December 2011

EPL, 96 (2011) 50001 www.epljournal.org

doi: 10.1209/0295-5075/96/50001
Covariant renormalizable anisotropic theories and o-diagonal
Einstein-Yang-Mills-Higgs solutions
S. I. Vacaru
(a)
Science Department, University Al. I. Cuza Iasi - 54 Lascar Catargi street, 700107, Iasi, Romania, EU
received 20 September 2011; accepted in nal form 10 October 2011
published online 4 November 2011
PACS 04.60.-m Quantum gravity
PACS 04.50.Kd Modied theories of gravity
PACS 04.20.Jb Exact solutions
Abstract We use an important decoupling property of gravitational eld equations in the
general relativity theory and modications, written with respect to nonholonomic frames with
2 +2 spacetime decomposition. This allows us to integrate the Einstein equations (eqs) in very
general forms with generic o-diagonal metrics depending on all spacetime coordinates via generat-
ing and integration functions containing (un-)broken symmetry parameters. We associate families
of o-diagonal Einstein manifolds to certain classes of covariant gravity theories which have a nice
ultraviolet behavior and seem to be (super-) renormalizable in the sense of covariant modications
of Horava-Lifshitz gravity. The apparent breaking of Lorentz invariance is present in some part-
ner anisotropically induced theories due to nonlinear coupling with eective parametric interac-
tions determined by nonholonomic constraints and generic o-diagonal gravitational and matter
elds congurations. Finally, we show how the constructions can be extended to include exact
solutions for conjectured covariant reonormalizable models with Einstein-Yang-Mills-Higgs elds.
Copyright c EPLA, 2011
Introduction. One of the main diculties in elab-
orating viable models of quantum gravity (QG) is that
perturbations in the general relativity theory (GR) from
a at Minkowski spacetime result in nonrenormalizable
divergences from the ultraviolet region in momentum
space. This problem is caused by the dimension of grav-
itational (Newton) constant and seems impossible to be
solved for a unitary theory with four fundamental inter-
actions which has Lorentz invariance even if some higher-
derivative, spin-foam, loop-type etc. models of gravity may
be renormalizable, see [13] for reviews and references.
A recent new approach to QG is based on the idea [4]
to construct Lorentz noninvariant theories with scaling
properties of space, x
i
, and time, t, coordinates consid-
ered in the form (x, t) (bx, b
z
t), where z =2, 3, . . . . In
such cases, the ultraviolet (UV) behavior of the gravi-
ton propagator changes as 1/|k|
2
1/|k|
2z
, where k is
the spacial momenta. It is possible to elaborate models of
QG which are UV renormalizable, for instance, for z =3,
but with the price of introducing some terms breaking
the Lorentz invariance explicitly. Due to the lack of full
dieomorphysm invariance and, for instance, the impos-
sibility to exclude completely certain unphysical modes,
(a)
E-mail: sergiu.vacaru@uaic.ro
such theories where criticized in a series of works (see, for
example, [5,6]).
In [7], a covariant renormalizable gravity model devel-
oping the Horava-like gravity to full dieomorphysm
invariance was elaborated. In such an approach, the
Lorentz invariance of the graviton propagator is broken
via a nonstandard coupling with an unknown uid. Some
versions of such theories seem to be (super-) renormaliz-
able, when physical transverse modes may appear. Certain
applications in modern cosmology, with accelerating solu-
tions and possible power law inationary stage, seem to
be of substantial interest.
The results of research in [8,9] conclude that the
gravitational eld eqs in GR and various modications
can be decoupled with respect to certain classes of
nonholonomic frames. It is possible to integrate the
system of Einstein-Yang-Mills and Higgs (EYMH) eqs
in very general forms. For such solutions, the generic
o-diagonal metrics cannot be diagonalized via coordinate
transforms and depend on all spacetime coordinates via
corresponding generating and integration functions and
possible (broken, or preserving) symmetry parameters.
Choosing the necessary types of such functions and
parameters via corresponding deformations of frame,
metric and connection structures, and imposing noninte-
grable (nonholonomic) transforms, we can model eective
50001-p1
S. I. Vacaru
nonlinear interactions, with scaling properties and local
anisotropies, which can be renormalizable.
We start our constructions with a set of actions
[i]
S =
1

d
4
u

|g|
[i]
L on a four-dimensional (4-d) pseudo-
Riemannian manifold V for the Lagrangian
[1]
L =
s

R

=
[2]
L=
s

R+

L=
[3]
L
= R+L(T

, R

) =
[4]
L=F(

R). (1)
In our works, we use left up and low labels to geomet-
ric/physical objects which may have the same coe-
cients in a system of reference but that are subjected to
dierent types of eqs for certain analogous/eective theo-
ries. All terms in such formulas are stated by the same
metric structure g ={g

} for standard and/or modied

models of gravity theory generated by dierent (eective)
Lagrangians with curvature scalars (
s

R,
s

R, R,

R), cosmo-
logical constant

, nonstandard coupling of Ricci, R

,
and energy-momentum, T

, tensors via L(. . .) and F(. . .)

etc. The above values and other possible constants and
terms are postulated in such a way that
[1]
L is used for
constructing general classes of generic o-diagonal solu-
tions in GR [8,9] but
[3]
L is of necessary type to get
a covariant renormalization gravity as in [7]. For corre-
sponding conditions on F the value
[4]
L is related to
modied theories [10]. To study QG models, the space-
time metric for theories derived from
[3]
L and/or
[4]
L
will be taken, for simplicity, of type

g

+h

for
perturbative models with at background

. Such solu-
tions are with breaking of the Lorentz symmetry and, for
well-stated conditions, result into eective models with
covariant renormalization. Using frame transforms g

=
e

(see footnote
1
), we shall search for classi-
cal and quantum perturbative solutions for the theories
determined by
[1]
L, or
[2]
L.
In this paper, we show how to construct o-diagonal
solutions in GR (in general, with nontrivial eective
anisotropic polarizations of cosmological constants), when
in a theory for
[1]
S there are modelled eects with broken
Lorentz invariance, nonstandard eective anistoropic uid
coupling and behavior of the polarized propagator in
the ultraviolet/infrared region. Such eects are derived
to be similar to those for some (super-) renormalizable
theories
[3]
S and/or
[4]
S. The solutions for
[2]
S will be
used as bridges between generating functions determin-
ing certain classes of generic o-diagonal Einstein mani-
folds and eective models with anisotropies and resulting
violations for Lorentz symmetry. We state the conditions
when the Einstein eqs transform nonlinearly, by impos-
ing corresponding classes of nonholonomic constraints, to
the necessary types of eective systems of partial dier-
ential eqs (PDE) with parametric dependence of solutions
which under quantization survive and stabilize in some
1
We label local coordinates and u

=(x
i
, y
a
) u

=
(x
i

(u

), y
a

(u

)), for indices i, j, k, . . . =1, 2 and a, b, c, . . . =3, 4

(in brief, we write u =(x, y), u

=(x

, y

)); the Einstein rule on

summation on repeating up-low indices will be applied if a
contrary statement is not emphasized.
rescaled/anisotropic and renormalized forms. Finally, we
shall generalize the constructions for exact and approxi-
mate anisotropic solutions for generic o-diagonal EYMH
systems and their covariant renormalizable models.
Decoupling property and formal integration of
Einstein eqs. We briey review the anholonomic
deformation method for constructing o-diagonal exact
solutions in the GR theory and modications [8,9]. Any
spacetime metric can be parametrized in the form
g =g

(u)du

du

=g

(u)e

=g
ij
e
i
e
j
+ g
ab
e
a
e
b
, (2)
g

g
ij
+N
a
i
N
b
j
g
ab
N
e
j
g
ae
N
e
i
g
be
g
ab

, (3)
e

=[e
i
=/x
i
N
b
i
(u)
b
, e
a
=
a
=/y
a
], (4)
e

=[e
j
=dx
j
, e
b
=dy
b
+ N
b
i
(u)dx
i
]. (5)
We associate the coecients N={N
a
i
(u)} to a 2 +2
splitting of V stated explicitly for the tangent space
N: TV =hV vV for a nonintegrable (nonholonomic)
distribution with conventional horizontal, h, and vertical,
v, subspaces
2
. The decoupling property for the Einstein
eqs and their solutions can proved for any metric g (2)
and parametrization with nonunderlined (in general,
depending on variables (x
i
, y
3
)) and underlined
multiples (in general, depending on variables (x
i
, y
4
)),
g
i
= g
i
(x
k
), g
a
=
2
(x
i
, y
c
)h
a
(x
k
, y
3
)h
a
(x
k
, y
4
),
N
3
i
= w
i
(x
k
, y
3
) +w
i
(x
k
, y
4
), N
4
i
=n
i
(x
k
, y
3
)
+n
i
(x
k
, y
4
). (6)
Such functions of necessary smooth class have to be
dened in a form generating solutions of Einstein eqs.
A conformal v-factor (x
i
, y
c
) may depend on all coor-
dinates. We may simplify substantially the constructions
if we take =h
a
=1 and w
i
=n
i
=0 resulting in generic
o-diagonal metrics with Killing symmetry on /y
4
.
There will be used brief denotations for partial deriva-
tives: a

=a/x
1
, a

=a/x
2
, a

=a/y
3
, a

=a/y
4
(see footnote
3
).
The system of Einstein eqs derived for the rst action in
(1), where the Ricci tensor

R

is computed for (6), can

2
Such a conventional decomposition can be performed addition-
ally/alternatively/ in parallel to the well-known 3 +1 splitting and
Arnowit-Deser-Misner (ADM) formalism. Parametrizations of type
(3) are used, for instance, in Kaluza-Kelin gravity (see review [11]).
In this paper, we consider 4-d models with nonlinear dependences of
N
a
i
(u

) on all coordinates and do not set compactication on some

extra dimensions.
3
In this work, we restrict our constructions to some classes of
spacetime metrics, which can be generated from a set of nontrivial
data [g
i
, h
a
, w
i
, n
i
] with Killing symmetry by nonholonomic defor-
mations depending on certain small parameters resulting into general
non-Killing data [g
i
,
2
h
a
h
a
, w
i
+w
i
, n
i
+n
i
]. In a perturbative
approach to QG, we can consider that the contributions from
quadratic terms for products of the coecients of metrics and/or
connections, of type , are small for a xed open region U V
endowed with normal coordinates.
50001-p2
Covariant renormalizable anisotropic theories and o-diagonal EYMH solutions
be transformed equivalently into a system of PDE with
h-v decoupling, see details in [8,9],
g

2

g

1
g

2
2g
1

(g

2
)
2
2g
2
g

1

g

1
g

2
2g
2

(g

1
)
2
2g
1
=2g
1
g
2

,
1
2h
3
h
4

4

(h

4
)
2
2h
4

3
h

4
2h
3

+
1
2h
3
h
4
[h

3

(h

3
)
2
2h
3

3
h

4
2h
4
] =

, (7)
w
k
2h
4

4

(h

4
)
2
2h
4

3
h

4
2h
3

+
h

4
4h
4

k
h
3
h
3
+

k
h
4
h
4

k
h

4
2h
4
+
h
3
2h
4
n

k
+

h
3
h
4
h

3
2
h

k
2h
4
=0, (8)
w
k
2h
3

3

(h

3
)
2
2h
3

3
h

4
2h
4

+
h

3
4h
3

k
h
3
h
3
+

k
h
4
h
4

k
h

3
2h
3
+
h
4
2h
3
n

k
+

h
4
h
3
h

3
2
h

k
2h
3
=0, (9)
w

i
=(
i
w
i
) ln |h
4
|, (
k
w
k
)w
i
=(
i
w
i
)w
k
, (10)
n

i
=0,
i
n
k
=
k
n
i
,
w

i
=0,
i
w
k
=
k
w
i
, n

i
=(
i
n
i
) ln |h
3
|,
(
k
n
k
)n
i
=(
i
n
i
)n
k
,
e
k
=
k
(w
i
+w
i
)

(n
i
+n
i
)

=0.
(11)
The values

R

and

R
ak
are equal to the corresponding
ones computed for the Levi-Civita connection if the
conditions (10) are satised. In order to exclude certain
degenerate classes of solutions of the above nonlinear
systems of PDEs, we can impose the conditions h

4
=0
and h

3
=0. Such conditions can be satised always if
the corresponding frame/coordinate transforms to the
necessary ansatz are considered.
At the next step, we show that we can construct
in general form a class of exact solutions with generic
o-diagonal metrics determined by coecients with one
Killing symmetry, [g
i
, h
a
, w
i
, n
i
] with h

4
=0. For
(x
k
, y
3
) =ln

4
/

|h
3
h
4
|

,
i
=h

i
, =h

,
(12)
we can write respectively the eqs (7), (8) in the forms

4
=2h
3
h
4

=0, (13)
w
i
+
i
=0. (14)
For (9), we must take any trivial solution given by a
function n
i
= n
i
(x
k
) satisfying the conditions
i
n
j
=
j
n
i
in order to solve the constraints (10). Using coecients
(12) with
i
=0 and =0, the solution of the above
system of Einstein eqs with arbitrary o-diagonal coef-
cients for one Killing symmetry can be expressed in the
form determined by generating functions (x
k
), (x
k
, y
3
),

=0, n
i
(x
k
) and h
4
(x
k
, y
4
), and the integration function
0
(x
k
) following recurrent formulas and conditions,
h
3
=
(

)
2
4

, h
4
=
1
4

e
2[
0
]
, w
i
=

. (15)
In the above formulas, we should take respective values

i
=1 and in (15) in order to x a necessary space-
time signature. The generating/integration functions
may depend also on arbitrary nite sets of parame-
ters =(
1
,
2
, . . .) that we found in refs. [8,9]. In
general, such parametric functions are of type (x
k
, ),
(x
k
, y
3
, ) etc. and their explicit form have to be dened
from certain boundary/asymptotic conditions and/or
experimental data. In some limits, the resulting solu-
tions may describe congurations with eective broken
Lorentz symmetry, anisotropies, deformed symmetries
etc. For simplicity, we shall not write in explicit form the
parametric dependence of values if that will not result
in ambiguities. Re-scaling the generating function in
such a form that

for

=0, we can generate
in a similar form solutions with Killing symmetry on
/y
3
with data [g
i
, h
a
, w
i
, n
i
] for rescaled gener-
ating/integration functions (x
k
, y
4
),

=0, w
i
(x
k
),
0
(x
k
), when h
3
=e
2[
0
]
, h
4
=(

)
2
, n
i
=
i
/

.
We conclude that the Einstein eqs

R

, where

is the Ricci tensor and

=0, decouple in very general
forms with respect to N-adapted frames (4) and (5), which
allows us to express their solutions (up to corresponding
frame transforms) in the form (6)
g =
i
e
(x
k
)
dx
i
dx
i
+
2
[
3
(

)
2
e
2(
0
)
e
3
e
3
+
4
(

)
2
e
2(
0
)
e
4
e
4
], (16)
e
3
= dy
3
+(w
i

i
/

)dx
i
, e
4
=dy
4
+(n
i

i
/

)dx
i
,
where

=1 x a corresponding signature and is

subjected to constraints of type (11),
e
k
=
k
+(
i
/

w
i
)

n
i
+
i
/

=0.
Such classes of metrics are generic o-diagonal and dened
by corresponding sets of generating/integration functions
and parameters. Other classes of solutions with h

4
=0,
or h

3
=0, and/or

=0 consist of some special cases
analyzed in [8,9]. Physically important solutions for black
holes/ellipsoids, with singularities and horizons, can be
constructed for corresponding classes of coecients of
metrics, generating and integration functions etc.
Solutions mimicking nonstandard perfect uid
coupling in at backgrounds. The scalar curvature
s

R:=g

R
4k
derived from the Ricci tensors computed for
ansatz g (6) with Killing symmetry on /y
4
and nontriv-
ial [g
i
, h
a
, w
i
, n
i
] is computed
4 s

R=2

4
h
3
h
4
=4

.
Let us state the conditions for generating functions
when a metric g

is dened as a solution of eld eqs

in theories for
[1]
L and/or
[2]
L is also equivalent to a
solution derived for
[3]
L. Considering that

L is a matter
Lagrangian density depending only on the metric tensor
components g

but not on its derivatives, for the energy-

momentum tensor

T

=
2

|g|
(

|g|

L)
g

=g

L2

L
g

,
4
For simplicity, hereafter, we shall use such classes of o-diagonal
solutions even if technically it will be always possible to extend the
constructions to general ones (16).
50001-p3
S. I. Vacaru
we can construct sources with nontrivial
5
T
1
1
=T
2
2
=
(x
i
, y
3
) and T
3
3
=T
4
4
=

. For the eld eqs derived from
[2]
L=
s

R+

L, for the same g and
s

R=
s

R, with respect
to N-adapted frames, the rst eq from (12) and (13)
modify respectively as

|h
3
h
4
| =h

4
e

and

4
=2h
3
h
4
(17)
for a new generating function

(x
i
, y
3
) and (x
i
, y
3
).
The theories for
[1]
L and
[2]
L are equivalent if
their generating functions and sources are related
as ||
1
(e
2

1
(e
2
)

. Using this formula

and solutions (15), we can compute the func-
tional dependences h
4
[,

] =(4

)
1
(e
2
e
2
0

) =
h
4
[

, ], h
3
[

, ] =4[(

|h
4
|)

]
2
e
2

, when

e
2

=
e
2
0

dy
3
||(e
2
)

(we can use the inverse formula,

e
2
e
2
0

dy
3
||
1
(e
2

) with
0

(x
i
) and integra-
tion functions
0
(x
i
).
At the next step, we state the condition when a
source transforms a theory for
[2]
L into a model for
[3]
L. With respect to coordinate frames and for a at
background metric

, we consider a generic o-diagonal

metric

g

+h

(x
i
, t), where y
3
=t is the time-
like coordinate, with chosen gauge conditions h
tt
=
h
t

i
=h

it
=0, for

i,

j, =1, 2, 4 (on a manifold V, we can
consider a double splitting (3 +1) and (2 +2)). The
corresponding Ricci tensor and scalar curvature are

j
=
1
2
(h

j
+

k
h

k
+

k
h

k
h

j
),

R
33
=
1
2

j
h

j
;

R=

j
(h

k
h

j
) +

j
h

j
.
We chose a generating function

(x
i
, t),

=0, when
[1]
L =
s
R

= 3

4
h
3
h
4
=
[2]
L =
s

R +

L =
[3]
L =

R+

L, with

L taken for

g

and an eective nonstan-

dard coupling with a uid conguration, is induced
by for a generic o-diagonal solution of the system
(13) and (14). Einstein manifolds encoding uid-like
congurations are generated if

contains parameters ,
, , introduced into eqs

4
2h
3
h
4
=

L=
2
{[(31) +
1
2
]

j
h

j
+(+3 )(

j
h

k
h

j
)}
2
. (18)
In above formulas, we wrote

=

L in order to
emphasize that such a source is determined for a metric

and Lagrangian

L. Using this expression and
formulas (15) redened for (17), we nd recurrently
h
4
=
1
4

dy
3
|

|
1
(e
2

, h
3
=4

|h
4
|

2
e
2

,
for

e
2

=e
2
0

dy
3
|

|(e
2
)

and w
i
=

used for a
[1]
L-theory.
5
We may consider

L=(g
1
1
+g
1
2
)P
1
+(g
1
3
+g
1
4
)P
2
and
(g
1
1
+g
1
2
2)P
1
+(g
1
3
+g
1
4
)P
2
=

, (g
1
1
+g
1
2
)P
1
+(g
1
3
+
g
1
4
2)P
2
=.
In the limit (1 )/2(31), we can express

2
4
(+1)
2
[(

j
h

k
h

j
)]
2
and generate a
class of Einstein manifolds,
g =
i
e
(x
k
)
dx
i
dx
i
+
3

dy
3
|

|
1
(e
2

2
e
2

e
3
e
3
+
4

dy
3
|4

|
1
(e
2

e
4
e
4
,
e
3
= dy
3
(
i
/

)dx
i
, e
4
=dy
4
+n
i
dx
i
, (19)
Such manifolds are with Killing symmetry on /y
4
and
broken Lorentz invariance because the source

does
not contain the derivative with respect to /y
3
=
t
.
Non-Killing congurations of type (16) can be constructed
for sources

(x
i
, y
3
) +

(x
i
, y
4
) and nontrivial factors
(x
i
, y
a
), as more general classes of solutions.
The fact that such locally anisotropic spacetimes are
not Lorentz invariant is not surprising. We have a similar
case for the Schwarzschild-de Sitter solutions which
are dieomorphysm invariant and with broken Lorentz
symmetry. For the class of solutions (19), in the ultraviolet
region where momentum k is large, the second term for
the equivalent theory
[3]
L gives the propagator |k|
4
. The
longitudinal modes do not propagate being allowed the
propagation of the transverse one. Such a behavior is simi-
lar to that in a theory with nonstandard coupling of grav-
ity with perfect uid when the energy-momentum tensor
T

j
=p

j
=

j
and T
33
= (if we treat p, and as
standard uid parameters and eq of state) in the at back-
ground is computed for

L=(T

R

+T

)
2
.
Considering o-diagonal solutions for theories with
[1]
L and/or
[2]
L derived for generating functions and
sources of type (18), we obtain a situation in which
generic o-diagonal interactions in gravity induce a kind
of spontaneous violation of symmetry which is typical in
quantum eld theories. Using such nonholonomic congu-
rations, we model a vacuum gravitational aether via anal-
ogous coupling with nonstandard uid which breaks the
Lorentz symmetry for an eective equivalent theory
[3]
L.
This allows us to elaborate a (power-counting) renormal-
izable model of QG
6
.
Eective renormalizable EYMH congurations
and modied gravity. The constructions with
nonstandard uid coupling, eective renormalizability
etc. performed after (17) were provided, for simplicity,
for a at background. Nevertheless, the approach can be
extended to curved backgrounds and generic o-diagonal
gravitational-eld interactions. Using the principle of
relativity, we can work in a local Lorentz frame when,
for instance, the eective uid does not ow. This way
we preserve unitarity and axioms of GR working with
general class of solutions for theories
[1]
L and/or
[2]
L.
Via respective parametric dependence and nonholonomic
transforms we mimic some models
[3]
L with anisotropic
coupling. Moreover, even if the value

can be xed to
6
In this letter, we do not prove explicitly the renormalizability of
o-diagonal solutions but show that they can be associated/related
to certain renormalizable quantum models studied by other
authors.
50001-p4
Covariant renormalizable anisotropic theories and o-diagonal EYMH solutions
not contain time derivatives, it will get such terms in arbi-
trary frames of reference. As we mentioned in [8,9], the
solutions of type (19) and (16) in the diagonal spherical
symmetry limit (here we include the condition T

=0)
contain the Schwarzschild and Kerr black hole/ellipsoid
metrics. The analyzed models with =

L (18)
result in z =2 Horava-Lifshitz theories.
Let us show how in the scheme of Lagrangian densities
(1) we can include z =3 theories, which allows us to
generate ultraviolet power counting renormalizable 3 +1
and/or 2 +2 quantum models. Instead of

L, we can take
more general sources and generating functions,

n
L = {(T

+T

)
n
(T

R

+T

)}
2
, (20)
where n and are constants. Let us introduce =

n
L
in formulas (18) in order to state a dierent class of
generating functions
n

from
n

4
2h
3
h
4
=

n
L and construct
o-diagonal solutions of type (19) for data


n

,
n

and

n
L. In general, we can consider noninteger
values for n, for instance, n =1/2, 3/2 etc. If n is nega-
tive, we can generate nonlocal theories determined by
a respective o-diagonal polarization of vacuum and
cosmological constant in GR (the question of renor-
malizability is more complex for such cases). Following
the analysis provided in [7], the analogous
[3]
L and
[4]
L
models derived for (20) are renormalizable if n =1 and
super-renormalizable for n =2. The values induced by a
nontrivial source/Lagrangian density

n
L contains higher
derivative terms and eectively break the Lorentz symme-
try for high energies, i.e. in UV region. In the IR, we
positively get the usual Einstein gravity. Nevertheless,
in our approach, GR is not only a limit from certain
modications with nonstandard coupling of type (18)
and/or (20) but a model with possible branches of
complexity, anisotropies, inhomogeneities and Lorentz
violations depending on parameters and generating
functions. Certain families of solutions are (super-) renor-
malizable because of o-diagonal nonlinear interactions
of gravitational and eective matter elds.
It is important to analyze two special cases for
sources

n
L, for instance, in the uid approximation
for matter. We model two theories
[1]
L and
[3]
L
with cosmological constant when both =1 and

=0. This corresponds to diagonal solutions with

T

R

+T

=0 when 1/|k|
4
cannot be obtained
for the propagator. To correct the model, we have to
introduce additional o-diagonal terms. The case =1/3
corresponds to the radiation of conformal matter with
divergent . This also does no provide good solutions
if other anisotropic contributions are not considered.
The approach with generic o-diagonal solutions
modelling eective covariant renormalizable theories
can be extended for gravitational interactions with a
non-Abelian SU(2) gauge eld A=A

coupled to
a triplet Higgs eld , see details in [8,9]. In order to
prove the decoupling property, we have to use a covariant
operator

D which is adapted to the N-splitting. It is
not possible to see a general splitting of Einstein and
matter eld eqs if we do not consider 2 +2 spacetime
decompositions with such nonintegrable distributions.
The linear connection

D is equivalent to the Levi-Civita
connection if the conditions (10) are satised and
all computations are performed with respect N-adapted
bases. In terms of

D, such nonholonomic interactions are
described,

1
2
g
s

R=8G

H
T

+
Y M
T

, (21)
D

|g|F

) =
1
2
ie(

|g|) [, D

], (22)
D

|g|) =(

|g|) (
2
[0]

2
), (23)
where the source of the Einstein eqs is
H
T

=Tr

1
4
(

)
1
4
g

(24)
g

V(),
Y M
T

=2 Tr(g

1
4
g

).
(25)
The curvature of gauge eld A

is F

=e

+
ie[A

, A

], where e is the coupling constant, i

2
=1,
and [, ] is used for the commutator.
[0]
in (23) is the
vacuum expectation of the Higgs eld which determines
the mass
H
M =

, when is the constant of scalar eld

self-interaction with potential V() =
1
4
Tr(
2
[0]

2
)
2
.
The gravitational constant G denes the Planck mass
M
Pl
=1/

G; the is a nontrivial mass of gauge boson,

W
M =ev.
A class of diagonal solutions of the system (21)(23),
see ref. [12], is given by the metric ansatz

g =

g
i
(x
1
)dx
i
dx
i
+

h
a
(x
1
, x
2
)dy
a
dy
a
= q
1
(r)dr dr +r
2
d d +r
2
sin
2
dd

2
(r)q(r)dt dt, (26)
where the coordinates and metric coecients are
parametrized, respectively, u

=(x
1
=r, x
2
=, y
3
=
, y
4
=t) and

g
1
=q
1
(r),

g
2
=r
2
,

h
3
=r
2
sin
2
,

h
4
=

2
(r)q(r), for q(r) =1 2m(r)/r r
2
/3, where is
a cosmological constant. The function m(r) is usually
interpreted as the total mass-energy within the radius r
which for m(r) =0 denes an empty de Sitter, dS, space
written in a static coordinate system with a cosmological
horizon at r =r
c
=

3/. An ansatz for the solution of

the Yang-Mills (YM) eqs on (21) on a spherically symmet-
ric background in GR is dened by a single magnetic
potential v(r),

A =

A
2
dx
2
+

A
3
dy
3
=
1
2e
[v(r)
1
d +
(cos
3
+v(r)
2
sin ) d], where
1
,
2
,
3
are Pauli
matrices, and for eqs (23) of the Higgs (H) eld it is given
by =

=(r)
3
. The functions (r), q(r), v(r), (r)
can be computed for the data [

g(r),

A(r),

(r)].
For instance, the diagonal Schwarzschild-de Sitter
solution of (21)(23) is that determined by data
v(r) =1, (r) =1, (r) =0, q(r) =1 2M/r r
2
/3 de-
ning a black-hole conguration inside a cosmological
50001-p5
S. I. Vacaru
horizon because q(r) =0 has two positive solutions and
M <1/3

.
A prime diagonal solution

g (26) can be transformed
into target o-diagonal metrics g =

g,

g

g,

g =
i
e

dx
i
dx
i
+
2
[
3
(

)
2
e
2(
0
)
e
3
e
3
+
4
(

)
2
e
2(
0
)
e
4
e
4
], (27)
e
3
= dy
3
+
3
i
(w
i
+w
i
)dx
i
, e
4
=dy
3
+
4
j
(n
j
+n
j
)dx
j
,
where the gravitational -polarizations
i
(x
k
),
b
(x
k
, y
a
)
and
c
j
(x
k
, y
a
) have to be found from the condition that
such metrics generate solutions of (21)(23). With respect
to N-adapted frames, the gauge elds are deformed
A

(x
i
, y
3
) =

(x
1
) +

(x
i
, y
a
), (28)
where

A

is a function for which F

=

F

(x
1
) +

(x
i
, y
a
) =s

|g|

, for s =const and

being the
absolute antisymmetric tensor. Such a tensor always solve
the eqs D

|g|F

) =0, which always give us the possi-

bility to determine the distortions

F

, and

A

, for any
given

A

and/or

F

. The scalar eld is nonholonomi-

cally modied

(x
1
) (x
i
, y
a
) =

(x
i
, y
a
)

(x
1
) by

is such a way that D

=0 and (x
i
, y
a
) =
[0]
.
This nonholonomic conguration of the nonlinear scalar
eld is not trivial even with respect to N-adapted frames
V() =0 and
H
T

=0, see (24)

7
.
The gauge elds with potential A

(28) modied
nonholonomically by determine exact solutions for the
YMH system (22) and (23) for spacetime metrics of type
(27). The corresponding stress energy-momentum tensor
is computed (see details in sect. 3.2 and 6.51 from [13])
Y M
T

=4s
2

, which means that nonholonomically

interacting gauge and Higgs elds, with respect to
N-adapted frames, result in an eective cosmological
constant
s
=8s
2
which should be added to a respec-
tive sources of Einstein eqs. Using the above-presented
constructions, we conclude that various classes of nonholo-
nomic EYMH systems can be modeled, respectively, as
theories of type
[1]
L,
[2]
L and
[3]
L using an eective
cosmological constant

= +
s
and o-diagonal ansatz
for metrics (27) when -polarizations are treated as
additional generating functions. Instead of such classes
of equivalence of conjectured covariantly renormalizable
theories with locally anisotropic interactions, we can
reformulate the results in the language of F(R)-gravity
of Horava-Lifshitz type [10]. The eective Lagrangian
density
[4]
L=F(

R) can be chosen for

R=R +

L,
see (18), when z =2, or

R=R+

n
L, see (20), when
z =2n+2. EYMH congurations are expected to be
covariantly renormalizable when z 3 which may be
demonstrated using arguments presented in [7,14] and in
this paper and/or other variants of anisotropic renormal-
ization. A rigorous proof is possible if we chose generating
7
For the ansatz (27), the Higgs eqs are (/x
i
A
i
)=(w
i
+
w
i
)

+(n
i
+n
i
)

, (/y
3
A
3
)=0, (/y
4
A
4
)=0, i.e. a
scalar eld modies the o-diagonal components of the metric via
w
i
+w
i
and n
i
+n
i
and nonholonomic conditions for A

=

A

.
functions when our o-diagonal metrics induce models
with covariant power-counting renormalization [1517];
such constructions will be provided in our future works.
Concluding remarks. In summary, we have used
the possibility to decouple in very general forms the grav-
itational eld eqs in GR and construct various classes of
exact generic o-diagonal solutions in order to elaborate
covariant (super-) renormalizable theories of gravity. We
also briey considered how EYMH interactions can be
encoded into nonholonomic Einstein manifolds and eec-
tive models of interactions with broken Lorentz symmetry
and covariant renormalization.
We emphasize that in this work we followed an ortho-
dox approach to modeling classical and quantum geome-
tries and physical theories of interactions keeping the
constructions and physical paradigm to be maximally
closed to Einstein gravity. Our opinion is that the bulk
of experimental data for modern gravity and cosmology
can be explained/predicted using generic o-diagonal solu-
tions in GR and their quantized versions.

I thank E. Elizalde and S. Odintsov for hospital-
ity and important critical remarks and references. The
research in this paper is partially supported by the
Program IDEI, PN-II-ID-PCE-2011-3-0256.
REFERENCES
[1] Alvarez E., Rev. Mod. Phys., 61 (1989) 561.
[2] Buchbinder I. L., Odintsov S. and Shapiro I., Eec-
tive Action in Quantum Gravity (IOP, London) 1992.
[3] Oriti D. (Editor), Approaches to Quantum Gravity
Toward a New Understanding of Space Time and Matter
(Cambridge University Press, Cambridge) 2009.
[4] Ho rava P., Phys. Rev. D, 79 (2009) 084009.
[5] Lu H., Mei J. and Pope C. N., Phys. Rev. Lett., 103
(2009) 091301.
[6] Moffat J. W., Class. Quantum Grav., 27 (2010)
135016.
[7] Nojiri S. and Odintsov S., Phys. Rev. D, 81 (2010)
043001.
[8] Vacaru S., Int. J. Geom. Methods Mod. Phys., 8 (2011) 9.
[9] Vacaru S., arXiv:1108.2022.
[10] Carloni S., Chaichian M., Nonjiri S., Odintsov S.
D., Oksanen M. and Tureanu A., Phys. Rev. D, 82
(2010) 065020.
[11] Overduin J. M. and Wesson P. S., Phys. Rep., 283
(1997) 303.
[12] Brihaye Y., Hartmann B., Radu E. and Stelea C.,
Nucl. Phys. B, 763 (2007) 115.
[13] Lidsey J. E., Wands D. and Copeland E. J., Phys.
Rep., 337 (2000) 343.
[14] Vacaru S., Int. J. Geom. Methods Mod. Phys., 7 (2010)
713.
[15] Nojiri S. and Odintsov S. D., arXiv:1004.3613.
[16] Kluson J., Nojiri S. and Odintsov S., Phys. Lett. B,
701 (2011) 117.
[17] Nojiri S. and Odintsov S., Phys. Rev. D, 83 (2011)
023001.
50001-p6