1. Introduction
General relativity is a Lagrangian theory, i.e., the Einstein equations are derived as the Euler–Lagrange equation of the Einstein–Hilbert functional
where
,
, is a globally hyperbolic Lorentzian manifold,
is the scalar curvature, and
is a cosmological constant. We also omitted the integration density in the integral. In order to apply a Hamiltonian description of general relativity, one usually defines a time function
and considers the foliation of
N given by the slices
We may, without loss of generality, assume that the spacetime metric splits
cf. [
1] (Theorem 3.2). Then, the Einstein equations also split into a tangential part
and a normal part
where the naming refers to the given foliation. For the tangential Einstein equations, one can define equivalent Hamilton equations due to the groundbreaking paper by Arnowitt, Deser, and Misner [
2]. The normal Einstein equations can be expressed by the so-called Hamilton condition
where
is the Hamiltonian used in defining the Hamilton equations. In the canonical quantization of gravity, the Hamiltonian is transformed to a partial differential operator of a hyperbolic type
and the possible quantum solutions of gravity are supposed to satisfy the so-called Wheeler–DeWitt equation
in an appropriate setting, i.e., only the Hamilton condition, (6) was quantized, or equivalently, the normal Einstein equation, while the tangential Einstein equations were ignored.
In [
1], we solved the Equation (7) in a fiber bundle
E with the base space
,
and fibers
,
,
the elements of which are the positive definite symmetric tensors of order two, the Riemannian metrics in
. The hyperbolic operator
is then expressed in the form
where
is the Laplacian of the DeWitt metric given in the fibers,
R the scalar curvature of the metrics
, and
is defined by
where
is a fixed metric in
, such that, instead of densities, we consider functions. The Wheeler–DeWitt equation could be solved in
E but only as an abstract hyperbolic equation. The solutions could not be split into corresponding spatial and temporal eigenfunctions.
In a recent paper [
3], we overcame this difficulty by quantizing the Hamilton equations instead of the Hamilton condition.
As a result, we obtained the equation
in
E, where the Laplacian is the Laplacian in (10). The lower order terms of
were eliminated during the quantization process. This result was valid for all dimensions
, provided
.
The fibers add additional dimensions to the quantized problem, namely,
The fiber metric, the DeWitt metric, which is responsible for the Laplacian in (12), can be expressed in the form
where the coordinate system is
The
,
, are coordinates for the hypersurface
We also assumed that
and that the metric
in (11) is the Euclidean metric
. It is well-known that
M is a symmetric space
It is also easily verified that the induced metric of M in E is isometric to the Riemannian metric of the coset space .
Now, we were in a position to use the separation of variables, namely, we wrote a solution of (12) in the form
where
v is a spatial eigenfunction of the induced Laplacian of
M
and
w is a temporal eigenfunction satisfying the ODE
with
The eigenfunctions of the Laplacian in are well-known and we chose the kernel of the Fourier transform in in order to define the eigenfunctions. This choice also allowed us to use the Fourier quantization similar to the Euclidean case, such that the eigenfunctions were transformed to Dirac measures and the Laplacian to a multiplication operator in the Fourier space.
In the present paper, we quantize the Einstein–Hilbert functional combined with the functionals of the other fundamental forces of nature, i.e., we look at the Lagrangian functional
where
is a positive coupling constant,
and
N is a globally hyperbolic spacetime with metric
,
, where the metric splits as in (3).
The functional J consists of the Einstein–Hilbert functional, the Yang–Mills and Higgs functional, and a massive Dirac term.
The Yang–Mills field
corresponds to the adjoint representation of a compact, semi-simple Lie group
with Lie algebra
. The
,
are the structure constants of
.
We assume the Higgs field to have complex valued components.
The spinor field
has a spinor index
A,
, and a color index
I,
. Here, we suppose that the Lie group has a unitary representation
R, such that
are anti-Hermitian matrices acting on
. The symbol
is now defined by
There are some major difficulties in achieving a quantization of the functional in (23). Quantizing the Hamilton equations, to avoid the problem with the scalar curvature term, runs into technical difficulties, even if the required quantization of the matter fields in the curved spacetimes could be achieved since the resulting operator would no longer be hyperbolic because the elliptic parts of the gravitational resp. matter Hamiltonians would have different signs in the case of
. This particular problem would not occur when the Hamilton condition would be quantized. The Hamilton condition has the form
where the subscripts refer to gravity, Yang–Mills, Dirac, and Higgs. On the left-hand side are the Hamilton functions of the respective fields. They depend on the Riemannian metrics
, the Yang–Mills connections, and the spinor and Higgs fields. The main part of the quantized gravitational Hamiltonian is a second-order hyperbolic differential operator with respect to the variables
while the scalar curvature term
R is of zero-order. With this in mind, we also shall apply these categories to the gravitational Hamilton function where the main part, quadratic in the conjugate momenta, is said to be of the second-order and the zero-order terms consist of the scalar curvature and the cosmological constant
. Similarly, we consider the matter Hamilton functions to be zero-order terms with respect to the metric
, i.e., there is no qualitative difference by assuming
to be flat or non-flat, or more precisely, quantizing a matter Hamiltonian in a curved spacetime (when
is a given, fixed metric and not a variable) is qualitatively the same as quantizing it for the Euclidean metric, though the task is certainly more difficult.
Thus, the difficulties arising from quantizing the Hamilton condition can best be explained by considering the Wheeler–DeWitt equation
cf. (7), where we wrote
instead of
. This is a hyperbolic differential equation, which can be expressed by
where the Laplacian is the Laplacian of the fiber metric (15). In the coordinate system (16), we have
where
M is the hypersurface (17). Since
M is isometric to the symmetric space (18) it is mathematically irresistible to solve (31) by applying separation of variables and using the functions of the Fourier kernel of
M as spatial eigenfunctions
v, where
,
are the elements of
M. Since
the critical term
R can be expressed as
due to the relation between the scalar curvatures of conformal metrics.
Thus, it is obvious that the ansatz
where
solves an ODE is only possible if
is constant
The constant is arbitrary but determined by the metrics we consider to be important, e.g., in the case of a black hole, we would choose
to be the limit metric of a converging sequence of Cauchy hypersurfaces of the interior region of the black hole, which converge to the event horizon topologically but the induced metrics of which converge to a Riemannian metric, cf. [
4,
5] or [
6] (Chapters 4 and 5). In the present case, where we want to include the matter fields of the standard model, we could choose
.
However, this ansatz implies that the Wheeler–DeWitt equation is not solved for all but only for the satisfying (35). Given the simplicity and mathematical beauty of the solution, we are inclined to accept this restriction.
Let us now consider the quantization of the Hamilton condition (28) taking all Hamilton functions into account. In view of the relation (32), let us propose the following model: If we were able to express the non-gravitational Hamiltonians as
where the embellished Hamiltonians depend on
, then, by choosing in addition
and
, these Hamiltonians could be quantized by the known methods of QFT, if the Lie groups would be chosen appropriately. The Wheeler–DeWitt equation would then not be solved for all
but only for
. However, the spatial eigendistributions of the Hamilton operator
, i.e., the eigendistributions of the Laplacian of
M, cf. (20), would still be used but they would be evaluated at
.
In
Section 4, we prove that the expressions in (36) are indeed valid with
provided
and provided that the mass term in the Dirac Lagrangian and the Higgs Lagrangian is slightly modified. The embellished Hamiltonians are then standard Hamiltonians without any modifications, for details, we refer to
Section 4. The Hamilton constraint then has the form
where the subscript
refers to the fields of the standard model or a corresponding subset of fields. The solutions of the Wheeler–DeWitt equation
can then be achieved by using the separation of variables. We proved:
Theorem 1. Let , and let ψ be an eigendistribution of when such thatand let w be a solution of the ODEthenis a solution of the Wheeler–DeWitt equationwhere is evaluated at and where we note that . We shall refer to and as the spatial eigenfunctions and w as the temporal eigenfunction.
Remark 1. We could also apply the respective Fourier transforms to resp. and consideras the solution in Fourier space, where would be expressed with the help of the ladder operators. The temporal eigenfunctions are analyzed in
Section 5. They must satisfy an ODE of the form
where
For simplicity, we shall only state the result when , which is tantamount to setting .
Theorem 2. Assume and , then the solutions of the ODE (45) are generated byandwhere is the Bessel function of the first kind. Lemma 1. The solutions in the theorem above diverge to complex infinity if t tends to zero and they converge to zero if t tends to infinity.
4. Quantization of the Lagrangian
We consider the functional
where
is a positive coupling constant and
.
We use the action principle that, for an arbitrary as above, a solution should be a stationary point of the functional with respect to compact variations. This principle requires no additional surface terms for the functional.
As we proved in [
1], we may only consider metrics
that split with respect to some fixed globally defined time function
, such that
where
are Riemannian metrics in
,
The first functional on the right-hand side of (101) can be written in the form
where
is the DeWitt metric,
R the scalar curvature of the slices
with respect to the metric
, and where we also assumed that
is a cylinder
such that
for some
, where the
are special coordinate patches of
N, such that there exists a local trivialization in
with the properties that there is a fixed Yang–Mills connection
satisfying
cf. [
8] (Lemma 2.5). We may then assume that the Yang–Mills connections
are of the form
where
is a tensor, see [
8] (Section 2).
The Riemannian metrics
are elements of the bundle
. Denote by
E the fiber bundle with base
where the fibers
consist of the Riemannian metrics
. We shall consider each fiber to be a Lorentzian manifold equipped with the DeWitt metric. Each fiber
F has the dimension
Let
,
, be coordinates for a local trivialization, such that
is a local embedding. The DeWitt metric is then expressed as
where a comma indicates partial differentiation. In the new coordinate system, the curves
can be written in the form
and we infer
Hence, we can express (104) as
where we now refrain from writing down the density
explicitly, since it does not depend on
and, therefore, should not be part of the Legendre transformation. Here, we follow Mackey’s advice in [
9] (p. 94) to always consider rectangular coordinates when applying canonical quantization, which can be rephrased that the Hamiltonian has to be a coordinate invariant, hence no densities are allowed.
Denoting the Lagrangian function in (118) by
L, we define
and we obtain for the Hamiltonian function
where
is the inverse metric. Hence,
is the Hamiltonian that will enter the Hamilton constraint, for details see [
6] (Chapter 1.4).
Let us recall that the fibers
F can be considered Lorentzian manifolds, even globally hyperbolic manifolds, equipped with the DeWitt metric
, where
is a time function, cf. [
6] (Theorem 1.4.2). In the fibers, we can introduce new coordinates,
,
, and
, such that
and
are coordinates for the hypersurface
The Lorentzian metric in the fibers can then be expressed in the form
where
is a Riemannian metric on
M, which is independent of
t. When we work in a local trivialization of the bundle
E, the coordinates
are independent of
x. The time coordinate
t is also independent of
x, cf. [
1] (Lemma 1.8). Moreover, the fiber elements
can be expressed in the form
where
is an element of
M, i.e.,
or equivalently,
Next, let us look at the Yang–Mills Lagrangian, which can be expressed as
Let
be the adjoint bundle
with base space
, where the gauge transformations only depend on the spatial variables
. Then the mappings
can be looked at as curves in
, where the fibers of
are the tensor products
which are vector spaces equipped with the metric
For our purposes, it is more convenient to consider the fibers to be Riemannian manifolds endowed with the above metric. Let
,
, where
, be local coordinates and
be a local embedding, then the metric has the coefficients
Hence, the Lagrangian
in (128) can be expressed in the form
and we deduce
yielding the Hamilton function
Thus, after introducing a normal Gaussian coordinate system, such that
, the Hamiltonian that will enter the Hamilton constraint equation is
Combining, now, (122), (125) and (133) we infer that the Yang–Mills Hamiltonian can be expressed as
where the indices in the last term are raised with respect to the metric
, i.e.,
In the case of
, the exponents of
t in (138) are equal
and we can write
Moreover, if as well as are equal to the Euclidean metric , then the quantization of would be achieved by known methods of QFT.
Hence, we shall attempt to express the Hamiltonians of the other physical forces, such as the Dirac and Higgs Hamiltonians, when evaluated for
and in the case of
in the form
resp.
such that the quantization of the spatial Hamiltonian
would be well known, and in the end, all spatial Hamiltonians of the standard model could be incorporated.
Let us first consider the Dirac Hamiltonian. In the Dirac Lagrangian , defined in Equation (100) on page 11, the volume density is missing, i.e., in order to define the Hamiltonian, we have to multiply the Lagrangian with , or, since we work with functions instead of densities, we have to multiply the Lagrangian with .
In addition, we shall also consider—at least locally—a normal Gaussian coordinate system, such that
. Then, the final Dirac Lagrangian has the form
The spinorial variables are anti-commuting Grassmann variables. They are elements of a Grassmann algebra with involution, where the involution corresponds to the complex conjugation and will be denoted by a bar.
The
are complex variables and we define the real resp. imaginary parts as
resp.
With these definitions, we obtain
Casalbuoni quantized the Bose–Fermi system in [
10] (Section 4), the results of which can be applied to spin
fermions. The Lagrangian in [
10] is the same as the main part our Lagrangian in (146) on page 15, and the left derivative is used in that paper; hence, we use left derivatives as well such that the conjugate momenta of the odd variables are, e.g.,
and, thus, the conclusions in [
10] can be applied.
The Lagrangian has been expressed in real variables—at least the important part of it—and it follows that the odd variables
satisfy, after introducing anti-commutative Dirac brackets as in [
10] (equ. (4.11)),
and
cf. [
10] (equ. (4.19)).
In view of (149), (150) we then derive
where
are the conjugate momenta.
Canonical quantization—with
—then requires that the corresponding operators
satisfy the anti-commutative rules
and
cf. [
11] (equ. (3.10)) and [
10] (equ. (5.17)).
From (146), we then deduce that the spinorial Hamilton function is equal to
When we attempt to quantize this Hamilton function, then the vielbein
and its inverse
will correspond to a given element
in the fiber
F, which can be expressed as in (125), and we deduce that the vielbein
and its inverse
correspond to the metric
. Furthermore, the covariant derivative
is independent of
t, in view of (97) and (98) on page 10. Thus, the Hamilton function
can be expressed as
i.e., the main part already has the form that we looked for in (143), provided
, only the mass term spoils the necessary configuration. To overcome this setback, we either have to omit the mass term or modify it by multiplying the mass term in (23) on page 3 with the factor
where
is defined in (87) on page 10. Note that
if
as is the case in QFT. Either by omitting or by modifying the mass term, the Dirac Hamilton function can be expressed in the required form
where the underlying Riemannian metric is
, provided
.
The remaining Hamiltonian is the Hamiltonian of the Higgs field. The Higgs Lagrangian is defined by
where
V is a smooth potential. We assume that in a local coordinate system
has real coefficients. The covariant derivatives of
are defined by a connection
in
E0As in the preceding section, we work in a local trivialization of
E0 using the temporal gauge, i.e.,
hence, we conclude
Expressing the density
g as in (87) on page 10, we obtain Lagrangian
where, again, we use local coordinates, such that
. In order to apply our approach, outlined in (144), we have to modify the Lagrangian. Instead of the above Lagrangian, we have to consider
Let us define
then we obtain the Hamilton function
After quantization, the
are elements of the fiber
F, i.e.,
If
, then
has to be chosen, such that
which is the case if
For
, we obtain
yielding
Thus, the Hamilton function of the modified Higgs field has the required form
where
is a standard Hamiltonian of a Higgs field in QFT by choosing
and
,
as well as the Yang–Mills connection appropriately.
Combining the four Hamilton functions in (120), (138), (179) and (162), the Hamilton constraint has the form
where we omit the subscript
and where
refers to the fields of the standard model or to a corresponding subset of fields.
The Hamiltonian
we quantize, as in our former papers [
1,
12], to obtain
where the Laplacian is the Laplacian of the metric (124) acting in the fibers
F of
E. The Laplacian acts on smooth functions
u of the form
. Choosing the Gaussian coordinate system
, such that the fiber metric has form as in (124), then, the hyperbolic term
can be expressed as
where
is the Laplacian of the hypersurface
Using the separation of variables we consider functions
u which are products
where
v is a spatial eigenfunction, or eigendistribution, of the Laplacian
The hypersurface
can be considered a subbundle of
E, where each fiber
is a hypersurface in the fiber
of
E. We shall use the same notation
M for the subbundle as well as for the hypersurface, and in general, we shall omit the reference to the base point
. Furthermore, we specify the metric
, which we used to define
, to be equal to the Euclidean metric, such that in Euclidean coordinates
Then, it is well-known that each
with the induced metric
is a symmetric space, namely, it is isometric to the coset space
cf. [
13] (equ. (5.17), p. 1123) and [
14] (p. 3). The eigenfunctions in symmetric spaces, and especially of the coset space in (190), are well-known; they are the so-called
spherical functions. One can also define a Fourier transformation for functions in
and prove a Plancherel formula, similar to the Euclidean case, cf. [
15] (Chapter III). Moreover, similar to the Euclidean case, we shall use the Fourier kernel to define the eigenfunctions or eigendistributions, cf. [
3] (Section 5).
Let
be an Iwasawa decomposition of
G, where
N is the subgroup of unit upper triangle matrices,
A is the abelian subgroup of the diagonal matrices with strictly positive diagonal components and
. The corresponding Lie algebras are denoted by
The Iwasawa decomposition is unique, when
we define the maps
by
We also use the expression
, where log is the matrix logarithm. In the case of diagonal matrices,
with positive entries
hence,
Remark 3. (i) The Lie algebra is a (n − 1)-dimensional real algebra, which, as a vector space, is equipped with a natural real, symmetric scalar product, namely, the trace form (ii) Let be the dual space of . Its elements will be denoted by Greek symbols, some of which have special meanings in the literature. The linear forms are also called additive characters.
(iii) Let , then there exists a unique matrix , such that This definition allows defining a dual trace form in by setting for The Fourier theory in
, which we summarized in [
3] (Section 6), uses the functions
as the Fourier kernel, where
Here,
M is the centralizer of
A in
K and
is a special character with the norm
cf. [
3] (Lemma 1). If
, then
For a precise definition of
, we refer to [
3] (p. 19), which also contains references to the corresponding mathematical literature given, especially to Helgason’s book [
15] (Chapter III).
The Fourier transform for functions
is then defined by
for
and
, or, if we use the definition in (202)
by
The functions
are real analytics in
x and are eigenfunctions of the Laplacian, cf. [
15] (Prop. 3.14, p. 99),
where
cf. (201), and similarly for
. We also denote the Fourier transform by
, such that
In Equation (209), we identified
In [
3], we finally dropped the embellishment and simply wrote
when referring to the above Laplacian, but at the moment we refrain from doing so to avoid confusion.
We shall consider the eigenfunctions
as tempered distributions of the Schwartz space
and shall use their Fourier transforms
as the spatial eigenfunctions of
which is a multiplication operator, such that
cf. [
3] (Section 6) for details.
Looking at the Fourier transformed eigenfunctions
it is obvious that the dependence on
b has to be eliminated, since there is neither a physical nor a mathematical motivation to distinguish between
and
. We discard the integration over
B in [
3] (Section 6) and pick instead a special element
, namely,
and only consider the eigenfunctions
with corresponding Fourier transforms
For justification, see [
3] (Lemma 4) and the arguments preceding the referenced Lemma.
The eigenfunctions
depend on the characters
but not all characters are physically relevant. For a definition of the physically relevant characters, let us rephrase [
3] (Remark 2, p. 18):
Remark 4. There are characters , , that will represent the elementary gravitons stemming from the degrees of freedom in choosing the coordinatesof a metric tensor. The diagonal elements offer, in general, additional n degrees of freedom, but in our case, where we consider metrics satisfyingonly diagonal components can be freely chosen, and we shall choose the first entries, namely, The corresponding additive characters are named .
The characters , , and will represent the elementary gravitons at the character level. We shall normalize the characters by definingandsuch that the normalized characters have unit norm, cf. (201). We can now define the corresponding forms in with arbitrary energy levels:
Definition 1. Let be arbitrary. Then we consider the characterswhere we recall that the terms embellished by a tilde refer to the corresponding unit vectors. Then the eigenfunctions representing the elementary gravitons are and . The corresponding eigenvalue with respect to is , where, by a slight abuse of notation, and . Note that if , cf. (205).
We define a zero-point energy eigenfunction by choosing . The corresponding eigenfunction would be , satisfyingif . We are now able to quantize the Hamiltonian
H in (181). For brevity we denote the quantized Hamiltonians, which are operators, by using the same symbols as for the Hamilton functions. For the Hamilton operator
, we express as in (183)
where we use the separation of variables in (186), the form of the metric in (125), namely,
and the relation between the scalar curvatures of conformal metrics
Let us recall that for the quantization of
we shall specify
, such that the spatial eigendistributions, or approximate eigendistributions,
, satisfying
can be derived by applying standard methods of QFT. We then solve the Wheeler–DeWitt equation
not for all
but only for
, where
is arbitrary. Thus, we shall solve
by using
for arbitrary
, but we shall evaluate
only at
. Furthermore, we observe that for
and
, we have
cf. [
3] (equ. (202), p. 18), hence, if
, i.e., if
, then
Moreover, let
be the isometry, then
where
is the adjoint. Thus, if
, we infer
and we have proved:
Theorem 3. Let , , and let ψ be an eigendistribution of when such thatand let w be a solution of the ODEthenis a solution of the Wheeler–DeWitt equationwhere is evaluated at and where we note that . We shall refer to and as the spatial eigenfunctions and to w as the temporal eigenfunction.
Remark 5. We could also apply the respective Fourier transforms to resp. and consideras the solution in the Fourier space, where would be expressed with the help of the ladder operators. In the next section, we shall analyze the temporal eigenfunctions.