Physical Interpretation of Coordinates for the Schwarzschild Metric

Tarun Biswas

arXiv:0809.1452v1 [gr-qc] 8 Sep 2008

State University of New York at New Paltz,

New Paltz, NY 12561, USA.
(Dated: September 8, 2008)
Since its first introduction, the Schwarzschild metric has been written in various coordinate systems. This has been done primarily to understand the nature of the coordinate singularity at the
event horizon. However, very often, the mathematics of a coordinate system does not provide a
clear physical interpretation. In Schwarzschilds original work, the origin of the radial coordinate
was at the event horizon. Hence, there was no black hole. The generally accepted current definition
of the radial coordinate has an origin beyond the horizon. This necessitates the discussion of black
holes. Here, some well-known and some not-so-well-known coordinate systems will be visited in
search of a physical interpretation. It will be noted that they all agree at large radial distances.
However, the location of the origin of the radial coordinate can be different for different systems.
Some natural coordinate systems can exclude the entirety of the interior of a black hole from the
physical manifold. Such coordinate systems maybe physically more acceptable as they avoid the
issue of the metric signature change across the horizon. Mathematically, the metric signature alone
can distinguish time from space coordinates. Hence, coordinate systems that include the interior of
the black hole need to switch the physical meanings of time and one of the space coordinates.
PACS numbers: 04.70.-s, 97.60.Lf

I.

INTRODUCTION

It is customary to use the standard spherical polar coordinates (t, r, , ) to describe the spherically symmetric
vacuum solution to Einsteins equations. This coordinate
system is meant for a R4 topology. However, the solution to Einsteins equations distorts not only the flatness
of R4 but also its topology. As a result, the spatial R3
part needs to have a region around the origin amputated
to provide the physical manifold. Standard practice is
to exclude only the singular point at the origin (r = 0).
However, here it will be argued that a finite spherical
region needs to be excluded to maintain the physical nature of the manifold. In either case, what is left of the
spatial part has the topology of a semi-infinite cylinder.
The Schwarzschild line element in standard polar coordinates is given as

rs  2 
rs 1 2
d 2 = 1
dr r2 d2 ,
(1)
dt 1
r
r
where

d2 = d2 + sin2 d2 ,

itself meaningless for such an observer as t is used in its

definition. Although the observer takes a finite amount
of time to reach r = rs in his/her own frame, it is not
clear what he/she sees for r < rs as the above metric
can no longer be used. There is no real coordinate transformation available between and t for r < rs . This is
why, if the above metric is forced for r < rs , the metric
signature changes. In other words, t needs to be imaginary!
In this light, it is useful to consider coordinate systems that exclude the r rs region (interior of a black
hole) from the physical manifold in a natural manner.
First I shall review some well-known coordinate systems
that include the interior of the black hole (EddingtonFinkelstein and Kruskal). Then I shall discuss some others that effectively exclude the interior of the black hole.
Among these will be the somewhat well-known isotropic
coordinates. But there are several others that produce
the same effect in a much simpler fashion. The original
system used by Schwarzschild is among them[2, 3]. An
indirect reason for excluding the interior of a black hole
has been seen recently in some quantum effects[4].

(2)

the speed of light c = 1, rs = 2GM/c2 , G is the universal gravitational constant and M is the mass of the
source. An observer starting at a point r = r0 > rs
and falling in a radial direction in this metric is known
to take infinite coordinate time t to reach r = rs (the
Schwarzschild radius)[1]. Hence, once the observer has
fallen past rs , the coordinate time t exceeds which is
meaningless. This might render the above line element

II.

THE EDDINGTON-FINKELSTEIN
COORDINATES

First, let us look at the ingoing Eddington-Finkelstein

coordinates[1]. Here the coordinate t is replaced by the
null coordinate V given by
V = t + r + rs ln |r/rs 1|.

(3)

Thus, the line element is given by

Electronic

address: [email protected]

d 2 = (1 rs /r)dV 2 2dV dr r2 d2 .

(4)

2
This removes the singularity in the metric component
at r = rs by using a coordinate transformation that is
singular at the same point. Hence, it is often argued
that the (V, r) coordinates represent reality better at the
Schwarzschild horizon (r = rs ) than the original (t, r)
coordinates. However, t is directly measured by some
observers clock while V , being a null coordinate, has no
direct observational meaning from any observers point
of view. Besides, removing the singularity in a metric
component through a coordinate transformation is only
a cosmetic advantage as the components of the curvature
tensor are well-behaved in either coordinate system. So,
on physical grounds, t might still be the better coordinate.
The real problem of the Schwarzschild metric is not
the coordinate singularity, but the metric signature in
the interior of the black hole. It is the metric signature
alone that mathematically distinguishes space from time.
Physically, space and time are very different. But the
mathematics of relativity makes them look very similar.
The only mathematical tag that tells us which is which is
the metric signature. Within the black hole this metric
signature is switched for t and r and this plays havoc
with physical interpretation. Hence, it would be useful
to eliminate the interior of a black hole in a natural
way if possible.
The ingoing Eddington-Finkelstein coordinates are often used to show that objects can fall into a black hole
but they cannot fall out. This is a bit perplexing as
the original metric is time reversal symmetric! To reconstruct this original symmetry, one has to look at the other
Eddington-Finkelstein coordinate system the outgoing
one. For this, one defines
U = t r rs ln |r/rs 1|.

(5)

The resulting form for the line element is

d 2 = (1 rs /r)dU 2 + 2dU dr r2 d2 .

(6)

This describes the case of objects that fall out of a

black hole.
III.

.
....
..........
..
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
.
.
s ......... ....... ....... ...
..
....
...
...
..
...
...
...
...
.
..................................................................................................................................................................................................................................................................
.

...
...
...
...
...
..............
...
................
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
...
.........
...
................
.................
.
.
.
...
.
.
.
.
.
.
.
.
.
.
.
.
.
.......
.......
..........................................

rs /4

FIG. 1: r vs. the isotropic coordinate.

III
IV

II


u = (1 r/rs )1/2 er/2rs sinh(t/2rs ), for r < rs

(8)
v = (1 r/rs )1/2 er/2rs cosh(t/2rs ), for r < rs

u = (r/rs 1)1/2 er/2rs cosh(t/2rs ), for r > rs

(9)
v = (r/rs 1)1/2 er/2rs sinh(t/2rs ), for r > rs

u = (1 r/rs )1/2 er/2rs sinh(t/2rs ), for r < rs

(10)
v = (1 r/rs )1/2 er/2rs cosh(t/2rs ), for r < rs

The resulting form for the line element is

d 2 = (4rs3 /r)er/rs (dv 2 du2 ) r2 d2 ,

(11)

where r is implicitly a function of u and v.

The metric signature problem (discussed in the last
section) is well hidden in these coordinates. Notice that
the forms for u and v are effectively
switched in patches

I and II (with a factor of i = 1 removed). There is a

similar switch in patches III and IV. This is like switching
r and t of the standard Schwarzschild coordinates. It is
not clear what the physical meaning of such a switch
would be. If r is the physical time inside a black hole,
then what would be the significance of r = 0?
IV.

THE ISOTROPIC COORDINATES NO

BLACK HOLE

THE KRUSKAL COORDINATES

The Kruskal coordinates[1] u and v are defined to replace r and t of the standard coordinates. Like the
Eddington-Finkelstein coordinates, they effectively remove the singularity of the metric components by using
a singular coordinate transformation. But the Kruskal
coordinates have no null coordinates. This might make
physical interpretation somewhat easier. However, the
switching of the metric signature within the black hole
is still a problem. Also, the Kruskal coordinates cover
the physical space twice! This is an added challenge to
their physical interpretation. The coordinates u and v
are defined in four patches:

u = (r/rs 1)1/2 er/2rs cosh(t/2rs ), for r > rs
I
(7)
v = (r/rs 1)1/2 er/2rs sinh(t/2rs ), for r > rs

In the isotropic coordinates[1], r is replaced by which

is defined as follows.
2

rs
.
(12)
r = 1+
4
The resulting form for the line element is
2

d =

1+

rs
4
rs
4

!2


4
rs
dt 1 +
(d2 + 2 d2 ). (13)
4
2

Figure 1 shows a graph of r vs. . It shows that

does not cover the interior of the black hole (r < rs )
at all. However, it covers the exterior twice! So, if the
range rs /4 < < is defined as the physical manifold,

3
imum value of r is the Schwarzschild radius rs . But unlike of the isotropic coordinates, r2 = 0 at the minima. Physically, it is possible to define r2 in the range
< r2 < +. As a result, it covers the physical space
twice.
The form for the line element is


rs
2
d = 1 2
dt2
(r2 + rs2 )1/2

..
.....
.........
....
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
..
...........................................................................
s
...
...
...
...
...
..
......................................................................................................................................................................................................................................................................
...
...
2
...
...
...
...
..

r
.....
.....
.....
.....
.....
.
.
.
.
.....
..
.....
....
.....
....
.
.
.
.....
.
..
.....
.....
......
.....
.
.
.
......
.
.
.......
....
.........
.......
............................... r = r

(r22 + rs2 rs (r22 + rs2 )1/2 )1 r22 dr22

(r22 + rs2 )d2 .

Once again the smallest sphere around the origin does

not have zero surface area. So, next we shall consider the
physical acceptability of this phenomenon.

FIG. 2: r vs. r2 : a hyperbolic coordinate.

there would be only one coordinate patch for the complete physical manifold and the interior of the black hole
would be naturally eliminated.
However, there is one unexpected side effect of the
isotropic coordinates. The smallest sphere around the
origin does not have a zero surface area! This effect will
be discussed in greater detail shortly.
V.

A CLASS OF SIMPLE COORDINATE

SYSTEMS

The success of the isotropic coordinates in naturally

eliminating the interior of the black hole leads one to
look for other simple coordinate systems that would do
the same. A class of such coordinate systems is given by
the transformation from r to some rn defined as follows.
rnn = rn rsn , for n = 1, 2, 3, . . . .

(14)

Each value of n corresponds to a coordinate system. The

range 0 < rn < is completely outside the black hole.
This may seem like a rather ad hoc way of eliminating
the interior of the black hole. However, it is interesting to
note that in Schwarzschilds original work[2, 3] the physical radial coordinate used was not r but r3 as given by
the n = 3 case of the above class of coordinate systems.
Such ad hoc coordinate systems are given legitimacy by
the unphysical nature of the interior of the black hole. No
probe can be sent by an outside observer to a black hole
to examine its interior and return with the information.
A quantum theoretical analysis has also produced similar
results[4].
Hence, a closer look at this class of coordinate systems
is warranted. The n = 1 case is quite trivial and the
n = 3 case is already known from Schwarzschilds work.
So, let us look at the n = 2 case. For n = 2 we get
r22 = r2 rs2 .

(16)

(15)

Figure 2 shows the r vs. r2 graph. Qualitatively, this

is similar to the case of isotropic coordinates. The min-

VI.

THE SURFACE AREA OF SPHERES

AROUND THE ORIGIN

Historically, it is not clear why Schwarzschilds original

radial coordinate was abandoned in favor of the currently
accepted r. It is very likely because the Schwarzschild
radial coordinate (called r3 here[5]) has a rather unusual
feature the surface area of the smallest spherical surface around the origin (r3 = 0) is not zero! However, one
needs to note that the system under consideration here is
also unusual. Gravity is expected to be very strong near
the horizon and hence space is expected to be significantly distorted due to curvature. So, the usual relation
of surface area S and radius r of a sphere should not
be expected. However, in standard coordinates this is
exactly what is assumed:
S = 4r2 .

(17)

This assumption is built into the choice of the angular

part of the metric r2 d2 . Note that in a standard textbook presentation of the derivation of the Schwarzschild
metric the following ansatz is used for the metric.
d 2 = A(r)dt2 B(r)dr2 r2 d2 ,

(18)

The undetermined functions A(r) and B(r) are determined from Einsteins equations. Schwarzschild used an
ansatz with three undetermined functions as follows.
d 2 = A(r)dt2 B(r)dr2 C(r)r2 d2 ,

(19)

This does not agree with equation 17 unless C(r) = 1.

The standard argument for choosing C(r) = 1 is the freedom of coordinate choice for r. Although such a choice
is legitimate, it results in a loss of the physical meaning
of the origin of r. There is no reason to believe that
the physical manifold starts at r = 0 and not at r = rs .
Schwarzschild assumed the physical manifold starts at
r3 = 0 which is the same as r = rs . In fact, all coordinate systems defined in the last two sections have this
property.

4
To appreciate the significance of not assuming equation 17 for curved spaces, one can consider some simple
2-dimensional curved surfaces. The circumference C of
a circle drawn on such a surface can be related to the
radius r in very different ways compared to flat space.
For example, a circle drawn on a 2-dimensional conical
surface with the apex as center has a circumference of
C = (2 sin )r,

(20)

where is the half angle of the cone and r is the radius of

the circle. Another example is that of a circle drawn on a
2-dimensional spherical surface. Here the circumference
is related to radius as follows.
C = 2R sin(r/R),

(21)

where R is the radius of the spherical surface. This relationship of C and r is not even linear. Hence, it should
not be surprising to find the surface area of a sphere in
4-dimensional curved space-time to be related to the ra-

[1] C. W. Misner, K. S. Thorne and J. A. Wheeler, Gravitation, (W. H. Freeman and Company, 1973).
[2] K. Schwarzschild, Sitzungsber. K. Preuss. Akad. Wiss.,
Phys.-Math. Kl. 1916, 189-196 (1916).
[3] K. Schwarzschild, (translation by S. Antoci and
A. Loinger) eprint arXiv:physics/9905030 (1999).

dial coordinate in ways different from the flat space case

given in equation 17.

VII.

CONCLUSION

It is customary to assume that the standard radial

coordinate r used for the Schwarzschild metric has the
physical range 0 < r < . Here it is shown that there is
no fundamental reason for assuming this range. Indeed,
there exist natural coordinate systems that exclude the
region inside the black hole. Such coordinates consider
the physical range to be rs < r < . Besides, there are
strong physical reasons for excluding the interior of the
black hole. One reason being the switching of the metric
signature within the black hole. Another is the physical
inaccessibility of interior black hole information to the
outside observer. This makes it impossible for an outside
observer to determine for sure if an object is indeed a
black hole.

[4] T. Vachaspati, D. Stojkovic and L. M. Krauss Phys. Rev.

D, 76, 024005 (2007).
[5] What is called r3 here was called r by Schwarzschild. Here,
we call it r3 to distinguish it from the currently accepted
r.