri Poincar´ e 19, 163 (2018). arXiv:1512.05512 [math-ph] Marco Benini Fachbereich Mathematik Universit¨ at Hamburg Bundesstr. 55 20146 Hamburg Germany e-mail: marco.benini@uni-hamburg.de Claudio Dappiaggi INFN, Sezione di Pavia Universit` a di Pavia Via Bassi 6 27100 Pavia Italy e-mail: claudio.dappiaggi@unipv.it and AQFT on Spacetimes with Timelike Boundary Dipartimento di Fisica INFN, Sezione di Pavia Via Bassi 6 27100 Pavia Italy Alexander Schenkel School of Mathematical Sciences University of Nottingham University Park Nottingham NG7 2RD UK e-mail: alexander.schenkel@nottingham.ac.uk Communicated by Karl Henning Rehren. Received: January 3, 2018. Accepted: May 3, 2018. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Annales Henri Poincaré Springer Journals

Algebraic Quantum Field Theory on Spacetimes with Timelike Boundary

Free
33 pages
Loading next page...
 
/lp/springer_journal/algebraic-quantum-field-theory-on-spacetimes-with-timelike-boundary-V3XAYNi0uk
Publisher
Springer International Publishing
Copyright
Copyright © 2018 by The Author(s)
Subject
Physics; Theoretical, Mathematical and Computational Physics; Dynamical Systems and Ergodic Theory; Quantum Physics; Mathematical Methods in Physics; Classical and Quantum Gravitation, Relativity Theory; Elementary Particles, Quantum Field Theory
ISSN
1424-0637
eISSN
1424-0661
D.O.I.
10.1007/s00023-018-0687-1
Publisher site
See Article on Publisher Site

Abstract

Ann. Henri Poincar´ e Online First c 2018 The Author(s) Annales Henri Poincar´ e https://doi.org/10.1007/s00023-018-0687-1 Algebraic Quantum Field Theory on Spacetimes with Timelike Boundary Marco Benini , Claudio Dappiaggi and Alexander Schenkel Dedicated to Klaus Fredenhagen on the occasion of his 70th birthday. Abstract. We analyze quantum field theories on spacetimes M with time- like boundary from a model-independent perspective. We construct an adjunction which describes a universal extension to the whole spacetime M of theories defined only on the interior intM . The unit of this adjunc- tion is a natural isomorphism, which implies that our universal extension satisfies Kay’s F-locality property. Our main result is the following char- acterization theorem: Every quantum field theory on M that is additive from the interior (i.e., generated by observables localized in the interior) admits a presentation by a quantum field theory on the interior intM and an ideal of its universal extension that is trivial on the interior. We shall illustrate our constructions by applying them to the free Klein–Gordon field. 1. Introduction and Summary Algebraic quantum field theory is a powerful and far developed framework to address model-independent aspects of quantum field theories on Minkowski spacetime [18] and more generally on globally hyperbolic spacetimes [7]. In addition to establishing the axiomatic foundations for quantum field theory, the algebraic approach has provided a variety of mathematically rigorous con- structions of non-interacting models, see e.g., the reviews [1, 3, 4], and more in- terestingly also perturbatively interacting quantum field theories, see e.g., the recent monograph [26]. It is worth emphasizing that many of the techniques involved in such constructions, e.g., existence and uniqueness of Green’s opera- tors and the singular structure of propagators, crucially rely on the hypothesis that the spacetime is globally hyperbolic and has empty boundary. M. Benini et al. Ann. Henri Poincar´ e Even though globally hyperbolic spacetimes have plenty of applications to physics, there exist also important and interesting situations which require non-globally hyperbolic spacetimes, possibly with a non-trivial boundary. On the one hand, recent developments in high energy physics and string theory are strongly focused on anti-de Sitter spacetime, which is not globally hyperbolic and has a (conformal) timelike boundary. On the other hand, experimental setups for studying the Casimir effect confine quantum field theories between several metal plates (or other shapes), which may be modeled theoretically by introducing timelike boundaries to the system. This immediately prompts the question whether the rigorous framework of algebraic quantum field theory admits a generalization to cover such scenarios. Most existing works on algebraic quantum field theory on spacetimes with a timelike boundary focus on the construction of concrete examples, such as the free Klein–Gordon field on simple classes of spacetimes. The basic strategy em- ployed in such constructions is to analyze the initial value problem on a given spacetime with timelike boundary, which has to be supplemented by suitable boundary conditions. Different choices of boundary conditions lead to different Green’s operators for the equation of motion, which is in sharp contrast to the well-known existence and uniqueness results on globally hyperbolic spacetimes with empty boundary. Recent works addressing this problem are [19, 20, 33], the latter extending the analysis of [31]. For specific choices of boundary con- ditions, there exist successful constructions of algebraic quantum field theories on spacetimes with timelike boundary, see e.g., [8, 10–12]. The main message of these works is that the algebraic approach is versatile enough to account also for these models, although some key structures, such as for example the notion of Hadamard states [11, 32], should be modified accordingly. Unfortunately, model-independent results on algebraic quantum field the- ory on spacetimes with timelike boundary are more scarce. There are, how- ever, some notable and very interesting works in this direction: On the one hand, Rehren’s proposal for algebraic holography [25] initiated the rigorous study of quantum field theories on the anti-de Sitter spacetime. This has been further elaborated in [13] and extended to asymptotically AdS spacetimes in [27]. On the other hand, inspired by Fredenhagen’s universal algebra [15–17], a very interesting construction and analysis of global algebras of observables on spacetimes with timelike boundaries has been performed in [30]. The most notable outcome is the existence of a relationship between maximal ideals of this algebra and boundary conditions, a result which has been of inspiration for this work. In the present paper, we shall analyze quantum field theories on space- times with timelike boundary from a model-independent perspective. We are mainly interested in understanding and proving structural results for whole categories of quantum field theories, in contrast to focusing on particular the- ories. Such questions can be naturally addressed by using techniques from the recently developed operadic approach to algebraic quantum field theory [5]. Let us describe rather informally the basic idea of our construction and its implications: Given a spacetime M with timelike boundary, an algebraic AQFT on Spacetimes with Timelike Boundary quantum field theory on M is a functor B : R → Alg assigning algebras of observables to suitable regions U ⊆ M (possibly intersecting the boundary), which satisfies the causality and time-slice axioms. We denote by QFT(M ) the category of algebraic quantum field theories on M . Denoting the full sub- category of regions in the interior of M by R ⊆R , we may restrict int M M any theory B ∈ QFT(M)toatheory res B ∈ QFT(int M ) defined only on the interior regions. Notice that it is in practice much easier to analyze and construct theories on int M as opposed to theories on the whole spacetime M . This is because the former are postulated to be insensitive to the boundary by Kay’s F-locality principle [22]. As a first result, we shall construct a left adjoint of the restriction functor res : QFT(M ) → QFT(int M ), which we call the universal extension functor ext : QFT(int M ) → QFT(M ). This means that given any theory A ∈ QFT(int M ) that is defined only on the interior regions in M , we obtain a universal extension ext A ∈ QFT(M ) to all regions in M , including those that intersect the boundary. It is worth to emphasize that the adjective universal above refers to the categorical concept of universal properties. Below we explain in which sense ext is also “universal” in a more physical meaning of the word. It is crucial to emphasize that our universal extension ext A ∈ QFT(M ) is always a bona fide algebraic quantum field theory in the sense that it satisfies the causality and time-slice axioms. This is granted by the operadic approach to algebraic quantum field theory of [5]. In particular, the ext  res adjunction investigated in the present paper is one concrete instance of a whole fam- ily of adjunctions between categories of algebraic quantum field theories that naturally arise within the theory of colored operads and algebras over them. A far reaching implication of the above mentioned ext  res adjunction is a characterization theorem that we shall establish for quantum field theories on spacetimes with timelike boundary. Given any theory B ∈ QFT(M)ona spacetime M with timelike boundary, we can restrict and universally extend to obtain another such theory ext res B ∈ QFT(M ). The adjunction also provides us with a natural comparison map between these theories, namely the counit  :extres B → B of the adjunction. Our result in Theorem 5.6 and Corollary 5.7 is that  induces an isomorphism ext res B/ ker  B B B of quantum field theories if and only if B is additive from the interior as formalized in Definition 5.5. The latter condition axiomatises the heuristic idea that the theory B has no degrees of freedom that are localized on the boundary of M , i.e., all its observables may be generated by observables supported in the interior of M . Notice that the results in Theorem 5.6 and Corollary 5.7 give the adjective universal also a physical meaning in the sense that the extensions are sufficiently large such that any additive theory can be recovered by a quotient. We strengthen this result in Theorem 5.10 by constructing an equivalence between the category of additive quantum field theories on M and a category of pairs (A, I) consisting of a theory A ∈ QFT(int M)onthe interior and an ideal I ⊆ ext A of the universal extension that is trivial on the interior. More concretely, this means that every additive theory B ∈ QFT(M ) may be naturally decomposed into two distinct pieces of data: (1) A theory M. Benini et al. Ann. Henri Poincar´ e A ∈ QFT(int M ) on the interior, which is insensitive to the boundary as postulated by F-locality, and (2) an ideal I ⊆ ext A of its universal extension that is trivial on the interior, i.e., that is only sensitive to the boundary. Specific examples of such ideals arise from imposing boundary conditions. We shall illustrate this fact by using the free Klein–Gordon theory as an example. Thus, our results also provide a bridge between the ideas of [30] and the concrete constructions in [8, 10–12]. The remainder of this paper is structured as follows: In Sect. 2, we recall some basic definitions and results about the causal structure of spacetimes with timelike boundaries, see also [9, 29]. In Sect. 3, we provide a precise definition of the categories QFT(M)and QFT(int M ) by using the ideas of [5]. Our universal boundary extension is developed in Sect. 4, where we also provide an explicit model in terms of left Kan extension. Our main results on the characterization of additive quantum field theories on M are proven in Sect. 5. Section 6 illustrates our construction by focusing on the simple example of the free Klein–Gordon theory, where more explicit formulas can be developed. It is in this context that we provide examples of ideals implementing boundary conditions and relate to analytic results, e.g., [12]. We included “Appendix A” to state some basic definitions and results of category theory which will be used in our work. 2. Spacetimes with Timelike Boundary We collect some basic facts about spacetimes with timelike boundary, following [29, Section 3.1] and [9, Section 2.2]. For a general introduction to Lorentzian geometry, we refer to [2, 24], see also [1, Sections 1.3 and A.5] for a concise presentation. We use the term manifold with boundary to refer to a Hausdorff, second countable, m-dimensional smooth manifold M with boundary, see e.g., [23]. This definition subsumes ordinary manifolds as manifolds with empty bound- ary ∂M = ∅. We denote by int M ⊆ M the submanifold without the boundary. Every open subset U ⊆ M carries the structure of a manifold with (possibly empty) boundary and one has int U = U ∩ int M . Definition 2.1. A Lorentzian manifold with boundary is a manifold with bound- ary that is equipped with a Lorentzian metric. Definition 2.2. Let M be a time-oriented Lorentzian manifold with boundary. The Cauchy development D(S) ⊆ M of a subset S ⊆ M is the set of points p ∈ M such that every inextensible (piecewise smooth) future directed causal curve stemming from p meets S. The following properties follow easily from the definition of Cauchy de- velopment. Proposition 2.3. Let S, S ⊆ M be subsets of a time-oriented Lorentzian man- ifold M with boundary. Then, the following holds true: (a) S ⊆ S implies D(S) ⊆ D(S ); AQFT on Spacetimes with Timelike Boundary (b) S ⊆ D(S)= D(D(S)); (c) D(D(S) ∩ D(S )) = D(S) ∩ D(S ). We denote by J (S) ⊆ M the causal future/past of a subset S ⊆ M , i.e., the set of points that can be reached by a future/past directed causal curve stemming from S. Furthermore, we denote by I (S) ⊆ M the chronological future/past of a subset S ⊆ M , i.e., the set of points that can be reached by a future/past directed timelike curve stemming from S. Definition 2.4. Let M be a time-oriented Lorentzian manifold with boundary. + − We say that a subset S ⊆ M is causally convex in M if J (S) ∩ J (S) ⊆ S. M M We say that two subsets S, S ⊆ M are causally disjoint in M if (J (S) ∪ J (S)) ∩ S = ∅. The following properties are simple consequences of these definitions. Proposition 2.5. Let S, S ⊆ M be two subsets of a time-oriented Lorentzian manifold M with boundary. Then, the following holds true: (a) D(S) and D(S ) are causally disjoint if and only if S and S are causally disjoint; (b) Suppose S and S are causally disjoint. Then, the disjoint union S S ⊆ M is causally convex if and only if both S and S are causally convex. The following two definitions play an essential role in our work. Definition 2.6. A spacetime with timelike boundary is an oriented and time- oriented Lorentzian manifold M with boundary, such that the pullback of the Lorentzian metric along the boundary inclusion ∂M → M defines a Lorentzian metric on the boundary ∂M . Definition 2.7. Let M be a spacetime with timelike boundary. (i) R denotes the category whose objects are causally convex open subsets U ⊆ M and whose morphisms i : U → U are inclusions U ⊆ U ⊆ M . We call it the category of regions in M . (ii) C ⊆ Mor R is the subset of Cauchy morphisms in R , i.e., inclusions M M M i : U → U such that D(U)= D(U ). (iii) R ⊆R is the full subcategory whose objects are contained in the int M M interior int M . We denote by C ⊆C the Cauchy morphisms between int M M objects of R . int M Proposition 2.8. Let M be a spacetime with timelike boundary. For each subset S ⊆ M and each object U ∈R , i.e., a causally convex open subset U ⊆ M, the following holds true: ± ± (a) I (S) is the largest open subset of J (S); M M ± ± ± ± ± (b) J (I (S)) = I (S)= I (J (S)); M M M M M (c) S ⊆ int M implies D(S) ⊆ int M; (d) D(U ) ⊆ M is causally convex and open, i.e., D(U ) ∈R . M M. Benini et al. Ann. Henri Poincar´ e Proof. (a) and (b): These are standard results in the case of empty bound- ary, see e.g., [1, 2, 24]. The extension to spacetimes with non-empty timelike boundary can be found in [29, Section 3.1.1]. (c): We show that if D(S) contains a boundary point, so does S: Suppose p ∈ D(S) belongs to the boundary of M . By Definition 2.6, the boundary ∂M of M can be regarded as a time-oriented Lorentzian manifold with empty boundary; hence, we can consider a future directed inextensible causal curve γ in ∂M stemming from p. Since ∂M is a closed subset of M , γ must be inextensible also as a causal curve in M , hence γ meets S because it stems from p ∈ D(S). Since γ lies in ∂M by construction, we conclude that S contains a boundary point of M . (d): D(U ) ⊆ M is causally convex by the definition of Cauchy develop- ment and by the causal convexity of U ⊆ M . To check that D(U ) ⊆ M is open, we use quasi-limits as in [24, Definition 14.7 and Proposition 14.8]: First, ob- serve that I (U ) ⊆ M is open by (a). Hence, it is the same to check whether + − a subset of I := I (U ) ∪ I (U)isopenin M or in I (with the induced topol- M M ogy). Indeed, U ⊆ D(U ) ⊆ I because U is open in M . From now on, in place of M , let us therefore consider I, equipped with the induced metric, orientation and time-orientation. By contradiction, assume that there exists p ∈ D(U )\U such that all of its neighborhoods intersect the complement of D(U ). Then, there exists a sequence {α } of inextensible causal curves in I never meeting U such that {α (0)} converges to p. We fix a convex cover of I refining the open + − cover {I (U ),I (U )}. Relative to the fixed convex cover, the construction M M of quasi-limits allows us to obtain from {α } an inextensible causal curve λ through p ∈ D(U ). Hence, λ meets U,say in q. By the construction of a quasi- limit, q lies on a causal geodesic segment between p and p , two successive k k+1 limit points for {α } contained in some element of the fixed convex cover. It + − follows that either p or p belongs to J (U ) ∩ J (U ), which is contained k k+1 I I in U by causal convexity. Hence, we found a subsequence {α } of {α } and n n a sequence of parameters {s } such that {α (s )} converges to a point of U j n j (either p or p ). By construction the sequence {α (s )} is contained in k k+1 n j I\U ; however, its limit lies in U . This contradicts the hypothesis that U is open in I. The causal structure of a spacetime M with timelike boundary can be affected by several pathologies, such as the presence of closed future directed causal curves. It is crucial to avoid these issues in order to obtain concrete examples of our constructions in Sect. 6. The following definition is due to [9, Section 2.2] and [29, Section 3.1.2]. Definition 2.9. A spacetime M with timelike boundary is called globally hy- perbolic if the following two properties hold true: (i) Strong causality Every open neighborhood of each point p ∈ M contains a causally convex open neighborhood of p. + − (ii) Compact double-cones J (p) ∩ J (q) is compact for all p, q ∈ M . M M AQFT on Spacetimes with Timelike Boundary Remark 2.10. In the case of empty boundary, this definition agrees with the usual one in [1, 2, 24]. Simple examples of globally hyperbolic spacetimes with m−1 m non-empty timelike boundary are the half space {x ≥ 0}⊆ R , the spatial m−1 m 1 2 m−1 2 m slab {0 ≤ x ≤ 1}⊆ R and the cylinder {(x ) +···+(x ) ≤ 1}⊆ R in Minkowski spacetime R ,for m ≥ 2, as well as all causally convex open subsets thereof. The following results follow immediately from Definition 2.9 and Propo- sition 2.8. Proposition 2.11. Let M be a globally hyperbolic spacetime with timelike bound- ary. (a) M admits a cover by causally convex open subsets. (b) For each U ∈R , i.e., a causally convex open subset U ⊆ M, both U and D(U ) are globally hyperbolic spacetimes with (possibly empty) boundary when equipped with the metric, orientation and time-orientation induced by M.Ifmoreover U ⊆ int M is contained in the interior, then both U and D(U ) have empty boundary. 3. Categories of Algebraic Quantum Field Theories Let M be a spacetime with timelike boundary. (In this section we do not have to assume that M is globally hyperbolic in the sense of Definition 2.9). Recall the category R of open and causally convex regions in M and the subset C M M of Cauchy morphisms (cf. Definition 2.7). Together with our notion of causal disjointness from Definition 2.4, these data provide the geometrical input for the traditional definition of algebraic quantum field theories on M . Definition 3.1. An algebraic quantum field theory on M is a functor A : R −→ Alg (3.1) with values in the category Alg of associative and unital ∗-algebras over C, which satisfies the following properties: (i) Causality axiom For all causally disjoint inclusions i : U → U ← U : i , 1 1 2 2 the induced commutator A(i )(−), A(i )(−) : A(U ) ⊗ A(U ) −→ A(U ) (3.2) 1 2 1 2 A(U ) is zero. (ii) Time-slice axiom For all Cauchy morphisms (i : U → U ) ∈C , the map A(i): A(U ) −→ A(U ) (3.3) is an Alg-isomorphism. We denote by qft(M ) ⊆ Alg the full subcategory of the category of func- tors from R to Alg whose objects are all algebraic quantum field theories on M , i.e., functors fulfilling the causality and time-slice axioms. (Morphisms in this category are all natural transformations). M. Benini et al. Ann. Henri Poincar´ e We shall now show that there exists an alternative, but equivalent, de- scription of the category qft(M ) which will be more convenient for the tech- nical constructions in our paper. Following [5, Section 4.1], we observe that the time-slice axiom in Definition 3.1(ii) is equivalent to considering functors −1 B : R C → Alg that are defined on the localization of the category R at the set of Cauchy morphisms C . See Definition A.4 for the defini- M M tion of localizations of categories. By abstract arguments as in [5, Section 4.6], one observes that the universal property of localizations implies that the cat- egory qft(M ) is equivalent to the full subcategory of the functor category −1 R C −1 M [ ] Alg whose objects are all functors B : R C → Alg that satisfy −1 the causality axiom for the pushforward orthogonality relation on R C . Loosely speaking, this means that the time-slice axiom in Definition 3.1(ii) −1 can be hard-coded by working on the localized category R C instead of using the usual category R of regions in M . The aim of the remainder of this section is to provide an explicit model −1 for the localization functor R →R C . With this model, it will be- M M come particularly easy to verify the equivalence between the two alternative descriptions of the category qft(M ). Let us denote by −1 R C ⊆R (3.4a) M M the full subcategory of R whose objects V ⊆ M are stable under Cauchy de- velopment, i.e., D(V )= V where D(V ) ⊆ M denotes the Cauchy development (cf. Definition 2.2). In the following, we shall always use letters like U ⊆ M for generic regions in R and V ⊆ M for regions that are stable under Cauchy −1 development, i.e., objects in R C . Recall from Definition 2.7 that each object U ∈R is a causally convex open subset U ⊆ M , hence the Cauchy de- velopment D(U ) ⊆ M is a causally convex open subset by Proposition 2.8(d), which is stable under Cauchy development by Proposition 2.3(b). This shows −1 that D(U ) is an object of R C . Furthermore, a morphism i : U → U in R is an inclusion U ⊆ U , which induces an inclusion D(U ) ⊆ D(U )by −1 Proposition 2.3(a) and hence a morphism D(i): D(U ) → D(U )in R C . We define a functor −1 D : R −→ R C (3.4b) M M by setting on objects and morphisms U −→ D(U ), (i : U → U ) −→ D(i): D(U ) → D(U ) . (3.4c) Furthermore, let us write −1 I : R C −→ R (3.5) M M for the full subcategory embedding. Lemma 3.2. D and I form an adjunction (cf. Definition A.1) −1 D : R  R C : I (3.6) M M M AQFT on Spacetimes with Timelike Boundary −1 whose counit is a natural isomorphism (in fact, the identity), hence R C is a full reflective subcategory of R . Furthermore, the components of the unit are Cauchy morphisms. Proof. For U ∈R ,the U -component of the unit η :id −→ ID (3.7) is given by the inclusion U ⊆ D(U)of U into its Cauchy development, which is a Cauchy morphism, see Proposition 2.3(b) and Definition 2.7(ii). For V ∈ −1 R C ,the V -component of the counit : DI −→ id −1 (3.8) R C M [ ] is given by the identity of the object D(V )= V . The triangle identities hold trivially. −1 −1 Proposition 3.3. The category R C and the functor D : R →R C M M M M M defined in (3.4) provide a model for the localization of R at C . M M Proof. We have to check all the requirements listed in Definition A.4. (a) By Definition 2.7, for each Cauchy morphism i : U → U one has D(U)= −1 D(U ) and hence D(i)=id is an isomorphism in R C . D(U ) M (b) Let F : R → D be any functor to a category D that sends morphisms in C to D-isomorphisms. Using Lemma 3.2, we define F := FI : M W −1 R C → D and consider the natural transformation Fη : F → FI D = F D obtained by the unit of the adjunction D  I. Because all components of η are Cauchy morphisms (cf. Lemma 3.2), Fη is a natural isomorphism. −1 (c) Let G, H : R C → D be two functors. We have to show that the map Hom G, H −→ Hom R GD, HD (3.9) −1 R C D [ ] is a bijection. Let us first prove injectivity: Let ξ, ξ : G → H be two natural transformations such that ξD = ξD. Using Lemma 3.2, we obtain commutative diagrams ξDI ξDI GDI HDI GDI HDI (3.10) G H G H G H G H where the vertical arrows are natural isomorphisms because the counit is an isomorphism. Recalling that by hypothesis ξD = ξD, it follows that ξ = ξ. Hence, the map (3.9) is injective. M. Benini et al. Ann. Henri Poincar´ e It remains to prove that (3.9) is also surjective. Let χ : GD → HD be any natural transformation. Using Lemma 3.2, we obtain a commutative diagram GD HD (3.11) GDη HDη GDID HDID χID where the vertical arrows are natural isomorphisms because the components of the unit η are Cauchy morphisms and D assigns isomorphisms to them. Let us define a natural transformation ξ : G → H by the commutative diagram G H (3.12) −1 GDI HDI χI where we use that  is a natural isomorphism (cf. Lemma 3.2). Combining the last two diagrams, one easily computes that ξD = χ by using also the triangle identities of the adjunction D  I. Hence, the map (3.9) is surjective. We note that there exist two (a priori different) options to define an −1 orthogonality relation on the localized category R C , both of which are provided by [5, Lemma 4.29]: (1) The pullback orthogonality relation along −1 the full subcategory embedding I : R C →R and (2) the pushforward M M −1 orthogonality relation along the localization functor D : R →R C . M M In our present scenario, both constructions coincide and one concludes that −1 two R C -morphisms are orthogonal precisely when they are orthogonal in R . Summing up, we obtain −1 Lemma 3.4. We say that two morphisms in the full subcategory R C ⊆ R are orthogonal precisely when they are orthogonal in R (i.e., causally M M −1 disjoint, cf. Definition 2.4). Then, both functors D : R →R C and M M −1 I : R C →R preserve (and also detect) the orthogonality relations. M M Proof. For I this holds trivially, while for D see Proposition 2.5(a). With these preparations, we may now define our alternative description of the category of algebraic quantum field theories. −1 R [C ] Definition 3.5. We denote by QFT(M ) ⊆ Alg the full subcategory −1 whose objects are all functors B : R C → Alg that satisfy the following version of the causality axiom: For all causally disjoint inclusions i : V → 1 1 −1 V ← V : i in R C , i.e., V , V and V are stable under Cauchy develop- 2 2 M 1 2 ment, the induced commutator B(i )(−), B(i )(−) : B(V ) ⊗ B(V ) −→ B(V ) (3.13) 1 2 1 2 B(V ) is zero. AQFT on Spacetimes with Timelike Boundary Theorem 3.6. By pullback, the adjunction D  I of Lemma 3.2 induces an adjoint equivalence (cf. Definition A.2) ∗ ∗ I : qft(M )  ∼ QFT(M): D . (3.14) In particular, the two categories qft(M ) of Definition 3.1 and QFT(M ) of Definition 3.5 are equivalent. −1 Proof. It is trivial to check that the adjunction D : R  R C : I M M induces an adjunction −1 ∗ R R [C ] ∗ M M I : Alg  Alg : D (3.15) ∗ ∗ between functor categories. Explicitly, the unit η  :id R → D I has Alg M components ∗ ∗ η  := Aη : A −→ D (I (A)) = AID, (3.16) where A : R → Alg is any functor and η :id → ID denotes the unit of M R ∗ ∗ D  I. The counit   : I D → id has components −1 R C [ ] Alg ∗ ∗ := B : I (D (B)) = BDI −→ B, (3.17) −1 where B : R C → Alg is any functor and  : DI → id −1 denotes M R C M [ ] ∗ ∗ the counit of D  I. The triangle identities for I  D follow directly from those of D  I. Next, we have to prove that this adjunction restricts to the claimed source and target categories in (3.14). Given A ∈ qft(M ) ⊆ Alg , the functor ∗ −1 I (A)= A I : R C → Alg satisfies the causality axiom of Definition 3.5 because of Lemma 3.4. Hence, I (A) ∈ QFT(M ). Vice versa, given B ∈ −1 R C M [ ] ∗ QFT(M ) ⊆ Alg , the functor D (B)= B D : R → Alg satisfies the causality axiom of Definition 3.5 because of Lemma 3.4 and the time-slice axiom of Definition 3.5 because D sends by construction morphisms in C to isomorphisms. Hence, D (B) ∈ qft(M ). Using Lemma 3.2, we obtain that the counit   of the restricted adjunc- tion (3.14) is an isomorphism. Furthermore, all components of η are Cauchy morphisms, hence η  = Aη is an isomorphism for all A ∈ qft(M ), i.e., the unit η  is an isomorphism. This completes the proof that (3.14) is an adjoint equivalence. Remark 3.7. Theorem 3.6 provides us with a constructive prescription of how to change between the two equivalent formulations of algebraic quan- tum field theories given in Definitions 3.1 and 3.5. Concretely, given any A ∈ qft(M ), i.e., a functor A : R → Alg satisfying the causality and time-slice axioms as in Definition 3.1, the corresponding quantum field the- ory I (A) ∈ QFT(M ) in the sense of Definition 3.5 reads as follows: It ∗ −1 is the functor I (A)= A I : R C → Alg on the category of regions V ⊆ M that are stable under Cauchy development, which assigns to V ⊆ M the algebra A(V ) ∈ Alg and to an inclusion i : V → V the algebra map M. Benini et al. Ann. Henri Poincar´ e A(i): A(V ) → A(V ). More interestingly, given B ∈ QFT(M ), i.e., a func- −1 tor B : R C → Alg satisfying the causality axiom as in Definition 3.5, the corresponding quantum field theory D (B) ∈ qft(M ) in the sense of Def- inition 3.1 reads as follows: It is the functor D (B)= B D : R → Alg defined on the category of (not necessarily Cauchy development stable) re- gions U ⊆ M , which assigns to U ⊆ M the algebra B(D(U )) corresponding to the Cauchy development of U and to an inclusion i : U → U the al- gebra map B(D(i)) : B(D(U )) → B(D(U )) associated with the inclusion D(i): D(U ) → D(U ) of Cauchy developments. Remark 3.8. It is straightforward to check that the results of this section still hold true when one replaces R with its full subcategory R of regions M int M contained in the interior of M and C with C (cf. Definition 2.7). This M int M follows from the observation that the Cauchy development of a subset of the interior of M is also contained in int M , as shown in Proposition 2.8(c). We denote by −1 R C int M [ ] int M QFT(int M ) ⊆ Alg (3.18) the category of algebraic quantum field theories in the sense of Definition 3.5 on the interior regions of M . Concretely, an object A ∈ QFT(int M ) is a functor −1 A : R [C ] → Alg that satisfies the causality axiom of Definition 3.5 int M int M for causally disjoint interior regions. 4. Universal Boundary Extension The goal of this section is to develop a universal construction to extend quan- tum field theories from the interior of a spacetime M with timelike boundary to the whole spacetime. (Again, we do not have to assume that M is globally hyperbolic in the sense of Definition 2.9). Loosely speaking, our extended quan- tum field theory will have the following pleasant properties: (1) It describes precisely those observables that are generated from the original theory on the interior, (2) it does not require a choice of boundary conditions, (3) specific choices of boundary conditions correspond to ideals of our extended quantum field theory. We also refer to Sect. 5 for more details on the properties (1) and (3). The starting point for this construction is the full subcategory inclusion R ⊆R defined by selecting only the regions of R that lie in the int M M M interior of M (cf. Definition 2.7). We denote the corresponding embedding functor by j : R −→ R (4.1) int M M and notice that j preserves (and also detects) causally disjoint inclusions, i.e., j is a full orthogonal subcategory embedding in the terminology of [5]. Making use of Proposition 3.3, Lemma 3.2 and Remark 3.8, we define a functor J : −1 −1 R [C ] →R C on the localized categories via the commutative int M M int M M AQFT on Spacetimes with Timelike Boundary diagram −1 −1 R [C ] R [C ] (4.2a) int M M int M M I D R R int M M Notice that J is simply an embedding functor, which acts on objects and morphisms as V ⊆ int M −→ V ⊆ M, (i : V → V ) −→ (i : V → V ). (4.2b) From this explicit description, it is clear that J preserves (and also detects) causally disjoint inclusions, i.e., it is a full orthogonal subcategory embedding. The constructions in [5, Section 5.3] (see also [6] for details how to treat ∗- algebras) then imply that J induces an adjunction ext : QFT(int M )  QFT(M):res (4.3) between the category of quantum field theories on the interior int M (cf. Re- mark 3.8) and the category of quantum field theories on the whole spacetime M . The right adjoint res := J : QFT(M ) → QFT(int M ) is the pullback −1 −1 along J : R [C ] →R C , i.e., it restricts quantum field theories int M M int M M defined on M to the interior int M . The left adjoint ext : QFT(int M ) → QFT(M ) should be regarded as a universal extension functor which extends quantum field theories on the interior int M to the whole spacetime M.The goal of this section is to analyze the properties of this extension functor and to develop an explicit model that allows us to do computations in the sections below. An important structural result, whose physical relevance is explained in Remark 4.2 below, is the following proposition. Proposition 4.1. The unit η :id −→ res ext (4.4) QFT(int M ) of the adjunction (4.3) is a natural isomorphism. Proof. This is a direct consequence of the fact that the functor J given in (4.2) is a full orthogonal subcategory embedding and the general result in [5, Proposition 5.6]. Remark 4.2. The physical interpretation of this result is as follows: Let A ∈ QFT(int M ) be a quantum field theory defined only on the interior int M of M and let B := ext A ∈ QFT(M ) denote its universal extension to the whole spacetime M.The A-component η : A −→ res ext A (4.5) of the unit of the adjunction (4.3) allows us to compare A with the restriction res B of its extension B =ext A. Since η is an isomorphism by Propo- sition 4.1, restricting the extension B recovers our original theory A up to M. Benini et al. Ann. Henri Poincar´ e isomorphism. This allows us to interpret the left adjoint ext : QFT(int M ) → QFT(M ) as a genuine extension prescription. Notice that this also proves that the universal extension ext A ∈ QFT(M ) of any theory A ∈ QFT(int M)on the interior satisfies F-locality [22]. We next address the question how to compute the extension functor ext : QFT(int M ) → QFT(M ) explicitly. A crucial step toward reaching this goal is to notice that ext may be computed by a left Kan extension. Proposition 4.3. Consider the adjunction −1 −1 R [C ]  R [C ] int M M ∗ int M M Lan : Alg Alg : J , (4.6) −1 corresponding to left Kan extension along the functor J : R [C ] → int M int M −1 R C . Then, the restriction of Lan to QFT(int M ) induces a functor M J Lan : QFT(int M ) −→ QFT(M ) (4.7) that is left adjoint to the restriction functor res : QFT(M ) → QFT(int M ) in (4.3). Due to uniqueness (up to unique natural isomorphism) of adjoint functors (cf. Proposition A.3), it follows that ext Lan , i.e., (4.7) is a model for the extension functor ext in (4.3). Proof. A general version of this problem has been addressed in [5, Section 6]. Using in particular [5, Corollary 6.5], we observe that we can prove this propo- −1 sition by showing that every object V ∈R C is J -closed in the sense of [5, Definition 6.3]. In our present scenario, this amounts to proving that for all −1 causally disjoint inclusions i : V → V ← V : i with V ,V ∈R [C ] 1 1 2 2 1 2 int M int M −1 in the interior and V ∈R C not necessarily in the interior, there ex- ists a factorization of both i and i through a common interior region. Let 1 2 us consider the Cauchy development D(V V ) of the disjoint union and 1 2 the canonical inclusions j : V → D(V V ) ← V : j . As we explain be- 1 1 1 2 2 2 −1 low, D(V V ) ∈R [C ] is an interior region and j ,j provide the 1 2 int M 1 2 int M desired factorization: Since the open set V V ⊆ int M is causally con- 1 2 vex by Proposition 2.5(b), D(V V ) is causally convex, open and contained 1 2 in the interior int M by Proposition 2.8(c–d). It is, moreover, stable under Cauchy development by Proposition 2.3(b), which also provides the inclusion V ⊆ V V ⊆ D(V V ) inducing j ,for k =1, 2. Consider now the chain k 1 2 1 2 k of inclusions V ⊆ V V ⊆ V corresponding to i ,for k =1, 2. From the k 1 2 k stability under Cauchy development of V , V and V , we obtain also the chain 1 2 of inclusions V ⊆ D(V V ) ⊆ V ,for k =1, 2, that exhibits the desired k 1 2 factorization V (4.8) i i 1  2 D(i i ) 1 2 V D(V V ) V 1 1 2 2 j j 1 2 which completes the proof.  AQFT on Spacetimes with Timelike Boundary We shall now briefly review a concrete model for left Kan extension along full subcategory embeddings that was developed in [5, Section 6]. This model is obtained by means of abstract operadic techniques, but it admits an intuitive graphical interpretation that we explain in Remark 4.4 below. It allows us to compute quite explicitly the extension ext A =Lan A ∈ QFT(M)ofa quantum field theory A ∈ QFT(int M ) defined on the interior int M to the −1 whole spacetime M . The functor ext A : R C → Alg describing the −1 extended quantum field theory reads as follows: To V ∈R C it assigns a quotient algebra ext A(V )= A(V ) ∼ (4.9) i:V →V that we will describe now in detail. The direct sum (of vector spaces) in (4.9) −1 runs over all tuples i : V → V of morphisms in R C , i.e., i =(i : V → M 1 1 V,...,i : V → V ) for some n ∈ Z , with the requirement that all sources n n ≥0 −1 V ∈R [C ] are interior regions. (Notice that the regions V are not k int M k int M assumed to be causally disjoint and that the empty tuple, i.e., n = 0, is also allowed). The vector space A(V ) is defined by the tensor product of vector spaces |V | A(V ):= A(V ), (4.10) k=1 where |V | is the length of the tuple. (For the empty tuple, we set A(∅)= C). This means that the (homogeneous) elements (i,a) ∈ A(V ) (4.11) i:V →V are given by pairs consisting of a tuple of morphisms i : V → V (with all V in the interior) and an element a ∈ A(V ) of the corresponding tensor product vector space (4.10). The product on (4.11) is given on homogeneous elements by (i,a)(i ,a ):= (i,i ),a ⊗ a , (4.12) where (i,i )=(i ,...,i ,i ,...,i ) is the concatenation of tuples. The unit 1 n 1 m element in (4.11)is 1 := (∅, 1), where ∅ is the empty tuple and 1 ∈ C, and the ∗-involution is defined by ∗ ∗ (i ,...,i ),a ⊗ ··· ⊗ a := (i ,...,i ),a ⊗ ··· ⊗ a (4.13) 1 n 1 n n 1 n 1 and C-antilinear extension. Finally, the quotient in (4.9) is by the two-sided ∗-ideal of the algebra (4.11) that is generated by i i ,...,i ,a ⊗ ··· ⊗ a − i, A(i ) a ⊗ ··· ⊗ A(i ) a ∈ A(V ) , 1 n 1 n 1 1 n n i:V →V (4.14) for all tuples i : V → V of length |V | = n ≥ 1 (with all V in the interior), all −1 tuples i : V → V of R [C ]-morphisms (possibly of length zero), for k int M k k int M M. Benini et al. Ann. Henri Poincar´ e k =1,...,n, and all a ∈ A(V ), for k =1,...,n. The tuple in the first term k k of (4.14) is defined by composition i i ,...,i := i i ,...,i i ,...,i i ,...,i i (4.15a) 1 11 1 1|V | n n1 n n|V | 1 n 1 n and the expressions A(i) a in the second term are determined by A(i): A(V ) −→ A(V ),a ⊗ ··· ⊗ a −→ A(i ) a ··· A(i ) a , (4.15b) 1 n 1 1 n n where multiplication in A(V ) is denoted by juxtaposition. To a morphism −1 −1 i : V → V in R C , the functor ext A : R C → Alg assigns the M M M M algebra map ext A(i ) : ext A(V ) −→ ext A(V ), [i,a] −→ i (i),a , (4.16) where we used square brackets to indicate equivalence classes in (4.9). Remark 4.4. The construction of the algebra ext A(V ) above admits an intu- itive graphical interpretation: We shall visualize the (homogeneous) elements (i,a)in(4.11) by decorated trees ··· a a 1 n (4.17) where a ∈ A(V ) is an element of the algebra A(V ) associated to the interior k k k region V ⊆ V , for all k =1,...,n. We interpret such a decorated tree as a formal product of the formal pushforward along i : V → V of the family of observables a ∈ A(V ). The product (4.12) is given by concatenation of the k k inputs of the individual decorated trees, i.e., V V V ··· ··· ··· a a a 1 n 1 a a a 1 m m (4.18) where the decorated tree on the right-hand side has n + m inputs. The ∗- involution (4.13) may be visualized by reversing the input profile and applying ∗ to each element a ∈ A(V ). Finally, the ∗-ideal in (4.14) implements the fol- k k lowing relations: Assume that (i,a) is such that the sub-family of embeddings (i ,i ,...,i ):(V ,V ,...,V ) → V factorizes through some common in- k k+1 l k k+1 l terior region, say V ⊆ V . Using the original functor A ∈ QFT(int M ), we may form the product A(i )(a ) ··· A(i )(a ) in the algebra A(V ), which we k k l l AQFT on Spacetimes with Timelike Boundary denote for simplicity by a ··· a ∈ A(V ). We then have the relation k l V V ··· · ·· ··· ··· · ·· a a a a a a ··· a a 1 k l n 1 k l n (4.19) which we interpret as follows: Whenever (i ,i ,...,i ):(V ,V ,...,V ) → k k+1 l k k+1 l V is a sub-family of embeddings that factorizes through a common interior re- gion V ⊆ V , then the formal product of the formal pushforward of observables is identified with the formal pushforward of the actual product of observables on V . 5. Characterization of Boundary Quantum Field Theories In the previous section, we established a universal construction that allows us to extend quantum field theories A ∈ QFT(int M ) that are defined only on the interior int M of a spacetime M with timelike boundary to the whole spacetime. The extension ext A ∈ QFT(M ) is characterized abstractly by the ext  res adjunction in (4.3). We now shall reverse the question and ask which quantum field theories B ∈ QFT(M)on M admit a description in terms of (quotients of) our universal extensions. Given any quantum field theory B ∈ QFT(M ) on the whole spacetime M , we can use the right adjoint in (4.3) in order to restrict it to a theory res B ∈ QFT(int M ) on the interior of M . Applying now the extension func- tor, we obtain another quantum field theory ext res B ∈ QFT(M)onthe whole spacetime M , which we would like to compare to our original theory B ∈ QFT(M ). A natural comparison map is given by the B-component of the counit  :extres → id of the adjunction (4.3), i.e., the canonical QFT(M ) QFT(M )-morphism :extres B −→ B. (5.1a) Using our model for the extension functor given in (4.9) (and the formulas −1 following this equation), the R C  V -component of  explicitly reads M B as ( ) : B(V ) ∼−→ B(V ), [i,b] −→ B(i) b . (5.1b) B V i:V →V In order to establish positive comparison results, we have to introduce the concept of ideals of quantum field theories. Definition 5.1. An ideal I ⊆ B of a quantum field theory B ∈ QFT(M)isa −1 functor I : R C → Vec to the category of complex vector spaces, which satisfies the following properties: M. Benini et al. Ann. Henri Poincar´ e −1 (i) For all V ∈R C , I(V ) ⊆ B(V ) is a two-sided ∗-ideal of the unital ∗-algebra B(V ). −1 (ii) For all R C -morphisms i : V → V , the linear map I(i): I(V ) → I(V ) is the restriction of B(i): B(V ) → B(V ) to the two-sided ∗-ideals () () I(V ) ⊆ B(V ). Lemma 5.2. Let B ∈ QFT(M ) and I ⊆ B any ideal. Let us define B/I(V ):= −1 B(V )/I(V ) to be the quotient algebra, for all V ∈R C ,and B/I(i): B/I(V ) → B/I(V ) to be the Alg-morphism induced by B(i): B(V ) → −1 B(V ), for all R C -morphisms i : V → V .Then B/I ∈ QFT(M ) is a quantum field theory on M which we call the quotient of B by I. −1 Proof. The requirements listed in Definition 5.1 ensure that B/I : R C → Alg is an Alg-valued functor. It satisfies the causality axiom of Definition 3.5 because this property is inherited from B ∈ QFT(M ) by taking quotients. Lemma 5.3. Let κ : B → B be any QFT(M )-morphism. Define the vec- tor space ker κ(V):=ker κ : B(V ) → B (V ) ⊆ B(V ), for all V ∈ −1 R C ,and ker κ(i):ker κ(V ) → ker κ(V ) to be the linear map induced −1 by B(i): B(V ) → B(V ), for all R C -morphisms i : V → V .Then −1 ker κ : R C → Vec is an ideal of B ∈ QFT(M ), which we call the kernel of κ. Proof. The fact that ker κ defines a functor follows from naturality of κ.Prop- erty (ii) in Definition 5.1 holds true by construction. Property (i) is a conse- quence of the fact that kernels of unital ∗-algebra morphisms κ : B(V ) → B(V ) are two-sided ∗-ideals. Remark 5.4. Using the concept of ideals, we may canonically factorize (5.1) according to the diagram ext res B B (5.2) ext res B ker where both the projection π and the inclusion λ are QFT(M )-morphi- B B sms. As a last ingredient for our comparison result, we have to introduce a suitable notion of additivity for quantum field theories on spacetimes with timelike boundary. We refer to [14, Definition 2.3] for a notion of additivity on globally hyperbolic spacetimes. Definition 5.5. A quantum field theory B ∈ QFT(M ) on a spacetime M with timelike boundary is called additive (from the interior) at the object −1 V ∈R C if the algebra B(V ) is generated by the images of the Alg- −1 morphisms B(i ): B(V ) → B(V ), for all R C -morphism i : V → int int M int int −1 V whose source V ∈R [C ] is in the interior int M of M . We call int int M int M B ∈ QFT(M ) additive (from the interior) if it is additive at every object AQFT on Spacetimes with Timelike Boundary −1 V ∈R C . The full subcategory of additive quantum field theories on M add is denoted by QFT (M ) ⊆ QFT(M ). We can now prove our first characterization theorem for boundary quan- tum field theories. Theorem 5.6. Let B ∈ QFT(M ) be any quantum field theory on a (not necessarily globally hyperbolic) spacetime M with timelike boundary and let −1 V ∈R C . Then the following are equivalent: (1) The V -component (λ ) :extres B(V ) ker  (V ) −→ B(V ) (5.3) B V B of the canonical inclusion in (5.2) is an Alg-isomorphism. −1 (2) B is additive at the object V ∈R C . −1 Proof. Let i : V → V be any R C -morphism whose source V ∈ int int M int −1 R [C ] is in the interior int M of M . Using our model for the extension int M int M functor given in (4.9) (and the formulas following this equation), we obtain an Alg-morphism [i , −]: B(V ) −→ ext res B(V ),b −→ [i ,b]. (5.4) int int int Composing this morphism with the V -component of  given in (5.1), we obtain a commutative diagram ( ) B V ext res B(V ) B(V ) (5.5) [i ,−] B(i ) int  int B(V ) int for all i : V → V with V in the interior. int int int Next we observe that the images of the Alg-morphisms (5.4), for all i : V → V with V in the interior, generate ext res B(V ). Combining int int int this property with (5.5), we conclude that ( ) is a surjective map if and B V only if B is additive at V . Hence, (λ ) given by (5.2), which is injective by B V construction, is an Alg-isomorphism if and only if B is additive at V . Corollary 5.7. λ :extres B ker  → B given by (5.2) is a QFT(M )- B B add isomorphism if and only if B ∈ QFT (M ) ⊆ QFT(M ) is additive in the sense of Definition 5.5. We shall now refine this characterization theorem by showing that add QFT (M ) is equivalent, as a category, to a category describing quantum field theories on the interior of M together with suitable ideals of their uni- versal extensions. The precise definitions are as follows. Definition 5.8. Let B ∈ QFT(M ). An ideal I ⊆ B is called trivial on the −1 interior if its restriction to R [C ] is the functor assigning zero vector int M int M −1 spaces, i.e., I(V ) = 0 for all V ∈R [C ] in the interior int M of M . int int int M int M M. Benini et al. Ann. Henri Poincar´ e Definition 5.9. Let M be a spacetime with timelike boundary. We define the category IQFT(M ) as follows: • Objects are pairs (A, I) consisting of a quantum field theory A ∈ QFT(int M ) on the interior int M of M and an ideal I ⊆ ext A of its universal extension ext A ∈ QFT(M ) that is trivial on the interior. • Morphisms κ :(A, I) → (A , I )are QFT(M )-morphisms κ :ext A → ext A between the universal extensions that preserve the ideals, i.e., κ restricts to a natural transformation from I ⊆ ext A to I ⊆ ext A . There exists an obvious functor add Q : IQFT(M ) −→ QFT (M ), (5.6a) which assigns to (A, I) ∈ IQFT(M ) the quotient add Q(A, I):=ext A I ∈ QFT (M ). (5.6b) Notice that additivity of ext A I follows from that of the universal exten- sion ext A (cf. the arguments in the proof of Theorem 5.6) and the fact that quotients preserve the additivity property. There exists also a functor add S : QFT (M ) −→ IQFT(M ), (5.7a) which ‘extracts’ from a quantum field theory on M the relevant ideal. Explic- add itly, it assigns to B ∈ QFT (M ) the pair SB := res B, ker  . (5.7b) Notice that the ideal ker  ⊆ ext res B is trivial on the interior: Applying the restriction functor (4.3)to  , we obtain a QFT(int M )-morphism res  : res ext res B −→ res B. (5.8) Proposition 4.1 together with the triangle identities for the adjunction ext res in (4.3) then imply that (5.8) is an isomorphism with inverse given by η :res B → res ext res B. In particular, res  has a trivial kernel and res B B hence ker  ⊆ ext res B is trivial on the interior (cf. Definition 5.8). Our refined characterization theorem for boundary quantum field theories is as follows. Theorem 5.10. The functors Q and S defined in (5.6) and (5.7) exhibit an equivalence of categories add QFT (M ) IQFT(M ). (5.9) add Proof. We first consider the composition of functors QS : QFT (M ) → add add QFT (M ). To B ∈ QFT (M ), it assigns QSB =ext res B ker  . (5.10) The QFT(M )-morphisms λ :ext res B/ ker  → B given by (5.2) define a B B natural transformation λ : QS → id add , which is a natural isomorphism QFT (M ) due to Corollary 5.7. AQFT on Spacetimes with Timelike Boundary Let us now consider the composition of functors SQ : IQFT(M ) → IQFT(M ). To (A, I) ∈ IQFT(M ), it assigns SQ(A, I)= res ext A I , ker  = res ext A, ker  , (5.11) ext A/I ext A/I where we also used that res ext A I = res ext A because I is by hypothesis trivial on the interior. Using further the QFT(int M )-isomorphism η : A → res ext A from Proposition 4.1, we define a QFT(M )-morphism q via the (A,I) diagram (A,I) ext A ext A I (5.12) ext η A = ext A/I ext res ext A ext res ext A I Using the explicit expression for  given in (5.1) and the explicit expres- ext A/I sion for η given by (η ) : A(V ) −→ res ext A(V )= A(V ) ∼, int int A V int i:V →V int a −→ [id ,a], (5.13) int −1 for all V ∈R [C ], one computes from the diagram (5.12) that q is int int M (A,I) int M the canonical projection π :ext A → ext A/I. Hence, the QFT(M )- isomorphisms ext η :ext A → ext res ext A induce IQFT(M )-isomorphisms ext η :(A, I) −→ SQ(A, I), (5.14) which are natural in (A, I), i.e., they define a natural isomorphism ext η : id → SQ. IQFT(M ) Remark 5.11. The physical interpretation of this result is as follows: Every add additive quantum field theory B ∈ QFT (M ) on a (not necessarily glob- ally hyperbolic) spacetime M with timelike boundary admits an equivalent description in terms of a pair (A, I) ∈ IQFT(M ). Notice that the roles of A and I are completely different: On the one hand, A ∈ QFT(int M)isa quantum field theory on the interior int M of M and as such it is independent of the detailed aspects of the boundary. On the other hand, I ⊆ ext A is an ideal of the universal extension of A that is trivial on the interior, i.e., it only captures the physics that happens directly at the boundary. Examples of such ideals I arise by imposing specific boundary conditions on the universal ex- tension ext A ∈ QFT(M ), i.e., the quotient ext A/I describes a quantum field theory on M that satisfies specific boundary conditions encoded in I. We shall illustrate this assertion in Sect. 6 below using the explicit example given by the free Klein–Gordon field. Let us also note that there is a reason why our universal extension cap- tures only the class of additive quantum field theories on M . Recall that ext A ∈ QFT(M ) takes as an input a quantum field theory A ∈ QFT(int M ) on the interior int M of M . As a consequence, the extension ext A can only M. Benini et al. Ann. Henri Poincar´ e have knowledge of the ‘degrees of freedom’ that are generated in some way out of the interior regions. Additive theories in the sense of Definition 5.5 are precisely the theories whose ‘degrees of freedom’ are generated out of those localized in the interior regions. 6. Example: Free Klein–Gordon Theory In order to illustrate and make more explicit our abstract constructions de- veloped in the previous sections, we shall consider the simple example given by the free Klein–Gordon field. From now on M will be a globally hyperbolic spacetime with timelike boundary, see Definition 2.9. This assumption implies that all interior regions R are globally hyperbolic spacetimes with empty int M boundary, see Proposition 2.11. This allows us to use the standard techniques of [1, Section 3] on such regions. −1 6.1. Definition on R [C ]: int M int M Let M be a globally hyperbolic spacetime with timelike boundary, see Def- −1 inition 2.9. The free Klein–Gordon theory on R [C ] is given by the int M int M following standard construction, see e.g., [3, 4] for expository reviews. On the interior int M , we consider the Klein–Gordon operator 2 ∞ ∞ P :=  + m : C (int M ) −→ C (int M ) , (6.1) where  is the d’Alembert operator and m ≥ 0 is a mass parameter. When −1 restricting P to regions V ∈R [C ], we shall write int M int M ∞ ∞ P : C (V ) −→ C (V ). (6.2) It follows from [1] that there exists a unique retarded/advanced Green’s oper- ator ± ∞ ∞ G : C (V ) −→ C (V ) (6.3) V c −1 for P because every V ∈R [C ] is a globally hyperbolic spacetime V int M int M with empty boundary, cf. Proposition 2.11. The Klein–Gordon theory K ∈ QFT(int M ) is the functor −1 K : R [C ] → Alg given by the following assignment: To any V ∈ int M int M −1 R [C ] it assigns the associative and unital ∗-algebra K(V ) that is freely int M int M generated by Φ (f ), for all f ∈ C (V ), modulo the two-sided ∗-ideal gener- ated by the following relations: • Linearity Φ (αf + βg)= α Φ (f)+ β Φ (g), for all α, β ∈ R and V V V f, g ∈ C (V ); ∗ ∞ • Hermiticity Φ (f ) =Φ (f ), for all f ∈ C (V ); V V • Equation of motion Φ (P f ) = 0, for all f ∈ C (V ); V V • Canonical commutation relations (CCR) Φ (f)Φ (g) − Φ (g)Φ (f ) V V V V =i τ (f, g) 1, for all f, g ∈ C (V ), where τ (f, g):= fG (g)vol (6.4) V V V V AQFT on Spacetimes with Timelike Boundary + − with G := G − G the causal propagator and vol the canonical vol- V V V V ume form on V . (Note that τ is antisymmetric, see e.g., [1, Lemma 4.3.5]). −1 −1 To a morphism i : V → V in R [C ], the functor K : R [C ] → int M int M int M int M Alg assigns the algebra map that is specified on the generators by pushforward along i (which we shall suppress) K(i): K(V ) −→ K(V ), Φ (f ) −→ Φ  (f ). (6.5) V V The naturality of τ (i.e., naturality of the causal propagator, cf. e.g., [1, Sec- tion 4.3]) entails that the assignment K defines a quantum field theory in the sense of Definition 3.5. 6.2. Universal Extension Using the techniques developed in Sect. 4, we may now extend the Klein– Gordon theory K ∈ QFT(int M ) from the interior int M to the whole space- time M . In particular, using (4.9) (and the formulas following this equa- tion), one could directly compute the universal extension ext K ∈ QFT(M ). The resulting expressions, however, can be considerably simplified. We there- fore prefer to provide a more convenient model for the universal extension ext K ∈ QFT(M ) by adopting the following strategy: We first make an ‘ed- ext ucated guess’ for a theory K ∈ QFT(M ) which we expect to be the uni- versal extension of K ∈ QFT(int M ). (This was inspired by partially simpli- fying the direct computation of the universal extension). After this, we shall ext prove that K ∈ QFT(M ) satisfies the universal property that characterizes ext ext K ∈ QFT(M ). Hence, there exists a (unique) isomorphism ext K = K in ext QFT(M ), which means that our K ∈ QFT(M ) is a model for the universal extension ext K. −1 ext Let us define the functor K : R C → Alg by the following assign- −1 ment: To any region V ∈R C , which may intersect the boundary, we ext assign the associative and unital ∗-algebra K (V ) that is freely generated by Φ (f ), for all f ∈ C (int V ) in the interior int V of V , modulo the two-sided ideal generated by the following relations: • Linearity Φ (αf + βg)= α Φ (f)+ β Φ (g), for all α, β ∈ R and V V V f, g ∈ C (int V ); ∗ ∞ • Hermiticity Φ (f ) =Φ (f ), for all f ∈ C (int V ); V V • Equation of motion Φ (P f ) = 0, for all f ∈ C (int V ); V int V • Partially-defined CCR Φ (f)Φ (g) − Φ (g)Φ (f)=i τ (f, g) 1,for V V V V V int −1 all interior regions V ∈R [C ] with V ⊆ int V and f, g ∈ int int M int int M C (int V ) with supp(f ) ∪ supp(g) ⊆ V . int Remark 6.1. We note that our partially-defined CCR are consistent in the () −1 following sense: Consider V ,V ∈R [C ] with V ⊆ int V and int int M int int M int () f, g ∈ C (int V ) with the property that supp(f ) ∪ supp(g) ⊆ V .Us- c int ing the partially-defined CCR for both V and V , we obtain the equality int int ext ext i τ (f, g) 1 =i τ  (f, g) 1 in K (V ). To ensure that K (V ) is not the V V int int zero-algebra, we have to show that τ (f, g)= τ (f, g). This holds true V V int int due to the following argument: Consider the subset V ∩ V ⊆ int M . This int int M. Benini et al. Ann. Henri Poincar´ e is open, causally convex and by Proposition 2.3(c) also stable under Cauchy −1 development, hence V ∩ V ∈R [C ]. Furthermore, the inclusions int int M int int M () −1 V ∩ V → V are morphisms in R [C ]. It follows by construction int int M int int int M that supp(f ) ∪ supp(g) ⊆ V ∩ V and hence due to naturality of the τ ’s we int int obtain τ (f, g)= τ (f, g)= τ (f, g). (6.6) V V ∩V V int int int int Hence, for any fixed pair f, g ∈ C (int V ), the partially-defined CCR are independent of the choice of V (if one exists). int −1 −1 ext To a morphism i : V → V in R [C ], the functor K : R C → M M M M Alg assigns the algebra map that is specified on the generators by the push- forward along i (which we shall suppress) ext ext ext K (i): K (V ) −→ K (V ), Φ (f ) −→ Φ (f ). (6.7) V V ext Compatibility of the map (6.7) with the relations in K is a straightforward check. −1 −1 Recalling the embedding functor J : R [C ] →R C given in int M M int M M (4.2), we observe that the diagram of functors −1 R [C ] Alg (6.8) int M int M ext J K −1 R [C ] ext commutes via the natural transformation γ : K → K J with components specified on the generators by the identity maps ext γ : K(V ) −→ K (V ), Φ (f ) −→ Φ (f ), (6.9) int int V V V int int int −1 for all V ∈R [C ]. Notice that γ is a natural isomorphism because int int M int M int V = V and the partially-defined CCR on any interior region V coin- int int int cides with the CCR. −1 Theorem 6.2. (6.8) is a left Kan extension of K : R [C ] → Alg along int M int M −1 −1 J : R [C ] →R C . As a consequence of uniqueness (up to unique int M M int M M natural isomorphism) of left Kan extensions and Proposition 4.3,itfollows ext ext that K ext K, i.e., K ∈ QFT(M ) is a model for our universal extension ext K ∈ QFT(M ) of the Klein–Gordon theory K ∈ QFT(int M ). Proof. We have to prove that (6.8) satisfies the universal property of left Kan −1 extensions: Given any functor B : R C → Alg and natural transfor- mation ρ : K → B J , we have to show that there exists a unique natural ext transformation ζ : K → B such that the diagram K B J (6.10) ζJ ext K J AQFT on Spacetimes with Timelike Boundary commutes. Because γ is a natural isomorphism, it immediately follows that ζJ is uniquely fixed by this diagram. Concretely, this means that the components −1 ζ corresponding to interior regions V ∈R [C ] are uniquely fixed V int int M int int M by −1 ext ζ := ρ γ : K (V ) −→ B(V ). (6.11) int int V V V int int int It remains to determine the components ext ζ : K (V ) −→ B(V ) (6.12) −1 ext for generic regions V ∈R C . Consider any generator Φ (f)of K (V ), M V where f ∈ C (int V ), and choose a finite cover {V ⊆ int V } of supp(f)by −1 interior regions V ∈R [C ], together with a partition of unity {χ } α int M α int M subordinate to this cover. (The existence of such a cover is guaranteed by the assumption that M is a globally hyperbolic spacetime with timelike boundary, see Proposition 2.11). We define ζ Φ (f ) := B(i ) ζ Φ (χ f ) , (6.13) V V α V V α α α where i : V → V is the inclusion. Our definition (6.13) is independent of the α α choice of cover and partition of unity: For any other {V ⊆ int V } and {χ }, β β we obtain B(i ) ζ Φ (χ f ) = B(i ) ζ Φ (χ χ f ) β V V αβ V ∩V V ∩V α β α α β β β β β β α,β = B(i ) ζ Φ (χ f ) , (6.14) α V V α α α where i : V → V and i : V ∩ V → V are the inclusions. In particular, this β αβ α β β implies that (6.13) coincides with (6.11) on interior regions V = V .(Hint: int Choose the cover given by the single region V together with its partition of int unity). ext We have to check that (6.13) preserves the relations in K (V ). Preserva- tion of linearity and Hermiticity is obvious. The equation of motion relations are preserved because ζ Φ (P f ) = B(i ) ζ Φ (χ P f ) V V α V V α int V α α V = B(i ) ζ Φ (χ P χ f ) αβ V ∩V V ∩V α β α β α β V ∩V α β α,β = B(i ) ζ Φ (P χ f ) =0. (6.15) β V V β β β V −1 Regarding the partially-defined CCR, let V ∈R [C ] with V ⊆ int int M int int M int V and f, g ∈ C (int V ) with supp(f ) ∪ supp(g) ⊆ V .Wemay choose int the cover given by the single region i : V → V together with its partition of int M. Benini et al. Ann. Henri Poincar´ e unity. We obtain for the commutator ζ Φ (f ) ,ζ Φ (g) = B(i) ζ Φ (f ) , B(i) ζ Φ (g) V V V V V V V V int int int int =i τ (f, g) 1, (6.16) int which implies that the partially-defined CCR are preserved. Naturality of the components (6.13) is easily verified. Uniqueness of the ext resulting natural transformation ζ : K → B is a consequence of uniqueness of ext ext ζJ and of the fact that the Alg-morphisms K (i ): K (V ) K(V ) → int int int ext ext K (V ), for all interior regions i : V → V , generate K (V ), for all V ∈ int int −1 R C . This completes the proof. 6.3. Ideals from Green’s Operator Extensions The Klein–Gordon theory K ∈ QFT(int M ) on the interior int M of the glob- ally hyperbolic spacetime M with timelike boundary and its universal exten- ext sion K ∈ QFT(M ) depend on the local retarded and advanced Green’s oper- ± −1 ∞ ∞ ators G : C (V ) → C (V ) on all interior regions V ∈R [C ] int int int M c int V int M int in M . For constructing concrete examples of quantum field theories on globally hyperbolic spacetimes with timelike boundary as in [12], one typically imposes suitable boundary conditions for the field equation in order to obtain also global retarded and advanced Green’s operators on M . Inspired by such examples, we + − shall now show that any choice of an adjoint-related pair (G ,G ) consisting of a retarded and an advanced Green’s operator for P on M (see Definition 6.3 ext below) defines an ideal I ± ⊆ K ∈ QFT(M ) that is trivial on the interior ext (cf. Definition 5.8). The corresponding quotient K I ± ∈ QFT(M ) then may be interpreted as the Klein–Gordon theory on M , subject to a specific choice of boundary conditions that is encoded in G . Definition 6.3. A retarded/advanced Green’s operator for the Klein–Gordon ± ∞ ∞ operator P on M is a linear map G : C (int M ) → C (int M ) which satisfies the following properties, for all f ∈ C (int M ): (i) PG (f)= f , (ii) G (Pf)= f,and (iii) supp(G (f )) ⊆ J (supp(f )). + − A pair (G ,G ) consisting of a retarded and an advanced Green’s operator + − for P on M is called adjoint-related if G is the formal adjoint of G , i.e., + − G (f ) g vol = fG (g)vol , (6.17) M M M M for all f, g ∈ C (int M ). Remark 6.4. In contrast to the situation where M is a globally hyperbolic spacetime with empty boundary [1], existence, uniqueness and adjoint- relatedness of retarded/advanced Green’s operators for the Klein–Gordon op- erator P is in general not to be expected on spacetimes with timelike boundary. Positive results seem to be more likely on globally hyperbolic spacetimes with AQFT on Spacetimes with Timelike Boundary non-empty timelike boundary, although the general theory has not been de- veloped yet to the best of our knowledge. Simple examples of adjoint-related pairs of Green’s operators were constructed, e.g., in [12]. −1 Given any region V ∈R C in M , which may intersect the boundary, we use the canonical inclusion i : V → M to define local retarded/advanced Green’s operators ∞ ∞ C (int V ) C (int V ) (6.18) ∗ i ∞ ∞ C (int M ) C (int M ) where i denotes the pushforward of compactly supported functions (i.e., ex- tension by zero) and i the pullback of functions (i.e., restriction). Since V ⊆ M ± ± is causally convex, it follows that J (p) ∩ V = J (p) for all p ∈ V . Therefore M V G satisfies the axioms of a retarded/advanced Green’s operator for P on V . V V (Here, we regard V as a globally hyperbolic spacetime with timelike bound- ary, see Proposition 2.11. J (p) denotes the causal future/past of p in the −1 spacetime V ). In particular, for all interior regions V ∈R [C ]in M , int int M int M by combining Proposition 2.11 and [1, Corollary 3.4.3] we obtain that G int as defined in (6.18) is the unique retarded/advanced Green’s operator for the restricted Klein–Gordon operator P on the globally hyperbolic spacetime int V with empty boundary. int + − Consider any adjoint-related pair (G ,G ) of Green’s operator for P on −1 ext M . For all V ∈R C ,weset I ± (V ) ⊆ K (V ) to be the two-sided ∗-ideal generated by the following relations: ± ∞ • G -CCR Φ (f)Φ (g) − Φ (g)Φ (f)=i τ (f, g) 1, for all f, g ∈ C V V V V V (int V ), where τ (f, g):= fG (g)vol (6.19) V V V + − with G := G − G the causal propagator and vol the canonical V V V V volume form on V . + − The fact that the pair (G ,G ) is adjoint-related (cf. Definition 6.3) implies −1 that for all V ∈R C the causal propagator G is formally skew-adjoint, M V hence τ is antisymmetric. ext Proposition 6.5. I ± ⊆ K is an ideal that is trivial on the interior (cf. Definition 5.8). −1 Proof. Functoriality of I ± : R C → Vec is a consequence of (6.18), G M ext hence I ± ⊆ K is an ideal in the sense of Definition 5.1. It is trivial on the −1 interior because for all interior regions V ∈R [C ], the Green’s op- int M int int M erators defined by (6.18) are the unique retarded/advanced Green’s operators for P and hence the relations imposed by I ± (V ) automatically hold true G int int ext in K (V ) on account of the (partially-defined) CCR. int M. Benini et al. Ann. Henri Poincar´ e Remark 6.6. We note that the results of this section still hold true if we slightly weaken the hypotheses of Definition 2.9 by assuming the strong causality and the compact double-cones property only for points in the interior int M of M . In fact, int M can still be covered by causally convex open subsets and any causally convex open subset U ⊆ int M becomes a globally hyperbolic space- time with empty boundary once equipped with the induced metric, orientation and time-orientation. m−1 m Example 6.7. Consider the sub-spacetime M := R × [0,π] ⊆ R of m- dimensional Minkowski spacetime, which has a timelike boundary ∂M = m−1 + − R ×{0,π}. The constructions in [12] define an adjoint-related pair (G ,G ) of Green’s operators for P on M that corresponds to Dirichlet boundary conditions. Using this as an input for our construction above, we obtain a ext quantum field theory K I ± ∈ QFT(M ) that may be interpreted as the Klein–Gordon theory on M with Dirichlet boundary conditions. It is worth to emphasize that our theory in general does not coincide with the one con- structed in [12]. To provide a simple argument, let us focus on the case of m = 2 dimensions, i.e., M = R × [0,π], and compare our global algebra BDS ext DNP A (M):= K I ± (M ) with the global algebra A (M ) constructed in [12]. Both algebras are CCR-algebras, however the underlying symplec- tic vector spaces differ: The symplectic vector space underlying our global BDS ∞ ∞ algebra A (M)is C (int M ) PC (int M ) with the symplectic structure c c (6.19). Using that the spatial slices of M = R × [0,π] are compact, we observe DNP that the symplectic vector space underlying A (M ) is given by the space Sol (M)of all solutions with Dirichlet boundary condition on M (equipped Dir with the usual symplectic structure). The causal propagator defines a sym- ∞ ∞ plectic map G : C (int M ) PC (int M ) → Sol (M ), which however is not Dir c c surjective for the following reason: Any ϕ ∈ C (int M ) has by definition com- pact support in the interior of M , hence the support of Gϕ ∈ Sol (M)is Dir schematically as follows supp Gϕ supp ϕ x = π x =0 (6.20) 2 2 The usual mode functions Φ (t, x)=cos( k + m t)sin(kx) ∈ Sol (M ), k Dir ∞ ∞ for k ≥ 1, are clearly not of this form, hence G : C (int M ) PC (int M ) → c c Sol (M ) cannot be surjective. As a consequence, the models constructed in Dir ext [12] are in general not additive from the interior and our construction K I ± should be interpreted as the maximal additive subtheory of these examples. AQFT on Spacetimes with Timelike Boundary It is interesting to note that there exists a case where both constructions m−1 m coincide: Consider the sub-spacetime M := R × [0, ∞) ⊆ R of Minkowski spacetime with m ≥ 4 even and take a massless real scalar field with Dirichlet boundary conditions. Using Huygens’ principle and the support properties of BDS the Green’s operators, one may show that our algebra A (M ) is isomorphic to the construction in [12]. Acknowledgements We would like to thank Jorma Louko for useful discussions about boundaries in quantum field theory and also Didier A. Solis for sending us a copy of his PhD thesis [29]. The work of M. B. is supported by a research grant funded by the Deutsche Forschungsgemeinschaft (DFG, Germany). Furthermore, M. B. would like to thank Istituto Nazionale di Alta Matematica (INdAM, Italy) for the financial support provided during his visit to the School of Mathematical Sciences of the University of Nottingham, whose kind hospitality is gratefully acknowledged. The work of C. D. is supported by the University of Pavia. C. D. gratefully acknowledges the kind of hospitality of the School of Mathematical Sciences of the University of Nottingham, where part of this work has been done. A. S. gratefully acknowledges the financial support of the Royal Society (UK) through a Royal Society University Research Fellowship, a Research Grant and an Enhancement Award. He also would like to thank the Research Training Group 1670 “Mathematics inspired by String Theory and Quantum Field Theory” for the invitation to Hamburg, where part of this work has been done. Open Access. This article is distributed under the terms of the Creative Com- mons Attribution 4.0 International License (http://creativecommons.org/licenses/ by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Appendix A: Some Concepts from Category Theory Adjunctions This is a standard concept, which is treated in any category theory textbook, e.g., [28]. Definition A.1. An adjunction consists of a pair of functors F : C D : G (A.1) M. Benini et al. Ann. Henri Poincar´ e together with natural transformations η :id → GF (called unit)and  : FG → id (called counit ) that satisfy the triangle identities Fη ηG F FGF G GF G (A.2) F G id id F G F G We call F the left adjoint of G and G the right adjoint of F , and write F  G. Definition A.2. An adjoint equivalence is an adjunction F : C  ∼ D : G (A.3) for which both the unit η and the counit  are natural isomorphisms. Existence of an adjoint equivalence in particular implies that C = D are equivalent as categories. Proposition A.3. If a functor G : D → C admits a left adjoint F : C → D, then F is unique up to a unique natural isomorphism. Vice versa, if a functor F : C → D admits a right adjoint G : D → C, then G is unique up to a unique natural isomorphism. Localizations Localizations of categories are treated for example in [21, Section 7.1]. In our paper we restrict ourselves to small categories. Definition A.4. Let C be a small category and W ⊆ Mor C a subset of the set −1 of morphisms. A localization of C at W is a small category C[W ] together −1 with a functor L : C → C[W ] satisfying the following properties: (a) For all (f : c → c ) ∈ W , L(f): L(c) → L(c ) is an isomorphism in −1 C[W ]. (b) For any category D and any functor F : C → D that sends morphisms −1 in W to isomorphisms in D, there exists a functor F : C[W ] → D and a natural isomorphism F F L. −1 C[W ] (c) For all objects G, H ∈ D in the functor category, the map Hom −1 G, H −→ Hom C GL,H L (A.4) C[W ] is a bijection of Hom-sets. −1 Proposition A.5. If C[W ] exists, it is unique up to equivalence of categories. References [1] B¨ ar, C., Ginoux, N., Pf¨ affle, F.: Wave Equations on Lorentzian Manifolds and Quantization. European Mathematical Society, Zuric ¨ h (2007) [2] Beem, J.K., Ehrlich, P.E., Easley, K.L.: Global Lorentzian geometry. Marcel Dekker, New York (1996) AQFT on Spacetimes with Timelike Boundary [3] Benini, M., Dappiaggi, C.: Models of free quantum field theories on curved back- grounds. In: Brunetti, R., Dappiaggi, C., Fredenhagen, K., Yngvason, J. (eds.) Advances in Algebraic Quantum Field Theory, pp. 75–124. Springer, Heidelberg (2015). arXiv:1505.04298 [math-ph] [4] Benini, M., Dappiaggi, C., Hack, T.-P.: Quantum field theory on curved backgrounds—a primer. Int. J. Mod. Phys. A 28, 1330023 (2013). arXiv:1306.0527 [gr-qc] [5] Benini, M., Schenkel, A., Woike, L.: Operads for algebraic quantum field theory. arXiv:1709.08657 [math-ph] [6] Benini, M., Schenkel, A., Woike, L.: Involutive categories, colored ∗-operads and quantum field theory. arXiv:1802.09555 [math.CT] [7] Brunetti, R., Fredenhagen, K., Verch, R.: The generally covariant locality prin- ciple: A new paradigm for local quantum field theory. Commun. Math. Phys. 237, 31 (2003). arxiv:math-ph/0112041 [8] Bussola, F., Dappiaggi, C., Ferreira, H.R.C., Khavkine, I.: Ground state for a massive scalar field in the BTZ spacetime with Robin boundary conditions. Phys. Rev. D 96(10), 105016 (2017). arXiv:1708.00271 [gr-qc] [9] Chru´ sciel, P.T., Galloway, G.J., Solis, D.: Topological censorship for Kaluza– Klein space-times. Ann. Henri Poincar´ e 10, 893 (2009). arXiv:0808.3233 [gr-qc] [10] Dappiaggi, C., Ferreira, H.R.C.: Hadamard states for a scalar field in anti-de Sit- ter spacetime with arbitrary boundary conditions. Phys. Rev. D 94(12), 125016 (2016). arXiv:1610.01049 [gr-qc] [11] Dappiaggi, C., Ferreira, H.R.C.: On the algebraic quantization of a massive scalar field in anti-de-Sitter spacetime. Rev. Math. Phys. 30, 1850004 (2018). arXiv:1701.07215 [math-ph] [12] Dappiaggi, C., Nosari, G., Pinamonti, N.: The Casimir effect from the point of view of algebraic quantum field theory. Math. Phys. Anal. Geom. 19, 12 (2016). arXiv:1412.1409 [math-ph] [13] Dutsc ¨ h, M., Rehren, K.-H.: Generalized free fields and the AdS-CFT correspon- dence. Ann. Henri Poincar´ e 4, 613 (2003). arXiv:math-ph/0209035 [14] Fewster, C.J.: Endomorphisms and automorphisms of locally covariant quantum field theories. Rev. Math. Phys. 25, 1350008 (2013). arXiv:1201.3295 [math-ph] [15] Fredenhagen, K.: Generalizations of the theory of superselection sectors. In: Kastler, D. (ed.) The Algebraic theory of Superselection Sectors: Introduction and Recent Results, p. 379. World Scientific Publishing, Singapore (1990) [16] Fredenhagen, K.: Global observables in local quantum physics. In: Araki, H., Ito, K.R., Kishimoto, A., Ojima, I. (eds.) Quantum and Non-commutative Analysis: Past, Present and Future Perspectives, pp. 41–51. Kluwer Academic Publishers, Dordrecht (1993) [17] Fredenhagen, K., Rehren, K.-H., Schroer, B.: Superselection sectors with braid group statistics and exchange algebras II: geometric aspects and conformal co- variance. Rev. Math. Phys. 4, 113 (1992) [18] Haag, R., Kastler, D.: An algebraic approach to quantum field theory. J. Math. Phys. 5, 848 (1964) [19] Ishibashi, A., Wald, R.M.: Dynamics in non-globally-hyperbolic static space- times II: general analysis of prescriptions for dynamics. Class. Quantum Gravity. 20, 3815 (2003). arXiv:gr-qc/0305012 M. Benini et al. Ann. Henri Poincar´ e [20] Ishibashi, A., Wald, R.M.: Dynamics in non-globally-hyperbolic static space- times III: anti-de Sitter space-time. Class. Quantum Gravity 21, 2981 (2004). arXiv:hep-th/0402184 [21] Kashiwara, M., Schapira, P.: Categories and Sheaves. Springer, Berlin (2006) [22] Kay, B.S.: The principle of locality and quantum field theory on (non globally hyperbolic) curved spacetimes. Rev. Math. Phys. 4, 167 (1992) [23] Lee, J.M.: Introduction to Smooth Manifolds. Springer, New York (2013) [24] O’Neill, B.: Semi-Riemannian Geometry. Academic Press, New York (1983) [25] Rehren, K.-H.: Algebraic holography. Ann. Henri Poincar´ e 1, 607 (2000). arXiv:hep-th/9905179 [26] Rejzner, K.: Perturbative Algebraic Quantum Field Theory: An Introduction for Mathematicians. Springer, Heidelberg (2016) [27] Ribeiro, P.L.: Structural and Dynamical Aspects of the AdS/CFT Correspon- dence: A Rigorous Approach. Ph.D. Thesis, University of Sao Paulo (2007). arXiv:0712.0401 [math-ph] [28] Riehl, E.: Category Theory in Context. Dover Publications Inc, New York (2016) [29] Solis, D.A.: Global Properties of Asymptotically de Sitter and Anti de Sitter Spacetimes. Ph.D. Thesis, University of Miami (2006). https:// scholarlyrepository.miami.edu/dissertations/2414/ [30] Sommer, C.: Algebraische Charakterisierung von Randbedingungen in der Quan- tenfeldtheorie. Diploma Thesis in German, University of Hamburg (2006). http://www.desy.de/uni-th/lqp/psfiles/dipl-sommer.ps.gz [31] Wald, R.M.: Dynamics in nonglobally hyperbolic, static space-times. J. Math. Phys. 21, 2802 (1980) [32] Wrochna, M.: The holographic Hadamard condition on asymptotically anti-de Sitter spacetimes. Lett. Math. Phys. 107, 2291 (2017). arXiv:1612.01203 [math- ph] [33] Zahn, J.: Generalized Wentzell boundary conditions and quantum field theory. Ann. Henri Poincar´ e 19, 163 (2018). arXiv:1512.05512 [math-ph] Marco Benini Fachbereich Mathematik Universit¨ at Hamburg Bundesstr. 55 20146 Hamburg Germany e-mail: marco.benini@uni-hamburg.de Claudio Dappiaggi INFN, Sezione di Pavia Universit` a di Pavia Via Bassi 6 27100 Pavia Italy e-mail: claudio.dappiaggi@unipv.it and AQFT on Spacetimes with Timelike Boundary Dipartimento di Fisica INFN, Sezione di Pavia Via Bassi 6 27100 Pavia Italy Alexander Schenkel School of Mathematical Sciences University of Nottingham University Park Nottingham NG7 2RD UK e-mail: alexander.schenkel@nottingham.ac.uk Communicated by Karl Henning Rehren. Received: January 3, 2018. Accepted: May 3, 2018.

Journal

Annales Henri PoincaréSpringer Journals

Published: May 30, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off