See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/278019492

How to cancel the central singularity of the

Schwarzschild solution. Natural mass inversion
process
Research June 2015
DOI: 10.13140/RG.2.1.2884.6243

READS

40

All in-text references underlined in blue are linked to publications on ResearchGate,

letting you access and read them immediately.

Available from: Jean-Pierre Petit

Retrieved on: 26 June 2016

How to cancel the central singularity of the Schwarzschild solution.

Natural mass inversion process
Jean-Pierre Petit1 G. DAgostini2

Key words : black hole, space bridge, Schwarzschild metric, Kerr metric, central
singularity, Janus cosmological model, Gaussian coordinates, mass inversion process
____________________________________________________________________________________________________
Abstract : We reconsider the classical Schwarzschild solution in the context of a Janus
cosmological model. We show that the central singularity can be eliminated and that the
transit from one fold to the other is accompanied by mass inversion.
____________________________________________________________________________________________________


I Introduction through 2D illustrations.

A metric solution of a field equation relies on symmetry hypotheses that follow from a
underlying group structure, linked to the local or global topology of space-time. In 1916
Karl Schwarzschild found [1] his famous:
(1)
Rs 2 2
dr 2
2
ds = (1
)c dt
r 2 ( d 2 + sin 2 d 2 )
R
r
(1 s )
r

which is a solution of the Einstein equation with zero right and second hand term :
(2)
R = 0

In order to build this solution it was necessary to express the different components of
the Ricci tensor and then write the solutions that resulted from that.

In this solution, spherically symmetric, with signature
, the letter was
identified to a radial variable. The angles and expressed spherical symmetry.
Moreover, when the radial distance approaches infinity the metric tended to the
Lorentz metric expressed in spherical coordinates :
(3)
d s 2 = c 2 dt 2 dr 2 r 2 ( d 2 + sin 2 d 2 )

The letter denoted the speed of light, while t was identified to an universal time .


1
Former Research Director at the french National Center for Scientific Research, CNRS.
Private mail : jppetit1937@yahoo.fr
2 Private mail : dagostini.gilles@laposte.net

was nothing but a simple interaction constant and was only much later considered to
be positive or so negative.

Keep in mind that, as will be shown further, given a so-called coordinate invariant
metric
, the choice of coordinates in not innocent. It determines how a distant
observer will build his mental image of a geometrical object, including its topology.

There are four coordinates that can be written in the more general form as :
(4)


The metric would then correspond to the bilinear form :
(5)
d s 2 = g d x d x v

The Lorentz metric, expressed in a set of natural coordinates without crossed-terms :
(6)
d s 2 = goo (d x o )2 g11 (d x 1 )2 g2 2 (d x 2 )2 g33 (d x 3 )2

In considering the metric signature, that is a part of the fundamental hypotheses,
then a time marker , while the set

is

refers to space .

It is necessary not to forget that a solution of a field equation, in the form of a 4-surface,
is fundamentally coordinate-invariant. It is easy to show how the fact of setting
coordinates in a 2-surface gives it properties, and may bring pathologies, which it had
not at the beginning. The simplest example is the sphere S2. Nothing simpler that this
object where any point doesnt seem to play a particular role. However, given the fact
that its Euler-Poincar characteristic is equal to 2, we cannot create a complete mapping
of it with the aid of a single metric and without poles.

Fig. 1 : A sphere S2 with metric in longitude-latitude coordinates

Fig. 2 : polar singularities due to the choice of coordinates

By the way, we can merge the two poles, the two coordinate singularities, into a single
one :

Fig. 3 : one-pole cartography of a S2-sphere

In reference [3], in the chapter devoted to singularities, the reader will find that a S2-
sphere can be mapped with even more singularities, the whole being the consequence
of a peculiar choice of coordinates. The only thing that we know is that in a S2-sphere
poles cannot be avoided, due to the non-zero value of the Euler-Poincar characteristic.
On the contrary a mapping, without poles, can be set on a torus, whose Euler-Poincar
characteristic is zero.

As a general rule, the elements that form the intrinsic properties of a surface or a
hypersurface described by a Riemannian metric are :


- Its signature

- Its topology

- The different families of geodesics inscribed on it

- The length that can be measured along a path, geodesical or not, separating

two different points A and B.

As for the singularities, it is worth examining if they are intrinsic (that is, true
singularities like the top of a cone), or if they come from a bad choice of coordinates or
by the association of the metric with a bad topology, as we will show later. We will also
give some examples of attempts to describe a surface with the aid of a metric expressed
in a coordinate system that produces singular situations which result from the
association of the metric with a bad topology, all due to a simple change of coordinates.

Consider the following metric :
(7)
dr 2
ds 2 =
+ r 2 d 2
r2
1 2
Rs

There are the letters and . By instinct, one thinks immediately of polar coordinates,
being a strictly positive distance and an angle. Given that the terms of the metric
are invariant as
, one judges that this 2D object must have a rotational
symmetry. And without questioning, one associates the object with a M2 manifold
without boundaries, provided with the topology of R2. Its signature is (+ +) if 0 r < Rs .
It becomes (- +) when r > Rs and the term
becomes infinite when
. Does a
solution consist in limiting the validity of the metric, as in the figure below ?

Fig. 4 : R2 representation space

Lets now make the change of variable :
(8)
r = Rs sin

The metric is now expressed as follows :
(9)
d s 2 = Rs2 ( d 2 + sin 2 d 2 )

This is the metric of the sphere S2. The signature problem has disappeared, due to a
more suitable choice of coordinates and subsequent topologic representation. How to
link what is evoked on figure 4 to a sphere ? We can deal with a two-fold cover of a disk,
whose folds F (+ ) and F ( ) are jointed along their common border, represented by the
circle of perimeter
.

The passage from representation (9) to representation (7) by means of the change of
variable (8) is a simple projection of the sphere S2 on its equatorial plane. This
projection implies that geodesics, which on the sphere S2 are great circles , will be
projected as ellipses on the equatorial plane. Half of the ellipses will be inscribed on the
F (+ ) fold, whilst the other half will be on the F (+ ) fold. The fact that the semi-ellipses are
tangent to the circle boundary ensures the continuous linking of the geodesics of the two
folds.

Fig. 5 : The sphere S2 figured as a two-fold cover of a disk

Lets take another example. Consider the metric :
(10)

ro2 dr 2
2
ds =
+ r 2 d 2
with R > ro > 0
2
2
2
r + 2 r R + ro R

- Here we find again the variables and .
-
invariance suggests a radially symmetric object.
- A mental image arises, composed by a plane, with polar coordinates

The signature is (+ +) if :
(11)
R ro < r < R + ro

When r = R ro
or
r = R + ro the term gr r of the metric becomes infinite. Would it
correspond to a bounded manifold M2 with a non-simply connected boundary ?

Fig. 6 : Mental image corresponding to expression (10)

Lets suggest a simpler solution through a simple coordinate change :
(12)
r = R + ro cos

We get immediately the new form of the metric :
(13)
d s 2 = ro 2 d 2 + ( R + ro cos )2 d 2

whose (constant) signature is (+ +) and which is perfectly regular for any values of
and . This 2-surface can be isometrically embedded in R3. It is a torus.

Fig. 7 : A torus isometrically embedded in R3

Of course, we could figure out this torus as the two-fold cover of a bounded manifold,
with a non-simply connected boundary composed by two circles. Here again, in such
representation, the geodesics correspond to a projection on the plane of symmetry of
the object when embedded in R3. The tangency of the curves ensures the continuity of
the two geodesic systems, one for the fold
and the other for the fold
.

Fig. 8 : Geodesics in a torus. Right : the torus as a two-fold cover
of a non-simply connected manifold

This example also illustrates the errors one can make when associating a metric with an
inappropriate topology. Focusing on representation (10) many people would try to find
at any cost the properties of the tangent plane to the surface in the neighbourhood of
r = 0 , which is similar to try to study the properties of the rubber of an air tube in the
vicinity of a wheel axis.

Remark :

Imagine the metric of the torus would be some solution of a field equation, written in
a tensor for, so that this solution should be coordinate invariant . Imagine the first
for of the solution would be :

(a)
d s 2 = ro2 d 2 + ( R + ro cos )2 d 2

A good looking solution, with signature ( + + ) . No problem, no pathology . This
metric is well defined for all values of the coordinates ( ( , ) . Then, somebody says :

- Hey, guys, why dont you try the following coordinate change :

(b)
rR
= arc cos (
)
ro

- Hmmmm.... complicated. But, why not ? Let us see....

The result :

8
(c)

ds 2 =

ro 2
+ r 2 d 2
r 2 + 2 r R R 2 + ro 2

Whats that stuff ? Lets try to arrange the denominator :

(d)
ro 2 dr 2
ds 2 =
+ r 2 d 2
[ r ( R ro ) ][( R + ro ) r ]

-
-

Nice, isnt ?
You think so !? In (a) we had a nice metric, good looking, with a constant signature
( + + ) . Now we have this sort of monster (d), whose signature changes when
r < ( R ro ) and r > ( R + ro )
- So ... where is the problem ?
- When ( R ro ) < r < ( R + ro ) everthing goes well. But outside ...
- Outside, what ?
- Take = constant. Outside, the length ds becomes imaginary !
- Right. The first set is perhaps more suitable.

We will now consider a third example explicitly linked to the Schwarzschild metric. We
will in fact consider a metric built with two of the terms of its line element.
(14)
dr 2
d 2 =
+ r 2 d 2
Rs
1
r

Note that if r tends to infinity this metric tends towards enclidean metric expressed in
polar coordinates. Same problem for r < Rs : the signature (+ +) is changed into (- +).

Lets make the change of variable :
(15)
r = Rs ( 1+ Log ch )

which gives :
(16)
( 1 + Log ch ) 2

d 2 = Rs2
th d 2 + ( 1 + Log ch ) 2 d 2
Log ch

All singularities disappear,.

. In this point the determinant of the metric is no longer
r = Rs corresponds to
zero :
(17)

det g = Rs4

( 1 + Log ch )2 2
th
Log ch

The metric is well defined for all values of . The determinant does not go to zero when
. If we embed the surface in a 3D-Euclidean space we can define the meridians
corresponding to :
(18)
dr 2
d 2 =
+ d z 2
Rs
1
r

And immediatly get the meridians as :
(19)

r
z2
z = 2 Rs
1
r = Rs +

Rs
4 Rs

The surface is a space bridge linking two 2D-Euclidean surfaces. Let us call it 2D
diabolo .

Fig. 9 : The 2D diabolo embedded in R3

The problem of the signature has disappeared.

10

Here again we can reverse the line :

Let us start fromollowing form : the metric expressed into the following form :
(e)
( 1 + Log ch ) 2

d 2 = Rs2
th d 2 + ( 1 + Log ch ) 2 d 2
Log ch

The signature is ( + + )

- Why dont you try :
(f)
r

= arg ch e Rs

- No kitting ! This is definitively ugly. The metric will become terribly complicated !
- Try !

The result :

r
r
1
1
r
= arg ch e Rs ch = e Rs Log ch =
1
Rs

dr
d ( Log ch ) = d 1 th d =

Rs
Rs

(g)

dr 2
d =
+ r 2 d 2
R
1 s
r
2

-
So what ?
-
Look. In (e) your metric was four inches long. Now its simpler and shorter.
Less that two inches ! Its in progress, no ?
-
Yes, but we have problems with signature. It is modified when r < Rs
-
Nothing is perfect.
-
If you dont mind, I prefer a 2D object with real length.
-
I see. Youre not ready for metaphysics.
-
Not yet.


From Lagrange equations we can compute the geodesics in the
coordinate
system. If embedded, the surface has a throat circle whose perimeter is 2 Rs . As
before, we can shape the surface as a two-fold cover of a M2 manifold with a hole. We
can associate adjacent points M (+ ) and M ( ) .

11

Fig. 10 : Two-fold cover of a manifold with a circle as common boundary

Figure 9 gives an image of a geodesic AC crossing the throat circle at B. When projected,
it is tangent to the common boundary, and this ensures the continuity of the geodesics of
the fold F (+ ) with the ones of the fold F ( ) . In figure 11 we see a set of calculated
geodesics of that surface. When the curves pass to the other fold we have represented
the portion inscribed on F ( ) with dotted lines. A geodesical path evokes a radial path.
In fact, it is the image of a meridian curve of the surface that is itself a geodesic. This
remark is important for what follows.

Fig. 11 :Set of geodesics

II 2D two-fold covers with common boundary and enantiomorphic relationship

If we consider the surface as a two-fold cover of a bounded manifold, it creates a
mapping between adjacent points M (+ ) and M ( ) . The neighbourhoods of those points
are involved in an enantiomotrphic relationship.

12

In figure 12 we can see an observer represented as an oriented triangle. And in this

same figure we show that when this triangle crosses the common boundary, its
orientation is reversed.

Fig. 12 : When the triangle crosses the common boundary,

its orientation is reversed


III Extension to 3D-hypersurfaces

Introduce the metric :
(20)

dr 2
+ r 2 ( d 2 + sin 2 d 2 )
Rs
1
r

which is Euclidean at infinity. Here again, when r < Rs the signature
is changed
into
. Introducing the same change of coordinate as in (15) we get :
(21)

d 2 =

( 1 + Log ch ) 2
d s 2 = Rs2
th d 2 + ( 1 + Log ch ) 2 ( d 2 + sin 2 d 2 )
Log
ch

Lets go back to the enantiomorphic relationship. In 3D we must use an oriented
tetrahedron.

13

Fig. 13 : Oriented tetrahedron

We have so far used representations in the form of drawings, given that they are 2D
objects which one can embed in the only mental space of representation we have at
disposal: an Euclidean 3D space, which is in turn oriented and simply-connected. When
it comes to manipulate non-Euclidean 3D objects, that in addition do not have the
usual topology , one faces considerable mental difficulties.
The problem we evoked in 2D came from our immediate mental image of a null-
homotopic space, i.e. of a space 2D or 3D composed by contractible cells. The figure 14
illustrates the concept.

Fig. 14 : Null-homotopy concept in 2D and 3D

This concept can obviously be extended to higher dimensions. The following figure
corresponds to a 2D space which presents a region that does not satisfy this criterion.
The metric that led us to figure 9 leads now to a surface which has what we could call a
space bridge. We have also drawn a throat circle on it.

Fig. 15 : Non-null homotopic region in a 2D space with throat circle

The metric (14) has led us, through an isometric embedding in R3, to visualize a surface,
figure 9, that represents a space bridge joining two planes. But given that it is a 2D

14

object, it is possible to devise a representation in a higher dimensional space. In

supposing that we forbid this artifice we would find ourselves confined to the
representation of figure 10, in which two planes are represented joined along a common
boundary that is a circle. In this figure 10 the surface exhibits a pleat along a circle. But
we know that such pleat doesnt exist in figure 9, which is an isometrical representation
of the object. The 2D-representation gives rise to other paradoxical situations shown in
figure 11. If the geodesical curves at the top of this figure appear acceptable to us, this is
because mentally we reconstruct in our minds the 3D structure of the object. The
straight line at the bottom of the figure is still more disconcerting. We imagine that this
geodesic is going to encounter the common boundary circle and then suddenly pleats
and departs in the opposite direction. But we know that this is a 2D representation of a
geodesic that is a meridian curve of the surface, which in turn does not have accidents of
this type when it gets over the throat circle.

We will now try to give a representation of a 3D hypersurface with a space bridge
structure. The latter will join two 3D Euclidean spaces but will in itself have a curvature
that well be incapable to represent. By the way we will loose the isometry.

It is possible to represent this isometry loss with the aid of the 2D diabolo figuring
that its non-isometrical representation arises from its flattening by an iron .

Fig. 16 : Flattening of the 2D diabolo

One sees that this operation generates a distension of the object. If it is made of a non-
elastic material, it will crack (see right). We need now to imagine that we could make a
similar operation with a 3D diabolo , and our line of reasonig will be the analogy. The
circle is a S1-sphere. When we operate a 3D flattening of our 3D structure we obtain
a throat sphere S2. If we look at figure 16 we could say that there isnt anything in the
circle interior . Its simply because we are outside the 2D space we are studying.

Now, consider a non-isometric embedding of our 3D-diabolo in an Euclidean 3D
space. What would we see ? The throat circle becomes a throat sphere.

- As evoked above, inside the throat circle you are out of the 2D surface.

- Similarly, inside the throat sphere you are out of the considered 3D
hypersurface.

- What appears to us as straight lines, normal to the sphere, are only

15
representations of a peculiar family of geodesics, equivalent to the meridian
lines in the 2D-model (which rebound on the circle).

Similar rebound on the sphere. This abrupt change of direction is an illusion

du to our mental representation (the normal representation in a normal human
brain).

Fig. 17a : Images of meridian geodesics of the 3D hypersurface

In the next figure we have presented some non-radial geodesics :

Fig. 17b : Images of meridian geodesics of the 3D hypersurface


Its a 3D space with a pleat. I admit it is not easy to see .

Every test particle following these geodesics would set out again following an
identical trajectory (right dotted line), but this breaking in the hypersurface, this pleat in
the sphere S2, does not exist. Lets imagine an object made of points that form a
geometrical figure, in our case the tetrahedron, that would straightly run towards that
throat sphere. When it has crossed , how should we representate it ? In the same way
we have done previously in 2D using a grey outline. It allows us to understand that we
will rencounter the same enantimorphic relationship between adjacent structures, as
shown in figure 18. A tetrahedron is a set of four vertices joined by segments. We will

16

continue these vertices with normal paths. In the left side of the figure vertex A has
already crossed the throat and emerges, accompanied by grey outlines. In the right side
the tetrahedron has completely re-emerged. One sees that its orientation is inverted.
And so one finds again the enantiomorphic relationship between adjacent regions.

Fig. 18 : in crossing the throat sphere, the tetrahedron is inverted

In the
representation the adjacent points in 2D and 3D are defined by the
relation :
2D M :( , ) M ':( , )

3D M :( , , ) M ':( , , )

The association of points M and M goes hand in hand with an enantiomorphic relation
between their corresponding neighbourhoods.


IV- Building a 3D-spacetime on 2D-space, considered as two-fold cover of a bounded 2D-
manifold with adjacent points (
)and common circular border.


Consider the 2D-diabolo. We can transform this 2D-surface into a 3D-spacetime through
a { t , r , } reprzsentation, as shown on figure 19. On the border the line element
reduces to :
(22)
d 2 = Rs2 d 2

17

Fig. 19 : Building a 3D-spacetime

On the common boundary (the pleat ) the orientations of space and time cannot be
further defined. When crossing this common boundary, space and time are inverted with
respect to their initial (arbitrary) orientations. It is worth to imagine that this 3D-
spacetime, a two-fold cover of a 3D-manifold, is not limited to the representation given
here. On the contrary, the two enantiomorphic and time-arrow opposed spacetimes
intepenetrate each other as suggests the following figure.

Fig. 20 : In turning the throat circle, the arrow of time is inverted

This structuring of the spacetime is managed in Relativity using Gaussian coordinates.


V The problem of Gaussian coordinates


Have a close look to reference [2], section 2.4, entitled Gaussian coordinates . By
letting a family of geodesics play a particular role among the coordinates lines Gauss
introduced a useful coordinate system. Consider a four dimensional space with

18

hyperbolic metric with signature

. Assume we can imbed a three
dimensional hypersurface , imbedded in the four-dimensional space . Assume, in
some place, we can define a vector n normal to , which satisfies :
(23)
n o no + ( n1 n1 + n1 n1 + n1 n1 ) > 0

which, in the familiar langage of special relativity theory, implies that is oriented in
space ( whereas the vector n, normal to s, is ortiented in time ).
We introduce in the surface
three coordinates x *1 , x *2 , x *3 wich serve to
characterise the point P * 3 . Through each point P * of the the three-dimensional
surface we draw the geodesic which is orthogonal to at P * . These geodesic will
form a non intersecting curves in some neighbourhood M of such that , through each
point P of M there will be exactly one of the geodesics constructed. We introduce now,
in the entire four-dimensional domain M , coordinates as follows : Given P , we
consider the geodesic passing through P and its original point P * 3 . We define the
coordinate x i of P in terms of the arc length P * P of the geodesic and of the coordinate
x *i of P * :

Fig.21 : Figure 2.1 after reference [2]

(24)

x = arc length P * P along the geodesic

x1 = x *1
x 2 = x *2
x 3 = x *3

In this manner, the three coordinates x1 , x 2 , x 3 remain constant alogs any geodesic
perpendicular to . It follows that, along such a geodesic,
(25)
d s 2 = (d x)2
goo = 1

19

This is the classical way Gaussian coordinates are defined [2]. This is possible if and only
if goo 0 . As we will see further, through various choices of coordinate systems, in
Schwarzschild solution this term

vanishes on the so-called horizon (r = Rs ) , or

throat surface
.

It means that on this peculiar portion of the hypersurface, normal vector and space
orientation cannot be defined, as didactically illustrated on figure 20. Through this
border the arrow of time and the space orientation are reversed. Extending this scheme
one is led to envisage, with a space bridge structure, an interwoven of two spacetime
folds F (+ ) and F ( ) that imply a PT-symmetry ( time and space inversion ).

At this point we must give an outline of the following sections of this paper.

- In section 6 we will consider again the classical exterior Schwarzschild
solution and complete it with a Schwarzschild interior solution, both of them
forming a description deprived of any pathology and in accordance with
observations. Geometrical criticality will be evoked.

- In section 7 we will recall the Schwarzschild solution in its different
representations, indicating that no one of them has devised the elimination of the
central singularity , always considered as a true singularity.

- In section 8 we will evoke the indispensable change of geometrical paradigm
(that is, the passage to a system of two coupled field equations) that will allow
the construction of a solution as a 4D-space bridge.

- In section 9 we will evoke the reinterpretation of the Schwarzschild solution in
the context of a Janus geometry.

- In section 10 we will derive conclusions with regard to physics.





VI Schwarzschilds solution and astronomical observation


What is known classically as Schwarzschilds solution (1), that is, a solution of equation
(2), which is Einsteins equation with zero right hand side, refers to a portion of
spacetime where the matter-energy density is zero, where the tensor T is zero. Under
these conditions, it could seem strange that such equation could produce other that
straight line geodesics and a metric solution different from Minkowskis (3). Things
become much more consistent if one completes the metric in (1), called Schwarzschild
exterior metric, with the Schwarzschild interior metric :

20

(26)

The latter describes the spacetime geometry in a domain that is a sphere or radius ro
filled with a material of uniform density = Cst . This metric is then a solution of the
Einstein equation with a non-zero right hand side term (but with the cosmological
constant equal to zero) :
(27)
1
R R g = T
2
where :
(28)
c 2 0
0
0

p
0
0
8 G

=
T =
4
c
0
0 p 0

0
0
0 p

The connection of these two metrics is straightforward since the two families of
geodesics connect without difficulty. This set of two solutions, Schwarzschild interior
metric plus Schwarzschild exterior metric describes perfectly the geometry inside and
outside the mass M of a sphere of radius filled with a material of constant density .
The density inside an ordinary star is not to be considered as constant. However, a
constant density is a very good approximation for a body like a neutron star. Using the
Lagrange equations it is extremely easy to construct the interior geodesics. But the
astrophysicist doesnt proceed in this manner because he can hardly imagine that a
particle following one geodesic can cross without being obstacled by such massive
object. In the limit, if the object of constant density is identified with a telluric planet, the
whole thing would correspond to the path of a neutrino crossing the planet without
interactig and experiencing a gravitational lensing effect.

The two Schwarzschild solutions are subject to constraints. If one wants the
Schwarzschild exterior solution to describe a portion of space such that the
Schwarzschild sphere of radius :
(29)
2G M

Rs =
c2

is located at the interior of the body of mass M, this metric will not raise any problem at
all. The Schwarzschild interior solution has equally a characteristic magnitude :

21

(30)

R = c

3

8 G

If ro < R , the interior metric will be deprived as well of any pathology. That will be the
case for the majority of celestial bodies we will consider, for example the Earth.

Fig. 22 : Subcritical conditions

The geodesics constructed from the exterior metric give rise to quasi-Keplerian
trajectories with differences more or less important. In figure 23, the advance of the
perihelion affecting a test mass orbiting around a neutron star.

Fig. 23 : Relativistic effect. Advance of perihelion in the vicinity of a neutron star

Lets consider a celestial body of constant density, like a neutron star, such that goes
from 1015 g/cm3 to 1016 g/cm3, and lets suppose that its mass increases by a matter
input in the form of stellar wind provided by a companion star. The two metrics will
tend to a critical situation, as shown in figure 24.

22

Fig. 24: Towards geometrical criticality

Several neutron stars are known which form a tight couple with a star of great mass.
Such condition leads at the end to the critical situation evoked in figure 24.
But if one stays apart from this criticality both metric solutions are free of singularities.

Didactic image :

A blunt cone is a good didactid image of a constant density object (grey) surrounded
by void (the truncated cone).

2D didactic image

A blunt cone can be made with a truncated cone, glued on a portion of a sphere. How
to manage geodesic adjustment ? Its easy. First, you build your truncated cone with a
piece of paper and cisors :

23

Top : how to make a truncated cone, which is built around an angular curvature .
To finish the blunt cone we have to add a portion of sphere which contains the
same angular curvature and own the same perimeter p. Then the continuity of
geodesics is ensured.

24

VII Schwarzschilds solution in its different forms


Faced to the problem linked to criticality, from the middle of the Sixties the answer of
the theoretician was to give a new interpretation of the solution found by Karl
Schwarzschild in 1916, known as the exterior metric . One can wonder how
theoreticians had the idea of reassigning a stationary solution to the description of the
ultra-fast implosion of a body where the interneutronic forces of repulsion couldnt do
other than oppose themselves to the extraordinary gravitational pressure.
Schwarzschilds metric allows to trace the curve corresponding to the free fall of a test
particle by giving the expressions of the laws r (s) and r (t) , according to whether the
chronology of the phenomenon is expressed in proper time or in the time t of an
observer located at a certain distance from the object. The calculation is straightforward.
The result is given, for example, in reference [2], section 6.8. We reproduce here a figure
taken from it.

Fig. 25 : Inward free fall of a test particle (fig. 6.2 from reference [2])

As one can see, the progress of the test particle in proper time along a radial path is
extremely fast. It reaches the Schwarzschild sphere in a finite time. On the contrary, if
one considers that the variable t in expression (1) is the time tied to a distant observer,
then the fall appears to have an infinite duration to him.. The observer is supposed to see
the implosion of a body, for example a destabilized neutron star, whose mass, according
to present knowledge, suddenly exceeds 2,5 or 3 solar masses. But this ultra brief
phenomenon whose duration doesnt exceed some thousandths of a second in proper
time, and that leads the mass to a central singularity , appears to him, as outside
observer, to occur in an infinite time. He is somehow in a state of freezing .

A glance at expression (1) reveals two configurations potentially singular, one of them
2G M
for r = Rs =
, the other for r = 0 , a situation clearly identified for the first time by
c2
the mathematician David Hilbert [4]. Classically, one considers that the first one is linked
to a bad choice of coordinates. It is called a coordinate singularity. The second one is
considered as impossible to eliminate, that is, a truly physical singularity.

Lets start by listing different changes of variable that allow to avoid the singular
situation at r = Rs .

25

Fig. 26 : Schwarzschilds geometry. Alternative coordinates

In 1921 Paul Painlev [5], and in 1922 Allvar Gullstrand [6] independently produced a
metric, a spherically symmetric solution of Einstein's equation that was later identified
with a simple transformation of the Schwarzschild metric. In the GullstrandPainlev
coordinates there is no singularity at r = .

In 1924 Arthur Eddington [7] showed for the first time that the singularity at r = was
indeed a coordinate singularity, an artifice. His work was later extended by D.
Finkelstein [8]. In 1932 Georges Lematre, using another coordinate change, explained
that this singularity was unphysical [9]. In 1939 Howard Robertson computed the free
fall time of a test particle using the Lagrange equations, and showed that it was finite if
measured in proper time, even if this free fall time was found to be infinite when
measured in the t coordinate frame (see above, figure 25).
In certain presentations of the Schwarzschild metric one finds, on the basis of symmetry
considerations, arguments like that the crossed terms should not exist, as it is shown in
many of the figures we have draw above. But it is worth remembering that these terms
do exist in the Kerr metric [10].
(31)
2m
2 + a 2 cos 2
2 2
ds 2 = (1 2
)c
d
t

d 2 ( 2 + a 2 cos 2 ) d 2
+ a 2 cos 2
2 + a2 2m

2 m a 2 sin 4 2
4ma
( 2 + a 2 )sin 2 + 2
d 2
c d t d

2
2
2
2

+
a
cos

+
a
cos

See section 7.6 of reference [2], equation (7.110). If in this expression one studies the
geodesics of zero length, that is the trajectories of photons, and if one considers circular
paths having as symmetry axis the axis of the system, one will find two different values
of the speed of light according to the sense of the circular movement, if accompanying
the rotation movement or inversely. This follows from the term d t d which is then
interpreted as a frame dragging : in a Machian sense the source competes (we
quote) with the Lorentzian boundary conditions at infinity in the establishment of a local
inertial frame.

26

It is to be remarked that keeping crossed terms of the type dr dt in the metric

expressions shown above implies a radial frame dragging in the sense that, with
regard to radial trajectories, the speed of light is different according to the consideration
of inward or outward trajectories. This aspect will be reconsidered and developed in
another contribution.

Being it an object with spherical symmetry or an axially symmetric one, the black hole
model is full of disconcerting and bizarre features. The most serious criticism of the
black hole model is as follows : when considering the initial formulation by K.
Schwarzschild, wouldnt it be possible that the signature of the metric in not conserved
and that in passing through the Schwarzschild sphere it goes from
to
? Some people pretend that the nature of the coordinates is simply modified
and becomes space-like and time-like. We have seen before that a bad choice of the
local topology could transform a simple torus into an object with multiple pathologies. A
modification in the conception of the local topology will be the basis of our revision of
the Schwarzschild solution.

A vast consensus supposedly based on the progress made in differential geometry has
emerged aimed at considering the singularity in the origin of the radial coordinate r
as belonging to the world of physics, presenting it as a point beyond which geodesics
cannot be extended . We will treat things in a different manner.



VIII First of all : a change in geometrical paradigm

In two recent papers ([11], [12]) we have proposed a complete revision of the
Einsteinian geometrical paradigm. This last is in fact based (provided the cosmological
constant is zero) on a description of the universe where it is considered as a M4
manifold with a unique metric that is a solution of Einsteins field equation. In 1957, H.
Bondi tried to introduce negative energy particles (and negative mass, whenever they
possessed one) in the model. We know that the Newtonian approximation, when applied
to Einsteins equation, gives the classical Newtons laws. When Bondi tries to apply this
method by considering a mixture of particles of positive and negative masses he obtains
the following dynamics :

- Positive masses attract every other mass, whatever its sign
- Negative masses repel every other mass, whatever its sign.

This conclusion leads in consequence to a physically unmanageable runaway
phenomenon. If one considers, for example, a couple of particles of the same mass but
with opposite signs, the positive mass particle will run away, pursued by the negative
one. And the pair will exhibit a uniform accelerated movement conserving energy
because one of the two masses is negative !

27

Fig. 27 : A physically unmanageable phenomenon

This feature dissuaded all theoreticians from considering negative masses for more than
half a century. But the problem presents itself in a different way if instead of one metric
one considers two,
and
, the first describing particles of positive energy (and
mass, whenever they possess one), the second describing particles of negative energy
(and mass, whenever they possess one). Its not the first time this has been done.
Several bimetric approaches have been suggested in the past. Some of them ([13],
[14]) refer the second metric to the behaviour of gravitons, eventually provided with
mass. The first attempt was a paper in 1994 [15]. The resulting system of coupled field
equations was then :
(32a)
1
(+ )
(+ )
R
R(+ ) g
= ( T(+ ) + T( ) )
2
(32b)
1
( )
( )
R
R( ) g
= ( T(+ ) + T( ) )
2

The Newtonian approximation applied to this system produces the complete
disappearing of the absurde runaway phenomenon. Such new bimetric view, in
terms of the Newtonian approximation, gives the following dynamics :

- Masses of the same sign attract according to Newtons law
- Masses of opposite signs repel according to anti-Newton

It is to be noted that this bimetric reformulation is extremely disconcerting for journal
referees because of three aspects which it is worth mentioning. In this model one
assumes, and it is clearly stated, that the two species of opposite energies do not interact
by means of electromagnetic forces, neither strong nor weak, and neither by strong
interactions. The interaction operates exclusively through the contribution given by
each of the species to the gravity field.

In General Relativity, sensu stricto, particles are not included in the model. Equations
refer only to continuous functions and produce only geodesic curves. So one makes the
hypothesis that particles with non-zero mass move along geodesics of non-zero length
and that photons move along geodesics of zero length. Moreover, one assumes that

28

observation is possible, that is, objects formed by particles with a mass can emit photons
which in turn can be captured by measurement instruments, or simply by the human
eye made of positive masses.

In this new model one assumes instead that negative masses emit negative energy
photons that move along zero length geodesics of the metric, and that in turn these same
negaphotons can be absorbed by structures formed by negative mass particles and
solely by them. Devices of positive mass (including our eyes) that use optics will only
capture positive energy photons and hence will not be able to react to the presence of
negative energy photons. And vice versa.

As we have assumed that particles of opposite masses do not interact either by
electromagnetic forces or strong forces, they could not enter into a collision. Referees
have critized this idea arguing that the particles are on the same space-time . In fact, if
one considers the problem from purely geometrical grounds, those encounters would
be geometrically impossible because the two subsets move along disjoint families of
geodesics.

The second criticism relies on the immediate instability of a quantum vacuum which
could create pairs (+m, -m). But it is based on the theoretical framework of quantum
gravity that, however, still today remains purely hypothetical. How can one base a
criticism on a phenomenon (the creation and annihilation of pairs of particles of
opposite mass) that is not adequately described ?

The third criticism issues from Quantum Field Theory, which excludes straight away
states of negative energy because a particle could not have an energy less than that of
the vacuum ([16], page 76). We quote :

If we suppose that T is linear and unitary then we should face the disastrous
conclusion that for any state of energy there is another state T
of energy
. To avoid this we are forced here to conclude that T is antilinear and
antiunitary.

In order to refute this statement we will say that it is bootstrap talk and that the
conclusion is contained in the hypothesis, as occurs with the CPT theorem .

Returning to the implications of our work, it had immediate interesting results [17]
based on numerical simulations that modelled the very large scale (VLS) structure of the
universe, according to which by the end of the radiative era the negative mass, supposed
to be denser (feature that will be the subject of a future paper devoted to a bi-
radiative era), would have formed the first spherical clusters, repelling immediately
the negative mass of the rest of the space, giving it a lacunar structure, properly
identified and described in [18], [19], [20] and [21].

29

Fig. 28: After Piran [18] : VLS bubble structure from IRAS survey
In the same way, the negative gravitational lensing effect, described in [17], leads to an
alternative interpretation of the weak lensing recently proposed by K. Izumi et al. [22].
In a more recent work ([11], [12]) the system of equations proposed from a Lagrangian
derivation becomes :
(33a)

1
g ( ) ( )
(+ )
(+ )
R
R(+ ) g
= T(+ ) +
T
2
g (+ )

(33b)
g (+ ) (+ )

1
( )
( )
( )
R
R( ) g
=
T
+
T

( )
2
g

By the way, we remark its similarity with the system of equations of reference [23].

The model has then proved capable of giving an exact solution describing the evolution
of the scale factors a (+ ) and a ( ) , accounting for the acceleration of the positive mass
entities ([24],[25],[26],[27],[28],[29],[30],[31],[32] and [33]), whereas the subset of
negative mass slows down.

Once again, it is a very organized vision of cosmic evolution where two 4-folds F (+ ) and
(+ )
( )
and g
, have different dynamical behaviours
F ( ) , constructed from the metrics g
even if they are built on the same manifold M4.

And this work, extended to a situation where there are two different speed lights (
, consequence of
), constitutes the first theoretical modelling giving
an explanation to the cosmic acceleration instead of a void word like dark energy .

Back to the work of S. Weinberg [16], lets quote his sentence on page 104 :

No examples are known of particles that furnish unconventional representation of
inversions, so these possibilities will not be pursued here. From now on, the inversions
will be assumed to have the conventional action assumed in Section 2.6

30

This sentence refers of course to the hypothesis expressed on page76 of reference [16]
about the anti-unitary and anti-linear character of the T operator. However, cosmic
acceleration implies the action of a negative pressure and hence of negative energy
(pressure is an energy density by unit volume). The discovery of such quite unforeseen
phenomenon makes it compelling for Quantum Field Theory to be extended in order to
include negative energy states.

The consequences of this change of geometrical paradigm analysed through Dynamical
Group Theory will the the subject of a future contribution. The approach introduces a
new isometry group tied to the new geometrical structure, with eight connected
components and whose matrix representation is :
(34)

0

with = 1 and = 1
0 Lo C
0
0
1

is the matrix representing the ortochronic Lorentz group. This group has been
extended to a pentadimensional Kaluza space
in a way permitting the
geometrical interpretation of matter-antimatter duality (
), showing
that it is present both in the positive mass and energy region and in the negative one.



IX - Schwarzschilds solution revisited in the context of a Janus geometry

It is known since long time ago that in the Schwarzschild interior and exterior solutions
the magnitudes and , simple integration constants, can be either be positive or
negative. For the geodesics to be able to join it is necessary that they have the same sign.
If the set
gives geodesics that evoke the attraction exerted by a mass
on a test particle of mass +1, a set
would indicate a repulsive
action on that same test particle, this time exerted by a negative mass
. For more
than half a century no one thought that this solution could have any physical sense at all.
But in the context of a Janus geometry it is just the contrary. The calculated geodesics
correspond then to the interaction laws derived from the Newtonian approximation. A
negative mass, being it concentrated or spread, exerts on a photon of positive energy a
negative gravitational lensing effect whose features, introduced already in [17], will be
detailed in a future contribution.



X Suppression of the central singularity and inversion of the mass

In the following we will not be concerned with charges and matter-antimatter duality,
limiting our isometry group to the subgroup :

31

(35)

Lo

1
0

with = 1

and

0
0
0

The Schwarzschild-like solution under consideration will be represented this time by a
couple of metrics (
,
) that are solution of a coupled system of equations. But it is
known since long time ago that the classical Schwarzschild exterior metrics that is
solution of Einsteins equation without second term and which describes a portion of
void spacetime works fine if one changes the integration constant from m to m. The
same holds for the Schwarzschild interior solution that is solution of Einsteins
equation with second term when one changes to - . We have then our couple of
solutions (
,
).

Lets go back to our Schwarzschild solution :
(36)
R
dr 2
d s 2 = ( 1 s ) c 2 dt 2
r 2 ( d 2 + sin 2 d 2 )
R
r
(1 s )
r

Introducing the coordinate change (15), we get :
(37)
( 1 + Log ch ) 2

Log ch
d s2 =
c 2 d t 2 Rs2
th d 2 + ( 1 + Log ch ) 2 ( d 2 + sin 2 d 2 )
1 + Log ch
Log ch

When tends to infinity, Log ch and th 1 . And the metric tends to :
(38)
d s 2 = c 2 d t 2 Rs2 d 2 + 2 ( d 2 + sin 2 d 2 )

that is, to the Lorentz metric.

When tends to zero :
(39)

And the metric tends to :

(40)
d s 2 Rs2 2 d 2 + 2 ( d 2 + sin 2 d 2 )

The determinant of this metric is :
(41)

32

g = c 2 Rs3 sin 2 ( 1 + Log ch )3 th 2

It is zero for = 0 because of the vanishing of the term
When calculating the geodesics on the plane
one finds (see reference [2], equation 6.90) :
(42)
1
d = 2
r
c2 l 2 1

h2

in the

representation

dr
R
1
R
+ 2s 2 + s 3
h r r
f

where and are the classical parameters of the quasi-Keplerian trajectory (see
reference [2], equations 6.80 and 6.81). On the Schwarzschilds sphere (r = Rs) one has :
(43)
d
h
tg = Rs
=

d r r = R Rs c l
s

In a { t , , , } representation with
be inscribed this time in the two folds
following figure :

Fig. 29 : (plane,

well obtain the curves

and

that will

. Anticipating things, well have the

) geodesical path in a

representation

33

With :
(44)

( t g ) 0

d
r d d r
=
d
dr r d

h
2 0
R cl
2
s

The throat sphere
reduces to a point. But the geodesics of the fold
can be
prolonged in the adjacent fold
. The central singularity disappears. For a better
understanding of figure 29 it is worth to consider a piece of paper with a hole in O. The
geodesical path there indicated inscribes it by half over the front of the fold until the
point O, and then over its reverse side. The front and the reverse side of the paper give a
good image of the two-fold cover of a M2 manifold.

We have seen that the transit to a geometry with a two-fold cover of a bounded manifold
created an enantiomorphic relation between two adjacent regions.

With respect to time, it is necessary to recall section 6 and the coupled field equations
(29a) and (29b). The folds
and
have only in common the variables and ,
being aware that the fold
corresponds to
and the fold
to
. The
respective time-arrows are opposite, which correspond to the variables :
(45)
t ( ) = t (+ )

the inversion being due to the mass and energy inversion (see reference [34]). In the
neighbourhood of
it is possible to write the nearby expressions :
(46a)

(46b)

The inversion of the time variable does not imply a change in the sign of the proper time,
and from one fold to the other d s ( ) takes the reins of d s (+ ) . Its not possible to have an
inversion in the length measure along a geodesic, or in other words to have
(d s (+ ))(d s ( )) > 0 . But t (+ ) and t ( ) are nothing more than time markers , simple
coordinates, and thus one will have : dt ( ) = dt (+ ) . So here one finds again the central
idea of differential geometry : that only the length element has an intrinsic reality.
Lagrange equations give always in the vicinity of
the relations :
(47)


These functions are linear and monotonic as a function of the proper time (lengths s (+ )
and s ( ) ) that enchain themselves in passing from one fold to the other without

34

inversion).
(48)

d t (+ )
C
= 2
(+ )
ds

d t ( ) C
=
d s ( ) 2

C = Cst

The sign of the constant C depends on the sense adopted for the passage from one fold to
the other. The object is thus a black hole and a white fountain at the same time.
But the passage is achievable only in an infinite time. So one is faced again with the
infinity ot the transit time, question that we will come up to in our next paper. If the
object results from the implosion of a neutron star, its mass would be transferred to the
negative energy region. But for the observers located in one of the folds such
phenomena of implosion-explosion will be freezed in time .


Conclusion

We have recalled some aspects of the joined Schwarzschild interior and exterior
solutions as well as the main features of the black hole model that arise from the joint
conditions of their geometrical criticality. We have hence shown that several
singularities may appear as a consequence of an inadequate choice of coordinates, and
that such choices may contain implicitly a bad topological choice giving rise to a
central singularity . We have also shown that this central singularity, classically
associated to the exterior Schwarzschild solution, can be eliminated by a simple change
of variable, leading to the concept of space bridge, which in turn completes the change in
geometrical paradigm that stems from a Janus geometry. Geodesics can be extended
from an ortochronic fold F (+ ) into an antichronic and enantiomorphic fold F ( ) . The
inversion of the time marker signals the inversion of mass in the process, always in
freezing mode .


Addendum to conclusion :

In 1916 Karl Schwarzschild built the first solution of Einsteins equation without
second member. All details of this calculation can be found in all books devoted to
cosmology. During 58 years all people believed that the associated central
singularity was a true singularity , that could not be avoided. The Kretschmann
1
invariant scalar 6 should sign the presence of such singularity, because the region
r
of space is singular in any coordinate system , they said.

In this paper we have seen that some magic change of coordinates :

r = Rs ( 1 + Log ch )

cancelled it. So, what about this ( r = 0 ) singular point ?

Nuts, its just out the hypersurface !

35
But what is a coordinate system ? Nothing but the arbitrary choice of an observer, a
distant one, who is supposed to live in void, so that, in its vicinity the metric tends to
Lorentz.

This work arises a foundamental question. Do singularities exist in nature ? This is a
great question. During decades, scientists tried to study that so-called central
singularity that was supposed to live in the core of black holes.

By the way, do stellar black holes really exist ? Another big question. In effect, when
an astrophysicist builds a model of an unkown object and if this object exists in
nature, late or soon it is discovered. An example : in the begining of the thirties Fritz
Zwicky predicted that the death of massive stars should be catastrophic and produce
supernova phenomenon that he described quite precisely. People were very skeptical.
But Zwicky insisted and finally discovered the first supernovae just before the second
wold war II. Now, noboby would try to count them. Same thing for exo-planets, by the
way ...

Neutrons stars were also modelized, as the remnant of the supernova phenomenon.
There were discovered too, as so-called pulsars , and their number grows
exponentially. Just because the universe is very large and contains so many things.

The stellar black hole model was built in the middle of the sixties. Almost half a
century have past. Why so few candidates ?

Imagine we go backwards in time, in 1915. Schwarzschild has not built his solution
yet. And, suddently, someone builds the following metric as a solution of Einsteins
equation, without second member :

d s2 =

( 1 + Log ch ) 2

Log ch
c 2 d t 2 Rs2
th d 2 + ( 1 + Log ch ) 2 ( d 2 + sin 2 d 2 )
1 + Log ch
Log ch

The signature of such Riemanian metric is ( + ) . Good ! In addition the solution

tends to Lorentz at infinite. The geodesics are plane and at distance identify to
Keplerian trajectories. When becomes a little bit maller one finds the advance of
perihelion of almost elliptic paths. The model produces gravitational lensing effect.

When calculated the free fall time, towards = 0 is found to be infinite, as well as the
escape time.

Suddently, a guys says :

- Hey, men, why dont we try the following coordinate change ?
r

= arg ch e Rs

Naturally, the metric becomes :

36

ds 2 = (1

Rs 2 2
dr 2
) c dt
r 2 ( d 2 + sin 2 d 2 )
R
r
(1 s )
r

What is the benefit ? Now we have two kinds of pathologies. We have troubles when
r tends to Rs. When r < Rs the signature becomes ( + ) , the lengths become
imaginary ? And when r tends to zero, its a mess. Whats that stuff ? Metaphysics ?

This is the beginging of the story. We are working on chapter II and we hope to
present it soon. By the way, if somebody could develop the straightforward
calculation, from Einsteins equation, giving the Schwarzschild geometry, expressed
in { t , , , } coordinates, he would be welcome.

References

[1] K. Schwarzschild : ber das Gravitational eines Massenpunktes nach der
Einsteineschen Theory Sitzber. Preuss. Akad. Wiss. Berlin, p.189-196 (1916)
[2] R. Adler, M. Bazin, M. Schiffer : Introduction to General Relativity, Mc Graw Hill Cie
1965-1975
[3] J.-P. Petit, 1983 : http://www.savoir-sans-
frontieres.com/JPP/telechargeables/English/topo_the_world_eng.htm
[4] Hilbert, David (1924). "Die Grundlagen der Physik". Mathematische Annalen
(Springer-Verlag) 92 (1-2): 132. doi:10.1007/BF01448427.
[5] Paul Painlev, La mcanique classique et la thorie de la relativit, C. R. Acad. Sci.
(Paris) 173, 677680(1921)
[6] Allvar Gullstrand, Allgemeine Lsung des statischen Einkrperproblems in der
Einsteinschen Gravitationstheorie, Arkiv. Mat. Astron. Fys. 16(8), 115 (1922).
[7] Eddington, A. S. (Feb 1924). Nature 113 (2832): 192
[8] Finkelstein, David (1958). Phys. Rev 110: 965967
[9] G. Lematre (1933). "L'Univers en expansion". Annales de la Socit Scientifique de
Bruxelles A53: 5185.
Freeman.
[10] Kerr, R. P. : Gravitational Field of a Spinning Mass as an Example of Algebraically
Special Metrics. Phys. Rev. Let. 11, 237-238, 1963.
[11] J.P.Petit, G.DAgostini : Negative mass hypothesis in cosmology and the nature of
dark energy. Astrophysics and Space Science, 2014 sept. 20 th
[12] J.P.Petit, G.DAgostini : Cosmological bimetric model with interacting positive and
negative masses and two different speeds of light, in agreement with the observed
acceleration of the Universe. Modern Physics Letters A, Vol. 29, N 34, 2014 November
10th
[13] Damour T. , Kogan I I. Effective Lagrangians and universality classes of nonlinear
bigravity Phys. Rev. D 66 (2002) 104024. hep-th/0206042.
[14] Damour T. , Kogan I. I. , Papazoglou A. Non-linear bigravity and cosmic acceleration
Phys. Rev. D 66 (2002) 104025. hep-th/0206044.
[15] J.-P.Petit : The missing mass problem. Il Nuovo Cimento B Vol. 109 July 1994, pp.
697-710

37

[16] Weinberg S. : The quantum theory of fields, Cambridge University Press, Vol.1,
2005.
[17] Petit J.-P. : Twin Universe Cosmology. Astrophysics and Space Science (226): 273
307. 1995
[18] Piran T. : On Gravitational Repulsion, Gen. Relat. and Gravit. Vol. 29, N 11, (1997)
[19] El-Ad H. , Piran T. , and da Costa L. N. , (1996) Astrophys. J. Lett. 462 L13
[20] H. El-Ad, T. Piran and L. N. da Costa, Mon. Not. R. Astron. Soc. 1997, Vol 287, pp 790-
798
[21] H. El-Ad and T. Piran : Voids in the large-scale structure Astrophys. J. (1997), Vol
491, pp 421-435
[22] Koki Izumi, Chizaki Hagiwara, Koki Nakajima ,Takao Kitamura and Hideki Asada :
Gravitational lensing shear by an exotic lens with negative convergence or negative
mass. Physical Review D 88, 024049 (2013)
[23] Hossenfelder S.: A bimetric Theory with Exchange Symmetry. Phys. Rev. D78,
044015, 2008
[24] Riess, A. G., et al. 1998, AJ, 116, 1009
[25] Perlmutter, S., et al. 1999, ApJ, 517, 565
[26] Riess A. G. 2000, PASP, 112, 1284
[27] Filippenko, A. V., & Riess, A. G. 2001, in AIP Conf. Proc. 540, Particle
[28] Leibundgut, B. 2001, ARA&A, 39, 67
[29] Knop, R., et al. 2003, ApJ, 598, 102
[30] Tonry, J. T., et al. 2003, ApJ, 594, 1
[31] Barris, B., et al. 2004, ApJ, 602, 571
[32] Riess, A. G. 2004. Apj : 607:665687, 2004 June 1
[33] Brian Schmidt, Nobel Lecture: Accelerating expansion of the Universe through
observations of distant supernovae. Rev. of Mod Phys. , Vol. 84, July-sept. 2012. DOI:
10.1103/RevModPhys.84.1151
[34] Souriau. J. M. : Structure des systmes dynamiques. Dunod Ed. France, 1970 and
Structure of Dynamical Systems. Boston, Birkhaser Ed. 1997, Ch. III, inversion of space
and time, eq. (14.67)