Professional Documents
Culture Documents
r
X
i
v
:
s
u
b
m
i
t
/
0
2
9
6
1
7
0
[
h
e
p
-
t
h
]
8
A
u
g
2
0
1
1
Decoupling of EYMH Equations, ODiagonal
Solutions, and Black Ellipsoids and Solitons
Sergiu I. Vacaru
a
= 0 . . . . . . . . . . . . . . . 15
3.1.2 Eective vacuum EYMH congurations . . . . . . . . . . . . 17
3.2 NonKilling EYMH congurations . . . . . . . . . . . . . . . . . . . 18
3.2.1 Nonvacuum odiagonal solutions . . . . . . . . . . . . . . . 19
3.2.2 Eective vacuum odiagonal solutions . . . . . . . . . . . . 19
4 The Cauchy Problem and Decoupling of EYMH Equations 20
4.1 The local Nadapted evolution problem . . . . . . . . . . . . . . . . 21
4.2 On initial data sets and global nonholonomic evolution . . . . . . . . 24
5 Anholonomic YMH Deformations of Black Holes 26
5.1 (Non) holonomic nonAbelian eective vacuum spaces . . . . . . . . 26
5.2 NonAbelian deformations of the Schwarzschild metric . . . . . . . . 27
5.3 Linear parametric polarizations induced by YMH elds . . . . . . . 29
6 Ellipsoidal EYMH Congurations and Solitons 30
6.1 Nonholonomic rotoid deformations . . . . . . . . . . . . . . . . . . . 30
6.2 Eective vacuum solitonic congurations . . . . . . . . . . . . . . . . 31
6.2.1 Solutions with solitonic factor (x
1
, y
3
, t) . . . . . . . . . . . 32
6.2.2 Solitonic metrics with factor (x
i
, t) . . . . . . . . . . . . . . 32
A Nonholonomic 2+2 Splitting of Lorentz Manifolds 33
B Proof of Theorem 2.1 37
B.1 Coecients of the canonical dconnection . . . . . . . . . . . . . . . 38
B.2 Coecients for torsion of
D . . . . . . . . . . . . . . . . . . . . . . . 39
B.3 Calculation of the Ricci tensor . . . . . . . . . . . . . . . . . . . . . 40
B.4 Zero torsion conditions . . . . . . . . . . . . . . . . . . . . . . . . . . 45
B.5 Geometric data for diagonal EYMH congurations . . . . . . . . . . 45
2
1 Introduction
The issue of constructing exact and approximate solutions of gravita-
tional and matter eld equations is of interest in mathematical relativity,
particle physics and for various applications in modern cosmology and as-
trophysics. The Einstein, YangMills, Higgs and other fundamental eld
equations in physics are sophisticate systems of nonlinear partial dieren-
tial equations (PDE) which are very dicult to be integrated and studied
in general forms. There were elaborated various geometric, analytic and
numeric methods, see [1, 2, 3, 4, 5, 6, 7, 8] and references therein.
In this work, we address again and develop a geometric approach (the
socalled anholonomic deformation method, see [9, 10, 11]) to constructing
exact solutions of PDE describing generic odiagonal and nonlinear gravita-
tional, gauge and scalar eld interactions. The goal is to provide new results
on decoupling property
1
of EinsteinYangMillsHiggs, EYMH, equations
and integration of such systems. Certain issues on the Cauchy problem and
decoupling of Einstein equations and the initial data sets and nonholonomic
evolution will be analyzed. Examples of odiagonal solutions for EYMH
black holes/ellpsoids and solitonic congurations will be given.
In Section 2 we provide the main Theorems on decoupling the Einstein
equations for generic odiagonal metrics with one Killing symmetry and
consider extensions to classes of nonKilling solutions with coecients
depending on all set of four coordinates. We prove also that the decoupling
property also holds true for certain classes of nonholonomic EYMH systems.
The theorems on generating odiagonal solutions are considered in Section
3. Section 4 contains a study of the Cauchy problem in connection to the
decoupling property of Einstein equations. Next two sections are devoted to
examples of generic odiagonal exact solutions of the EYMH equations. In
Section 5 there are constructed nonholonomic YMH vacuum deformations
of black holes. Ellipsoidsolitonic nonAbelian congurations are analyzed
in Section 6. Appendix A contains a survey on the geometry of nonholo-
nomic 2+2 splitting of Lorentz manifolds. The most important formulas
and computations which are necessary to prove the decoupling property of
Einstein equations are given in Appendix B.
1
it is used also the term separation of equations, which should not be confused with
separation of variables
3
2 Decoupling Property of Einstein Equations
In general relativity (GR), the curved spacetime (V, g) is dened by a
pseudoRiemannian manifold V endowed with a Lorentzian metric g as a
solution of Einstein equations.
2
The space of classical physical events is mod-
elled as a Lorenzian four dimensional, 4-d, manifold V (of necessary smooth
class, Housdorf and paracompact one) when the symmetric 2covariant ten-
sor g = {g
and e
}
on a chart U V, we can write e
= /u
and e
= du
and, for
instance, dene the coecients of a vector X and a metric g, respectively,
in the forms X = X
and
g = g
(u)e
, (1)
where g
:= g(e
, e
).
3
We consider bases with nonintegrable (equiva-
lently, nonholonomic/anholonomic) 2 + 2 splitting for conventional, hori-
zontal, h, and vertical, v, decomposition, when for the tangent bundle TV
:=
u
T
u
V a Whitney sum
N : TV = hV vV (2)
is globally dened. Such a nonholonomic distribution is determined locally
by its coecients N
a
i
(u), when N =N
a
i
(x, y)dx
i
/y
a
, where u
= (x
i
, y
a
)
splits into hcoordinates, x = (x
i
), and vcoordinates, y = (y
a
), with indices
running respectively values i, j, k, ... = 1, 2 and a, b, c, ... = 3, 4.
4
2
We assume that readers are familiar with basic concepts and results on mathematical
relativity and methods of constructing exact solutions outlined, for instance, in above
mentioned monographs and reviews.
3
The summation rule on repeating lowup indices will be applied if the contrary will
be not stated.
4
We note that the 2+2 splitting can be considered as an alternative to the well known
3+1 splitting. The rst one is convenient, for instance, for constructing generic o
diagonal solutions and elaborating models of deformation and/or Abrane quantization
of gravity, but the second one is more important for canonical/loop quantization etc.
4
Any spacetime (V, g) can be equipped with a nonintegrable bred struc-
ture (2) and such a manifold is called nonholonomic (equivalently, Nanholo-
nomic). We use boldface letters in order to emphasize that certain spaces
and geometric objects/constructions are Nadapted, i. e. adapted to
a hvsplitting. The geometric objects are called distinguished (in brief,
dobjects, dvectors, dtensors etc). For instance, we write a dvector as
X = (hX, vX) for a nonholonomic Lorentz manifold/spacetime (V, g).
On a spacetime (V, g) , we can perform/adapt the geometric construc-
tions using Nelongated local bases ( partial derivatives), e
= (e
i
, e
a
),
and cobases (dierentials), e
= (e
i
, e
a
), when
e
i
= /x
i
N
a
i
(u)/y
a
, e
a
=
a
= /y
a
, (3)
and e
i
= dx
i
, e
a
= dy
a
+ N
a
i
(u)dx
i
. (4)
Such (co) frame structures depend linearly on Nconnection coecients be-
ing, in general, nonholonomic. For instance, the basic vectors (3) satisfy
certain nontrivial nonholonomy relations
[e
, e
] = e
= W
, (5)
with (antisymmetric) nontrivial anholonomy coecients
W
b
ia
=
a
N
b
i
, W
a
ji
=
a
ij
= e
j
(N
a
i
) e
i
(N
a
j
). (6)
Any spacetime metric g = {g
= (dx
i
, dy
a
), as an o
diagonal metric
g = g
du
du
,
where
g
g
ij
+N
a
i
N
b
j
g
ab
N
e
j
g
ae
N
e
i
g
be
g
ab
. (8)
A metric g is generic odiagonal if (8) can not be diagonalized via coordi-
nate transforms. Ansatz of this type are used in KaluzaKlein gravity when
N
a
i
(x, y) =
a
bi
(x)y
a
and y
a
are compactied extradimensions coordi-
nates, or in Finsler gravity theories, see details in [12, 11]. In this work, we
restrict our considerations only to the 4d Einstein gravity theory. The prin-
ciple of general covariance in GR, allows us to consider any frame/coordinate
5
transforms and write a spacetime metric g equivalently in any above form
(1), (8) and/or (7). The last mentioned parametrization will allow us to
prove a very important property of decoupling of Einstein equations with
respect to Nadapted bases (3) and (4).
Via frame/ coordinate transforms e
= e
(x, y)e
, g
= e
,
a metric g (7) can be written in a form with separation of vcoordinates
and nontrivial vertical conformal transforms,
g = g
i
dx
i
dx
i
+
2
h
a
h
a
e
a
e
a
, (9)
e
3
= dy
3
+ (w
i
+w
i
) dx
i
, e
4
= dy
4
+ (n
i
+n
i
) dx
i
,
were
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
), (10)
are functions of necessary smooth class which will be dened in a form to
generate solutions of Einstein equations.
5
The aim of this section is to prove that the gravitational eld equa-
tions in GR, in the vacuum cases and certain vary general classes of matter
eld sources decouple for parametrizations of metrics in the form (10). For
convenience, we present in Appendix A a brief review on the geometry of
nonholonomic 2+2 splitting of Lorentz manifolds and Einstein equations.
2.1 Splitting of (non) vacuum gravitational eld equations
We shall use brief denotations for partial derivatives, a
= a/x
1
, a
=
a/x
2
, a
= a/y
3
, a
= a/y
4
. The equations will be written with
respect to Nadapted frames of type (3) and (4).
2.1.1 Odiagonal spacetimes with Killing symmetry
We use an ansatz (9) when = 1, h
3
= 1, w
i
= 0 and n
i
= 0 in data
(10) and = 0 for (A.15). Such a generic odiagonal metric does not
depend on variable y
4
, i.e. /y
4
is a Killing vector, if h
4
= 1. Nevertheless,
the decoupling property can proven for the same assumptions but arbitrary
h
4
(x
k
, y
4
) with nontrivial dependence on y
4
. We call this class of metrics
to be with weak Killing symmetry because they result in systems of PDE
(A.12) as for the Killing case but there are dierences in (A.13) if h
4
= 1.
5
There is not summation on repeating low indices a in formulas (10) but such a
summation is considered for crossing uplow indices i and a in (9)). We shall underline
a function if it positively depends on y
4
but not on y
3
and write, for instance, n
i
(x
k
, y
4
).
6
Theorem 2.1 The Einstein eqs (A.12) and (A.13) for a metric g (10) with
= h
3
= 1 and w
i
= n
i
= 0 and = 0 in matter source
(A.15) are
equivalent, respectively, to
R
1
1
=
R
2
2
=
1
2g
1
g
2
[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
] =
v
, (11)
R
3
3
=
R
4
4
=
1
2h
3
h
4
[h
4
(h
4
)
2
2h
4
3
h
4
2h
3
] = , (12)
R
3k
=
w
k
2h
4
[h
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
= 0, (13)
R
4k
=
h
4
2h
3
n
k
+
h
4
h
3
h
3
2
h
k
2h
3
= 0, (14)
and
w
i
= (
i
w
i
) ln |h
4
|, (
k
w
k
)w
i
= (
i
w
i
)w
k
, (15)
n
k
h
4
=
k
h
4
, n
i
= 0,
i
n
k
=
k
n
i
.
Proof. See Appendix B.
Let us discuss the decoupling (splitting) property of the Einstein equa-
tions with respect to certain classes of Nadapted frames which is contained
in the system of PDE (11) (15). For instance, the rst equation is for a
2d metric which always can be diagonalized, [g
1
, g
2
], and/or made to be
conformally at. Prescribing a function g
1
and source
v
, we can nd g
2
,
or inversely. The equation (12) contains only the rst and second deriva-
tives on /y
3
and relates two functions h
3
and h
4
. Prescribing one of such
functions and source , we can dene the second one taking, respectively,
one or two derivations on y
3
. The equation (13) is a linear algebraic system
for w
k
if the coecients h
a
have been already dened as a solution of (12).
Nevertheless, we have to solve a system of rst order PDE on x
k
and y
3
in
order to nd w
k
resulting in zero torsion conditions (15). The forth equa-
tions (14) became trivial for any n
i
= 0 if we wont to satisfy completely
such zero torsion conditions.
6
A nontrivial function h
4
is explicitly present
in the conditions (15). If such restrictions are satised, this allows us to
eliminate h
4
from (13) (see details in the proof of above theorem).
We conclude that the Einstein equations for metrics with one Killing
symmetry can be such way parametrized with respect to Nadapted frames
that they decouple and separate into quite simple PDE for hcomponents,
6
Nontrivial solutions and nonzero torsion congurations present interest in modied
theories of gravity, see such examples in Ref. [17]
7
g
i
, and then for vcomponents, h
a
. The Nconnection coecients also sep-
arate and can be dened from corresponding algebraic and/or rst order
PDE. The zero torsion conditions impose certain additional constraints
(as some simple rst order PDE with possible separation of variables) on
Ncoecients and coecients of vmetric.
In a similar form, we can decouple the Einstein equations for spacetimes
with one Killing symmetry on /y
3
.
Corollary 2.1 The Einstein eqs (A.12) and (A.13) for a metric g (10) with
= h
4
= 1 and w
i
= n
i
= 0 and = 0 in matter source
(A.15) are
equivalent, respectively, to
R
1
1
=
R
2
2
=
1
2g
1
g
2
[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
] =
v
, (16)
R
3
3
=
R
4
4
=
1
2h
3
h
4
[h
3
(h
3
)
2
2h
3
3
h
4
2h
4
] = , (17)
R
3k
= +
h
3
2h
4
w
k
+
h
3
h
4
h
3
2
h
k
2h
4
= 0, (18)
R
4k
=
n
k
2h
3
[h
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
= 0, (19)
and
n
i
= (
i
n
i
) ln |h
3
|, (
k
n
k
)n
i
= (
i
n
i
)n
k
, (20)
w
k
h
3
=
k
h
3
, w
i
= 0,
i
w
k
=
k
w
i
.
Proof. It is similar to that for Theorem 2.1 provided in Appendix B.
We do not repeat such computations.
Using above Theorem and Corollary and mutual transforms of the sys-
tems of equations, we can formulate:
Conclusion 2.1 The nonlinear systems of PDE corresponding to Einstein
equations (A.12) and (A.13) for metrics g (10) with Killing symmetry on
/y
4
, when = h
3
= 1 and w
i
= n
i
= 0 and = 0 in matter source
(A.15), can be transformed into respective systems of PDE for data with
Killing symmetry on /y
3
, when = h
4
= 1 and w
i
= n
i
= 0 and = 0,
if h
3
(x
i
, y
3
) h
4
(x
i
, y
4
), h
4
(x
i
, y
3
) h
3
(x
i
, y
4
), w
k
(x
i
, y
3
) n
k
(x
i
, y
4
)
and n
k
(x
i
, y
3
) w
k
(x
i
, y
4
).
The above presented method of nonholonomic deformations can be used
for decoupling the Einstein equations even some metrics do not possess, in
8
general, any Killing symmetries. The generic nonlinear character of such
systems of PDE does not allow us to use a principle of superposition of
solutions. Nevertheless, certain classes of conformal transforms for the v
components of dmetrics and nonholonomic constraints of integral varieties
give us the possibility to extend the anholonomic deformation method to
nonKilling vacuum and nonvacuum gravitational interactions. In next
two subsections, we analyze two possibilities to decouple the Einstein equa-
tions for metrics with coecients depending on all spacetime coordinates.
2.1.2 Preserving decoupling under vconformal transforms
This property is stated by
Lemma 2.1 The Einstein equations (A.12) for geometric data (B.1), i.e.
the system of PDE (12)(14), do not change under a vertical conformal
transform with nontrivial (x
k
, y
a
) to a dmetric (10) if there are satised
the conditions
k
w
i
n
i
= 0 and
T
a
kb
= 0. (21)
Proof. It follows from straightforward computations when coecients
g
i
(x
k
), g
3
= h
3
(x
k
, y
3
), g
4
= h
4
(x
k
, y
3
)h
4
(x
k
, y
4
), N
3
i
= w
i
(x
k
, y
3
), N
4
i
=
n
i
(x
k
, y
3
)) are generalized to a nontrivial (x
k
, y
a
) with
g
3
=
2
h
3
and
g
4
=
2
h
4
h
4
. Using respectively formulas (A.2), (A.3), (A.6) and (A.7), we
get distortion relations for the Ricci dtensors (A.11),
R
a
b
=
R
a
b
+
Z
a
b
and
R
bi
=
R
bi
= 0,
where
R
a
b
and
R
bi
are those computed for = 1, i.e. (12)(14). The values
R
a
b
and
Z
a
b
are dened by a nontrivial and computed using the same
formulas. We do not repeat here such details provided in Refs. [10, 11] for
h
4
= 1 because a nontrivial h
4
does not modify substantially the proof that
Z
a
b
= 0 if the conditions (21) are satised.
Using Theorem 2.1, Corollary 2.1, Conclusion 2.1, Lemma 2.1, we prove
Theorem 2.2 A dmetric
g = g
i
(x
k
)dx
i
dx
i
+
2
(x
k
, y
a
)
h
3
e
3
e
3
+h
4
h
4
e
4
e
4
,
e
3
= dy
3
+w
i
(x
k
, y
3
)dx
i
, e
4
= dy
4
+n
i
(x
k
)dx
i
, (22)
satisfying the PDE (11) (15) and
k
w
i
n
i
= 0, or a dmetric
g = g
i
(x
k
)dx
i
dx
i
+
2
(x
k
, y
a
)
h
3
h
3
e
3
e
3
+h
4
e
4
e
4
,
e
3
= dy
3
+w
i
(x
k
)dx
i
, e
4
= dy
4
+n
i
(x
k
, y
4
)dx
i
, (23)
9
satisfying the PDE (16) (20) and
k
w
i
n
i
= 0, dene, in general,
two dierent classes of generic odiagonal solutions of Einstein equations
(A.12) and (A.13) with respective sources of type (A.15).
Both ansatz of type (22) and (23) consist particular cases of parametriza-
tions of metrics in the form (9). Via frame/coordinate transform into a nite
region of a point
0
u V any spacetime metric in GR can be represented
in an above mentioned dmetric form. If only one of coordinates y
a
is time-
like, the solutions of type (22) and (23) can not be transformed mutually
via nonholonomic frame deformations preserving causality.
2.1.3 Decoupling with eective linearization of Ricci tensors
The explicit form of eld equations for vacuum and nonvacuum gravi-
tational interactions depends on the type of frames and coordinate systems
we consider for decoupling such PDE. We can split such systems for more
general parameterizations (than ansatz (22) and (23)) in a form (9) with
nontrivial and Ncoecients in (10). This is possible in any open region
U V where for computing the Nadapted coecients of the Riemann and
Ricci dtensors, see formulas (A.6) and (A.7), we can neglect contributions
from quadratic terms of type
but preserve values of type
. For
such constructions, we have to introduce a class of Nadapted normal coor-
dinates when
(u
0
) = 0 for points u
0
, for instance, belonging to a line on
U. Such conditions can be satised for decompositions of metrics and con-
nections on a small parameter like it is explained in details in Ref. [9] (see
decompositions on a small eccentricity parameter in Section 5). Other
possibilities can be found if we impose nonholonomic constraints, for in-
stance, of type h
4
= 0 but for nonzero h
4
(x
k
, y
3
) and/or h
4
(x
k
, y
3
); such
constraints can be solved in nonexplicit form and dene a corresponding
subclass of Nadapted frames. Considering further nonholonomic deforma-
tions with a general decoupling with respect to a convenient system of
reference/coordinates, we can deform the equations and solutions to cong-
urations when terms of type
became important.
Theorem 2.3 (Nonquadratic decoupling) The Einstein equations
in GR (for instance, (A.12) and (A.13)), via nonholonomic frame deforma-
tions to a metric g (10) and matter source
R
1
1
=
R
2
2
=
1
2g
1
g
2
[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
] =
v
, (24)
10
R
3
3
=
R
4
4
=
1
2h
3
h
4
[h
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
]
= , (25)
R
3k
=
w
k
2h
4
[h
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
(26)
+
h
3
2h
4
n
k
+
h
3
h
4
h
3
2
h
k
2h
4
= 0,
R
4k
=
w
k
2h
3
[h
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, (27)
w
i
= (
i
w
i
) ln |h
4
|, (
k
w
k
)w
i
= (
i
w
i
)w
k
, n
i
= 0,
i
n
k
=
k
n
i
, (28)
and
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. (29)
Proof. It is a consequence of Conclusion 2.1 and Theorems 2.1 and 2.2
for superpositions of ansatz (22) and (23) resulting into (9). If we repeat
the computations from Appendix B for geometric data (10) considering that
contributions of type
1
2
g
s
R = 8G
H
T
+
Y M
T
, (30)
(
|g|)
1
D
|g|F
) =
1
2
ie[, D
], (31)
(
|g|)
1
D
|g|) = (
2
[0]
2
), (32)
11
where the source is determined by the stressenergy tensor
H
T
= Tr[
1
4
(D
+D
)
1
4
g
] g
V(),(33)
Y M
T
= 2Tr
1
4
g
. (34)
These equations can be derived following a variational principle for a grav-
itating nonAbelian SU(2) gauge eld A = A
= e
+ie[A
=
D
+ie[A
is
F
= e
+ie[A
, A
], (35)
where e is the coupling constant, i
2
= 1, and [, ] is used for the commuta-
tor. The value
[0]
in (32) is the vacuum expectation of the Higgs eld which
determines the mass
H
M =
N
a
i
= 0] and matter elds
A
(x
1
) and
(x
1
), for instance, of type con-
structed in Ref. [18] (see also Appendix B.5)). We suppose that there are
satised the following conditions:
1. The target dmetric
g with nontrivial Ncoecients, for
g
g
is parametrized in a form (9),
g =
i
(x
k
)
g
i
(x
1
)dx
i
dx
i
+
a
(x
k
, y
a
)
h
a
(x
1
, x
2
)e
a
e
a
= g
i
(x
k
)dx
i
dx
i
+
2
(x
k
, y
b
)h
a
(x
k
, y
a
)e
a
e
a
, (36)
e
3
= dy
3
+ [w
i
+w
i
]dx
i
, e
4
= dy
3
+ [n
i
+n
i
]dx
i
.
12
2. The gauge elds are nonholonomically deformed as
A
(x
i
, y
3
) =
A
(x
1
) +
A
(x
i
, y
a
), (37)
where
A
(x
1
) is dened by formula (B.8) and
A
(x
i
, y
a
) are any
functions for which
F
=
F
(x
1
) +
F
(x
i
, y
a
) = s
|g|
, (38)
for s = const and
,
F
and
F
(38)
satises the condition D
|g|F
,
A
,
F
.
3. The scalar eld is nonholonomically modied
(x
1
) (x
i
, y
a
) =
(x
i
, y
a
)
(x
1
) by a polarization
is such way that
D
= 0 and (x
i
, y
a
) =
[0]
. (39)
This nonholonomic conguration of the nonlinear scalar eld is not
trivial even with respect to Nadapted frames V() = 0 and
H
T
= 0,
see formula (33). For ansatz (36), the equations (39) transform into
(/x
i
A
i
) = (w
i
+w
i
)
+ (n
i
+n
i
)
, (40)
/y
3
A
3
= 0,
/y
4
A
4
= 0.
So, a nonolonomically constrained/deformed Higgs eld (depending
in nonexplicit form on two variables because of constraint (39)) mod-
ies indirectly the odiagonal components of the metric via w
i
+ w
i
and n
i
+n
i
and conditions (40) for
A
.
4. The gauge elds (38) with the potential A
= 4s
2
= diag[
+
Y M
T
= diag[
4s
2
].
Such a statement is a straightforward consequence of above assumptions 1-4
when in Nadapted frames the contributions of matter elds is dened by
an eective cosmological constant
s
. In a particular case, we can state
a
= 0 and certain special cases with zero and nonzero
sources (A.15). In general, such generic odiagonal metrics are determined
by generating functions depending on three/four coordinates. The bulk of
known exact solutions with diagonalizable metrics and coecients depending
on two coordinates (in certain special frames of references) can be included
as special cases for more general nonholonomic congurations.
3.1.1 (Non) vacuum metrics with h
a
= 0
For ansatz (9) with data = 1, h
3
= 1, w
i
= 0 and n
i
= 0 for (10), when
h
a
= 0, and the condition that the source
= diag[
:
1
1
=
2
2
= (x
k
, y
3
) 4s
2
;
3
3
=
4
4
=
v
(x
k
) 4s
2
],
(41)
is not zero, the solutions of Einstein eqs can be constructed following
Theorem 3.1 The EYHM equations of type (11) (14) with source (41)
can be integrated in general forms by metrics
g =
i
e
(x
k
)
dx
i
dx
i
+h
3
(x
k
, y
3
)e
3
e
3
+h
4
(x
k
, y
3
)h
4
(x
k
, y
4
)e
4
e
4
,
e
3
= dy
3
+w
i
(x
k
, y
3
)dx
i
, e
4
= dy
4
+n
i
(x
k
)dx
i
, (42)
with coecients determined by generating functions (x
k
), (x
k
, y
3
),
=
0, n
i
(x
k
) and h
4
(x
k
, y
4
), and integration functions
0
(x
k
) following recur-
rent formulas and conditions
+
2
= 2 [
v
4s
2
]; (43)
h
4
=
1
4
4s
2
e
2
dy
3
, or (44)
=
1
4( 4s
2
)
e
2[
0
]
, if = = const = 4s
2
;
h
3
=
|h
4
|
2
e
2
=
2 |4s
2
|
(ln|h
4
|)
or
=
(
)
2
4( 4s
2
)
, if = = const = 4s
2
; (45)
w
i
=
i
/
, (46)
where constraints
w
i
= (
i
w
i
) ln |h
4
|, (
k
w
k
)w
i
= (
i
w
i
)w
k
, (47)
n
k
h
4
=
k
h
4
,
i
n
k
=
k
n
i
.
15
must be imposed in order to satisfy the zero torsion conditions (15); we
should take respective values
i
= 1 and in (44) and/or (45) if we wont
to x a necessary spacetime signature.
Proof. We sketch a proof which transforms into similar ones in [10, 11, 9]
if h
4
= 1.
A horizontal metric g
i
(x
2
) is for 2d and can be always represented
in a conformally at form
i
e
(x
k
)
dx
i
dx
i
. For such a hmetric, the
equation (11) is a 2-d Laplace/wave equation (43) which can be solved
exactly if a source
v
(x
k
) 4s
2
is prescribed from certain physical
conditions.
If h
4
= 0, we can dene nontrivial functions
= ln
|h
3
h
4
|
, :=
ln
|h
4
|
3/2
|h
3
|
,
i
= h
i
, = h
(48)
for a function (x
k
, y
3
). If
4
= 2h
3
h
4
4s
2
. (49)
w
i
+
i
= 0, (50)
n
i
+n
i
= 0.
For the last equation, we must take any trivial solution given by func-
tions n
i
(x
k
) satisfying the conditions
i
n
j
=
j
n
i
in order to solve
the constraints (15). Using coecients (48) with
i
= 0 and = 0,
we can always express w
i
via derivatives of , i.e. in the form (46).
We can chose any (x
k
, y
3
) with
. (51)
A possible dependence on y
4
is present in function h
4
which must
satisfy conditions of type (B.6) in order to be compatible with (15). It
16
is not possible to write in explicit form the solutions for the zero torsion
condition if the source
i
= (
i
w
i
) ln |h
4
| and
i
w
j
=
j
w
i
, (52)
transformed into
(
i
)
= 0. By straightforward computa-
tion, we can check that (15) are satised by (47) when n
i
= 0 and w
i
is determined by (46).
The solutions constructed in Theorem 3.1, and those which can be de-
rived following Corollary 2.1 are very general ones and contain as particular
cases (pehaps) all known exact solutions for (non) holonomic Einstein spaces
with Killing symmetries. They also can be generalized to include arbitrary
nite sets of parameters as it is proven in Ref. [9].
3.1.2 Eective vacuum EYMH congurations
We can consider a subclass of generic odiagonal EYMH interactions
which can be encoded as eective Einstein manifolds with nontrivial cos-
mological constant = 4s
2
. In general, such classes of solutions depend
parametrically on 4s
2
and do not have a smooth limit from non-vacuum
to vacuum models.
Corollary 3.1 The eective vacuum solutions for the EYHM equations
with ansatz for metrics of type (42) with vanishing source (41) are paramet-
rized in the form
g =
i
e
(x
k
)
dx
i
dx
i
+h
3
(x
k
, y
3
)e
3
e
3
+h
4
(x
k
, y
3
)h
4
(x
k
, y
4
)e
4
e
4
,
e
3
= dy
3
+w
i
(x
k
, y
3
)dx
i
, e
4
= dy
4
+n
i
(x
k
)dx
i
, (53)
where coecients are dened by solutions of the system
= 0, (54)
4
= 0, (55)
w
i
+
i
= 0, (56)
where coecients are computed following formulas (48) for nonzero
and
h
4
with possible further zero limit; such coecients and h
3
and h
4
are sub-
jected additionally to the zerotorsion conditions (47).
17
Proof. Considering the system of equations (11) (14) with zero right
sides, we obtain respectively the equations (54) (56). For positive signa-
tures on hsubspace and equation (54), we can take = 0, or consider a
trivial 2-d wave equation if one of coordinates x
k
is timelike. There are two
possibilities to satisfy the condition (55). The rst one is to consider that
h
4
= h
4
(x
k
), i.e. h
4
which states that the equation (55) has solutions with
zero source for arbitrary function h
3
(x
k
, y
3
) and arbitrary Ncoecients
w
i
(x
k
, y
3
) as follows from (48). For such vacuum congurations, the func-
tions h
3
and w
i
can be taken as generation ones which should be constrained
only by the conditions (47). Equations of type (52) constrain substantially
the class of admissible w
i
if h
4
depends only on x
k
. Nevertheless, h
3
can be
an arbitrary one generating solutions which can be extended for nontrivial
sources and systems (16) (20) and/or (24) (29).
A dierent class of solutions can be generated if we state, after corre-
sponding coordinate transforms, = ln
4
/
|h
3
h
4
|
=
0
= const,
=
0. For such congurations, we can consider h
4
= 0, and solve (55) as
|h
3
| =
0
h(
|h
4
|)
, (57)
for
0
h = const = 0. Such vmetrics are generated by any f(x
i
, y
3
), f
= 0,
when
h
4
= f
2
x
i
, y
3
and h
3
= (
0
h)
2
f
x
i
, y
3
2
, (58)
where the sines are such way xed that for N
a
i
0 we obtain diagonal
metrics with signature (+, +, +, ). The coecients
i
= = 0 in (56) and
w
i
(x
k
, y
3
) can be any functions solving (47). This is equivalent to
w
i
= 2
i
ln |f| 2w
i
(ln |f|)
, (59)
k
w
i
i
w
k
= 2(w
k
i
w
i
k
) ln |f|,
for any n
i
(x
k
) when
i
n
k
=
k
n
i
. Constraints of type n
k
h
4
=
k
h
4
(B.6)
have to be imposed for a nontrivial multiple h
4
.
Using Corollary 2.1, the ansatz (53) can be dualized to generate ef-
fective vacuum solutions with weak Killing symmetry on /y
3
. Finally, we
note that the signature of the generic odiagonal metrics generated in this
subsection depend on the fact which coordinate x
1
, x
2
, y
3
or y
4
is chosen to
be a timelike one.
3.2 NonKilling EYMH congurations
The Theorem 2.2 can be applied for constructing nonvacuum and eec-
tive vacuum solutions of the EYMH equations depending on all coordinates
18
without explicit Killing symmetries.
3.2.1 Nonvacuum odiagonal solutions
We can generate such YMH Einstein manifolds following
Corollary 3.2 An ansatz of type (22) with dmetric
g =
i
e
dx
i
dx
i
+
2
[
(
)
2
4( 4s
2
)
e
3
e
3
1
4( 4s
2
)
e
2[
0
]
h
4
e
4
e
4
],
e
3
= dy
3
(
i
/
)dx
i
, e
4
= dy
4
+n
i
(x
k
)dx
i
,
where the coecients are subjected to conditions (43)(47) and
k
+
(
i
/
n
i
=
( 4s
2
)g
+(
i
/
= 0 with respective
dual generating functions and when the data (43)(47) are redened
for solutions with weak Killing symmetry on /y
3
.
3.2.2 Eective vacuum odiagonal solutions
Vacuum Einstein spaces encoding nonholonomic interactions of EYMH
elds can be constructed using
Corollary 3.3 An ansatz of type (22) with dmetric
g =
i
e
(x
k
)
dx
i
dx
i
+
2
(x
k
, y
a
)[(
0
h)
2
f
x
i
, y
3
2
e
3
e
3
f
2
x
i
, y
3
h
4
(x
k
, y
4
)e
4
e
4
],
e
3
= dy
3
+w
i
(x
k
, y
3
)dx
i
, e
4
= dy
4
+n
i
(x
k
)dx
i
,
where the coecients are subjected to conditions (57)(59), (47) and
k
w
i
n
i
= 0.
Proof. It is a consequence of Theorem 2.2 and Corollary 3.1.
Solutions of type (23) can be dened if the conformal factor is a solution
of
k
w
i
(x
k
)
+ n
i
(x
k
, y
4
)
= ( 4s
2
)g
, (60)
which can be found in very general forms with respect to Nadapted frames
for certain nonintegrable spacetime 2+2 splitting of type N(2). An eective
cosmological constant ( 4s
2
) encodes a gravitational vacuum cosmo-
logical constant in GR and s
2
is induced by of nonholonomic dynamics
of YMH elds. PseudoRiemanian manifolds with metrics g
adapted to
chosen nonintegrable distribution with 2 + 2 splitting and satisfying (60)
are called nonholonomic Einstein manifolds. Hereafter we shall refer to such
8
such formulas are, for instance, of type (43)(47), with functions h3(., y
3
) and h4(., y
3
)
with further dualization h3 h
4
(., y
4
) and h4 h
3
(., y
4
) which corresponding re
denition of Nconnection coecients
20
systems of PDEs as nonholnomic vacuum spacetimes, regardless of wether
or not an (eective) cosmological constant vanishes or can be polarized by
gravitational and/or matter eld interactions to some Nadapted diagonal
sources admitting formal integration of gravitational eld equations.
The equations (60), and their Nadapted equivalents (A.12)(A.13), con-
stitute a secondorder system quasilinear PDE for the coecients of space-
time metric g = {g
(the coecients of
such PDE are rational functions of g
= ( x
i
, y
a
)} is canonically Nhar-
monic, i.e. it both harmonic and adapted to a splitting N (2), if each of the
functions u
= 0, (61)
where the canonical dAlambert operator
:=
D
acts on a scalar
21
f(x, y) in the form
f := (
|g
|)
1
e
|g
|g
= (
|g
kl
|)
1
e
i
|g
kl
|g
ij
e
j
f
+ (
|g
cd
|)
1
|g
kl
|g
ab
e
b
f
,
for a dmetric g
= (g
ij
, g
ab
) (7) dened with respect to Nadapted (co)
frames (3)(4) and canonical dconnection
D
(A.3).
We can say that four such coordinates u
= ( x
i
, y
a
) are Nadapted
wavecoordinates.
Lemma 4.1 In canonical Nharmonic coordinates, the Einstein equations
(60) redened in canonical dconnection variables (A.12) can be written
equivalently
=
g
(g
+g
+ 2g
2( 4s
2
)g
= 0, (62)
i.e. such PDE for g
) form a system
of secondorder quasilinear Nadapted wavetype equations.
Proof. It is a standard computation with respect to Nadapted frames
by using formulas (A.3), (A.9) and (A.7). If the zero torsion conditions
(A.13) are imposed, we get well known results from GR but (in our case)
adapted to 2 + 2 nonintegrable splitting.
This Lemma allows us to apply the standard theory of hyperbolic PDE
(see, for instance, [20]). Let us denote by H
k
loc
the Sobolev spaces of func-
tions which are in L
2
(K) for any compact set K when their distributional
derivatives are considered up to an integer order k also in L
2
(K). We shall
also use Nadapted wave coordinates with additional formal 3 +1 splitting,
for instance, in a form u
= (
t
u, u
i
), where
t
u is used for the timelike
coordinate and u
i
are for 3 spacelike coordinates. Hats can be eliminated
if such a splitting is considered for arbitrary local coordinates. Standard
results from the theory of PDE give rise to this
Theorem 4.1 The eld equations (A.12) for nonholonomic Einstein man-
ifolds have a unique solution g
(0, u
i
) H
k+1
loc
and
g
(0, u
i
)
(
t
u)
H
k+1
loc
, k > 3/2. (63)
22
The set U can be chosen in a form that
U, g
= (g
t
, g
i
) and e
= (e
t
, e
i
) and consider the dvector eld
n
|g
|g
ti
= e
i
|g
|g
ti
(65)
with Nelongated operators. So, the initial data from Theorem 4.1 can not
be xed in arbitrary form if we wont to establish a hyperbolic (evolutionary)
character for nonholonomic vacuum Einstein equations, i.e. to satisfy both
systems (62) and (61). Really, the last system of st order PDE allows us to
compute the timelike derivatives g
ti
(0, u
i
)/
t
u |
{
t
u=0}
if g
ij
|
{
t
u=0}
and
g
ij
/
t
u |
{
t
u=0}
have been dened. We conclude that the essential data for
formulating a Nadapted evolution problem should be formulated for a 3d
space dmetric
[3]
g := g
ij
(x, y)e
i
e
j
, (66)
9
Choosing corresponding classes of nonholonomic distributions N (2), we can relax the
conditions of dierentiability as in Refs. [21, 22] (we omit such constructions in this work).
10
We do not use labels for coordinates like 0, 1, 2, 3 because the decoupling property
of the Einstein equations and general solutions can be proven for arbitrary signature, for
instance (+ ++) , (+ ++) etc and for any set of local coordinates u
with = 1, 2, 3, 4.
23
where e
j
are Nelongated dierentials of type (4), and its Nelongated (3)
timelike derivatives.
Using the Theorem 4.1 and above presented considerations, we get the
proof of
Theorem 4.2 If the initial data (63) satisfy the conditions (64) and (65)
and the socalled Einstein Nadapted constraint equations,
+ g
= 0, (67)
are computed for zero distortion in (A.1), then the dmetric stated by The-
orem 4.1 satised the nonholonomic vacuum equations (60) and/or (A.12)
(A.13).
This theorem gives us the possibility to state the Cauchy data for de-
coupled EYMH systems and their generic odiagonal solutions in order to
generate Nadapted evolutions.
4.2 On initial data sets and global nonholonomic evolution
We adopt this convention for spacetime nonholonomic manifolds V =
R
3
V, were
3
V is a Nadapted 3-d manifold, when a Whitney sum
T
3
V = h
3
V v
3
V is stated by a space like nonholnomic distribution
3
N
of type (2) and there is an embedding e :
3
V V.
Using the dmetric
[3]
g (66), we can dene the second fundamental form
K of a spacelike hypersurface
3
V in V,
K(X, Y) := g(
D
X
n, Y), X T
3
V .
The unity dvector n = n
= n
i
e
i
= n
t
e
t
=
|g
tt
|
1
e
t
, with nor-
malization g(n, n) = g
= g
tt
(n
t
)
2
= 1, is timefuture and normal
to
3
V. The value
K = {
K
ij
} is the extrinsic canonical curvature dtensor
of
3
V. Imposing the zero torsion conditions (A.13),
K {K
ij
}, where
K
ij
=
1
2
g
t
e
j
g
i
+e
i
g
j
e
g
ij
n
t
are components computed in stan-
dard form using the LeviCivita connection, but with respect to Nadapted
frames. We can invert the last formula and write
t
g
ij
= 2(g
tt
n
t
)
1
K
ij
+
{terms determined by g
D
t
,
D
i
and dmetric
[3]
g in-
duced on a spacelike hypersurface in a Lorentzian nonholonomic manifold
V, we can derive a Nadapted variant of GaussCodazzi equations
[3]
R
i
jkl
=
R
i
jkl
+
K
i
l
K
jk
K
i
k
K
jl
,
D
i
K
jk
D
j
K
ik
=
R
i
jkl
n
l
.
24
In these formulas,
[3]
R
i
jkl
is the canonical curvature dtensor of
[3]
g,
R
i
jkl
is computed following formulas (A.6) as the spacetime canonical dcurvature
tensor, n is the timelike normal to hypersurface
3
V when the local N
adapted coordinate system is such way chosen that dvectors e
j
are tangent
to
3
V. Contracting indices, introducing divergence operator
div determined
by Nelongated partial derivatives (3), trace operator tr and absolute dier-
ential d, we derive from above equations and (A.12) this system of equations:
div
K d(tr
K) = 8
J, momentum constraint;(68)
s
[3]
R 2
[3]
g
lj
K
i
l
K
ji
+ (tr
K)
2
= 16 , Hamiltonian constraint;
C(
F,
[3]
g) = 0, Einstein constraint eqs;
where
s
[3]
3
V,
[3]
g,
K,
F
, we have
J := (n, ) and := (n, n) . The
term C(
F,
[3]
g) denotes the set of additional constraints resulting from the
nongravitational part, including nonholonomic distributions N (2). If in
such a set there are included the zero torsion conditions (A.13), we can
omit hats on geometric/physical objects if they are written in not N
adapted frames of reference. The equations (68) form an undetermined
system of PDE. For 3-d, there are locally four equations for twelve unknown
values give by the components of dtensors
[3]
g and
K. Using the conformal
method with the LeviCivita connection, see details and references in Ref.
[23], we can study the existence and uniqueness of solutions to such systems.
Above considerations motivate
Denition 4.2 A canonical vacuum initial Nadapted data set is dened
by a triple
3
V,
[3]
g,
, where (
[3]
g,
K) are equivalent
to similar ones (
[3]
g, K) with K computed for the LeviCivita connection.
Covering
3
V by coordinate neighborhoods O
u
of points u
3
V, we can
use Theorem 4.1 to construct globally hyperbolic Nadapted developments
(U
u
, g
u
) of an initial data set
U
u
3
V,
[3]
g,
3
V,
[3]
g,
of dierentiability
class H
s+1
H
s
, s > 3/2, admits a globally hyperbolic, Nadapted and unique
development (in the sense of Theorem 4.1 and above considered assumptions
and proofs).
For the EYMH systems dened by assumptions in Condition 2.1, we
encode the data on such gravitational and gaugescalar interactions into
the nonholonomic structure of certain eective Einstein manifolds.
5 Anholonomic YMHDeformations of Black Holes
GaugeHiggs nonholonomic interactions can be parametrized in such
forms that they dene odiagonal deformations (for instance, of rotoid
type) of Schwarzschild black holes. In this section, we study such EYMH
congurations when +
s
= 0. Nonholonomic deformations can be de-
rived from any prime data
g,
A
=
A
+
A
(37), F
= s
|g|
(38)
and =
subjected to the conditions (39) are encoded as vacuum
odiagonal polarizations into solutions of equations (24)(25).
5.1 (Non) holonomic nonAbelian eective vacuum spaces
This class of eective vacuum solutions are generated not just as a simple
limit +
s
0, for instance, for a class of solutions (42) with coecients
(43)(45). We have to construct odiagonal solutions of the Einstein equa-
tions for the canonical dconnection taking the vacuum equations
R
= 0
and ansatz g (36) with coecients satisfying the conditions
(r, ) +
2
(r, ) = 0; (69)
h
3
= e
2
0
(h
4
)
2
h
4
for a given h
4
(r, , ), (r, , ) =
0
= const;
w
i
= w
i
(r, , ), for any such functions if = 0;
n
i
=
1
n
i
(r, ) +
2
n
i
(r, )
(h
4
)
2
|h
4
|
5/2
dv, if n
i
= 0;
1
n
i
(r, ), if n
i
= 0.
Eective vacuum solutions of the Einstein equations for the LeviCivita
connection, i.e of R
= 1, as solutions of (47),
h
3
= 4
|h
4
|
2
, h
4
= 0; (70)
w
1
w
2
ln |
w
1
w
2
|
= w
2
w
1
, w
i
= 0; w
2
w
1
= 0, w
i
= 0;
1
n
1
(r, )
1
n
2
(r, ) = 0, n
i
= 0. (71)
The constructed class of vacuum solutions with coecients subjected to
conditions (69)(71) is of type (53) for (54)(56). Such metrics consist a
particular case of vacuum ansatz dened by Corollary 3.3 with h
4
= 1 and
= 1.
Here we note that former analytic and numeric programs (for instance,
standard ones with Maple/ Mathematica) for constructing solutions in grav-
ity theories can not be directly applied for alternative verications of our
solutions. Those approaches do not encode the nonholonomic constraints
which we use for constructing integral varieties. Nevertheless, it is possi-
ble to check in general analytic form, see all details summarized in Refs.
[9, 10, 11] (and formulas from Appendix B), that the vacuum Einstein equa-
tions (24)(15) with zero eective sources,
v
4s
2
= 0 and 4s
2
= 0,
can be solved by above presented odiagonal ansatz for metrics.
5.2 NonAbelian deformations of the Schwarzschild metric
We can consider a prime metric which, in general, is not a solution of
Einstein equations,
g = d d r
2
() ddr
2
() sin
2
dd+
2
() dt dt. (72)
We shall deform nonholonomically this metric into a target odiagonal
one which will be a solution of the vacuum Einstein equations. The non-
trivial metric coecients in (72) are stated in the form
g
1
= 1, g
2
= r
2
(),
h
3
= r
2
() sin
2
,
h
4
=
2
(), (73)
for local coordinates x
1
= , x
2
= , y
3
= , y
4
= t, where
=
dr
1
2
0
r
+
r
2
1/2
and
2
(r) = 1
2
0
r
+
r
2
.
If we put = 0 with
0
considered as a point mass, the metric
g (72)
determines the Schwarzschild solution. For simplicity, we analyze only
27
the case of pure gravitational vacuum solutions, not considering a more
general construction when = e
2
can be related to the electric charge for
the ReissnerNordstrom metric. In our approach, is a small parameter
(eccentricity) dening a small deformation of a circle into an ellipse.
We generate an exact solution of the system (54)(56) with eective
+
s
= 0 via nonholonomic deformations
g
g, when g
i
=
i
g
i
and
h
a
=
a
h
a
and w
i
, n
i
dene a target metric
g =
1
()d d +
2
()r
2
() d d + (74)
3
(, , )r
2
() sin
2
4
(, , )
2
() t t,
= d +w
1
(, , )d +w
2
(, , )d, t = dt +n
1
(, )d +n
2
(, )d.
The gravitational eld equations for zero source relate the coecients of
the vertical metric and polarization functions,
h
3
= h
2
0
(b
)
2
=
3
(, , )r
2
() sin
2
, h
4
= b
2
=
4
(, , )
2
(), (75)
for |
3
| = (h
0
)
2
|
h
4
/
h
3
|[(
|
4
|)
]
2
. In these formulas, we have to chose h
0
=
const ( h
0
= 2 in order to satisfy the rst condition (71)), where
h
a
are stated
by the Schwarzschild solution for the chosen system of coordinates and
4
can be any function satisfying the condition
4
= 0. We generate a class of
solutions for any function b(, , ) with b
4
= 0, and/or h
4
, for h
4
= 0.
The gravitational polarizations
1
and
2
, when
1
=
2
r
2
= e
(,)
, are
found from (24) with zero source, written in the form
= 0.
Introducing the dened values of the coecients in the ansatz (74), we
nd a class of exact odiagonal vacuum solutions of the Einstein equations
dening stationary nonholonomic deformations of the Schwarzschild metric,
g = e
(d d + d d) 4
|
4
|)
2
+
4
2
t t,
= d +w
1
d +w
2
d, t = dt +
1
n
1
d +
1
n
2
d. (76)
The Nconnection coecients w
i
(, , ) and
1
n
i
(, ) must satisfy the
conditions (71) in order to get vacuum metrics in GR.
It should be emphasized here that, in general, the solutions from the
set of target metrics do not dene black holes and do not describe obvious
physical situations. They preserve the singular character of the coecient
2
vanishing on the horizon of a Schwarzschild black hole if we take only
smooth integration functions for some small deformation parameters .
28
5.3 Linear parametric polarizations induced by YMH elds
We may select some locally anisotropic congurations with possible phys-
ical interpretation of gravitational vacuum congurations with spherical
and/or rotoid (ellipsoid) symmetry if it is considered a generating function
b
2
= q(, , ) +(, , ). (77)
For simplicity, we restrict our analysis only with linear decompositions on a
small parameter , with 0 < << 1.
Using (77), we compute (b
)
2
= [(
|q|)
]
2
[1 +
1
(
|q|)
(/
|q|)
] and
the vertical coecients of dmetric (76), i.e h
3
and h
4
(and corresponding
polarizations
3
and
4
), see formulas (75).
11
We model rotoid congurations
if we chose
q = 1
2(, , )
r
and =
q
0
(r)
4
2
sin(
0
+
0
), (78)
for (, , ) =
0
+
1
(, , ) (supposing that the mass is locally anisotrop-
ically polarized) with certain constants ,
0
and
0
and arbitrary func-
tions/polarizations
1
(, , ) and q
0
(r) to be determined from some bound-
ary conditions, with being the eccentricity.
12
This condition denes a
small deformation of the Schwarzschild spherical horizon into an ellipsoidal
one (rotoid conguration with eccentricity ).
The resulting odiagonal solution with rotoid type symmetry is
rot
g = e
(d d + d d) + (q +) t t
4
|q|)
2
[1 +
1
(
|q|)
(/
|q|)
] , (79)
= d +w
1
d +w
2
d, t = dt +
1
n
1
d +
1
n
2
d.
The functions q(, , ) and (, , ) are given by formulas (78) and the
Nconnection coecients w
i
(, , ) and n
i
=
1
n
i
(, ) are subjected to
conditions of type (71),
w
1
w
2
ln |
w
1
w
2
|
= w
2
w
1
, w
i
= 0; (80)
or w
2
w
1
= 0, w
i
= 0;
1
n
1
(, )
1
n
2
(, ) = 0
and (, ) being any function for which
= 0.
11
Nonholonomic deformations of the Schwarzschild solution (not depending on ) can
be generated if we consider = 0 and b
2
= q and (b
)
2
=
|q|)
2
.
12
We may treat as an eccentricity imposing the condition that the coecient h4 =
b
2
= 4(, , )
2
() becomes zero for data (78) if r+ 20/
1 +
q
0
(r)
4
2
sin(0 + 0)
.
29
6 Ellipsoidal EYMH Congurations and Solitons
We can consider nonholonomic deformations for the EYMH systems for
arbitrary signs of the cosmological constant and an eective nontrivial
source +
s
= 0 containing contributions of nonholonomic YMH cong-
urations. Such classes of solutions can be constructed in general form for
a system (24)(27) and (47) with coecients of metric of type (42). Such
metrics consist a particular case of nonvacuum ansatz dened by Corollary
3.2 with h
4
= 1 and = 1.
6.1 Nonholonomic rotoid deformations
Using the anholonomic frame method, we can generate a class of solu-
tions with nontrivial cosmological constant possessing dierent limits (for
large radial distances and small nonholonomic deformations) than the vac-
uum congurations considered in previous section.
Let us consider a diagonal metric of type
g = d d +r
2
() d d +r
2
() sin
2
dd+
2
() dt dt, (81)
where nontrivial metric coecients are parametrized in the form g
1
= 1, g
2
=
r
2
(),
h
3
= r
2
() sin
2
,
h
4
=
2
(), for local coordinates x
1
= , x
2
=
, y
3
= , y
4
= t, with =
dr/ |q(r)|
1
2
, and
2
(r) =
2
(r)q(r), for
q(r) = 1 2m(r)/r r
2
/3. In variables (r, , ) , the metric (81) is equiv-
alent to (B.7).
The ansatz for such classes of solutions is chosen in the form
g = e
(,)
(d d + d d) +h
3
(, , ) +h
4
(, , ) t t,
= d +w
1
(, , ) d +w
2
(, , ) d,
t = dt +n
1
(, , ) d +n
2
(, , ) d,
for h
3
= h
2
0
(b
)
2
=
3
(, , )r
2
() sin
2
, h
4
= b
2
=
4
(, , )
2
().
The coecients of this metric determine exact solutions if
(, ) +
(, ) = 2( +
s
); (82)
h
3
=
(
)
2
4 ( +
s
)
e
2
0
(,)
, h
4
=
1
4 ( +
s
)
e
2(
0
(,))
;
w
i
=
i
/
;
n
i
=
1
n
i
(, ) +
2
n
i
(, )
)
2
e
2(
0
(,))
d,
=
1
n
i
(, ) +
2
n
i
(, )
e
4
(h
4
)
2
h
4
d, if n
i
= 0;
1
n
i
(, ), if n
i
= 0;
30
for any nonzero h
a
and h
a
and (integrating) functions
1
n
i
(, ),
2
n
i
(, ),
generating function (, , ), and
0
(, ) to be determined from certain
boundary conditions for a xed system of coordinates.
For nonholonomic ellipsoid de Sitter congurations, we parametrize
rot
g = e
(,)
(d d + d d) +
q +
t t
h
2
0
|q|)
2
[1 +
1
(
|q|)
(/
|q|)
] ,
= d +w
1
d +w
2
d, t = dt +n
1
d +n
2
d, (83)
where q = 1
2
1
(r,,)
r
, =
q
0
(r)
4
2
0
sin(
0
+
0
), are chosen to generate
an anisotropic rotoid conguration for the smaller horizon (when h
4
=
0), r
+
2
1
/
1 +
q
0
(r)
4
2
0
sin(
0
+
0
)
, for a corresponding q
0
(r).
We have to impose the condition that the coecients of the above d
metric induce a zero torsion in order to generate solutions of the Einstein
equations for the LeviCivita connection. Using formula (80), for
= 0,
we obtain that (r, , ) = ln |h
4
/
|h
3
h
4
|| must be any function dened in
nonexplicit form from equation 2e
2
= +
s
. The set of constraints for
the Nconnection coecients is solved if the integration functions in (82) are
chosen in a form when w
1
w
2
ln |
w
1
w
2
|
= w
2
w
1
for w
i
= 0; w
2
w
1
= 0
for w
i
= 0; and take n
i
=
1
n
i
(x
k
) for
1
n
1
(x
k
)
1
n
2
(x
k
) = 0.
In a particular case, in the limit 0, we get a subclass of solutions of
type (83) with spherical symmetry but with generic odiagonal coecients
induced by the Nconnection coecients. This class of spacetimes depend
on cosmological constants polarized nonholonomically by YMH elds. We
can extract from such congurations the Schwarzschild solution if we select
a set of functions with the properties const, w
i
0, n
i
0 and
h
4
2
. In general, the parametric dependence on cosmological constants
and for eective YMH contributions, in nonholonomic congurations, is not
smooth.
6.2 Eective vacuum solitonic congurations
It is possible to construct odiagonal vacuum spacetimes generating by
3d solitons as examples of generic odiagonal solutions with nontrivial
vertical conformal factor . We consider that there are satised the con-
ditons of Corollary 3.3 with h
4
= 1 for eective vacuum solutions (such
congurations may encode EYMH elds) and the Cauchy problem is stated
as in Section 3.
31
6.2.1 Solutions with solitonic factor (x
1
, y
3
, t)
We take = (x
1
, y
3
, t), when y
4
= t is a time like coordinate, as a
solution of KdP equation [24]
+ (
t
+
= 0, (84)
with dispersion and possible dependencies on a set of parameters . It is
supposed that in the dispersionless limit 0 the solutions are independent
on y
3
and determined by Burgers equation
t
+
= 0. For such 3d
solitonic congurations, the conditions (47) are written in the form
e
1
=
+w
1
(x
i
, y
3
)
+n
1
(x
i
)
t
= 0.
For
(x
1
, y
3
, t)
2
h
a
(x
1
, y
3
) e
a
e
a
,
e
3
= dy
3
+w
1
(x
k
, y
3
)dx
1
, e
4
= dy
4
+n
1
(x
k
)dx
1
.
This class of metrics does not have (in general) Killing symmetries but may
possess symmetries determined by solitonic solutions of (84). Alternatively,
we can consider that is a solution of any three dimensional solitonic and/
or other nonlinear wave equations; in a similar manner, we can generate
solutions for = (x
2
, y
3
, t).
6.2.2 Solitonic metrics with factor (x
i
, t)
There are eective vacuum metrics when the solitonic dynamics does not
depend on anisotropic coordinate y
3
. In this case = (x
k
, t) is a solution
of KdP equation
+ (
t
+
= 0. (85)
In the dispersionless limit 0 the solutions are independent on x
1
and
determined by Burgers equation
t
+
= 0.
This class of vacuum solitonic EYMH congurations is given by
2
g = e
(x
k
)
(dx
1
dx
1
+dx
2
dx
2
) +
(x
k
, t)
2
h
a
(x
k
, y
3
) e
a
e
a
,
e
3
= dy
3
+w
1
(x
k
, y
3
)dx
1
, e
4
= dy
4
+n
1
(x
k
)dx
1
;
the conditions (47) are e
1
=
+n
1
(x
i
)
t
= 0, e
2
=
+n
2
(x
i
)
t
= 0.
32
It is possible to derive an innite number of vacuum gravitational 2-d and
3-d congurations characterized by corresponding solitonic hierarchies and
biHamilton structures, for instance, related to dierent KdP equations (85)
with possible mixtures with solutions for 2-d and 3-d sineGordon equations
etc, see details in Ref. [25].
A Nonholonomic 2+2 Splitting of Lorentz Mani-
folds
In a general case, a metricane manifold V is endowed with a metric
structure g and an ane (linear) connection structure D (as a covariant
dierentiation operator). A linear connection gives us with the possibility
to compute the directional derivative D
X
Y of a vector eld Y in the direction
of X. It is characterized by three fundamental geometric objects:
1. the torsion eld is (by denition) T (X, Y ) := D
X
Y D
Y
X [X, Y ];
2. the curvature eld is R(X, Y ) := D
X
D
Y
D
Y
D
X
D
[X,Y]
;
3. the nonmetricity eld is Q(X) := D
X
g.
Introducing X = e
and Y = e
} and
corresponding fundamental geometric objects,
T = {T
T
i
jk
, T
i
ja
, T
a
ji
, T
a
bi
, T
a
bc
};
R = {R
R
i
hjk
,R
a
bjk
,R
i
hja
,R
c
bja
,R
i
hba
, R
c
bea
}; Q = {Q
}.
Every (pseudo) Riemannian manifold (V, g) is naturally equipped with a
LeviCivita connection D = = {
} completely dened by g = {g
}
if and only if there are satised the metric compatibility,
Q(X) =
X
g =
0, and zero torsion,
T = 0, conditions. Hereafter, we shall write, for
simplicity,
}
for a prescribed N = {N
a
i
}, see details in [9, 10, 11].
Theorem A.1 With respect to Nadapted frames (3) and (4), the coe-
cient of distortion relation (A.1) are computed
+
Z
, (A.2)
where the canonical dconnection
D = {
= (
L
i
jk
,
L
a
bk
,
C
i
jc
,
C
a
bc
)} is
dened by coecients
L
i
jk
=
1
2
g
ir
(e
k
g
jr
+e
j
g
kr
e
r
g
jk
) , (A.3)
L
a
bk
= e
b
(N
a
k
) +
1
2
g
ac
e
k
g
bc
g
dc
e
b
N
d
k
g
db
e
c
N
d
k
C
i
jc
=
1
2
g
ik
e
c
g
jk
,
C
a
bc
=
1
2
g
ad
(e
c
g
bd
+e
b
g
cd
e
d
g
bc
) ,
and the distortion tensor
Z
is
Z
a
jk
=
C
i
jb
g
ik
g
ab
1
2
a
jk
, Z
i
bk
=
1
2
c
jk
g
cb
g
ji
ih
jk
C
j
hb
,
Z
a
bk
=
+
ab
cd
T
c
kb
, Z
i
kb
=
1
2
a
jk
g
cb
g
ji
+
ih
jk
C
j
hb
, Z
i
jk
= 0, (A.4)
Z
a
jb
=
ad
cb
T
c
jd
, Z
a
bc
= 0, Z
i
ab
=
g
ij
2
T
c
ja
g
cb
+
T
c
jb
g
ca
,
for
ih
jk
=
1
2
(
i
j
h
k
g
jk
g
ih
) and
ab
cd
=
1
2
(
a
c
b
d
+ g
cd
g
ab
). The nontrivial
coecients
a
jk
and
T
g, N,
} of
D is
determined in a unique form by the metric compatibility condition,
Dg = 0,
34
and zero horizontal and vertical torsion coecients,
T
i
jk
= 0 and
T
a
bc
= 0,
but with nontrivial hv coecients
T
i
jk
=
L
i
jk
L
i
kj
,
T
i
ja
=
C
i
jb
,
T
a
ji
=
a
ji
,
T
c
aj
=
L
c
aj
e
a
(N
c
j
),
T
a
bc
=
C
a
bc
C
a
cb
.
(A.5)
Proof. The coecients (A.5) are computed by introducing D =
D,
with coecients (A.3), and X = e
, Y = e
, Y = e
, Z = e
into
R(X, Y ) := D
X
D
Y
D
Y
D
X
D
[X,Y]
, we prove
Theorem A.3 The curvature
R = {
} of
D is characterized by N
adapted coecients
R
i
hjk
= e
k
L
i
hj
e
j
L
i
hk
+
L
m
hj
L
i
mk
L
m
hk
L
i
mj
C
i
ha
a
kj
,
R
a
bjk
= e
k
L
a
bj
e
j
L
a
bk
+
L
c
bj
L
a
ck
L
c
bk
L
a
cj
C
a
bc
c
kj
, (A.6)
R
i
jka
= e
a
L
i
jk
D
k
C
i
ja
+
C
i
jb
T
b
ka
,
R
c
bka
= e
a
L
c
bk
D
k
C
c
ba
+
C
c
bd
T
c
ka
,
R
i
jbc
= e
c
C
i
jb
e
b
C
i
jc
+
C
h
jb
C
i
hc
C
h
jc
C
i
hb
,
R
a
bcd
= e
d
C
a
bc
e
c
C
a
bd
+
C
e
bc
C
a
ed
C
e
bd
C
a
ec
.
We can redene the dierential geometry of a (pseudo) Riemannian
space V in nonholonomic form in terms of geometric data (g,
D) which is
equivalent to the standard formulation with (g, ).
Corollary A.1 The Ricci tensor
R
:=
R
(A.11) of
D is character-
ized by Nadapted coecients
= {
R
ij
:=
R
k
ijk
,
R
ia
:=
R
k
ika
,
R
ai
:=
R
b
aib
,
R
ab
:=
R
c
abc
}. (A.7)
Proof. The formulas for hvcomponents (A.7) are obtained by con-
tracting respectively the coecients (A.6). Using
D (A.3), we express such
formulas in terms of partial derivatives of coecients of metric g (1) and
any equivalent parametrization in the form (7), or (8).
The scalar curvature
s
R of
D is by denition
s
R := g
= g
ij
R
ij
+g
ab
R
ab
. (A.8)
Using (A.7) and (A.8), we can compute the Einstein tensor
E
of
D,
1
2
g
s
R. (A.9)
35
In general, this tensor is dierent from that constructed using (A.10) for the
LeviCivita connection .
Proposition A.1 The Nadapted coecients
of
D are identic to the
coecients
}, for = {
}, (A.10)
and
Ric = {
R
:=
R
}, for
D = {
}. (A.11)
Theorem A.4 The eld equations in GR can be rewritten equivalently
using the canonical dconnection
D,
1
2
g
s
R =
, (A.12)
L
c
aj
= e
a
(N
c
j
),
C
i
jb
= 0,
a
ji
= 0, (A.13)
where the scalar curvature
s
R := g
R
and source dtensor
is such
way constructed that
for
D , where T
is the energy
momentum tensor in GR with coupling gravitational constant .
36
Proof. The Einstein equations are written in standard form for ,
E
= R
1
2
g
R = T
, (A.14)
where R := g
R
. On spacetimes with conventional h- and vsplitting,
we can dene geometrically (or following a Nadapted variational calculus
with operators (3) and (4)) a system of eld equations with the Einstein
tensor
E
results
in R
=
R
and E
=
E
, Y = e
(see (3))
into standard formula T (X, Y ) := D
X
Y D
Y
X [X, Y ].
We consider matter eld sources in (A.12) which can be diagonalized
with respect to Nadapted frames,
= diag[
:
1
1
=
2
2
= (x
k
, y
3
) + (x
k
, y
4
);
3
3
=
4
4
=
v
(x
k
)].
(A.15)
Such a formal diagonalization can be performed via corresponding frame/
coordinate transforms for very general distributions of matter elds.
An very important result which can be obtained using the anholonomic
deformation method [9, 10, 11], and developed in this work, is that the Ein-
stein equations for
D (A.12) decouple for parametrizations of dmetrics (7).
The corresponding systems of PDE can be integrated in general forms with
one Killing symmetry and for nonKilling congurations. This allows us
to generate exact solutions of standard Einstein equations (A.14) in GR
imposing additional nonholonomic constraints (A.13). It should be empha-
sized here that if we work from the very beginning with the LeviCivita
connection for metrics (8) computed with respect to coordinate, or other
not Nadapted, frames, we can not see a possibility of general decoupling
and formal integration of the gravitational eld equations.
B Proof of Theorem 2.1
For = 1 and h
a
= const, such proofs can be obtained by straightfor-
ward computations [9]. The approach was extended for = 1 and higher
dimensions in [10, 11]. In this section, we sketch a proof for ansatz (9) with
nontrivial h
4
depending on variable y
4
when = 1 in data (10). At the
next step, the formulas will be completed for nontrivial values = 1.
37
If
R
1
1
=
R
2
2
and
R
3
3
=
R
4
4
, the Einstein equations (A.12) for
D and
data (B.1) (see below) can be written for any source (A.15) in the form
E
1
1
=
E
2
2
=
R
3
3
= (x
k
, y
3
) +(x
k
, y
3
, y
4
),
E
3
3
=
E
4
4
=
R
1
1
=
v
(x
k
).
The geometric data for the conditions of Theorem 2.1 are g
i
= g
i
(x
k
) and
g
3
= h
3
(x
k
, y
3
), g
4
= h
4
(x
k
, y
3
)h
4
(x
k
, y
4
), N
3
i
= w
i
(x
k
, y
3
), N
4
i
= n
i
(x
k
, y
3
),
(B.1)
for h
3
= 1 and local coordinates u
= (x
i
, y
a
) = (x
1
, x
2
, y
3
, y
4
). For such
values, we shall compute respectively the coecients of
a
in (6), canonical
dconnection
(A.3), dtorsion
T
:=
R
(A.7)
and resulting
s
R (A.8) and
E
(A.3),
L
i
jk
=
1
2
g
i1
(e
k
g
j1
+e
j
g
k1
e
1
g
jk
) +
1
2
g
i2
(e
k
g
j2
+e
j
g
k2
e
2
g
jk
)
=
1
2
g
i1
(
k
g
j1
+
j
g
k1
1
g
jk
) +
1
2
g
i2
(
k
g
j2
+
j
g
k2
2
g
jk
) ,
i.e.
L
1
jk
=
1
2g
1
k
g
j1
+
j
g
k1
g
jk
,
L
2
jk
=
1
2g
2
k
g
j2
+
j
g
k2
g
jk
.
The hvcomponents
L
a
bk
are computed following formulas
L
3
bk
= e
b
(N
3
k
) +
1
2g
3
[e
k
g
b3
g
3
e
b
N
3
k
g
3b
(N
3
k
)
g
4b
(N
4
k
)
] = e
b
(N
3
k
)
+
1
2g
3
[
k
g
b3
N
3
k
g
b3
N
4
k
g
b3
g
3
e
b
N
3
k
g
3b
(N
3
k
)
g
4b
(N
4
k
)
]
= e
b
(w
k
+w
k
) +
1
2g
3
[
k
g
b3
(w
k
+ w
k
)g
b3
(n
k
+n
k
)g
b3
g
3
e
b
(w
k
+w
k
) g
3b
w
k
g
4b
n
k
],
L
4
bk
= e
b
(N
4
k
) +
1
2g
4
[e
k
g
b4
g
4
e
b
N
4
k
g
3b
(N
3
k
)
g
4b
(N
4
k
)
] = e
b
(N
4
k
)
+
1
2g
4
[
k
g
b4
N
3
k
g
b4
N
4
k
g
b4
g
4
e
b
N
4
k
g
3b
(N
3
k
)
g
4b
(N
4
k
)
]
= e
b
(n
k
+n
k
) +
1
2g
4
[
k
g
b4
(w
k
+w
k
)g
b4
(n
k
+n
k
)g
b4
g
4
e
b
(n
k
+n
k
) g
3b
w
k
g
4b
n
k
].
38
In explicit form, we obtain these nontrivial values
L
3
3k
= w
k
+
1
2g
3
[
k
g
3
w
k
)g
3
n
k
g
3
g
3
w
k
g
3
w
k
]
=
1
2g
3
[
k
g
3
w
k
g
3
] =
k
h
3
2h
3
w
k
h
3
2h
3
,
L
3
4k
=
1
2g
3
[g
4
n
k
] =
h
4
h
4
2h
3
n
k
,
L
4
3k
= n
k
+
1
2g
4
[g
4
n
k
] =
1
2
n
k
,
L
4
4k
=
1
2g
4
[
k
g
4
w
k
g
4
n
k
g
4
] =
k
(h
4
h
4
)
2h
4
h
4
w
k
h
4
2h
4
n
k
h
4
2h
4
.
For the set of hv Ccoecients, we get
C
i
jc
=
1
2
g
ik
g
jk
y
c
= 0. The v
components of Ccoecients are computed following formulas
C
3
bc
=
1
2g
3
(e
c
g
b3
+e
c
g
c3
g
bc
) ,
C
4
bc
=
1
2g
4
(e
c
g
b4
+e
b
g
c4
g
bc
) ,
i. e.
C
3
33
=
g
3
2g
3
=
h
3
2h
3
,
C
3
34
=
g
3
2g
3
= 0,
C
3
44
=
g
4
2g
3
=
h
4
h
4
h
3
,
C
4
33
=
g
3
2g
4
=
0,
C
4
34
=
g
4
2g
4
=
h
4
2h
4
,
C
4
44
=
g
4
2g
4
=
h
4
2h
4
. Putting together the above formulas,
we nd all nontrivial coecients,
L
1
11
=
g
1
2g
1
,
L
1
12
=
g
1
2g
1
,
L
1
22
=
g
2
2g
1
,
L
2
11
=
g
1
2g
2
,
L
2
12
=
g
2
2g
2
,
L
2
22
=
g
2
2g
2
, (B.2)
L
4
4k
=
k
(h
4
h
4
)
2h
4
h
4
w
k
h
4
2h
4
(n
k
+n
k
)
h
4
2h
4
,
L
3
3k
=
k
h
3
2h
3
w
k
h
3
2h
3
,
L
3
4k
=
h
4
h
4
2h
3
n
k
,
L
4
3k
=
1
2
n
k
,
C
3
33
=
h
3
2h
3
,
C
3
44
=
h
4
h
4
h
3
,
C
4
33
=
h
3
h
3
h
4
h
4
,
C
4
34
=
h
4
2h
4
,
C
4
44
=
h
4
2h
4
.
We shall need also the values
C
3
=
C
3
33
+
C
4
34
=
h
3
2h
3
+
h
4
2h
4
,
C
4
=
C
3
43
+
C
4
44
=
h
4
2h
4
. (B.3)
B.2 Coecients for torsion of
D
Using data (B.1) for w
i
= n
i
= 0, the coecients
a
ij
= e
j
(N
a
i
) e
i
(N
a
j
)
(6), are computed
a
ij
=
j
(N
a
i
)
i
(N
a
j
) N
b
i
b
N
a
j
+N
b
j
b
N
a
i
=
j
(N
a
i
)
i
(N
a
j
) N
3
i
(N
a
j
)
N
4
i
(N
a
j
)
+N
3
j
(N
a
i
)
+N
4
j
(N
a
i
)
=
j
(N
a
i
)
i
(N
a
j
) w
i
(N
a
j
)
+w
j
(N
a
i
)
.
39
We obtain such nontrivial values
3
12
=
3
21
=
2
w
1
1
w
2
w
1
w
2
+w
2
w
1
= w
1
w
2
w
1
w
2
+w
2
w
1
;
4
12
=
4
21
=
2
n
1
1
n
2
w
1
n
2
+w
2
n
1
= n
1
n
2
w
1
n
2
+w
2
n
1
. (B.4)
The nontrivial coecients of dtorsion (A.5) are
T
a
ji
=
a
ji
(B.4) and
T
c
aj
=
L
c
aj
e
a
(N
c
j
). We nd zero values for other types of coecients,
T
i
jk
=
L
i
jk
L
i
kj
= 0,
T
i
ja
=
C
i
jb
= 0,
T
a
bc
=
C
a
bc
C
a
cb
= 0.
We have such nontrivial Nadapted coecients of dtorsion:
T
3
3k
=
L
3
3k
e
3
(N
3
k
) =
k
h
3
2h
3
w
k
h
3
2h
3
w
k
,
T
3
4k
=
L
3
4k
e
4
(N
3
k
) =
h
4
h
4
2h
3
n
k
,
T
4
3k
=
L
4
3k
e
3
(N
4
k
) =
1
2
n
k
n
k
=
1
2
n
k
,
T
4
4k
=
L
4
4k
e
4
(N
4
k
) =
k
(h
4
h
4
)
2h
4
h
4
w
k
h
4
2h
4
n
k
h
4
2h
4
,
T
3
12
= w
1
w
2
w
1
w
2
+w
2
w
1
,
T
4
12
= n
1
n
2
w
1
n
2
+w
2
n
1
. (B.5)
If all coecients (B.5) are zero,
.
B.3 Calculation of the Ricci tensor
Let us compute the values
R
ij
=
R
k
ijk
from (A.11) using (A.6),
R
i
hjk
= e
k
L
i
.hj
e
j
L
i
hk
+
L
m
hj
L
i
mk
L
m
hk
L
i
mj
C
i
ha
a
jk
=
k
L
i
.hj
L
i
hk
+
L
m
hj
L
i
mk
L
m
hk
L
i
mj
,
where we put
C
i
ha
= 0 and
e
k
L
i
hj
=
k
L
i
hj
+N
a
k
L
i
hj
=
k
L
i
hj
+w
k
L
i
hj
+n
k
L
i
hj
=
k
L
i
hj
.
Taking derivatives of (B.2), we obtain
L
1
11
= (
g
1
2g
1
)
=
g
1
2g
1
(g
1
)
2
2 (g
1
)
2
,
1
L
1
12
= (
g
1
2g
1
)
=
g
1
2g
1
1
g
1
2 (g
1
)
2
,
L
1
22
= (
g
2
2g
1
)
=
g
2
2g
1
+
g
1
g
2
2 (g
1
)
2
,
1
L
2
11
= (
g
1
2g
2
)
=
g
1
2g
2
+
g
1
g
2
2 (g
2
)
2
,
L
2
12
= (
g
2
2g
2
)
=
g
2
2g
2
(g
2
)
2
2 (g
2
)
2
,
1
L
2
22
= (
g
2
2g
2
)
=
g
2
2g
2
2
g
2
2 (g
2
)
2
,
40
L
1
11
= (
g
1
2g
1
)
=
g
1
2g
1
1
g
1
2 (g
1
)
2
,
2
L
1
12
= (
g
1
2g
1
)
=
g
1
2g
1
(g
1
)
2
2 (g
1
)
2
,
L
1
22
= (
g
2
2g
1
)
=
g
2
2g
1
+
g
2
g
1
2 (g
1
)
2
,
2
L
2
11
= (
g
1
2g
2
)
=
g
1
2g
2
+
g
1
g
1
2 (g
2
)
2
,
L
2
12
= (
g
2
2g
2
)
=
g
2
2g
2
2
g
2
2 (g
2
)
2
,
2
L
2
22
= (
g
2
2g
2
)
=
g
2
2g
2
(g
2
)
2
2 (g
2
)
2
.
For these values, there are only 2 nontrivial components,
R
1
212
=
g
2
2g
1
1
g
2
4 (g
1
)
2
(g
2
)
2
4g
1
g
2
+
g
1
2g
1
1
g
2
4g
1
g
2
(g
1
)
2
4 (g
1
)
2
,
R
2
112
=
g
2
2g
2
+
g
1
g
2
4g
1
g
2
+
(g
2
)
2
4(g
2
)
2
g
1
2g
2
+
g
1
g
2
4(g
2
)
2
+
(g
1
)
2
4g
1
g
2
.
Considering
R
11
=
R
2
112
and
R
22
=
R
1
212
, when g
i
= 1/g
i
, we nd
R
1
1
=
R
2
2
=
1
2g
1
g
2
[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
],
which can be found in equations (25).
The next step is to derive the equations (26). We consider the third
formula in (A.6),
R
c
bka
=
L
c
bk
y
a
(
C
c
ba
x
k
+
L
c
dk
C
d
ba
L
d
bk
C
c
da
L
d
ak
C
c
bd
) +
C
c
bd
T
d
ka
=
L
c
bk
y
a
C
c
ba|k
+
C
c
bd
T
d
ka
.
Contracting indices, we get
R
bk
=
R
a
bka
=
L
a
bk
y
a
C
a
ba|k
+
C
a
bd
T
d
ka
. For
C
b
:=
C
c
ba
, we write
C
b|k
= e
k
C
b
L
d
bk
C
d
=
k
C
b
N
e
k
e
C
b
L
d
bk
C
d
=
k
C
b
w
k
C
b
n
k
C
b
L
d
bk
C
d
.
We split conventionally
R
bk
=
[1]
R
bk
+
[2]
R
bk
+
[3]
R
bk
, where
[1]
R
bk
=
L
3
bk
L
4
bk
,
[2]
R
bk
=
k
C
b
+w
k
C
b
+n
k
C
b
+
L
d
bk
C
d
,
[3]
R
bk
=
C
a
bd
T
d
ka
=
C
3
b3
T
3
k3
+
C
3
b4
T
4
k3
+
C
4
b3
T
3
k4
+
C
4
b4
T
4
k4
.
41
Using formulas (B.2), (B.5) and (B.3), we compute
[1]
R
3k
=
L
3
3k
L
4
3k
k
h
3
2h
3
w
k
h
3
2h
3
= w
k
h
3
2h
3
w
k
3
2h
3
+
1
2
k
h
3
h
3
,
[2]
R
3k
=
k
C
3
+w
k
C
3
+n
k
C
3
+
L
3
3k
C
3
+
L
4
3k
C
4
=
k
(
h
3
2h
3
+
h
4
2h
4
) +
w
k
(
h
3
2h
3
+
h
4
2h
4
)
+ (
k
h
3
2h
3
w
k
h
3
2h
3
)(
h
3
2h
3
+
h
4
2h
4
) +
1
2
n
k
h
4
2h
4
= w
k
[
h
3
2h
3
3
4
(h
3
)
2
(h
3
)
2
+
h
4
2h
4
1
2
(h
4
)
2
(h
4
)
2
1
4
h
3
h
3
h
4
h
4
] +n
k
h
4
4h
4
+
k
h
3
2h
3
+
k
h
3
2h
3
(
h
3
2h
3
+
h
4
2h
4
)
1
2
k
(
h
3
h
3
+
h
4
h
4
),
[3]
R
3k
=
C
3
33
T
3
k3
+
C
3
34
T
4
k3
+
C
4
33
T
3
k4
+
C
4
34
T
4
k4
=
h
3
2h
3
k
h
3
2h
3
w
k
h
3
2h
3
w
4
2h
4
k
(h
4
h
4
)
2h
4
h
4
w
k
h
4
2h
4
n
k
h
4
2h
4
= w
k
h
3
2h
3
+w
k
(h
3
)
2
4(h
3
)
2
+
(h
4
)
2
4(h
4
)
2
n
k
h
4
2h
4
h
4
2h
4
3
2h
3
k
h
3
2h
3
4
2h
4
k
h
4
2h
4
+
k
h
4
2h
4
.
Summarizing, we get
R
3k
= w
k
4
2h
4
1
4
(h
4
)
2
(h
4
)
2
1
4
h
3
h
3
h
4
h
4
+n
k
h
4
4h
4
+n
k
h
4
2h
4
h
4
2h
4
+
h
4
2h
4
k
h
3
2h
3
1
2
k
h
4
h
4
+
1
4
h
k
h
4
(h
4
)
2
h
4
2h
4
k
h
4
2h
4
.
which is equivalent to (26) if the conditions n
k
h
4
=
k
h
4
, see below formula
(B.6), are satised.
In a similar way, we compute
R
4k
=
[1]
R
4k
+
[2]
R
4k
+
[3]
R
4k
, where
[1]
R
4k
=
L
3
4k
L
4
4k
,
[2]
R
4k
=
k
C
4
+w
k
C
4
+n
k
C
4
+
L
3
4k
C
3
+
L
4
4k
C
4
,
[3]
R
4k
=
C
a
4d
T
d
ka
=
C
3
43
T
3
k3
+
C
3
44
T
4
k3
+
C
4
43
T
3
k4
+
C
4
44
T
4
k4
.
For the rst term, we use respective coecients of dconnection,
L
3
4k
and
42
L
4
4k
, from (B.2),
[1]
R
4k
=
L
3
4k
L
4
4k
= (
h
4
h
4
2h
3
n
k
)
+ (
k
(h
4
h
4
)
2h
4
h
4
w
k
h
4
2h
4
n
k
)
h
4
2h
4
)
= n
k
h
4
2h
3
h
4
n
k
(
h
4
2h
3
4
h
3
2(h
3
)
)h
4
n
k
(
h
4
2h
4
(h
4
)
2
2(h
4
)
2
) +
k
h
4
2h
4
k
h
4
2(h
4
)
2
.
In order to compute the second term, we use formulas (B.3), for
C
3
and
C
4
, and (B.2), for
L
3
4k
and
L
4
4k
,
[2]
R
4k
=
k
C
4
+w
k
C
4
+n
k
C
4
+
L
3
4k
C
3
+
L
4
4k
C
4
=
k
4
2h
4
+n
k
4
2h
4
+
h
4
h
4
2h
3
n
3
2h
3
+
h
4
2h
4
k
(h
4
h
4
)
2h
4
h
4
w
k
h
4
2h
4
n
k
h
4
2h
4
4
2h
4
= w
k
4
2h
4
h
4
2h
4
k
h
4
h
4
2h
3
h
3
3
2h
3
+
h
4
2h
4
+n
k
4
2h
4
4
2h
4
h
4
2h
4
+
k
h
4
2h
4
h
4
2h
4
+
k
h
4
2h
4
h
4
2h
4
k
h
4
2h
4
+
h
k
h
4
2(h
4
)
2
.
For the third term we use formulas (B.2), with
C
3
43
,
C
3
44
,
C
4
43
,
C
4
44
, and the
formulas (B.5), with
T
3
k3
,
T
4
k3
,
T
3
k4
,
T
4
k4
,
[3]
R
4k
=
C
3
43
T
3
k3
+
C
3
44
T
4
k3
+
C
4
43
T
3
k4
+
C
4
44
T
4
k4
=
h
3
2h
3
(w
k
w
k
h
3
2h
3
+
k
h
3
2h
3
) +
h
4
h
4
2h
3
(
1
2
n
k
)
h
4
2h
4
(
h
4
h
4
2h
3
n
k
)
4
2h
4
w
k
h
4
2h
4
n
k
h
4
2h
4
+
k
h
4
2h
4
+
k
h
4
2h
4
= w
k
h
4
2h
4
h
4
2h
4
+n
k
(
h
4
2h
4
)
2
k
h
4
2h
4
h
4
2h
4
4
2h
4
k
h
4
2h
4
.
Taking the last equalities in above three formulas, we get
[1]
R
4k
= n
k
h
4
2h
3
h
4
+n
4
2h
3
+
h
4
h
3
2(h
3
)
2
h
4
+
k
h
4
2h
4
k
h
4
2(h
4
)
2
,
[2]
R
4k
= w
k
4
2h
4
h
4
2h
4
k
h
4
h
4
2h
3
h
3
3
2h
3
+
h
4
2h
4
+n
k
[
4
2h
4
4
2h
4
h
4
2h
4
] +
k
h
4
2h
4
h
4
2h
4
+
k
h
4
2h
4
h
4
2h
4
k
h
4
2h
4
+
h
k
h
4
2(h
4
)
2
,
[3]
R
4k
= w
k
4
2h
4
h
4
2h
4
+n
k
(
h
4
2h
4
)
2
k
h
4
2h
4
h
4
2h
4
4
2h
4
k
h
4
2h
4
.
43
We summarize above three terms,
R
4k
= n
k
h
4
2h
3
h
4
+n
4
2h
3
+
h
4
h
3
2(h
3
)
4
h
3
4(h
3
)
4
4h
3
h
4
+n
k
4
2h
4
+
(h
4
)
2
2(h
4
)
2
+
4
2h
4
4
2h
4
h
4
2h
4
+ (
h
4
2h
4
)
2
+
k
h
4
2h
4
k
h
4
2(h
4
)
2
+
k
h
4
2h
4
h
4
2h
4
+
k
h
4
2h
4
h
4
2h
4
k
h
4
2h
4
+
h
k
h
4
2(h
4
)
2
k
h
4
2h
4
h
4
2h
4
4
2h
4
k
h
4
2h
4
,
and prove equations (27).
For
R
i
jka
=
L
i
jk
y
k
C
i
ja
x
k
+
L
i
lk
C
l
ja
L
l
jk
C
i
la
L
c
ak
C
i
jc
+
C
i
jb
T
b
ka
from
(A.6), we obtain zero values because
C
i
jb
= 0 and
L
i
jk
do not depend on y
k
.
So,
R
ja
=
R
i
jia
= 0.
Taking
R
a
bcd
=
C
a
.bc
y
d
C
a
.bd
y
c
+
C
e
.bc
C
a
.ed
C
e
.bd
C
a
.ec
from (A.6) and con-
tracting the indices in order to obtain the Ricci coecients,
R
bc
=
C
d
bc
y
d
C
d
bd
y
c
+
C
e
bc
C
e
C
e
bd
C
d
ec
, we compute
R
bc
= (
C
3
bc
)
+(
C
4
bc
)
c
C
b
+
C
3
bc
C
3
+
C
4
bc
C
4
C
3
b3
C
3
3c
C
3
b4
C
4
3c
C
4
b3
C
3
4c
C
4
b4
C
4
4c
. There are nontrivial values,
R
33
=
C
3
33
C
4
33
3
+
C
3
33
C
3
+
C
4
33
C
4
C
3
33
C
3
33
2
C
3
34
C
4
33
C
4
34
C
4
43
=
3
2h
3
3
2h
3
+
h
4
2h
4
+
h
3
2h
3
3
2h
3
+
h
4
2h
4
3
2h
3
4
2h
4
2
=
1
2
h
4
h
4
+
1
4
(h
4
)
2
(h
4
)
2
+
1
4
h
3
h
3
h
4
h
4
,
R
44
=
C
3
44
C
4
44
4
C
4
+
C
3
44
C
3
+
C
4
44
C
4
C
3
43
C
3
34
2
C
3
44
C
4
34
C
4
44
C
4
44
=
1
2
4
h
3
h
4
h
3
4
h
4
2h
3
h
3
3
2h
3
+
h
4
2h
4
4
h
4
=
1
2
h
4
h
3
h
4
+
1
4
h
3
h
4
(h
3
)
2
h
4
+
1
4
h
4
h
3
h
4
h
4
h
4
.
We get the nontrivial vcoecients of the Ricci dtensor,
R
3
3
=
1
h
3
h
3
R
33
=
1
2h
3
h
4
[h
4
+
(h
4
)
2
2h
4
+
h
3
h
4
2h
3
]
1
h
3
,
R
4
4
=
1
h
4
h
4
R
44
=
1
2h
3
h
4
[h
4
+
(h
4
)
2
2h
4
+
h
3
h
4
2h
3
]
1
h
3
,
i.e. the equations (25).
44
B.4 Zero torsion conditions
We analyze how to solve the equation
T
4
4k
=
L
4
4k
e
4
(N
4
k
) =
k
(h
4
h
4
)
2h
4
h
4
w
k
h
4
2h
4
n
k
h
4
2h
4
= 0,
which follows from formulas (B.5) for a vanishing torsion for
D. Taking any
h
4
for which
n
k
h
4
=
k
h
4
, (B.6)
the condition n
k
h
4
2h
4
h
4
2h
4
4
2h
4
k
h
4
2h
4
= 0 is satised. For instance, parametriz-
ing h
4
=
h
h
4
(x
k
)h(y
4
), the equation (B.6) is solved for any
h(y
4
) = e
y
4
and n
k
=
k
[
h
h
4
(x
k
)], for = cont.
We conclude that for any n
k
and h
4
related by conditions (B.6) the zero
torsion conditions (B.5) are the same as for h
4
= const. Using a similar
proof from [10, 11], it is possible to verify by straightforward computations
that
T
g =
g
i
(x
1
)dx
i
dx
i
+
h
a
(x
1
, x
2
)dy
a
dy
a
= (B.7)
= q
1
(r)dr dr +r
2
d d +r
2
sin
2
d d
2
(r)q(r)dt dt,
where the coordinates and metric coecients are parametrized, respectively,
u
= (x
1
= r, x
2
= , y
3
= , y
4
= t),
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 massenergy 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/.
The solution of YangMills equations (30) associated to the quadratic metric
element (B.7) is dened by a single magnetic potential (r),
A =
A
2
dx
2
+
A
3
dy
3
=
1
2e
[(r)
1
d + (cos
3
+(r)
2
sin ) d] ,
(B.8)
45
where
1
,
2
,
3
are Pauli matrices. The corresponding solution of (32) is
given by
=
= (r)
3
. (B.9)
Explicit values for the functions (r), q(r), (r), (r) have been found, for
instance, in Ref. [18] following certain considerations that the data (B.7),
(B.8) and (B.9), i.e. [
g(r),
A(r),
(r)] , dene physical solutions with
diagonal metrics depending only on radial coordinate. A well known diag-
onal Schwarzschildde Sitter solution of (30)(32) is that given by data
(r) = 1, (r) = 1, (r) = 0, q(r) = 1 2M/r r
2
/3
which denes a black hole conguration inside a cosmological horizon be-
cause q(r) = 0 has two positive solutions and M < 1/3
.
References
[1] Y. ChoquetBruhat and R. Geroch, Global aspects of the Cauchy prob-
lem in general relativity, Commun. Math. Phys. 14 (1969) 329335
[2] D. Kramer, H. Stephani, E. Herdlt and M. A. H. MacCallum, Exact
Solutions of Einsteins Field Equations (Cambridge University Press,
1980); 2d edition (2003)
[3] S. Klainerman and F. Nicolo, The Evolution Problem in General Rel-
ativity, Progress in Mathematical Physics 25 (Birkhouser, Boston Inc,
Boston, MA, 2003)
[4] P. Chrusciel and H. Friedrich (editors), The Einstein Equations and the
Large Scale Behaviour of Gravitational Fields. 50 Years of the Cauchy
Problem in General Relativity (Birkhauser, 2004)
[5] M. S anchez, Cauchy hypersurefaces and global Lorenzian geometry, in:
Proceeings of the XIV fall Workshop on Geometry and Physics, Bilbao,
September 14-15, 2005, Publ. de la RSME, vol. XX (2006) 2-22
[6] I. Rodnianski, The Cauchy problem in general relativity, in: Proceed-
ings of the Interantional Congress of Mathematicians, Madrid, Spain,
2006, Vol. III (2006) 421442
[7] D. Christodoulou, The formation of black holes in general relativity
(European Mathematical Society Publishing House, Z urich, Switzer-
land, 2009)
46
[8] P. Chrusciel, G. Galloway and D. Pollack, Mathematical general rela-
tivity: A sample. Bulletin (New Series) of the American Mathematical
Society, 47 (2010) 567638
[9] S. Vacaru, Parametric nonholonomic frame transforms and exact solu-
tions in gravity, Int. J. Geom. Methods. Mod. Phys. 4 (2007) 1285-1334
[10] S. Vacaru, On general solutions in Einstein gravity, Int. J. Geom. Meth.
Mod. Phys. 8 (2011) 9-21
[11] S. Vacaru,On general solutions in Einstein and high dimensional gravity,
Int. J. Theor. Phys. 49 (2010) 884-913
[12] J. M. Overduin and P. S. Wesson, KaluzaKlein gravity, Phys. Rept.
283 (1997) 303380
[13] G. t Hooft, Magnetic monopoles in unied gauge theories, Nucl. Phys.
B79 (1974) 276284
[14] R. Bartnik and J. McKinnon, Particlelike solutions of the Einstein
YangMills equations, Phys. Rev. Lett. 61 (1988) 141144
[15] Y. Brihaye, A. Chakbarati and D. H. Tchrakian, Particlelike solutions
to higher order curvature EinsteinYangMills systems in d dimensions,
Class. Quant. Grav. 20 (2003) 27652784
[16] Y. Brihaye, B. Hartmann and E. Radu, Cosmological solutions in a
spontaneously broken gauge theory, Phys. Rev. Lett. 96 (2006) 071101
[17] S. Vacaru and D. Singleton, Ellipsoidal, Cylindrical, Bipolar and
Toroidal Wormholes in 5D Gravity, J. Math. Phys. 43 (2002) 2486-
2504
[18] Y. Brihaye, B. Hartmann, E. Radu and C. Stelea, Cosmological
monopoles and nonAbelian black holes, Nucl. Phys. B 763 (2007)
115146
[19] J. E. Lidsey, D. Wands and E. J. Copeland, Superstryng Cosmology,
Phys. Rept. 337 (2000) 343492
[20] L. C. Evans, Partial Dierential Equations, Graduate Studies in Mathe-
matics, vol. 19 (American Mathematical Society, Providence, RI, 1998)
[21] S. Klainerman and I. Rodnianski, The causal structure of microlocalized
rough Einstein metrics, Ann. Math. (2) 161 (2005) 11951243
47
[22] H. F. Smith and D. Tataru, Sharp local wellposedness results for the
nonlinear wave equation, Ann. Math. (2) 162 (2005) 291366
[23] Y. ChoquetBurhat and J. York, The Cauchy problem, in: General
Relativity, ed. A. Held, (Plenum Press, NY, 1989), pp. 99172
[24] B. B. Kadomtsev and V. I. Petrviashvili, On the stability of solitary
waves in weakly dispersive media Doklady Akademii Nauk SSS 192
(1970) 753756 [in Russian]; Sov. Phys. Dokl. 15 (1970) 539541 [En-
glish translation]
[25] S. Vacaru, Curve ows and solitonic hierarchies generated by Einstein
metrics, Acta Applicandae Mathematicae 110 (2010) 73-107
48