We propose a new deformed Rieffel product for functions in de Sitter spacetimes and study the resulting deformation of quantum field theory in de Sitter using warped convolutions. This deformation is obtained by embedding de Sitter in a higher-dimensional Minkowski spacetime, deforming there using the action of translations and subsequently projecting back to de Sitter. We determine the two-point function of a deformed free scalar quantum field, which differs from the undeformed one, in contrast to the results in deformed Minkowski spacetime where they coincide. Nevertheless, we show that in the limit where the de Sitter spacetime becomes flat, we recover the well-known non-commutative Minkowski spacetime.

The core statement of the geometrical measurement problem given in Refs. 1 and 2 is that spacetime, locally modeled as a manifold, should break down at very short distances of the order of the Planck length. This is a consequence of the limitations in possible accuracies of localization of spacetime points due to the interplay of the quantum-mechanical uncertainty principle with the formation of black holes in general relativity. Hence, a true theory of quantum gravity has to take the loss of classical spacetime into account. A way out of this dilemma and a path to put these arguments into a mathematical framework is the proposition of quantizing spacetime itself. Hence, a typical starting point is to represent spacetime by (representations of) an algebra V with generators x̂ obeying the following commutation relations:

x̂μ,x̂ν=2iΘμν,
(1)

where Θ is a constant skew-symmetric matrix and μ, ν = 0, 1, …, n. In this way, the commutative smooth manifold structure becomes non-commutative, and one can derive an uncertainty principle for the measurement of coordinates.

In a next step, one has to define a quantum field theory (QFT) in the non-commutative spacetime described by V, which was also done in a mathematically rigorous fashion in Ref. 2. The representation space of a QFT in the given non-commutative spacetime is the tensor product of VH, where H is the Hilbert space of the quantum field ϕ under consideration. It was soon recognized in Ref. 3 (see also Ref. 4) that there exists a unitary operator mapping the tensor product space VH to the Hilbert space H. Not only did this mapping make computations with the quantum fields easier but also it could be used to prove a weakened locality property of the deformed quantum fields ϕΘ (the quantum fields in the non-commutative spacetime), known as wedge-locality, and enabled the non-perturbative calculation of the scattering of two deformed quantum fields.

The next advance in the area of strict deformations of QFT was the realization of warped convolutions.5,6 These are a well-defined mathematical prescription on how to obtain the deformed quantum field ϕΘ from the undeformed one ϕ, namely, as

ϕΘ=τΘx(ϕ)dE(x),
(2)

where dE is the spectral measure with respect to the translations on Rn and τ denotes the action of Rn on functions by translation: τaϕ(x) = ϕ(x + a). It was then shown6 that this prescription is connected to the Rieffel product,7 a noncommutative associative product of functions given by

f×Θg=(2π)nτΘx(f)τy(g)eixydnxdny,
(3)

where the oscillatory integral is well-defined for functions f,gD, as defined in  Appendix A. In particular, warped convolutions supply an isometric representation of Rieffel’s strict product,

AΘBΘ=A×ΘBΘ,
(4)

for operators A, B that belong to the subclass C of a C*-algebra, which is smooth with respect to the adjoint action generated by the translation group.

The scope of warped convolutions, however powerful, is unfortunately limited to smooth actions of Rn, i.e., mutually commutative generators of the deformation. Hence, to define the deformations, one needs a set of commuting operators, at least two. For a curved manifold, commuting Lie vector fields can supply such an action;8 another possibility would be a spacetime with commuting Killing vector fields.9 Yet, as is well known, such vector fields do not exist for arbitrary curved spacetimes, which prevents the application of deformations in general curved spacetimes.

A way out is to embed the curved spacetime manifold (or a patch thereof) into a higher dimensional flat (Minkowski) spacetime. To illustrate the concept, we consider an n-dimensional de Sitter spacetime, which is embedded into an (n + 1)-dimensional Minkowski spacetime R1,n in the standard straightforward way. One obtains in this way an n-dimensional group of translations that allow a straightforward definition of the Rieffel product and warped convolutions. Apart from this well-known and comparatively simple embedding, we choose de Sitter spacetime for its importance in early universe cosmology. In particular, de Sitter spacetime is a good description of the inflationary period in cosmology shortly after the Big Bang [see, for example, Ref. 10 (Chap. 11) or Ref. 11], where the universe expands exponentially, and one expects that the quantum nature of spacetime becomes important. The expansion rate is given by the Hubble parameter H, and in the limit H → 0, the de Sitter spacetime becomes the n-dimensional Minkowski spacetime R1,n1. We show that in this limit, our definition also recovers the non-commutative Minkowski spacetime (1), i.e., the deformation and the limit H → 0 commute.

The remainder of this article is structured as follows: in Sec. II, we review the embedding formalism for de Sitter, define the deformed (Rieffel) product, and determine the resulting non-commutative spacetime. In Sec. III, we determine the corresponding deformed quantum field theory of a free scalar field and compute the two-point function of the deformed quantum field ϕΘ. In contrast to the Minkowski spacetime, where the deformed two-point function coincides with the undeformed one, we show that in the de Sitter spacetime the two differ. We conclude in Sec. IV and collect various proofs in  Appendixes A–D.

Assumption I.1.

In the following, we assume that all functions that we consider are elements of the domainDdefined in  Appendix A.

To define the Rieffel product for functions on a manifold M, one needs an action of Rn. There are various ways to achieve such an action, for example, using commuting Killing vector fields, which, however, do not exist for a general curved spacetime. Therefore, we propose to embed the spacetime under consideration into a higher-dimensional Minkowski spacetime and deform the action of the (commutative) translations along the embedding coordinates, hence an action of Rn with which one can define a Rieffel product. In our case, the n-dimensional de Sitter spacetime can be embedded in the (n + 1)-dimensional Minkowski spacetime R1,n, which is the embedding with the least dimension possible, i.e., a minimal one.

Given thus R1,n with Cartesian coordinates XA, A = 0, 1, …, n and the flat metric ηAB, the de Sitter spacetime is defined as the hyperboloid

ηABXAXB=1H2,
(5)

where H is the Hubble parameter, a positive constant, and we use the Einstein summation convention throughout this paper. The part of the de Sitter spacetime that is relevant in cosmology is known as the (flat, expanding) Poincaré patch11 and has cosmological time t and spatial Cartesian coordinates xa, a = 1, …, n − 1, as coordinates. Those parameterize the hyperboloid (5) according to

X0=1Hsinh(Ht)+H2exp(Ht)x2,
(6a)
Xa=exp(Ht)xa,
(6b)
Xn=1Hcosh(Ht)H2exp(Ht)x2,
(6c)

where x2 = δabxaxb.38 The induced metric reads

g=dt2+exp2Htdx2,
(7)

and as H → 0, we have X0t, Xaxa, and gη such that the flat Minkowski space is recovered. In the following, we will consider exclusively the Poincaré patch but continue to call it de Sitter for brevity, in accordance with standard practice in cosmology.

By inverting the parameterization (6), which results in

t=1HlnH(X0+Xn),xa=XaH(X0+Xn),
(8)

functions f(x) defined on the de Sitter spacetime can be extended to the full embedding spacetime R1,n. We will make extensive use of this extension in the following modified form:

Definition II.1.

Given a functionfdefined on the de Sitter spacetime, we denote byf(X) its extension to the embedding spacetime using the inverse parameterization(8), withXnreplaced byH2+(X0)2δabXaXb. That is,f(X) is the (uniquely determined) function onR1,n, which is independent ofXnand agrees with the original functionfwhen restricted to the Poincaré patch(6).

To shorten the formulas, we will nevertheless continue to write Xn, with the implicit understanding that it is just a shorthand for H2+(X0)2δabXaXb. We note that this extension is, of course, not unique but very convenient for our purposes.

In the following, we define the deformed product using the embedding formalism. The definition is analogous to the case where a smooth action of the group Rn exists, i.e., à la Rieffel.7 In our case, the smooth action τ of Rn acts as a translation in the embedding coordinates on functions extended to the embedding space.

Definition II.2.
Let the smooth actionτof the groupRndenote translations with respect to the embedding coordinatesXμ = (X0, Xa). Moreover, let Θ be a constantn × nskew-symmetric matrix. The non-commutative product between two functionsf,gDis defined in analogy to Refs.6 and 7by
f×Θg(z)(2π)nlimϵ0+χ(ϵX,ϵY)τΘX(f(Z))τY(g(Z))eiXYdnXdnY=(2π)nlimϵ0+χ(ϵX,ϵY)f(Z+ΘX)g(Z+Y)eiXYdnXdnY,
(9)
whereZis the embedding point corresponding tozaccording to Definition II.1, the scalar productX · Yis with respect to the flat Minkowski metricημν, whereμ, ν = 0, …, n − 1, i.e.,X · Y = −X0Y0 + δabXaYb, and the integrations are with respect toX0andXa(respectively,Y0andYa)only. Moreover, the cutoff functionχS(R2n)is arbitrary except for the conditionχ(0, 0) = 1.

Note that the action τ on functions f extended from the de Sitter spacetime to the embedding space is well-defined, since the extension is constant along the Xn direction. For the same reason, the deformation matrix Θ only has size n × n and does not act on the last embedding coordinate Xn at all. While in embedding coordinates, the action τ is linear, in intrinsic coordinates, it is a highly non-linear operation. By its very definition, the deformed product (9) strongly depends on the exact choice of the extension of functions to the embedding space, but we will see later that our choice has phenomenological advantages.

Apart from the use of translations with respect to the embedding coordinates instead of intrinsic coordinates, this deformed product is completely analogous to the Rieffel product. Hence, we have the following result:

Proposition II.3.
The deformed product given in Definition II.2 is associative, i.e.,
f×Θg×Θh(z)=f×Θg×Θh(z)
(10)
for all functionsf,g,hD, and fulfills
  • limΘ→0f × Θg = fg,

  • f × Θ1 = 1 × Θf = f, where 1 is the constant function equal to 1.

Proof.

The proof is identical to the existing proofs of the Rieffel product using smooth actions of Rn {see Theorem 2.14 of Ref. 7 for the associativity condition (10), Corollary 2.8 of Ref. 7 for the limit of vanishing Θ [Proposition II.3 (i)], and Corollary 2.13 of Ref. 7 for the unit element condition [Proposition II.3 (ii)]}.□

Moreover, in the limit H → 0, it becomes the flat-space Rieffel product.

Proposition II.4.

If the limitH → 0 exists for functionsfandg, their deformed product turns into the Rieffel product in that limit.

Proof.

Choosing a cutoff function χ that is independent of H, the proposition follows from the fact that as H → 0, we have X0t and Xaxa (6) and that nothing in definition (9) of the deformed product depends on Xn such that the integrand has a smooth pointwise limit. For fixed ϵ, we can interchange the limit H → 0 and the integration in (9) because of the cutoff function, and the Rieffel product is the limit ϵ → 0 of the resulting integral.□

In the case of deformations on Rn or the Lorentzian analog thereof R1,n1, it is well-known that the Rieffel product leads to non-trivial but simple commutation relations between the coordinate functions. In particular, it leads to a constant non-commutative spacetime with commutation relations

xμ,xνΘ=2iΘμν,
(11)

where we defined the following:

Definition II.5.
The deformed commutator [·, ·]Θbetween two functionsf, gis given by the commutator between those two functions with the deformed instead of the pointwise product,
f,gΘf×Θgg×Θf.
(12)

For more details about the commutation relations on Minkowski space, see Ref. 12 (Lemma 5.2). Since we have defined the deformed product completely analogous but using the embedding coordinates, we have the same result for the n embedding coordinates Zμ and Zν,

Zμ,ZνΘ=2iΘμν.
(13)

Using that

(2π)nlimϵ0+χ(ϵX,ϵY)f(X)eiXYdnXdnY=f(0),
(14)

Eq. (13) follows from a straightforward computation,

Zμ×ΘZν(z)=(2π)nlimϵ0+χ(ϵX,ϵY)(Z+ΘX)μ(Z+Y)νeiXYdnXdnY=ZμZνi(2π)nlimϵ0+χ(ϵX,ϵY)(ΘX)μXνeiXYdnXdnY=ZμZν+iΘμν,
(15)

where integration by parts was permitted because χS(Rn), and the term with the derivative acting on χ vanishes in the limit ϵ → 0. Due to the skew symmetry of Θ, one then obtains the commutator relation (13).

However, since the embedding spacetime is only an auxiliary construct, we are more interested in the commutation relations that this new deformed product induces from the perspective of the intrinsic de Sitter coordinates. We obtain the following result:

Theorem II.6.
The deformed commutator relations of the de Sitter spacetime are (to first order in Θ) given by
t,xbΘ=2ie2HtΘ0bΘρbFρ+OΘ2,
(16a)
xa,xbΘ=2ie2HtΘabHxaΘ0bΘρbFρ+HxbΘ0aΘρaFρ+OΘ2,
(16b)
whereFis then-dimensional vector,
FμXμXn=cosh(Ht)H22exp(Ht)x21sinh(Ht)+H22exp(Ht)x2Hexp(Ht)xa.
(17)

Proof.
We first derive a formally equivalent representation of the deformed Rieffel product (Definition II.2) in terms of differential operators, which is useful for an expansion in powers of the non-commutativity matrix Θ. For this, we expand f(Z + ΘX) in Eq. (9) in a Taylor series in Θ, which gives
f×Θg(z)=k=01k!(2π)nlimϵ0+χ(ϵX,ϵY)(ΘX)kkf(Z)g(Z+Y)eiXYdnXdnY=k=0(i)kk!kf(Z)(2π)nlimϵ0+χ(ϵX,ϵY)g(Z+Y)(ΘY)keiXYdnXdnY,
(18)
where the sum over k is to be understood in the sense of formal power series in Θ. We then integrate the derivatives with respect to Y by parts, which is permissible again because χS(R2n), and note that derivatives acting on χ result in powers of ϵ and hence vanish in the limit ϵ → 0. Using then Eq. (14), we obtain
f×Θg(z)=k=0ikk!kf(Z)(Θ)kg(Z)=expiΘμνXμYνf(X)g(Y)X=Y=Z,
(19)
which is valid as formal power series in Θ. It becomes non-formal (i.e., strict) if the series of derivatives converges, i.e., if f and g are analytic functions. Note, however, that the original Definition II.2 is valid for all functions f,gD.
We then use the inverse parameterization (8) to compute
xa,xbΘ=xaxb+iXμxaΘμνYνybX=Y+OΘ2(ab)=i1H(X0+Xn)δμaXaH(X0+Xn)2δμ0XμXnΘμν×1H(X0+Xn)δνbXbH(X0+Xn)2δν0XνXn(ab)+OΘ2,
(20)
where we recall that Xn is a shorthand for H2+(X0)2δabXaXb=H2ημνXμXν according to Definition II.1. Using that from the relations (6), it follows H(X0 + Xn) = exp(Ht) and Xa = exp(Ht)xa, the result follows. In the same way, we obtain
t,xaΘ=iXμtΘμνYνyaXYX=Y+OΘ2=i1H(X0+Xn)δμ0XμXnΘμν1H(Y0+Yn)δνaYaH(Y0+Yn)2δν0YνYnXYX=Y+OΘ2,
(21)
which with H(X0 + Xn) = exp(Ht) and Xa = exp(Ht)xa give the stated result.□

Corollary II.7.
The resulting non-commutative spacetime given in Theorem II.6 becomes the Moyal–Weyl spacetime in the flat limitH → 0, i.e.,
limH0xμ,xνΘ=2iΘμν.
(22)

Proof.

To first order in Θ, this can be seen directly from Eq. (16), and in general, it follows from Proposition II.4.□

In particular, since deformation of the Minkowski spacetime using the Rieffel product leads to the non-commutative Moyal–Weyl spacetime, this shows that the deformation and the limit commute, which is shown in Fig. 1.

At this point, we want to mention that the deformed commutator relations fulfill the Jacobi identity as a direct consequence of the associativity of the deformed product, Proposition II.3.

Several motivations, both mathematical and physical, lead to the assumption of a non-commutative spacetime. One physical argument, for example, is the geometrical measurement problem,1,2 as explained in the Introduction. An example where a non-commutative space is physically realized is the Landau quantization. This physical situation occurs for non-relativistic electrons confined to a plane in the presence of a homogeneous magnetic field (see Ref. 13).

FIG. 1.

Commuting diagram for deformations of the de Sitter and Minkowski spacetime. Here, Θ(H) represents the deformed commutator relations given in Theorem II.6, and dSn is the expanding Poincaré patch of the de Sitter spacetime.

FIG. 1.

Commuting diagram for deformations of the de Sitter and Minkowski spacetime. Here, Θ(H) represents the deformed commutator relations given in Theorem II.6, and dSn is the expanding Poincaré patch of the de Sitter spacetime.

Close modal

If one, however, assumes that a non-commutative spacetime, a quantum spacetime, can resolve the Big Bang singularity, it is clear that a deformed flat spacetime (of Moyal–Weyl type) is not a satisfying model. Instead, one needs to consider curved (or generically dynamical) non-commutative spacetimes, such as the one we have obtained in Theorem II.6. Moreover, a striking feature of the non-commutativity has appeared naturally: as time passes and the universe expands, the non-commutativity dynamically vanishes. After the end of inflation, the non-commutativity is completely negligible even if at its beginning, it is of order one, since the overall multiplying factor exp(−Ht) decreases at least by e−60 from the beginning to the end of inflation [see Ref. 10 (Chap. 12)]. Since the assumption of a constant non-commutative spacetime, even for very small non-commutativity, leads to violations of Lorentz symmetry, which have not been observed,14 such models are unsustainable; even if one fundamentally violates Lorentz symmetry only at the Planck scale, quantum fluctuations of the metric enhance this breaking also for low energies.15 In contrast, the dynamical non-commutativity that results in our proposal can easily accommodate no detectable violations of Lorentz symmetry during most of the cosmological evolution, including today, and only modifies the inflationary period and the Big Bang singularity. Concrete predictions and results from this model are under investigation.

Rigorous deformations of quantum field theory in the Minkowski spacetime have been initiated by Refs. 3 and 4 and connected to the Rieffel product in Ref. 6 (see also Refs. 12, 16, and 17 and the reviews papers18,19). However, for deformations of QFT in the Minkowski spacetime, the two-point function of the deformed quantum field ϕΘ is identical to the one of the undeformed quantum field ϕ. Since the measured scalar power spectrum of the cosmic microwave background is directly related to the two-point function of the metric perturbations,11 this deformation thus has no observable effects, which is quite unsatisfactory. As we will see in this section, by deforming the QFT in the de Sitter spacetime using the deformed Rieffel product defined in Sec. II A, we obtain a deformed non-commutative model whose two-point function differs from the undeformed one, which could therefore give observable effects.

To construct the deformed model, we follow the same steps of the construction as in the Minkowski spacetime.3,4,12 That is, we first construct the undeformed QFT using canonical quantization. In a second step, we need to define a unitary group that implements the deformation action of Rn on the quantum fields, which we do by constructing the infinitesimal generators of translations and exponentiating them. In Minkowski space, these are just the global momentum operators, but we will see that in our case, they are not global but depend on the point where they act. Finally, we define the deformed quantum field by taking our deformed Rieffel product between the quantum field and the elements of the unitary group and compute its two-point function.

For simplicity, we consider a free massless and conformally coupled field ϕ that fulfills the Klein–Gordon equation

2n(n2)4H2ϕ(x)=0,
(23)

where the term proportional to H2 results from the conformal coupling.20 In the Poincaré patch with metric (7), this reads

t2(n1)Ht+exp(2Ht)n(n2)4H2ϕ(t,x)=0,
(24)

where δabxaxb. The canonical momentum is given by

π=detgtϕ=exp(n1)Httϕ,
(25)

and following canonical quantization, the mode expansion of the field operator reads

ϕ(x,t)=a(p)f(p,t)+a(p)f*(p,t)eipxdn1p(2π)n1,
(26)

with creation and annihilation operators fulfilling the commutation relation

a(p),a(q)=(2π)n1δn1(pq).
(27)

With the mode functions normalized to

f(p,t)tf*(p,t)f*(p,t)tf(p,t)=iexp(n1)Ht,
(28)

we have the canonical commutation relations

ϕ(x,t),π(y,t)=iδn1(xy).
(29)

Plugging expansion (26) into the Klein–Gordon equation (24), the mode functions are seen to fulfill

t2+(n1)Ht+exp(2Ht)p2+n(n2)4H2f(p,t)=0,
(30)

which has the positive-frequency solution [normalized according to Eq. (28)]

f(p,t)=12pexpn22Ht+iexp(Ht)1pH12pexpipt(H0).
(31)

The operator a(p) annihilates the vacuum vector 0, which with the above choice of mode functions is the Bunch–Davies vacuum.21 By standard methods, we obtain the one-particle Hilbert space H,39 from which we the define the bosonic Fock space F as the completion of F0=k=0Hk, i.e., FF0̄ with H0=C and Hk=HSk, where ⊗S denotes the symmetric tensor product.

Since our deformed product is obtained from the action of Rn in the embedding space, we now construct the operators that implement infinitesimal translations in embedding space. Because these translations commute, we obtain in this way n linearly independent commuting operators Pμ acting as

Pμ(X),ϕ(X)=iXμϕ(X),
(32)

where we anticipated that they depend on the point where they act. Here, we recall that ϕ(X) denotes the extension of ϕ(x) [defined by the expansion (26)] to the embedding space according to Definition II.1. From the inverse parameterization (8), we compute

tX0=1HXn,
(33a)
xaX0=xaXn,
(33b)
tXb=XbHXn(X0+Xn)=xbXn,
(33c)
xaXb=1H(X0+Xn)δab+HxaxbXn,
(33d)

where we recall that Xn=H2+(X0)2δabXaXb, and obtain in intrinsic coordinates

P0(X),ϕ(X)=ixμX0xμϕ(X)=iHXntϕ(x)Hxbxbϕ(x)
(34)

and

Pa(X),ϕ(X)=iHXnXnX0+Xnxaϕ(x)+H2xaxbxbϕ(x)Hxatϕ(x).
(35)

As H → 0, we have HXn → 1 and HX0 → 0 (6) and recover the correct flat limit. To determine the actual form of the Pμ(X), we use expansion (26), make an ansatz quadratic in creation and annihilation operators, and evaluate the unknown coefficients by requiring (34) and (35) to hold. The derivation and result are given in  Appendix B. By standard arguments,23 the so-defined operators Pμ are essentially self-adjoint operators on a dense domain D1F for each fixed point X, in general. From the explicit result of  Appendix B, Pμ do not annihilate the vacuum vector 0, contrary to the Minkowski case (even though their expectation value vanishes). We conclude that they do not coincide with the standard momentum operators that can be defined in de Sitter QFT24–26 and that do annihilate the vacuum vector. However, for us, this is a good result and the essential reason why we are able to obtain a two-point function that differs from the undeformed one in the following. From the operators Pμ(X), one obtains the unitary group implementing finite translations again in the standard way.

Definition III.1.
For all pointsXμ = (X0, Xa),Yμ, andZμin the embedding space, a two-parameter group of unitary operators is uniquely defined by
U(X,Y)=U(Y,X),U(X,X)=1,XμU(X,Y)=iPμ(X)U(X,Y).
(36)

Completely analogous to the Minkowski case,3,4,6 we then define the deformed field operators.

Definition III.2.
Let Θ,χ, and the dot product be as in the definition of the deformed product (Definition II.2), letϕ(x) be the free scalar field(26), and letτdenote the automorphism defined by the adjoint action of the unitary group defined in Definition III.1,
τYϕ(X)=U(X+Y,X)ϕ(X)U(X,X+Y)=ϕ(X+Y).
(37)
The deformed quantum fieldϕΘ(x) is then formally defined by warped convolution as
ϕΘ(z)(2π)nlimϵ0+χ(ϵX,ϵY)τΘXϕ(Z)U(Z+Y,Z)eiXYdnXdnY
(38)
whenever the limit exists, and the deformed fieldϕΘ(x)is defined by
ϕΘ(z)(2π)nlimϵ0+χ(ϵX,ϵY)U(Z,Z+Y)τΘXϕ(Z)eiXYdnXdnY.
(39)
In order to define the field (an operator-valued distribution) rigorously, one has to smear it with a real Schwartz functionfS(Rn)(the reality condition ensures the self-adjointness of the scalar field), andwe denote the smeared field byϕΘ(f).

It was proven in Lemma 2.2 of Ref. 5 that the two definitions ϕΘ and ϕΘ agree in Minkowski space, and the proof can be taken over. Moreover, comparing with the definition of the deformed product (Definition II.2), one has formally

ϕΘ(z)=ϕ×ΘU(,z)(z),ϕΘ(z)=U(z,)×Θϕ(z).
(40)

Furthermore, in analogy with Ref. 3 (Proposition 2.2 b) and Ref. 13 (Proposition 3.1), we want to prove that the (smeared) deformed operators ϕΘ(f) are well-defined operators in the Fock space F.

Proposition III.3.

Let us consider the smeared deformed fieldϕΘ(f), and letΦ,ΨFbe such thatPμ(X1)Pμ(Xk)Φ<for allk ≤ 2n + 2. Then, the scalar productΦ,ϕΘ(f)Ψ=Φ,ϕΘ(f)Ψis bounded. Hence, the deformation of the free field in the de Sitter spacetime, given by the deformed fieldϕΘ(f) of Definition III.2, is well-defined.

Proof.
We have
Φ,ϕΘ(f)Ψ=(2π)nlimϵ0+χ(ϵX,ϵY)bΘ(X,Y,f)eiXYdnXdnY
(41)
with bΘ(X, Y, f) formally given by
bΘ(X,Y,f)=f(Z)Φ,U(Z,Z+Y)τΘXϕ(Z)ΨdnZ.
(42)
Using Eq. 1.4 of Ref. 7 and choosing a suitable norm, the oscillatory integral in Eq. (41) converges absolutely, and we have for some k > n (e.g., k = n + 1) the estimate
Φ,ϕΘ(f)Ψcki+j2ksupX,YXiYjbΘ(X,Y,f)
(43)
for some constant ck. We use a standard trick and decompose the test function as
f(Z)=f(Z)(1+δμνZμZν)n(1+δμνZμZν)nf̃(Z)g(Z),
(44)
where f̃(Z)f(Z)(1+δμνZμZν)n is still a Schwartz test function. We then compute for i = j = 0 that
supX,YbΘ(X,Y,f)=supX,Yf(Z)Φ,U(Z,Z+Y)τΘXϕ(Z)ΨdnZ=supX,Yf̃(Z)g(Z)U(Z,Z+Y)Φ,τΘXϕ(Z)ΨdnZsupX,YU(g,Y)ΦτΘXϕ(f̃)Ψ,
(45)
where we used the Cauchy–Schwarz inequality, and we denote by U(g,Y) the smearing of U(Z,Z + Y) in Z with g. Since U is a unitary operator and thus of norm 1, we have
supYU(g,Y)Φg1Φ
(46)
with g1< the L1 norm of g, which is finite by construction. Furthermore, we have
supXτΘXϕ(f̃)Ψ=supXϕf̃τΘXΨc(Θ,f̃)Ψ,
(47)
where we used that (because the smeared field ϕ is a well-defined operator in Fock space)
ϕ(f)Ψc(f)Ψ
(48)
holds for some positive constant c depending on the Schwartz seminorms fα,β of the test function f and that (with Θ being the maximum modulus of the components of Θ)
supXf̃τΘXα,β=supX,YYβαf̃(YΘX)Θβf̃α,β.
(49)
For derivatives with respect to X, one obtains instead derivatives of ϕ and powers of Θ, which have again finite norm [possibly with a different constant c′(f)], while for derivatives with respect to Y acting on U, we obtain powers of Pμ. Requiring that the norm of these acting on Φ be finite, we obtain the stated restriction on the vector Φ and the conclusion.□

Finally, we compute the two-point function of the deformed field, up to first order in Θ. For this, we first expand the deformed field to first order in Θ. Using the definition of the unitary group (Definition III.1) and the automorphism τ (37), we obtain

τΘXϕ(Z)=ϕ(Z)iΘμνXνPμ(Z),ϕ(Z)+OΘ2.
(50)

Inserting this result into definition (38) and using (14), it follows that

ϕΘ(z)=ϕ(Z)Θμν(2π)nlimϵ0+χ(ϵX,ϵY)Pμ(Z),ϕ(Z)U(Z+Y,Z)YνeiXYdnXdnY+OΘ2=ϕ(Z)+ΘμνPμ(Z),ϕ(Z)YνU(Z+Y,Z)Y=0+OΘ2=ϕ(Z)iΘμνPμ(Z),ϕ(Z)Pν(Z)+OΘ2.
(51)

The computation of the two-point function is then long but straightforward, expressing ϕ and the Pμ in terms of annihilation and creation operators, acting with them on the vacuum vector 0 and computing the resulting momentum integrals. We obtain the following theorem:

Theorem III.4.
The two-point function of the non-commutative scalarϕΘin then-dimensional de Sitter space reads
0ϕΘ(x)ϕΘ(y)0=limϵ0+Γn222(2π)n2Hn21Z(X,Y)+iϵsgn(ts)2n2+i(n2)Γn24(2π)n2exp(Hs)Xn+exp(Ht)YnxaΘa0+Θ0bybHxaΘabyb×Hn1Z(X,Y)+iϵsgn(ts)n2+OΘ2,
(52)
with the de Sitter invariant
Z(X,Y)=H2ηABXAYB=cosHμ(x,y)=coshH(ts)H22exp[H(t+s)](xy)2,
(53)
whereμ(x, y) is the geodesic distance betweenxandyif they are connected by a geodesic.

Proof.

See  Appendix D.□

For applications in inflationary cosmology, one needs to consider the two-point function at equal times. We have the following corollary:

Corollary III.5.
Fort, the leading behavior of the two-point function of the non-commutative scalarϕΘat equal times is given by
0ϕΘ(x)ϕΘ(y)0Γn224πn2(x̃ỹ)22n2+i(n2)Γn2πn2Hexp(Ht)x̃aΘa0+Θ0bỹb+OΘ2,
(54)
wherex̃exp(Ht)xare the physical (co-moving) coordinates.11We see that the corrections coming from the non-commutativity are again exponentially suppressed.

Remark III.6.
Assuming that Θμνis independent ofH, we can also expand the result (Theorem III.4) for smallHand obtain up to first order in Θ and second order inHthe following result:
0ϕΘ(X)ϕΘ(Y)0=limϵ0+Γn22Hn22(2π)n21Z(X,Y)+iϵsgn(ts)2n2+i(n2)Γn24πn2H(xy)2+iϵsgn(ts)n22xaΘa0+2Θ0byb2HxaΘabyb+H(t+s)xaΘa0+Θ0byb1n(xy)2(xy)2+iϵsgn(ts),
(55)
where (xy)2 ≡ (xy)2 − (ts)2. Note that the non-commutativity vanishes forn = 2, already before expanding for smallH, which is due to the fact that in this case the operatorsPμannihilate the vacuum vector0, analogously to what happens in the Minkowski space. Furthermore, one sees very clearly that the non-commutative contributions also vanish in the limitH → 0.

At this point, we note an important difference to the Minkowski case: even though the deformation of the quantum fields is defined by the same warped convolution as in Ref. 6, it does not give an isometric representation of the deformed Rieffel product, i.e., in our case,

A×ΘBΘAΘBΘ,
(56)

in general. This results from the fact that the infinitesimal generators of translations Pμ(X) are not global but depend on a point X and do not commute at different points: [Pμ(X), Pν(Y)] ≠ 0 for XY (see  Appendix C).

We have proposed a new way of deforming quantum field theories on curved spacetimes by embedding them into a higher-dimensional Minkowski spacetime and constructing a Rieffel product and warped convolutions in the embedding spacetime. In our model, not only does the non-commutativity decrease exponentially as time progresses (such that constraints from Lorentz violation are fulfilled) but also the two-point function of the deformed scalar field differs from the undeformed case, with possible observable consequences. As an extension of the present work, it would be very interesting to determine potentially observable effects during the inflationary period and to compare with other approaches to non-commutative geometry, such as those of Refs. 2730.

Apart from the de Sitter spacetime, there exists a straightforward generalization of our deformation procedure to all n-dimensional maximally symmetric space(-times), with the embedding spacetime being Euclidean space (for the sphere), Minkowski spacetime (for hyperbolic space), or R2,n1 (for the anti-de Sitter spacetime). For more general curved spacetimes, the question of embedding is somewhat more complicated. It is known by the Nash embedding theorem that every Riemannian manifold (compact or not) can be isometrically embedded in Rm for some large enough m. In Ref. 31, this was extended to certain classes of Lorentzian manifolds (stably causal spacetimes with steep temporal function), which, in particular, include globally hyperbolic spacetimes in which quantum field theories can be defined rigorously. In this way, we can define a strict deformation of quantum field theories in any globally hyperbolic manifold, at least in principle.

While the chosen embedding for the de Sitter spacetime is standard, it is not unique, and we expect the noncommutative spacetime to change, in general, depending on the embedding. It is clear that nothing can change if one takes as the embedding spacetime the direct product R1,n×M with any space(-time) M, extends functions to be constant on M, and only deforms the first n coordinates of the R1,n factor as in our case. For a more general embedding, it is unclear what exactly might happen. Even if we keep the same embedding spacetime R1,n but change the way functions are extended from de Sitter to the full R1,n or which coordinates are deformed, the results will change. However, since Xμxμ as the Hubble parameter H → 0 while Xn diverges in this limit, any such change will destroy the commutative diagram shown in Fig. 1. That is, our choice is the only one where the limit H → 0 and the deformation commute, apart from the previously mentioned phenomenological advantages. Another criterion that one could employ to reduce the ambiguity in the embedding and the deformation is dynamical locality,32,33 which requires that descriptions of local physics based on kinematical and dynamical considerations should coincide. However, to use this criterion, one would first need to verify whether the deformed theories that we study correspond to local quantum field theories (we thank R. Verch for this comment).

For simplicity, we had considered only a massless, conformally coupled scalar field, where the de Sitter two-point function is given by a conformal rescaling of the Minkowski two-point function. However, already in this simplest example, this relation only holds for the undeformed two-point function but not for the deformed one, since in flat space, the deformed two-point function is equal to the undeformed one, whereas they differ in the de Sitter spacetime. For a scalar field with more general mass m or curvature coupling ξ, it is well-known20 that for m2 + ξR = m2 + ξn(n − 1)H2 > 0, the undeformed two-point function only depends on the de Sitter invariant Z, which for equal times only depends on the physical (co-moving) coordinates. Hence, at late times t, we would expect results very similar to Corollary III.5, namely, that the corrections coming from the non-commutativity are exponentially suppressed with respect to the undeformed two-point function.

Finally, it would be very interesting to connect our proposed deformation to the Weyl quantization for curved spacetimes defined in Ref. 34. We first remark that while heuristically the product of two Weyl quantized functions f and g should be related to the Weyl quantization of the deformed product (or the Rieffel product in the flat case) of the two functions, the relation cannot be totally straightforward since Weyl quantization concerns functions defined on phase space, i.e., depending on both coordinates and momenta, while the deformed product we presented only concerns functions defined on the manifold with no dependence on momenta. One possibility to make a comparison would be given by considering instead of just de Sitter space its direct product with R1,n1, which can be interpreted as momentum space. The embedding of the whole manifold into R2,2n1 is then straightforward, and one could define a deformed product in total analogy to what we have presented before. However, the physical interpretation of such a construction is not clear, since as we have seen the “momentum operators” Pμ(X) that induce translations of the quantum field operators depend on position, and no global momentum exists. Since for quantum field theory in curved spacetimes, the two-point function plays an essential role, a much more relevant method would be the comparison of the deformed two-point functions. For this, one would need to apply the Weyl quantization and the resulting star product of Ref. 34 to free scalar quantum field theory, with the scalar field and its conjugate momentum in canonical quantization playing the role of position and momenta. The resulting two-point function could then be compared with the one of our non-commutative field ϕΘ, and one could verify whether there is a choice of Θ that makes the result agree (at least to first order) with the one that would result using the methods of Ref. 34.

This work was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation)—Project Nos. 415803368 and 406116891 within the Research Training Group RTG 2522/1. We thank S. Franchino-Viñas for comments and the anonymous referee for various suggestions that helped improve the manuscript.

Data sharing is not applicable to this article as no new data were created or analyzed in this study.

The definitions and results of oscillatory integrals that we use are given in this section. First, we define an oscillatory integral as in Ref. 35 (Chap. 7.8).

Definition A.1.

LetXRnbe open, and let Γ be an open cone onX×RN\{0}for someN, i.e., for all (x, y) ∈ Γ also (x, λy) ∈ Γ for allλ > 0. We say that a functionϕC(Γ) is aphase functionin Γ if

  • ϕ(x, λy) = λϕ(x, y) for all (x, y) ∈ Γ andλ > 0;

  • Imϕ(x,y)0for all (x, y) ∈ Γ;

  • ϕ(x, y) ≠ 0 for all (x, y) ∈ Γ.

Then, an integral of the form (see Ref.35 , Eq. 7.8.1)
eiϕ(x,y)b(x,y)dNy
(A1)
is called anoscillatory integral.

Another definition that we use is that of a symbol [see Ref. 35 (Definition 7.8.1)].

Definition A.2.
Let m,ρ, andδbe real numbers with 0 < ρ ≤ 1 and 0 ≤ δ < 1. Then, we denote bySρ,δm(X×RN)the set of allbCX×RNsuch that for every compact setKXand allα,β, there exists some constantCα,β,Ksuch that the estimate
xβyαb(x,y)Cα,β,K(1+y)mρα+δβ
(A2)
is valid for all(x,y)K×RN. The elementsSρ,δmare called symbols of order m and typeρ, δ.

It is proven in Theorem 7.8.2 of Ref. 35 (see also Refs. 36 and 37) that if bSρ,δm and m < −N + 1, the oscillatory integral (A1) converges to a well-defined function. In the case m ≥ −N + 1, the oscillatory integral has to be defined in a distributional manner.

The Rieffel product (3), given as an oscillatory integral with phase function ϕ(x, y) = x · y, thus either converges to a well-defined function or can still be defined in a distributional manner if the integrand is a symbol. Hence, we define the domain D to be the set of all functions f, g such that

τΘx(f)τy(g)Sρ,δm.
(A3)

Here, we explicitly determine the operators defined by Eqs. (34) and (35) in terms of the creation and annihilation operators. We make the ansatz

Pμ(X)=pμ(q,X)a(q)a(q)+rμ(q,X)a(q)a(q)+rμ*(q,X)a(q)a(q)dn1q(2π)n1
(B1)

and compute

Pμ(X),ϕ(X)=pμ(p,X)f(p,t)rμ(p,X)f*(p,t)rμ(p,X)f*(p,t)a(p)+pμ(p,X)f*(p,t)+rμ*(p,X)f(p,t)+rμ*(p,X)f(p,t)a(p)eipxdn1p(2π)n1.
(B2)

Comparing with (34) and (35), we obtain

p0(p,X)=1H2Xn(X0+Xn)p+HXbpb,
(B3a)
r0(p,X)=n24iXnexp2i1H(X0+Xn)H(X0+Xn)pH,
(B3b)
pa(p,X)=1H(X0+Xn)pa+XaHXn(X0+Xn)p+HXbpb,
(B3c)
ra(p,X)=n24iXn(X0+Xn)Xaexp2i1H(X0+Xn)H(X0+Xn)pH=XaX0+Xnr0(p,X)=Hxar0(p,X),
(B3d)

and in the flat-space limit H → 0, this has the correct limit

p0(p,X)p,r0(p,X)0,pa(p,X)pa,ra(p,X)0.
(B4)

We compute the commutator of two operators Pμ at possibly different points X and Y,

Pμ(X),Pν(Y)=rμν(q,X,Y)a(q)a(q)rμν*(q,X,Y)a(q)a(q)+pμν(q,X,Y)a(q)a(q)+pμν(q,X,Y)a(q)a(q)dn1q(2π)n1,
(C1)

with

pμν(q,X,Y)[rμ(q,X)+rμ(q,X)]rν*(q,Y)[rμ*(q,X)+rμ*(q,X)]rν(q,Y),
(C2a)
rμν(q,X,Y)[rμ(q,X)+rμ(q,X)]pν(q,Y)pμ(q,X)[rν(q,Y)+rν(q,Y)],
(C2b)

which does not obviously vanish for any choice of points X and Y. However, we compute

p00(p,X,Y)=4iImr0(p,X)r0*(p,Y),
(C3a)
p0b(p,X,Y)=Hybp00(p,X,Y),
(C3b)
pa0(p,X,Y)=Hxap00(p,X,Y),
(C3c)
pab(p,X,Y)=H2xaybp00(p,X,Y),
(C3d)

so pμν vanish for X = Y, i.e., when the operators are taken at the same point X. We then further obtain

r00(p,X,X)=0,
(C4a)
r0b(p,X,X)=2H(X0+Xn)r0(p,X)pb,
(C4b)
ra0(p,X,X)=2H(X0+Xn)r0(p,X)pa,
(C4c)
rab(p,X,X)=2(X0+Xn)r0(p,X)xapbxbpa.
(C4d)

Since r0(p, X) is an even function of p, the rμν(p, X, X) are odd functions of p, and because the annihilation operators commute among themselves (as well as the creation operators), the integrand in the commutator (C1) is an odd function such that the integral vanishes. That is, we have

Pμ(X),Pν(X)=0.
(C5)

Proof.
To first order in the non-commutative parameter Θ, we have (51),
ϕΘ(X)=ϕ(X)iΘμνPμ(X),ϕ(X)Pν(X)+OΘ2=ϕ(X)iΘμνPμ(X)Pν(X),ϕ(X)Pμ(X)Pν(X),ϕ(X)+OΘ2=ϕ(X)+iΘμνPμ(X)Pν(X),ϕ(X)+OΘ2=ϕΘ(X)+OΘ2,
(D1)
where we used that ΘμνPμ(X)Pν(X)=ΘμνPμ(X),Pν(X)/2=0 (C5), which confirms the general equality ϕΘ=ϕΘ to first order.
We thus can compute the two-point function up to first order in Θ,
0ϕΘ(X)ϕΘ(Y)0=0ϕ(X)ϕ(Y)0+Θμν0Xμϕ(X)Pν(X)ϕ(Y)0+Θμν0ϕ(X)Yμϕ(Y)Pν(Y)0+OΘ2=expn22H(s+t)12pexpiexp(Ht)exp(Hs)pHeip(xy)dn1p(2π)n1n22Θ0bexp(Ht)1Xn+exp(Hs)1Ynexpn22H(s+t)×pb2pexpiexp(Ht)exp(Hs)pHeip(xy)dn1p(2π)n1+n22HΘabexp(Ht)xaXn+exp(Hs)yaYnexpn22H(s+t)×pb2pexpiexp(Ht)exp(Hs)pHeip(xy)dn1p(2π)n1+OΘ2.
(D2)
The p integral is easily done in spherical coordinates using that
f(p)eipxdn1p(2π)n1=1(2π)n120f(p)pn2px3n2Jn32pxdp,
(D3)
where J is the Bessel function, and gives for the undeformed two-point function
0ϕ(X)ϕ(Y)0=12(2π)n12rn2expn22H(s+t)limϵ0+0expiexp(Ht)exp(Hs)rHpϵppn32Jn32pdp=Γn224πn2limϵ0+r2expH(s+t)2H2coshH(st)1+iϵsgn(ts)2n2=Γn22Hn22(2π)n2limϵ0+1Z(X,Y)+iϵsgn(ts)2n2
(D4)
with rxy. This is, of course, the well-known result for the two-point function,20 which here merely serves as a check on the computation. Performing the remaining spatial derivatives in Eq. (D2) coming from the pb in the spatial integrals, we obtain after some rearrangements result (52).□

1.
D. V.
Ahluwalia
, “
Quantum measurements, gravitation, and locality
,”
Phys. Lett. B
339
,
301
303
(
1994
); arXiv:gr-qc/9308007.
2.
S.
Doplicher
,
K.
Fredenhagen
, and
J. E.
Roberts
, “
The quantum structure of space-time at the Planck scale and quantum fields
,”
Commun. Math. Phys.
172
,
187
220
(
1995
); arXiv:hep-th/0303037 [hep-th].
3.
H.
Grosse
and
G.
Lechner
, “
Wedge-local quantum fields and noncommutative Minkowski space
,”
J. High Energy Phys.
2007
,
11
(
2007
).
4.
H.
Grosse
and
G.
Lechner
, “
Noncommutative deformations of Wightman quantum field theories
,”
J. High Energy Phys.
2008
,
9
(
2008
).
5.
D.
Buchholz
and
S. J.
Summers
, “
Warped convolutions: A novel tool in the construction of quantum field theories
,” in
Quantum Field Theory and Beyond
, edited by
E.
Seiler
and
K.
Sibold
(
World Scientific
,
Singapore
,
2008
), pp.
107
121
.
6.
D.
Buchholz
,
G.
Lechner
, and
S. J.
Summers
, “
Warped convolutions, Rieffel deformations and the construction of quantum field theories
,”
Commun. Math. Phys.
304
,
95
123
(
2011
); arXiv:1005.2656 [math-ph].
7.
M.
Rieffel
, “
Deformation quantization for actions of Rd
,”
Mem. Am. Math. Soc.
506
,
1
93
(
1993
).
8.
E.
Morfa-Morales
, “
Deformations of quantum field theories on de Sitter spacetime
,”
J. Math. Phys.
52
,
102304
(
2011
).
9.
C.
Dappiaggi
,
G.
Lechner
, and
E.
Morfa-Morales
, “
Deformations of quantum field theories on spacetimes with killing vector fields
,”
Commun. Math. Phys.
305
,
99
130
(
2011
).
10.
J.
Hawley
and
K.
Holcomb
,
Foundations of Modern Cosmology
(
OUP
,
Oxford
,
2005
).
11.
F. V.
Mukhanov
,
H. A.
Feldman
, and
R. H.
Brandenberger
, “
Theory of cosmological perturbations. Part 1. Classical perturbations. Part 2. Quantum theory of perturbations. Part 3. Extensions
,”
Phys. Rep.
215
,
203
333
(
1992
).
12.
A.
Much
, “
Wedge-local quantum fields on a nonconstant noncommutative spacetime
,”
J. Math. Phys.
53
,
082303
(
2012
); arXiv:1206.4450 [math-ph].
13.
A.
Much
, “
Quantum mechanical effects from deformation theory
,”
J. Math. Phys.
55
,
022302
(
2014
); arXiv:1307.2609 [math-ph].
14.
A.
Bourgoin
,
A.
Hees
,
S.
Bouquillon
,
C.
Le Poncin-Lafitte
,
G.
Francou
, and
M.-C.
Angonin
, “
Testing Lorentz symmetry with lunar laser ranging
,”
Phys. Rev. Lett.
117
,
241301
(
2016
).
15.
B.
Knorr
, “
Lorentz symmetry is relevant
,”
Phys. Lett. B
792
,
142
148
(
2019
); arXiv:1810.07971 [hep-th].
16.
G.
Lechner
, “
Deformations of quantum field theories and integrable models
,”
Commun. Math. Phys.
312
,
265
302
(
2012
); arXiv:1104.1948 [math-ph].
17.
G.
Lechner
,
J.
Schlemmer
, and
Y.
Tanimoto
, “
On the equivalence of two deformation schemes in quantum field theory
,”
Lett. Math. Phys.
103
,
421
437
(
2013
); arXiv:1209.2547 [math-ph].
18.
M. R.
Douglas
and
N. A.
Nekrasov
, “
Noncommutative field theory
,”
Rev. Mod. Phys.
73
,
977
1029
(
2001
); arXiv:hep-th/0106048.
19.
R.
Szabo
, “
Quantum field theory on noncommutative spaces
,”
Phys. Rep.
378
,
207
299
(
2003
).
20.
B.
Allen
, “
Vacuum states in de Sitter space
,”
Phys. Rev. D
32
,
3136
(
1985
).
21.
N. D.
Birrell
and
P. C. W.
Davies
,
Quantum Fields in Curved Space
, Cambridge Monographs on Mathematical Physics (
Cambridge University Press
,
1982
).
22.
A.
Much
and
R.
Oeckl
, “
Self-adjointness in Klein–Gordon theory on globally hyperbolic spacetimes
,”
Math Phys. Anal. Geom.
24
,
5
(
2021
).
23.
M.
Reed
and
B.
Simon
,
Methods of Modern Mathematical Physics
, Fourier Analysis, Selfadjointness Vol. 2 (
Gulf Professional Publishing
,
1975
).
24.
U.
Moschella
, “
Quantum fields on dS and AdS
,” in
International Conference on Theoretical Physics
, edited by
D.
Iagolnitzer
,
V.
Rivasseau
and
J.
Zinn-Justin
(
Birkhäuser Basel
,
Basel
,
2004
), pp.
319
332
.
25.
H. J.
Borchers
and
D.
Buchholz
, “
Global properties of vacuum states in de Sitter space
,”
Ann. I.H.P.: Phys. Theor.
70
,
23
40
(
1999
), available http://www.numdam.org/item/AIHPA_1999__70_1_23_0/.
26.
J.
Bros
and
U.
Moschella
, “
Two-point functions and quantum fields in de Sitter spacetime
,”
Rev. Math. Phys.
08
,
327
391
(
1996
).
27.
R.
Brandenberger
and
P.-M.
Ho
, “
Noncommutative space-time, stringy space-time uncertainty principle, and density fluctuations
,”
Phys. Rev. D
66
,
023517
(
2002
); arXiv:hep-th/0203119.
28.
Q.-G.
Huang
and
M.
Li
, “
Power spectra in spacetime noncommutative inflation
,”
Nucl. Phys. B
713
,
219
234
(
2005
); arXiv:astro-ph/0311378.
29.
G.
Calcagni
and
S.
Tsujikawa
, “
Observational constraints on patch inflation in noncommutative spacetime
,”
Phys. Rev. D
70
,
103514
(
2004
); arXiv:astro-ph/0407543.
30.
G.
Oliveira-Neto
and
L. F.
Marcon
, “
Complete noncommutativity in a cosmological model with radiation
,” arXiv:2101.09369 [gr-qc] (
2021
).
31.
O.
Mueller
and
M.
Sanchez
, “
Lorentzian manifolds isometrically embeddable in LN
,”
Trans. Am. Math. Soc.
363
,
5367
5379
(
2011
).
32.
C. J.
Fewster
and
R.
Verch
, “
Dynamical locality of the free scalar field
,”
Ann. Henri Poincare
13
,
1675
1709
(
2012
).
33.
C. J.
Fewster
and
R.
Verch
, “
Dynamical locality and covariance: What makes a physical theory the same in all spacetimes?
,”
Ann. Henri Poincare
13
,
1613
1674
(
2012
).
34.
J.
Dereziński
,
A.
Latosiński
, and
D.
Siemssen
, “
Pseudodifferential Weyl calculus on (pseudo-)Riemannian manifolds
,”
Ann. Henri Poincare
21
,
1595
(
2020
).
35.
L.
Hörmander
,
The Analysis of Linear Partial Differential Operators I: Distribution Theory and Fourier Analysis
, 2nd ed. (
Springer
,
1989
).
36.
G.
Lechner
and
S.
Waldmann
, “
Strict deformation quantization of locally convex algebras and modules
,”
J. Geom. Phys.
99
,
111
144
(
2016
); arXiv:1109.5950 [math.QA].
37.
M. S.
Joshi
, “
Introduction to pseudo-differential operators
,” arXiv:math/9906155v1 [math] (
1999
).
38.

This parameterization can be characterized by X0 + Xn > 0, which covers half of the hyperboloid. The other half, known as the contracting Poincaré patch, is obtained by inverting the sign of Xn.

39.

One way is to apply the creation operator on the vacuum vector to create one-particle states and construct the one-particle Hilbert space as completion of their linear span. Another way to obtain this space is to construct it explicitly by using the symplectic structure and a complex structure, as in Ref. 22.