You are on page 1of 2

Singular General Relativity

How I learned to stop worrying and love the singularities


Cristi Stoica
Introduction
Do the singularities predict the breakdown of General Relativity?
It is often said that General Relativity predicts its own
breakdown, by predicting the occurrence of singularities.
In addition, when trying to quantize gravity, it appears to
be perturbatively nonrenormalizable.
Are these problems signs that we should give up Gen-
eral Relativity in exchange for more radical approaches
(superstrings, loop quantum gravity etc.)?
What if the limits are in fact not of GR, but of our tools?
What if understanding the singularities also helps with
the problem of quantization?*

*Usually is considered that when GR will be quantized, this will solve the singu-
larities too, by showing probably that quantum elds prevent the occurrence of
singularities. Here we will explore the opposite view.
The cure for singularities
Singular semi-Riemannian geometry
The methods of singular semi-Riemannian geometry (see [1, 2] and the appendix) allow us to:
nd mathematical descriptions of singularities, and understand them
replace the singular quantities with regular ones, which are equivalent to them if det g = 0
write non-singular eld equations, for example extensions of Einsteins equations
Singular General Relativity
Singular General Relativity is the application of the methods of singular semi-Riemannian ge-
ometry to General Relativity. For this, the metric is required to be benign, but often a metric
which is apparently malign can be put in a form which is benign, by a coordinate change.
Malign singularity: g
ab
. Benign singularity: g is smooth and det g = 0.
Results
Smoothening black hole singularities
Apparently, the black hole singularities are malign, because some of the components of the
metric are divergent (g
ab
). For the standard stationary black holes, we can nd transfor-
mations of coordinates which make them smooth. This also makes, for charged black holes,
the electromagnetic potential and eld smooth. The metric is put in a form which allows the
evolution equations to go beyond the singularities.
Malign singularities:
g
ab

Benign singularities:
g
ab
smooth, det g 0
Schwarzschild black hole (neutral):
In Schwarzschild coordinates
ds
2
=
r 2m
r
dt
2
+
r
r 2m
dr
2
+ r
2
d
2
,
where d
2
= d
2
+ sin
2
d
2
.
In non-singular coordinates [3]
ds
2
=
4
4
2m
2
d
2
+(2m
2
)
2T4
(Td + d)
2
+
4
d
2
,
where (r, t) (
2
,
T
), T 2.
Reissner-Nordstr om black hole (charged):
In Reissner-Nordstr om coordinates
ds
2
=

r
2
dt
2
+
r
2

dr
2
+ r
2
d
2
where = r
2
2mr + q
2
.
The electromagnetic potential is singular:
A =
q
r
dt
In non-singular coordinates [4]
ds
2
=
2T2S2
(d + Td)
2
+
S
2

4S2
d
2
+
2S
d
2
,
where (t, r) (
T
,
S
), T > S 1.
The electromagnetic potential is non-singular:
A = q
TS1
(d + Td)
A. If the singularity is malign, entering infor-
mation is lost. When accounting for Quantum
Mechanics, this causes in particular a violation
of unitarity, if a system which is entangled with
another system is lost in the singularity.
B. If the singularity is benign, the eld equa-
tions can be extended beyond the singularity,
and the information is preserved [3, 4, 5, 7].
Information paradox
Penrose-Carter diagrams for non-rotating and electrically neutral evaporating black holes.
A. B.
Malign
spacelike
singularity
Light ray joining the singularity
and the infinity
Benign
spacelike
singularity
Einstein equation at singularities
Singular semi-Riemannian geometry (see appendix and [1, 2]) introduces generalizations of
semi-Riemannian spacetimes which admit degenerate metric but have smooth Riemann cur-
vature R
abcd
. But the Einstein tensor is usually singular. Yet, in 4D we can rewrite the Einstein
equation in terms of quantities which remain smooth at singularities. The resulting equations
are equivalent to Einsteins so long as the metric is regular, but work as well at singularities.
Densitized Einstein equation
On 4D semi-regular spacetimes the Riemann
curvature R
abcd
is smooth. Hence the Einstein
tensor density Gdet g is smooth too, being:
G
ab
det g = g
kl
a
kst
n
lpq
R
stpq
.
We can write a densitized version of the Ein-
stein equation:
G
ab
det g + g
ab
det g = T
ab
det g,
For regular metric it is equivalent to Einsteins
equation, but works at singularities too [1].
Expanded Einstein equation
A quasi-regular spacetime is a 4D semi-regu-
lar manifold with smooth Ricci decomposition
R
abcd
=
1
12
R(g g)
abcd
+
1
2
(S g)
abcd
+ C
abcd
where S
ab
:= R
ab

1
4
Rg
ab
, and
(h k)
abcd
:= hack
bd
h
ad
k
bc
+ h
bd
kac h
bc
k
ad
.
In this case we can write the expanded Ein-
stein equation [9]:
(G g)
abcd
+ (g g)
abcd
= (T g)
abcd
.
Friedmann-Lematre-Robertson-Walker spacetimes
The FLRW spacetime is the warped product
I a , a : I R, a 0, with metric
ds
2
= dt
2
+ a
2
(t)d
2
where usually is S
3
, R
3
, or H
3
[2].
The energy density and pressure density p
are singular at the Big-Bang, where a(t) = 0:
=
3

a
2
+ k
a
2
, + 3p =
6

a
a
.
But in terms of the correct densities
=

g, p = p

g,
the equations become smooth:
=
3

a
2
+ k

, 3 p + =
6

a
2
a

.
Hence, and p are smooth, as it is the densi-
tized stress-energy tensor
T
ab

g = ( + p) uau
b
+ pg
ab
.
These singularities are quasi-regular [8, 13].
Weyl curvature hypothesis
The Weyl curvature hypothesis was proposed by R. Penrose, to account for the high homo-
geneity and low entropy characterizing the Big-Bang. It states that at the Big-Bang singularity,
the Weyl curvature tensor C
abcd
= 0.
At a quasi-regular Big-Bang singularity, C
abcd
is smooth. At the singularity it vanishes, because
it lives in T

M, which has dimension 3 at singularities [10].


Dimensional reduction and Quantum Gravity
Different results suggest that a dimension < 4
may act like a dimensional regulator for QFT
and for QG. In lower dimension, the Weyl ten-
sor vanishes, and the local degrees of free-
dom (gravitational waves and gravitons) disap-
pear, allowing QG to be renormalizable.
Benign singularities undergo a dimensional re-
duction, and may be the needed dimensional
regularizers for QG:
g
ab
is independent on some directions.


det g 0, reducing the contribution of
lower distance Feynman integrals.
The admissible elds live in lower dimension
spaces:
rank gp = dimTpM = dimTp

M < 4.
For quasi-regular singularities C
abcd
= 0.
The quantities det g and C
abcd
vanish as ap-
proaching singularities. But QG needs them
to decrease with the scale. They may do this:
As the energy approaches the UV limit, the
number of the particles in Feynman diagrams
increases. If particles are benign singularities,
this means lower det g and C
abcd
. We conjec-
ture: this gives the needed regularization [11].
References
[1] C. Stoica. On Singular Semi-Riemannian Manifolds. May 2011. arXiv:math.DG/1105.0201.
[2] C. Stoica. Warped Products of Singular Semi-Riemannian Manifolds. May 2011. arXiv:math.DG/1105.3404.
[3] C. Stoica. Schwarzschild Singularity is Semi-Regularizable. Eur. Phys. J. Plus, 127(83):18, 2012. arXiv:gr-qc/1111.4837.
[4] C. Stoica. Analytic Reissner-Nordstrom Singularity. Phys. Scr., 85:055004, 2012. arXiv:gr-qc/1111.4332.
[5] C. Stoica. Kerr-Newman Solutions with Analytic Singularity and no Closed Timelike Curves. November 2011. arXiv:gr-qc/1111.7082.
[6] C. Stoica. Spacetimes with Singularities. An. S t. Univ. Ovidius Constant a, 20(2):213238, July 2012.
[7] C. Stoica. Globally Hyperbolic Spacetimes with Singularities. August 2011. arXiv:math.DG/1108.5099.
[8] C. Stoica. Big Bang singularity in the Friedmann-Lemaitre-Robertson-Walker spacetime. December 2011. arXiv:gr-qc/1112.4508.
[9] C. Stoica. Einstein equation at singularities. March 2012. arXiv:gr-qc/1203.2140.
[10] C. Stoica. On the Weyl Curvature Hypothesis. March 2012. arXiv:gr-qc/1203.3382.
[11] C. Stoica. Quantum Gravity from Metric Dimensional Reduction at Singularities. May 2012. arXiv:gr-qc/1205.2586 .
[12] C. Stoica. An Exploration of the Singularities in General Relativity. Seminar held at JINR, Dubna, May 2012. arXiv:gr-qc/1207.5303.
[13] C. Stoica. Beyond the Friedmann-Lemaitre-Robertson-Walker Big Bang singularity. March 2012. arXiv:gr-qc/1203.1819.
Cristi Stoica holotronix@gmail.com
APPENDIX. Singular semi-Riemannian geometry
Cristi Stoica
Singular semi-Riemannian manifolds
A singular semi-Riemannian manifold (M, g) is a differentiable manifold M with a symmetric
bilinear form g, named metric, on the tangent bundle TM. The metric g may be indenite, and
may be degenerate (i.e. det g = 0 at some points). It may have constant or variable signature.
Malign singularity: g
ab
. Benign singularity: g is smooth and det g = 0.
The problem
The geometric quantities cant be dened:

c
ab
=
1
2
g
cs
(ag
bs
+
b
gsa sg
ab
)
R
d
abc
=
d
ac,b

d
ab,c
+
d
bs

s
ac
d
cs
s
ab
R
ab
= R
s
asb
, R = g
pq
Rpq
G
ab
= R
ab

1
2
Rg
ab
These quantities are not usually dened even
if g
ab
are nite, since g
ab
when det g 0.
The idea of a solution
Use non-singular quantities, equivalent to the
singular ones for non-degenerate metric g.
Singular Non-Singular When g is...

c
ab
(2
nd
)
abc
(1
st
) smooth
R
d
abc
R
abcd
semi-regular
R
ab
R
ab

[det g[
W
, W 2 semi-regular
R R

[det g[
W
, W 2 semi-regular
R
ab
Ric g quasi-regular
R Rg g quasi-regular
But how to dene them?
Degenerate metric and covariant contraction
(V,g) V*
u
u+w
w (V

,g

)
(V

,g

)
V

=V/V

(V, g) is an inner product vector space. The


morphism : V V

is dened by u
u

:= (u) = u

= g(u, ). The radical V :=


ker = V

is the set of isotropic vectors in V .


V

:= im V

is the image of . The inner


product g induces on V

an inner product de-


ned by g(u

1
, u

1
) := g(u
1
, u
2
), which is the in-
verse of g iff det g ,= 0. The quotient V := V/V
consists in the equivalence classes of the form
u + V . On V , g induces an inner product
g

(u
1
+ V , u
2
+ V ) := g(u
1
, u
2
).
Once we have dened the reciprocal inner product g(, ), we can dene covariant contrac-
tion for 1-forms from V

. Then we dene it for tensors with two covariant indices living in V

.
T(
1
, . . . , , . . . , , . . . ,
k
)
We extend these denitions to a singular semi-Riemannian manifold (M, g).
0 TpM (TpM, g) (V , g

) 0
0 T

pM T

p
M (T

pM, g) 0
i

T
p M
i


Covariant derivative
The covariant derivative
X
Y cant be de-
ned. We can use instead the Koszul form:
/(X, Y, Z) :=
1
2
XY, Z + Y Z, X ZX, Y
X, [Y, Z] + Y, [Z, X] + Z, [X, Y ]
For non-degenerate metric
X
Y = /(X, Y, )

.
The Christoffels symbols of the rst kind are

abc
= /(a,
b
, c) =
1
2
(ag
bc
+
b
gca cg
ab
)
The lower covariant derivative:
(

X
Y )(Z) := /(X, Y, Z), for any Z X(M).
A singular semi-Riemannian manifold is radical-
stationary if /(X, Y, ) /

(M) := (T

M).
Covariant derivative of covariant indices:
(
X
) (Y ) := X ((Y )) g(

X
Y, )
(
X
T) (Y
1
, . . . , Y
k
) = X (T(Y
1
, . . . , Y
k
))

k
i=1
/(X, Y
i
, )T(Y
1
, , . . . , , . . . , Y
k
)
A singular semi-regular manifold is a radical-
stationary manifold so that
X

Y
Z /

(M).
Isotropic singularities. A manifold (M,
2
g),
where g is a non-degenerate metric, and a
smooth function, is semi-regular.
Curvature
Let (M, g) be radical-stationary.
The lower Riemann curvature operator is:
1

XY
Z :=
X

Y
Z
Y

X
Z

[X,Y ]
Z
The Riemann curvature tensor is:
R(X, Y, Z, T) := (1

XY
Z)(T)
R
abcd
= a
bcd

acd
+ (ac

bd

bc

ad
)
The Ricci tensor: Ric(X, Y ) := R(X, , Y, )
The scalar curvature s := Ric(, ).
If (M, g) is semi-regular, the Riemann curva-
ture is a smooth tensor eld, but the Ricci and
scalar curvatures may be singular.
Singular warped products
The warped product of two singular semi-Rie-
mannian manifolds (B, g
B
) and (F, g
F
), with
warping function f F(B) is the singular semi-
Riemannian manifold
B
f
F :=

B F,

B
(g
B
) + (f
B
)

F
(g
F
)

,
where
B
: B F B and
F
: B F F
are the canonical projections.
The inner product on B
f
F takes the form
ds
2
= ds
2
B
+ f
2
ds
2
F
.
The usual denition of warped product requires
(B, g
B
) and (F, g
F
) to be non-degenerate, and
f > 0. Here we allow it to be f 0, including
by this possible singularities.
If (B, g
B
) and (F, g
F
) are radical-stationary
and df /

(B), the warped product manifold


B
f
F is radical-stationary.
If (B, g
B
) and (F, g
F
) are semi-regular, and
df A
1
(B), the warped product manifold
B
f
F is semi-regular.
Example of warped product: the Friedmann-
Lematre-Robertson-Walker spacetime.
References
Stoica, C. 2011a. Tensor Operations on Degenerate Inner Product Spaces . arXiv:gr-qc/1112.5864, December.
Stoica, C. 2011b. On Singular Semi-Riemannian Manifolds. arXiv:math.DG/1105.0201, May.
Stoica, C. 2011c. Warped Products of Singular Semi-Riemannian Manifolds. arXiv:math.DG/1105.3404, May.
Stoica, C. 2011d. Cartans Structural Equations for Degenerate Metric. arXiv:math.DG/1111.0646, November.
Cristi Stoica holotronix@gmail.com