Sunteți pe pagina 1din 11

History state formalism for scalar particles

N. L. Diaz,1 J. M. Matera,1 and R. Rossignoli1, 2


1
Departamento de Fı́sica-IFLP/CONICET, Universidad Nacional de La Plata, C.C. 67, La Plata (1900), Argentina
2
Comisión de Investigaciones Cientı́ficas (CIC), La Plata (1900), Argentina
We present a covariant quantum formalism for scalar particles based on an enlarged Hilbert space.
The particular physical theory can be introduced through a timeless Wheeler DeWitt-like equation,
whose projection onto four-dimensional coordinates leads to the Klein Gordon equation. The stan-
dard quantum mechanical product in the enlarged space, which is invariant and positive definite,
implies the usual Klein Gordon product when applied to its eigenstates. Moreover, the standard
three-dimensional invariant measure emerges naturally from the flat measure in four dimensions
when mass eigenstates are considered, allowing a rigorous identification between definite mass his-
tory states and the standard Wigner representation. Connections with the free propagator of scalar
field theory and localized states are subsequently derived. The formalism also allows the super-
arXiv:1910.04004v1 [quant-ph] 9 Oct 2019

position of different theories and remains valid in the presence of a fixed external field, revealing
special orthogonality relations. Other details such as extended identities for the current density, the
quantization of parameterized theories and the nonrelativistic limit, with its connection to the Page
and Wootters formalism, are discussed. A related consistent second quantization formulation is also
introduced.

I. INTRODUCTION theory through its mass eigenvalues, is also introduced,


while the Klein Gordon equation is obtained by project-
The introduction of the concept of time in a quan- ing the eigenvalue equation onto the |xi basis. It is then
tum mechanical framework [1, 2] has recently attracted shown that the 3d Klein Gordon product emerges from
renewed attention [3–11]. One persistent motivation is the 4d orthogonality of eigenstates with different eigen-
its connection with fundamental open problems, mainly values. This leads to Sec. II B where the relation with the
related to the quantization of gravity [12–19], whose clas- standard single particle representation of the Poincaré
sical description is a general covariant theory [20]. While group [25] is established, and the fixed mass states are
the Page and Wootters (PaW) formalism [21] has been identified with the histories of the corresponding states
able to provide a successful quantum treatment of time of the scalar Wigner representation. Since the history
[1, 3], it was mainly exploited to obtain non-relativistic states are more general, this correspondence only holds
equations, namely, the Schrödinger equation [3] and its in a particular mass subspace, excluding thus the states
discretized version [8]. However, the rigorous definition |xi. Yet, it is shown in Sec. II C that the space-time
of an hermitian time operator, enabled by this formalism localized states can be projected onto the “physical sub-
through an enlarged Hilbert space, has opened the pos- space” providing geometrical physical information. This
sibility to explore the construction of explicitly covariant result is employed to obtain the free propagation ampli-
Hilbert space representations, an idea which was recently tude of scalar field theory [26] within the present formal-
employed to obtain the Dirac equation [11]. ism. The proper action of covariant operators on physical
In this work we exploit these concepts further and de- subspaces is further clarified in Sec. II D by decompos-
velop the case of scalar particles, gaining new insight on ing the Hilbert space according to its different mass and
the subject. One of the main results is the definition of energy sectors. In particular, the unboundedness of P 0 ,
a consistent Hilbert space for the Klein Gordon equation the generator of the time translations, is discussed. The
[22, 23], where the inner product is the canonical prod- normalization in time is considered in the same section,
uct in four dimensions. Remarkably, this construction, where it is explicitly shown that a general normalizable
and the subsequent proper normalization of fixed mass state in the covariant Hilbert space is a superposition of
states, which are eigenstates of a Wheeler DeWitt-like the previous mass “improper” eigenstates.
equation [24], ensure the usual three-dimensional (3d) In Sec. III the universe equation is generalized to in-
norm. Moreover, the subspace of definite mass maps clude interactions with an external field. The Klein Gor-
onto the standard Wigner representation [25], directly don equation with a potential is again obtained by pro-
implying the standard 3d invariant measure. These re- jecting onto |xi the associated eigenvalue equation. It
sults provide a new perspective which could be suitable is then proved that the correct connection between the
to deal with the Hilbert space problem of the Wheeler canonical extended product and the Klein Gordon prod-
DeWitt framework of quantum gravity [15, 16, 19, 24]. uct holds for any mass and time independent external
The basic construction of the explicitly covariant field (for a given gauge choice and reference frame). It
Hilbert space is presented in Sec. II A, where event states is also remarked how the consideration of states with no
|xi are defined as eigenstates of the hermitian operators definite mass is already implicit when dealing with inter-
X µ , with X 0 introduced in accordance with the PaW for- actions.
malism. The universe operator, which specifies the free Some of the new insights which follow from the rela-
2

tivistic regime are transferred to the non relativistic case The equation
in Sec. IV. In particular, a proposal for the normaliza-
tion of states with infinite histories is derived in a self- J |Ψi = m2 |Ψi , (7)
contained non relativistic discussion. The case of a lin- has the general solution
early mass dependent potential is also briefly discussed. Z
The consistent construction of the single particle rep- |Ψm2 i = d4 p δ(pµ pµ − m2 )H + (p0 )α(p)|pi
resentation also allows a consistent definition of a Fock Z
space where the building block is the particle as a four-
dimensional entity. In Sec. V this “second quantization ⊕ d4 p δ(pµ pµ − m2 )H − (p0 )β(p)|pi (8)
of histories” is explored. The identification of Sec. II B
is extended to the standard Fock space of scalar field where H ± denotes the Heaviside function such that ±
theory through the definition of a proper subspace and corresponds to positive or negative p0 and m2 is a real
the generalization of the universe operator to a one-body eigenvalue of the hermitian operator J .
operator. Finally, conclusions and perspectives are dis- Defining Ψ(x) = hx|Ψi, Eq. (7) becomes the usual
cussed in Sec. VI. Klein Gordon equation [22, 23],
hx|(P µ Pµ − m2 )|Ψi = 0 ⇒ (∂ µ ∂µ + m2 )Ψ(x) = 0 (9)
II. SCALAR PARTICLE whose invariance is apparent since Ψ0 (x) = hx|U (Λ)|Ψi =
Ψ(Λ−1 x). Since
A. Quantum Formalism
δ(p0 − Epm )
δ(pµ pµ − m2 )H + (p0 ) = , (10)
2Epm
A general history state for a scalar particle can be writ-
ten as p
with Epm = p2 + m2 , an arbitrary solution with posi-
tive p0 can be written explicitly as
Z
|Ψi = d4 p Ψ(p)|pi (1) Z
1
|Ψm2 i = √ d4 x ψ(x)|xi ,
where |pi ∈ H are the improper eigenstates of the four 2π
operators Pµ . Here H = {S(R4 ), L2 (R4 ), S ∗ (R4 )} is the d3 p
Z
1 −ipx|p0 =Epm
ψ(x) = α(p)e , (11)
rigged Hilbert space constructed from L2 (R4 ), and S(R4 ) (2π) 3/2 2Epm
is the Schwartz space. Boost operators are defined by √
where ψ(x) = 2πΨ(x). Under a Lorentz transforma-
U (Λ)|pi = |Λpi, (2) d3 p
tion, α(p) → α(Λ−1 p) (Eq. (8)), implying 2E pm
invari-
wµν ant, in agreement with the well known result. The prod-
with Λµν = e and wµν = −wνµ an antisymmetric ten-
uct of two solutions corresponding to different eigenvalues
sor. The transformed state becomes
Z m2 and m̃2 yields
U (Λ)|Ψi = d4 p Ψ0 (p)|pi , d3 p
Z
(3)
hΨm̃2 |Ψm2 i = δ(Epm − Epm̃ )α̃∗ (p)α(p)
4Epm̃ Epm
with Z
d3 p ∗
= δ(m2 − m̃2 ) α̃ (p) α(p) (12)
Ψ0 (p) = hp|U (Λ)|Ψi = Ψ(Λ−1 p) . (4) 2Epm

1 since δ(Epm − Epm̃ ) = δ(m2 − m̃2 )2Epm . In the case


d4 p eipx |pi
R
We may also introduce the states |xi = (2π) 2
of two solutions with the same momenta distribution at
3
with px = pµ xµ = p0 x0 − i=1 pi xi , which transform as
P
equal mass, then
U (Λ)|xi = |Λxi. If |xi are eigenstates of operators X µ ,
d3 p
Z
the latter satisfy the commutation relations [X µ , Pν ] = hΨm̃2 |Ψm2 i = δ(m2 − m̃2 ) |α(p)|2 (13)
iδ µν . Clearly the operators Pµ , Lµν := Xµ Pν −Xν Pµ pro- 2Epm
vide a representation of the Lie algebra of the Poincaré with a similar expression in terms of β(p) for negative p0
group, where it is worth noting that P0 is not the Hamil- (solutions with positive and negative p0 are orthogonal).
tonian (see Sec. II D) and that the representation acts It is straightforward to see from Eq. (11) that
on H and not on a classical field. The representation is
d3 p
Z
manifestly unitary since
|α(p)|2 = Q(ψ, ψ)
Z 2Epm
hΨ̃|U (Λ)† U (Λ)|Ψi = d4 p Ψ̃0∗ (p)Ψ0 (p) = hΨ̃|Ψi . (5)
with
Z
Next we consider the operator Q(ϕ, ψ) := i d3 x (ϕ∗ (x, t)∂t ψ(x, t) − ψ(x, t)∂t ϕ∗ (x, t))
J = P µ Pµ . (6) (14)
3

and ψ(x, t) = ψ(x). Since √α(p) 3 , b(p) = √β(p) 3 , Epm → Ep and noting that
(2π) (2π)
2 2
hΨm̃2 |Ψm2 i = δ(m − m̃ ) Q(ψ, ψ) , (15) δ(pµ pµ − m2 )H ± (p0 ) = δ(p0 ∓ Epm )/2Ep ,
d3 p
Z
the proper normalization of these solutions in S ∗ (R4 ) |Ψm2 i = a(p)|Epm pi , (20)
(2π)3 2Ep
then implies, remarkably, the usual Klein Gordon nor-
d3 p
Z
malization [22] Q(ψ, ψ) = 1, i.e. ⊕ b(p)| − Epm pi , (21)
2 2
(2π)3 2Ep
hΨm̃2 |Ψm2 i = δ(m − m̃ ) ⇔
Z where we have introduced the states
i d3 x (ψ ∗ (x, t)∂t ψ(x, t) − ψ(x, t)∂t ψ ∗ (x, t)) = 1 . (16)
Z
3/2
| ± Epm pi := (2π) dp0 δ(p0 ∓ Epm )|p0 pi , (22)
The state of a particle at a given time t may
√ be identi- which satisfy (r, r0 = ±1)
fied with the “conditioned” state |ψ(t)i := 2πht|Ψm2 i,
with |ti = |x0 i for x0 = t, and thus ψ(x, t) with the hrEp0 m0 p0 |r0 Epm pi = (2π)3 2Ep δrr0 δ (3) (p−p0 )δ(m2 −m02 ) .
Klein Gordon wavefunction hx|ψ(t)i. In the case of mas- (23)
sive particles (positive m), the normalization hΨm̃ |Ψm i = The factor 2Ep now arises naturally from the mass or-
δ(m − m̃) can instead be chosen, in which case thogonality condition.
The one-to-one correspondence between the states
hΨm̃ |Ψm i = δ(m − m̃) ⇔
d3 p
Z
Z
i |Ψm2 i = a(p)|Epm pi ∈ H ,
d3 x (ψ ∗ (x, t)∂t ψ(x, t) − ψ(x, t)∂t ψ ∗ (x, t)) = 1, (17) (2π)3 2Ep
2m
d3 p
Z
i.e., d3 x ρ(x, t) = 1, with ρ(x, t) the usual Klein Gor- |ψi = a(p)|pi ∈ L2 (R3 , dµ(p)) , (24)
R
(2π)3 2Ep
don density [27, 28], which in the non relativistic-limit
reduces to the Schrödinger one for positive energy solu- is now explicit since in both cases
tions. Z
d3 p
More generally, it is now easy to prove the following |a(p)|2 = 1, (25)
(2π)3 2Ep
relations
and their transformation properties are identical. It shall
hΦ± ± 2 2
m̃2 |Ψm2 i = ±δ(m̃ − m )Q(ϕ, ψ) , (18) be noticed that while |ψi represents a particle at a fixed
hΦ± ∓ time (or equivalently, in the Heisenberg picture), |Ψm2 i
m̃2 |Ψm2 i = 0 , (19)
represents instead the whole history of the particle. Since
where the sign ± indicates the sign of p0R and hΦm̃2 |Ψm2 i one degree of freedom has been eliminated by the condi-
can be obviously also expressed as 2π 1
d4 x ϕ∗ (x)ψ(x). tion (7), providing the momentum distribution a(p) at a
It is important to notice that the previous relations pro- given time fixes the whole particle history. This allows
vide a positive normalization condition for both signs of the identification of |Ψm2 i with the history of |ψi.
p0 since hΨ± ± 2 2
m̃2 |Ψm2 i = δ(m̃ − m )|Q(ψ, ψ)|. The posi-
tivity follows from the canonical product in L2 (R4 ), yet
it implies the usual “norm”. The connection between C. Klein Gordon Propagator
both products can also be derived from extended rela-
tions satisfied by the current density. These relations are Given a general state in H, it can be projected onto the
obtained in the Appendix A using the present formalism. subspace of states satisfying (7) with a fixed eigenvalue
m2 by the operator
Πm2 : = δ(J − m2 ) . (26)
B. Relationship with Wigner representation
0
In general, this leaves both positive and negative p con-
The relation between the four and three-dimensional tributions. For the present discussionR it is useful to intro-
products provides a connection between a fixed mass duce additional projectors P ± := dp0 H ± (p0 )|p0 ihp0 | ⊗
solution of (7) and the usual (scalar) single parti- 1, satisfying [P ± , Πm ]=0, and define Π± ±
m2 := P Πm . In
2

cle representation in L2 (R3 , dµ(p)) where dµ(p) = particular it is interesting to project |xi onto the space
1 d3 p of “physical” particle states:
(2π)3 2Ep . The usual improper momentum eigenstates
√ √ Z
d4 p
|pi ∈ L2 (R3 , dµ(p)) are normalized as hp0 |pi = 2π Π+ |xi = 2π P +
δ(J − m 2
) eipx |pi
m2
p
(2π)3 2Ep δ (3) (p − p0 ). We notice that the standard in- (2π)4
variant normalization requires the addition of the factor Z
d4 p
2Ep in order to compensate the non invariance of the = p δ(pµ pµ − m2 )H + (p0 )eipx |pi
space volume [26, 29]. (2π)3
d3 p
Z
The connection with the present formalism becomes
= ei(Ep t−px) |Epm pi , (27)
apparent if we expand a solution (8) as, setting a(p) = (2π)3 2Ep
4

where the factor 2π in the first line was included
for normalization (see Eq. (32)). These states cor-
respond (in the sense discussed in Sec. II B) to
the single particle states φ(x)|0i, where φ(x) = FIG. 1. Pictorial representation of the two equivalent charac-
d3√
p terizations of the quantity hy|Π+ +
m0 Πm |xi. Each line represents
e−ipx ap + eipx a†p |p0 =Ep is the Klein Gor-
R 
3
(2π) 2Ep an amplitude D(y − x).
don field in the Heisenberg picture for the free theory
with mass m, and 2Ep a†p |0i = |pi. Moreover, from
p
(27) the following identity D. Normalization in Time

2πhy|Π+
m2 |xi = h0|φ(y)φ(x)|0i = D(y − x) , (28)
A state of the form
where Z
− − −
|Ψi = dm2 (γ + φ+ (m2 )|Ψ+ 2
m2 i + γ φ (m )|Ψm2 i) ,
d3 p
Z
D(y − x) = eip(x−y) |p0 =Ep , (29) (34)
(2π)3 2Ep
where |Ψ± m2 i are normalized states defined as in (18)
02
is the Klein Gordon propagator (or amplitude) [26] for (hΨ±
m 0 2 |Ψ ± 2
m2 i = δ(m − m )) with
the free theory with mass m, can be immediately shown. Z
This expression admits a straightforward interpretation: dm2 |φ± (m2 )|2 = 1 , (35)
by selecting the fixed mass particle contributions of an
event x (see also Sec. II D), we obtain a state whose prob- and
ability to be in another event y is essentially equal to
the amplitude for the particle to propagate from x to hΨ|Ψi = |γ + |2 + |γ − |2 = 1 , (36)
y. We notice that no unitary evolution was explicitly
introduced since the states contain all time information. belongs to L2 (R4 ). We will now prove that any state
Instead, a proper “selection” between possible histories |Ψi ∈ L2 (R4 ) admits the representation (34). This is in
was performed by employing the projector. principle apparent as the integral over all real values of
From Eq. (28) we see that we can rewrite the projection m2 covers the spectrum of the hermitian operator J and
of an event as |Ψ±m2 i are general states with definite mass and sign of
0
√ Z p . This also means that consideration of states which are
+ 1
2πΠm2 |xi = √ d4 z Dm2 (z − x)|zi , (30) normalizable in time (e. g. finite time history) is equiv-
2π alent to allow a mass/p0 sign uncertainty. The states
|Ψ+m2 i may be regarded as the idealization correspond-
where we added the index m2 to make the mass depen- ing to a particle with infinite history and infinitely well
dence explicit. We may also compute the overlap between defined dispersion relation, in which case the correspon-
two projected events as dence of Sec. II B follows.
Proof. An arbitrary normalized state |Ψi ∈ L2 (R4 )
Z
+ +
2πhy|Πm02 Πm2 |xi = 2π d4 z hy|Π+
m0 2
|zihz|Π+
m2 |xi can be expanded as
2
Z Z Z
= 2πδ(m2 − m0 ) hy|Π+
m2 |xi (31) |Ψi = d4 php|Ψi|pi = d4 p dm2 δ(pµ pµ − m2 )]hp|Ψi|pi
02
= δ(m2 − m )D(y − x) (32)
d3 p
Z Z
= dm2 hEp,m p|Ψi|Epm pi (37)
where in (31) we have employed Eq. (27). Thus, with (2π)3 2Epm
the normalization employed for the projected events their
d3 p
Z 
overlap is directly the propagator times the mass delta + h−E p,m p|Ψi| − E pm pi (38)
function. The identity (31) implies (2π)3 2Epm
R∞ R0
dm2 . . . = dm2 . . . + −∞ dm2 . . . in-
R
Z
2
where 0
d4 z Dm0 2 (y − z)Dm2 (z − x) = 2πδ(m2 − m0 )D(y − x) . cludes all real values of m2 . Using Eqs. (20)–
(33) (21), Eqs. (37)–(38) are seen to be of the form
The finite part is again essentially the propagator while (34) with a(p) = hEp,m p|Ψi/(γ + φ+ (m2 )), b(p) =
the presence of the delta function is in agreement with the h−Ep,m p|Ψi/(γ − φ− (m2 )) and
discussion of Sec. II. However, we see from Eq. (31) that sZ
we can reinterpret the appearance of the Dirac delta as ± ± 2 d3 p
γ φ (m ) = |h±Epm p|Ψi|2 .
the result of summing over all possibles space-time points (2π)3 2Epm
z of the propagation from x to y with the additional in-
termediate point z. This result is pictorially represented They involve four distinct terms, according to the signs of
in Fig. 1. m2 and Epm . For m2 < 0 the d3 p integration is restricted
to the region |p|2 > −m2 , as depicted in Fig. 2.
5

The four terms which arise from decomposing a general


state |Ψi ∈ L2 (R4 ) according to the signs of m2 and p0 in
Eqs. (37)–(38) belong to orthogonal subspaces which are
Hilbert space representations of the corresponding classes
of irreducible representations of the Poincaré group [25,
30]. This exhaustivity of H is precisely what allows to
represent events |xi and in particular the definition of a
time operator T such that TR := X 0 |xi = x0 |xi. The time
translation operator P 0 = d4 p p0 |pihp| is, as expected,
unbounded, however, this is not a problem in the present
formalism, in contrast with other approaches [13, 31]: By
writing (as in Eqs. (37)–(38))
d3 p
Z Z
P 0 = dm2 Epm |Epm pihEpm p| (39)
(2π)3 2Epm
d3 p
Z 
− E pm | − E pm pih−Epm p| (, 40) FIG. 2. Integration region in the variables m2 and p. Here
(2π)3 2Epm
|p| is the modulus of the three-momentum p.
it becomes clear that all four regions of H contribute to
its spectrum leading, as a consequence, to its unbound-
edness. Instead, on states which belong to a particular withR Aµ (X)|xi = Aµ (x)|xi. A state |Ψi =
irreducible representation, imposed ‘a posteriori’ by Eq. √1 d4 x ψ(x)|xi satisfies

(7) and by a given choice of the sign of p0 , P 0 acts prop-
erly: JA |Ψi = m2 |Ψi , (43)

P 0 | ± Epm pi = ±Epm | ± Epm pi . (41) iff ψ(x) satisfies the Klein Gordon equation
(−i∂µ + eAµ )(−i∂ µ + eAµ ) − m2 ψ(x) = 0 . (44)

The advantage of the present approach is apparent: co-
variant operators are defined independently of the par- Let us now consider the case where Aµ (X) does not
ticular theory, still, after a given theory, or superposition depend on T nor m2 . Considering solutions of definite
of them, is chosen, these operators, which are still de- energy E, ψ(x) = e−iEt ψ(x), Eq. (44) leads to the equa-
fined as before, act properly. This is precisely what we tion
have already found by projecting an event in Sec II B:
the state |xi, to which we associate a geometrical mean- [F (E) − m2 ]ψ(x) = 0 , (45)
ing, is “unphysical” for a theory with fixed mass m2 and
p0 > 0, however, the “closest” physical state corresponds where F (E) = (E − eA0 )2 − (P + eA)2 is an hermitian
to the well known state φ(x)|0i. Moreover, a perturba- operator with respect to the standard product in L2 (R3 ),
tive treatment of an interacting theory implicitly involves which does not depend on m2 . Then we may write a
states with an undefined mass when expanded in terms general solution of (44) as
of the free basis. This can already be discussed within X
ψm2 (x) = ck ψk (x, m2 ), (46)
a “first quantization” treatment of interactions as shown
k
in Sec. III.
2 −iEk (m2 )t
ψk (x, m ) = e ψk (x, m2 ) , (47)

III. KLEIN GORDON EQUATION IN AN where k labels the modes of definite energy Ek (m2 ) and
EXTERNAL FIELD ψk (x, m2 ) the corresponding eigenfunctions, obtained
from Eq. (45). They satisfy the Klein Gordon orthogo-
nality QA (ψk0 (m2 ), ψk (m2 )) = 0 for Ek (m2 ) 6= Ek0 (m2 ),
So far the discussion was centered in the case of a free
where (Dµ = ∂µ − ieAµ (x))
particle. In this section we discuss interactions at first Z
quantization level by treating fields as external entities.
This will follow from a straightforward extension of the QA (φ, ψ) = i d3 x (φ∗ (x, t)D0 ψ(x, t) − ψ(x, t)D0∗ φ∗ (x, t)) .
previous ideas which, remarkably, still provides the right (48)
connection between the invariant norm and the Klein
Gordon normalization, and more generally, between the The ensuing solution of (43) is
canonical product in L2 (R4 ) and the Klein Gordon prod- X
uct. |Ψm2 i = ck |Ψk (m2 )i , (49)
We replace J = P µ Pµ by k
Z
2 1
|Ψk (m )i = √ d4 x ψk (x, m2 )|xi . (50)
JA = (P µ + eAµ (X))(Pµ + eAµ (X)) , (42) 2π
6

2
We will prove in the first place that solutions with def- Note also that for m0 = m2 but Ek (m2 ) 6= Ek0 (m2 ) Eq.
inite energies Ek (m2 ) satisfy the orthogonality condition (56) leads to the standard Klein Gordon orthogonality
2 2
condition QA (ψk0 (m2 ), ψk (m2 )) = 0.
hΨk0 (m0 )|Ψk (m2 )i = δ(m2 − m0 )QA (ψk0 , ψk )sk (51) Previous results (53), (57) allows us to write, for modes
dEk
 of equal energies (Ek (m2 ) = Ek0 (m2 ) ∀m2 )
where sk = sgn dm 2 and the left hand side is the
canonical product in L2 (R4 ). This is a non trivial re- δ(m02 − m2 )
hΨk0 (m02 )|Ψk (m2 )i =
sult which follows from “special” orthogonality relations |dEk /dm2 |
of the usual solutions of Klein Gordon equation, as shown Z
below. While it warrants the expected orthogonality of × d3 x ψk∗0 (x, m2 )ψk (x, m2 ) . (58)
eigenstates with different masses, at equal mass it di-
rectly links the standard product in R4 with the Klein This second part of the proof involves finding an expres-
Gordon product in R3 , which in turn ensures orthogo- sion for dEk /dm2 . This is achieved by deriving Eq. (54)
nality of states with different energies at equal mass and with respect to m2 , which yields
implies QA (ψk , ψk )sk = |QA (ψk , ψk )|.
2
Secondly, we will show, choosing orthogonal modes dEk 2 dψk (x, m )
[F 0 (Ek ) 2
 
−1]ψ k (x, m ) = F (Ek ) − m .
ψk (m2 ) (QA (ψk0 (m2 ), ψk (m2 )) = 0 for k 6= k 0 , that rela- dm2 dm2
tion (51) implies
with F 0 (Ek ) = 2(Ek − eA0 ). We now multiply on the
left by a solution with the same energy ψk∗0 (x, m2 ) and
X ∗
hΨ0m02 |Ψm2 i = δ(m02 − m2 ) c0k ck |QA (ψk , ψk )| , (52)
k
integrate in space; we obtain
Z  
for general state with definite mass, which is identical dEk
d3 x ψk∗0 (x, m2 )ψk (x, m2 ) F 0 (Ek ) − 1 = 0,
with δ(m02 − m2 )|QA (ψ 0 , ψ)| when all QA (ψk , ψk ) have dm2
the same sign (i.e., all “positive” energy modes in stan- (59)
dard conditions). This is the sought extension of Eqs. and thus, for Ek (m2 ) = Ek0 (m2 ),
(18). Z
Proof. The overlap between two solutions (50) with dEk
d3 x ψk∗0 (x, m2 )ψk (x, m2 ) = QA (ψk0 (m2 ), ψk (m2 ))
definite energies yields dm2
(60)
hΨk0 (m02 )|Ψk (m2 )i = δ(Ek0 (m02 ) − Ek (m2 )) which is the natural extension extension of (56) for
Z
2
× d3 x ψk∗0 (x, m02 )ψk (x, m2 ) . (53) m2 = m0 and Ek (m2 ) = Ek0 (m2 ). Inserting this rela-
tion in (58) leads to the result (51). Eq. (59) also reveals
States with different energies are automatically orthogo- an additional orthogonality condition: orthogonal modes
nal while the equal energies condition can be separated at equal energies according to Klein Gordon product are
into two contributions: equal energies at equal masses, also orthogonal in the canonical product of L2 (R3 ), as-
dEk
or equal energies at different masses (and different k). suming dm 2 6= 0.

Consider first the second case: by writing Finally, we note from Eq. (51) that imposing the
2 2
normalization hΨk0 (m0 )|Ψk (m2 )i = δ(m2 − m0 )δkk0
F (Ek (m2 )) − m2 ψk (x, m2 ) = 0
 
(54) directly leads to the Klein Gordon normalization
02 02 02 |QA (ψk0 (m2 ), ψk (m2 ))| = δkk0 .
 
F (Ek0 (m )) − m ψk0 (x, m ) = 0 , (55)
The rigorous extension of the present results to a gen-
multiplying on the left by ψk∗0 (x, m02 ) (ψk∗ (x, m2 )) the eral potential Aµ (X) and curved space-times involves
first (second) equation, integrating in the whole space new concepts and will be presented elsewhere. Never-
and subtracting (conjugating one of the results) we ob- theless, general identities for the current density in the
tain presence of a general potential are discussed in the Ap-
pendix A. The case of a mass dependent Aµ is briefly dis-
Z
(m − m ) d3 x ψk∗0 (x, m02 )ψk (x, m2 ) =
02 2
cussed in the non relativistic limit for Newtonian gravity
2 2 in Sec. IV.
(Ek (m2 ) − Ek0 (m0 ))QA (ψk0 (m0 ), ψk (m2 )) (56) We finally remark that the mass eigenstates of JA in
Eq. (43) are obviously not eigenstates of the free particle
where we have used the hermiticity of (P + eA)2 . For
J of Eq. (7), since JA and J do not commute. There-
Ek0 (m02 ) = Ek (m2 ) then
fore, the expansion of eigenstates |ΨA m2 i of JA in terms of
those of J generally involves an expansion over different
Z
(m − m ) d3 x ψk∗0 (x, m02 )ψk (x, m2 ) = 0 ,
02 2
(57) masses (as well as positive and negative energies) as that
considered in Sec. II D. Thus, the consideration of states
implying an extended orthogonality condition for m02 6= with no definite mass in the free basis representation is
m2 when energies coincide. We conclude that no con- already implicit when dealing with an external field, i.e.,
tributions from different masses actually arise in (53). with interactions.
7

IV. NON RELATIVISTIC LIMIT eigenvalue of J corresponds to a global energy transla-


tion. On the other hand, since we are dealing with a free
It is well known that for positive energy solutions in particle it is wise to set the zero of the energy to its (pos-
the non relativistic limit E 0 /m  1 (order (v/c)2 , with itive) rest energy value mc2 (where we have momentarily
E 0 = E − m) the Klein Gordon equation reduces to the reintroduced the speed of light c). Then we have
Schrödinger equation [27]. In particular the Klein Gor-
−∇2
 
don norm for massive particles becomes the standard ∂
htx|(J +m)|Ψi = 0 ⇒ −i + + m ψ(x, t) = 0 ,
Schrödinger norm. It is then to be expected that a non ∂t 2m
relativistic version of Eq. (17) in terms of the usual quan- (66)
tum mechanical norm holds as a limit. Indeed this is the whose solutions are clearly of the form
case, but it is instructive to derive this result directly Z
from the non relativistic regime. |Ψm i = d3 p a(p)|p2 /2m + m, pi , (67)
We first recall that Schrödinger equation can be recov-
ered for states |ψ(t)i ∈ HS by imposing a global static implying
constraint on states |Ψi ∈ H = HT ⊗ HS . Here HT is
spanned by the eigenstates |ti of the operator T which |a(p)|2
Z
0
satisfies the canonical commutation [T, PT ] = i. In PaW hΨ m0 |Ψm i = δ(m − m ) d3 p p2
|1 − 2m2 |
interpretation [21] HT is regarded as the Hilbert space of  0 
Ep
Z 
a quantum clock such that the parameter t is a label of = δ(m − m0 ) d3 p |a(p)|2 × 1 + O .
states |ti of this particular system. m
The states |Ψi can be expanded as
Z Up to O(p2 /m2 ), this equation coincides with Eq. (13) af-
ter replacing a(p) = α(p)/(2Ep ). We find that solutions
|Ψi = dt |ti|ψ(t)i , (61)
with different “eigenvalues” m are orthogonal , even if J
is mass-dependent. Moreover, for states of non relativis-
while the state of the system at “time” t is |ψ(t)i = ht|Ψi. tic momenta, which is precisely the regime of validity of
By imposing the equation Schrödinger equation for a particle, the orthonormality
J |Ψi = 0 , (62) condition implies the Schrödinger norm (up to relativistic
corrections):
with
hΨm0 |Ψm i = δ(m − m0 ) ⇒ hψ(t0 )|ψ(t0 )i = 1 + O(v 2 /c2 ) .
J = PT ⊗ 1 + 1 ⊗ H , (63) (68)
This also means that the history states |Ψi can be nor-
where H is the Hamiltonian of the system, the standard malized according to the discussion of Sec. II D, a result
Schrödinger equation is obtained [3]: which provides (in the present case) a physical interpre-
d tation to the regularization proposal of [3].
ht|J |Ψi = 0 ⇒ i |ψ(t)i = H|ψ(t)i . (64) This result can be easily extended in this same context
dt
(but it also follows from the non relativistic limit of Sec.
The previous implication holds for arbitrary Hamilto- III) to Hamiltonians of the form
nians iff the spectrum of PT is the entire real line, which
also implies the same spectrum for T . Under this condi- (p − eA(x))2
tion the states |Ψi cannot be normalized in H [3]: roughly H= + V (x) + mφ(x) + m , (69)
2m
speaking, hΨ|Ψi is equal to hψ(t0 )|ψ(t0 )i times the (infi-
nite) length of time. On the other hand, if we focus on where A, V and φ are mass independent, employing a
the case of a scalar particle, from the discussion of Sec. similar strategy of Sec. III which was already employed
II D we can infer how to properly relate the norm of these for Dirac Hamiltonian in [11]. A minor modification fol-
global states with the norm of |ψ(t0 )i. It is also impor- lows from the mass dependent potential mφ(x): since
tant to notice that in the relativistic case the quantity now hx|(PT + H)|Ψm i = 0 yields
hΨ|Ψi is not related to the length of time as before since
the conditioned states |ψ(t)i are normalized according to 
∂ (p − eA(x))2

the Klein Gordon norm, not the Schrödinger one. We i − − V (x) ψ(x, t)
now focus then on the case HS = span{|xi} so that ∂t 2m
Z Z = m(1 + φ(x))ψ(x, t) , (70)
|Ψi = dtd3 x |ti|xihx|ψ(t)i ≡ dtd3 x ψ(x, t)|txi ,
the universe equation must be considered as a gener-
(65) alized eigenfunction equation (up to relativistic correc-
P2 tions coming from the mass dependent term on the left
and consider first the free case H = 2m . Notice that for
the quantum mechanical point of view, the zero eigen- hand side). In order to achieve orthogonality Rthe gener-
value in Eq. (62) plays no special role since a shifted alized product (Ψ|Ψ) := hΨ|(1 + φ(X))|Ψi = d4 x (1 +
8

φ(x))|ψ(x, t)|2 must in principle


R be employed. However, which satisfy
if c is reintroduced, (Ψ|Ψ) = d4 x (1 + φ(x)/c2 )|ψ(x, t)|2
†(m0 )
and we see that φ(x)/c2 must be dropped at this order [c(m)
p , cp’ ] = (2π)3 δ (3) (p − p0 )δ(m2 − m02 ) , (78)
[32]. This implies that potentials which depend on mass
linearly, as a Newtonian gravitational potential, do not and transform, according to (74), as
require a special treatment at the lowest order in c. It’s s
still remarkable that this simple analysis suggests a con- † (m) EΛpm (m)
nection between gravity and curvature since only Quan- U (Λ)cp U (Λ) = c . (79)
Epm Λp
tum Mechanical and Newtonian gravity considerations
were made together with the rest mass energy condition A single particle state of mass m is then written as
E = mc2 .
d3 p
Z
|Ψm2 i = 3
p a(p)c†(m)
p |0i (80)
(2π) 2Epm
V. EXTENDED FOCK SPACE
d3 p
Z
= a(p)|Epm pi ,
In this section we explore the construction of a Fock (2π)3 2Epm
space HEF where the building block is the single particle where
(sp) basis {|pi}, while the corresponding usual sp in HF
is {a†p |0iF = |pi}. The states |pi are reinterpreted as the 2Epm c†(m)
p
|Epm pi := p |0i . (81)
basis of a single particle space, i.e.
According to the discussion of Sec. II B, the state |Ψm2 i
|pi = c†p |0i , (71) can be identified with the history of
d3 p
Z
where the creation/annihilation operators satisfy, since |ψi = a(p)a†p |0iF (82)
hp0 |pi = δ 4 (p − p0 ), the algebra
p
(2π)3 2Epm
d3 p
Z
[cp , c†p0 ] = δ (4) (p − p0 ) , (72) = a(p)|pi ,
(2π)3 2Epm
[cp , cp0 ] = [c†p , c†p0 ] = 0. (73)
where ap are the standard 3d operators:
This algebra is explicitly preserved by boost operators
whose definition, [ap , a†p0 ] = (2π)3 δ (3) (p − p0 ) (83)

U (Λ)cp U † (Λ) = cΛp , (74) with

2Epm a†p |0iF .


p
follows from Eq. (2). Note that U (Λ) = exp[− 2i wµν Lµν ] |pi = (84)
is explicitly unitary and that
It is now straightforward to extend this identification
to many particles. From the usual transformation law of
Z  
∂ ∂
Lµν = i d4 p c†p pµ ν − pν µ cp , the operators ap , a†p , and Eq. (79) it follows that these
∂p ∂p
identifications are frame independent.
the generator of Lorentz transformations, is an hermitian It is now interesting to consider a two particle state
one-body operator.
d3 p1 d3 p 2
Z
Defining J as the one-body operator |Ψi = a(p1 , p2 )c†(m) †(m)
p p p1 cp2 |0i
Z (2π)3 2Ep1 (2π)3 2Ep2
J = d4 p (p2 − m2 )c†p cp , (75) †(m)
where cpi ∝ cp0 p with p0 = Epm are the operators
defined in (77). By a Fourier transform in p01 , p02 , we
which is the universe operator that corresponds to (6), obtain
the physical subspace is defined by those states built from
d p1 dt1 d3 p2 dt2 e−iEp1 t1 e−iEp2 t2
Z 3
creation operators commuting with J :
|Ψi = √ √
(2π)3 2π2Ep1 (2π)3 2π2Ep2
[J , c†p ] = (p2 − m2 )c†p = 0 ⇒ p2 = m2 . (76)
× a(p1 , p2 )c†t1 p1 c†t2 p2 |0i . (85)
As a basis of this subspace we can employ, for p0 > 0,
It is then seen that Rthis state has not a simple time
the operators
structure of the form dtd3 p1 d3 p2 ψ(t, p1 , p2 )c†tp1 c†tp2 |0i.
s This is relevant since such form cannot be preserved by
(2π)3
Z
(m)
cp := dp0 δ(p0 − Epm ) cp0 p , (77) Lorentz transformations. The more complex structure of
2Epm |Ψi is a novelty of the relativistic case which is required
to represent boosts.
9

VI. DISCUSSION space for the Wheeler DeWitt framework [15, 24], and
the rigorous treatment of quantum correlations in time
We have seen that it is possible to construct a consis- in quantum information and quantum optics [35–38].
tent history state formalism for a scalar particle whose
concept of time shares the underlying mathematical ideas
of the Page and Wootters mechanism [21]. The consider- ACKNOWLEDGMENTS
ation of a suited Hilbert space for the representation of
explicitly covariant operators, together with a timeless We acknowledge support from CONICET (NLD,
universe equation allows a simple derivation of the Klein JMM) and CIC (RR) of Argentina.
Gordon equation that complements the previous deriva-
tions of the Schrödinger [3] and Dirac equations [11]. The
canonical product of L2 (R4 ), which is invariant, provides Appendix A: Current density
a positive norm for this Hilbert space. Remarkably, the
subsequent proper normalization of “on-shell” states, di- In the present formalism, the Klein Gordon cur-
rectly ensures the usual 3-dimensional norm. In this rent Rdensity associated with an arbitrary state |Ψi =
way, we have found a straightforward connection with √1 d4 x ψ(x)|xi in the presence of a general potential

Klein Gordon’s wave equation, enhancing it with a posi- Aµ (X),
tive norm, and, for positive energy free particle solutions,
µ
with the standard irreducible single particle representa- jA (x) = i (ψ ∗ (x)Dµ ψ(x) − ψ(x)Dµ∗ ψ ∗ (x)) , (A1)
tion of the Poincaré group. Moreover, from relativistic
considerations we have inferred how to normalize states where Dµ = ∂ µ + ieAµ , can be written as
with infinite histories in a well defined non relativistic µ
limit, providing a physical interpretation to the previous j µ (x) = 2π hx|JA |xi , (A2)
regularization proposal for the Schrödinger equation [3].
In this sense, PaW mechanism reveals to be particularly where
adequate for the relativistic context. µ
At the same time, the new features of the resulting for- JA = −(PAµ |ΨihΨ| + |ΨihΨ|PAµ ) , (A3)
malism raise difficulties in the original relational interpre-
with PAµ = P µ + eAµ , is an hermitian operator. We can
tation: the time parameter ensuing from “conditioning
now express the 4-divergence of the current as
on the clock” is unequivocally identified with time in a
µ µ µ
given frame of reference by the Klein Gordon equation. A ∂µ jA (x) = 2πi hx|[Pµ , JA ]|xi = 2πi hx|[PµA , JA ]|xi
relational interpretation would lead us to the conclusion = 2πi hx|[|ΨihΨ|, JA ]|xi (A4)
that a single (quantum) clock is sufficient to describe the
evolution of a particle for any observer, in clear contrast where PµA = Pµ + eAµ and JA = PAµ PµA is the operator
with the necessity of a synchronization convention such (42). If |Ψi is an eigenvector of JA , i.e., a state with
as Einstein synchronization [33]. Moreover, this also re- definite mass |Ψm2 i, then [|ΨihΨ|, JA ] = 0 and we obtain
quires the spectrum of T to be continuous and unbounded the well known result
so it can hardly be associated with an observable of a
µ
clock other than a coordinate. These considerations sug- ∂µ jA (x) = 0 .
gest that in this context it is more adequate to simply
treat t as an additional coordinate of the particle itself, Previous relations can be immediately generalized to a
as Stueckelberg approach also suggests (see Appendix B). two-state current density
In the framework of general relativity, we would identify µ
jA (φ, ψ, x) = i (φ∗ (x)Dµ ψ(x) − ψ(x)Dµ∗ φ∗ (x))
the time parameter with “coordinate time” rather than µ
a time interval measured by a clock. = 2πhx|JA (Φ, Ψ)|xi , (A5)
On this basis, we have explored the construction of a
where
Fock space where the building block is the particle as
a 4-dimensional entity, extending thus the formalism to µ
JA (Φ, Ψ) = −(PAµ |ΨihΦ| + |ΨihΦ|PAµ ) . (A6)
a many particle scenario and defining a “second quanti-
zation” of histories. Through the definition of a proper Eq. (A4) now becomes
subspace, an identification with the standard many par-
µ
ticle states follows. At the same time, a richer time struc- ∂µ jA (φ, ψ, x) = 2πi hx|[|ΨihΦ|, JA ]|xi . (A7)
ture is revealed. This suggests a non-trivial extension to
quantum field theory, different from a direct application If |Ψi and |Φi are both solutions of the Klein Gordon
of the original PaW mechanism. The present formalism equation with the same mass, i.e., eigenstates of JA with
may thus provide a novel perspective for dealing with the same eigenvalue m2 , then [|ΨihΦ|, JA ] = 0, implying
different fundamental problems, like the concept of parti- µ
∂ µ jA (φ, ψ, x) = 0 . (A8)
cles in curved space-times [34], the definition of a Hilbert
10

On the other hand, for two eigenstates |Ψm2 i, |Φm0 2 i


2
with different masses m2 and m0 , we obtain instead Z τ2   Z τ2
q̇ 
S [q(τ ), t(τ )] = dτ ṫL q, ≡ dτ L̃ q, q̇, ṫ .
02 2 ṫ
[|Ψm2 ihΦm02 |, JA ] = (m − m )|Ψm2 ihΦm0 2 | , (A9) τ1 τ1
(B2)
implying the extended identity The momenta associated with L̃ are [13]:
µ 2
∂µ jA (φm02 , ψm2 , x) = i(m0 − m2 )ψm2 (x)φ∗m02 (x) , ∂ L̃
(A10) p̃q = = pq
∂ q̇
which holds for any mass-independent potential Aµ (X)
∂ L̃
(not necessarily time-independent). pt = = −H , (B3)
2
For m2 = m0 , integrating over d3 x and assuming that ∂ ṫ
µ 2
jA (φm2 , ψ m0 , x) vanishes for large |x|, Eq. (A10) leads while the Hamiltonian is H̃ = p̃q q̇ + pt ṫ − L̃ = ṫ(H + pt ).
to the well known result of Q(φ, ψ) constant in time, in If we define the “super Hamiltonian” Hs ≡ H + p̃t then
2
agreement with Eq. (15). For general m2 , m0 this rela- from Eq. (B3)
tion can be employed to re-obtain the previous relations
(15) and (56) (for a time and mass independent poten- Hs = H + pt ≈ 0 , (B4)
tial) by integration of (A10) over d3 x, assuming again
µ
the vanishing of jA for large |x|. where with ≈ we indicate this is a weak constraint [12].
The two-state current density can be also expressed as By applying canonical quantization to the extended con-
µ µ
figuration space, since t and pt are in phase space, an
jA (φ, ψ, x) = hΦ|JA (x)|Ψi (A11) enlarged Hilbert, which can be written as H = Ht ⊗ Hq ,
µ µ is obtained. The super Hamiltonian constraint (B4) de-
where JA (x) := JA (x, x) (Eq. (A6)) is the hermitian op- fines the subspace
erator
µ HS |Ψi = (Pt ⊗ I + I ⊗ H)|Ψi = 0 , (B5)
JA (x) = − (Π(x)PAµ + PAµ Π(x)) , (A12)

with Π(x) = |xihx|. Moreover, QA (φ, ψ) can be recast as which is precisely the universe equation of the PaW for-
Z malism for a one dimensional particle and continuum
µ time discussed in Sec. IV. We have obtained by this
QA (φ, ψ) = hΦ| d3 x JA (x)|Ψi = hΦ|QA (t)|Ψi , (A13) method the familiar notions of the non relativistic history
state formalism without considering a reference clock: t
QA (t) = − Π(t)PA0 + PA0 Π(t) ,

(A14)
is a coordinate.
where Π(t) = d x Π(x) = |tiht| ⊗ 1. All relations of It should be stressed that the conventional quantiza-
R 3
this appendix hold of course in the free case Aµ (X) = 0 tion procedure of a parameterized system doesn’t lead to
(PAµ → P µ , JA → J ). the present formalism where a time operator is defined
[13]. The key difference is that we are associating an en-
larged Hilbert space to the extended configuration space
Appendix B: Extended quantization of a such that the constraint (B5) has a precise mathematical
parameterized theory meaning instead of a formal one. The present proposal is
more close to Stueckelberg approach to relativistic quan-
Here we present an alternative version for deriving the tum mechanics [39]. In fact, the Hamiltonian R intro-
history state formalism of a particle which is closely re- duced by Stueckelberg under general relativistic consid-
lated to Stueckelberg proposal [39]. While the previous erations, which for a free particle reads R = 21 Pµ P µ ,
approach is self-contained, this different perspective fur- leads to the Stueckelberg equation [39]
ther clarifies that a relational interpretation is not strictly
needed for the parameter t. At the same time, recovering ∂
R Ψ(x, τ ) = i Ψ(x, τ ) , (B6)
the formalism in this way allows a first comparison with ∂τ
the conventional quantum gravity approach [13]. −im2
Consider the action of a one dimensional particle for a which for “τ stationary” solutions Ψ(x, τ ) = e 2 τ Ψ(x)
time independent Lagrangian yields
R 4 Eq. (7). The associated Stueckelberg norm
Z t2 d x |Ψ(x, τ )|2 , which is τ independent for a solution
of Eq. (B6), is precisely the one we have employed in
S [q(t)] = dt L(q, q̇) . (B1)
t1
Sec. II and related to Klein Gordon norm for fixed mass
solutions. The same considerations hold for the general
By promoting t to a coordinate and parameterizing the case R = 12 πµ π µ , where πµ = Pµ + eAµ , as follows im-
configuration space (t, q) with a variable τ we can write mediately form the results of Sec. III.
11

[1] R. Gambini, R. A. Porto, J. Pullin, and S. Torterolo, D 83, 125023 (2011).


Phys. Rev. D 79, 041501(R) (2009). [19] E. Anderson, Ann. Phys. (Berl.) 524, 757 (2012).
[2] J. G. Muga and C. R. Leavens, Phys. Rep. 338, 353 [20] A. Einstein, Ann. Phys. (Berl.) 49, 769 (1916).
(2000). [21] D. N. Page and W. K. Wootters, Phys. Rev. D 27, 2885
[3] V. Giovannetti, S. Lloyd, L. Maccone, Phys. Rev. D 92, (1983).
045033 (2015). [22] O. Klein, Z. Phys. 37, 895 (1926).
[4] A. Boette, R. Rossignoli, N. Gigena, and M. Cerezo, [23] W. Gordon, Z. Phys. 40, 117 (1926).
Phys. Rev. A 93, 062127 (2016). [24] B. S. DeWitt, Phys. Rev. 160, 1113 (1967).
[5] E. Moreva, M. Gramegna, G. Brida, L. Maccone, and [25] E. Wigner, Ann. of Math. 40, 149 (1939).
M. Genovese, Phys. Rev. D 96, 102005 (2017). [26] M. E. Peskin, An introduction to quantum field theory
[6] P. Erker, M. T. Mitchison, R. Silva, M. P. Woods, (CRC Press, 2018).
N. Brunner, and M. Huber, Phys. Rev. X 7, 031022 [27] W. Greiner et al., Relativistic quantum mechanics, Vol. 3
(2017). (Springer, 1990).
[7] E. O. Dias and F. Parisio, Phys. Rev. A 95, 032133 [28] F. Schwabl, Advanced Quantum Mechanics (Springer,
(2017). 2008).
[8] A. Boette and R. Rossignoli, Phys. Rev. A 98, 032108 [29] M. Maggiore, A modern introduction to quantum field
(2018). theory, Vol. 12 (Oxford university press, 2005).
[9] L. R. Mendes and D. O. Soares-Pinto, arXiv:1806.09669 [30] N. N. Bogolyubov, A. A. Logunov, and I. T. Todorov,
(2018). Introduction to Axiomatic Quantum Field Theory (1975).
[10] T. Martinelli and D. O. Soares-Pinto, Phys. Rev. A 99, [31] W. G. Unruh and R. M. Wald, Phys. Rev. D 40, 2598
042124 (2019). (1989).
[11] N. L. Diaz and R. Rossignoli, Phys. Rev. D 99, 045008 [32] H. Padmanabhan and T. Padmanabhan, Phys. Rev. D
(2019). 84, 085018 (2011).
[12] P. A. M. Dirac, Proc. Royal Soc. Lond. A 246, 333 [33] A. Einstein, Ann. Phys. (Berl.) 17, 50 (1905).
(1958). [34] L. Ford, in General Relativity and Gravitation (World
[13] C. Kiefer, Int. Ser. Monogr. Phys. 155, 1 (2004). Scientific, 2002) pp. 490–493.
[14] A. Peres, in On Einsteins path (Springer, 1999) pp. 367– [35] A. Peres and D. R. Terno, Rev. Mod. Phys 76, 93 (2004).
379. [36] I. Kull, P. A. Guérin, and Č. Brukner, npj Quantum Inf.
[15] K. V. Kuchař, Int. J. Mod. Phys. D 20, 3 (2011). 5, 48 (2019).
[16] M. Bojowald, P. A. Hoehn, and A. Tsobanjan, Class. [37] L. K. Shalm, D. R. Hamel, Z. Yan, C. Simon, K. J. Resch,
Quantum Gravity 28, 035006 (2011). and T. Jennewein, Nat. Phys. 9, 19 (2013).
[17] C. Rovelli, Found. Phys. 41, 1475 (2011). [38] M. Zych, F. Costa, I. Pikovski, and Č. Brukner, Nat.
[18] M. Bojowald, P. A. Höhn, and A. Tsobanjan, Phys. Rev. Commun. 10, 1 (2019).
[39] E. C. G. Stueckelberg, Helv. Phys. Acta 15, 23 (1942).

S-ar putea să vă placă și